Compare commits

..

85 Commits

Author SHA1 Message Date
Harmen Stoppels
594a376c52 Set version to v0.22.2 2024-09-21 12:39:55 +02:00
Harmen Stoppels
1538c48616 run-unit-tests: no xdist if coverage (#46480)
xdist only slows down unit tests under coverage
2024-09-21 12:39:55 +02:00
Massimiliano Culpo
683e50b8d9 Run unit test in parallel again in CI (#45793)
The --trace-config option was failing for linux unit-tests,
so we were running serial.
2024-09-21 12:39:55 +02:00
Harmen Stoppels
9b32fb0beb Revert "Change environment modifications to escape with double quotes (#36789)" (#42780)
This reverts commit 690394fabc, as it causes arbitrary code execution.
2024-09-21 12:39:55 +02:00
Harmen Stoppels
2c6df0d491 deal with TimeoutError from ssl.py (#45683) 2024-09-21 12:39:55 +02:00
Harmen Stoppels
ce7218acae buildcache: fix hard-coded, outdated layout version (#45645) 2024-09-21 12:39:55 +02:00
Dominic Hofer
246eeb2b69 Remove execution permission from setup-env.sh (#45641)
`setup-env.sh` is meant to be sourced, not executed directly.
By revoking execution permissions, users who accidentally execute
the script will receive an error instead of seeing no effect.

* Remove execution permission from `setup-env.sh` and friends
* Don't make output file executable in `spack commands --update-completion`

---------

Co-authored-by: Todd Gamblin <tgamblin@llnl.gov>
Signed-off-by: Todd Gamblin <tgamblin@llnl.gov>
2024-09-21 12:39:55 +02:00
Harmen Stoppels
cc47ee3984 unparser.py: remove print statements (#45235) 2024-09-21 12:39:55 +02:00
Harmen Stoppels
7b644719c1 Avoid duplicate detectable tag (#45160)
in case of inheritance the static tags prop may be updated multiple
times, and it turns out builder classes magically inherit from
traditional package classes
2024-09-21 12:39:55 +02:00
Harmen Stoppels
d8a6aa551e build_environment: explicitly disable ccache if disabled (#45275) 2024-09-21 12:39:55 +02:00
Massimiliano Culpo
ac7b18483a Bump archspec to latest commit (#46445)
This should fix an issue with Neoverse XX detection

Signed-off-by: Massimiliano Culpo <massimiliano.culpo@gmail.com>
2024-09-21 12:39:55 +02:00
Massimiliano Culpo
39f37de4ce Update archspec to v0.2.5-dev (7e6740012b897ae4a950f0bba7e9726b767e921f) (#45721) 2024-09-21 12:39:55 +02:00
Harmen Stoppels
703e153404 require spec in develop entry (#46485) 2024-09-21 12:39:55 +02:00
Harmen Stoppels
aa013611bc url join: fix oci scheme (#46483)
* url.py: also special case oci scheme in join

* avoid fetching keys from oci mirror
2024-09-21 12:39:55 +02:00
Harmen Stoppels
6a7ccd4e46 docs: refer to upstreams.yaml in chain.rst title (#46475) 2024-09-21 12:39:55 +02:00
Harmen Stoppels
1c6c4b4690 spack.util.url: fix join breakage in python 3.12.6 (#46453) 2024-09-21 12:39:55 +02:00
arezaii
68558b3dd0 Chapel package: updates post release (#45304)
* Fix +rocm variant, to ensure correct dependencies on ROCm packages
  and use of AMD LLVM
* Add a +pshm variant for comm=gasnet to enable fast shared-memory
  comms between co-locales
* Add logic to ensure we get the native CXI libfabric network provider
  on Cray EX
* Expand dependency type for package modules to encompass runtime
  dependencies
* Factor logic for setting (LD_)LIBRARY_PATH and PKG_CONFIG_PATH of
  runtime dependencies
* Workaround issue #44746 that causes a transitive dependency on lua
  to break SLURM
* Disable nonfunctional checkChplDoc test
* Annotate some variants as conditional, to improve spack info output
  and reduce confusion

---------

Co-authored-by: Dan Bonachea <dobonachea@lbl.gov>
2024-09-21 12:39:55 +02:00
arezaii
5440fe09cd update chapel package for v2.1 (#44931) 2024-09-21 12:39:55 +02:00
arezaii
03c22f403f Chapel package: major update (#42197)
* add cray detection taken from upcxx
* add CUDA/ROCm support
* add numerous pass-through options to Chapel build,
  like gpu_mem_strategy, comm_substrate, etc.; all variants are
  translated to analogous CHPL_* environment variables. As a side
  effect, this defines a number of environment variables that are
  not actually used by Chapel.
* Define LD_LIBRARY_PATH, LIBRARY_PATH, and PKG_CONFIG_PATH to
  help programs built with Chapel properly locate needed runtime
  dependencies

---------

Co-authored-by: bonachea <dobonachea@lbl.gov>
2024-09-21 12:39:55 +02:00
Greg Becker
f339225d22 include_concrete: read from older env formats properly (#45766)
* include_concrete: read from older env formats properly
* spack env rm: fix logic for checking env includes
* regression test

Signed-off-by: Todd Gamblin <tgamblin@llnl.gov>
2024-09-21 12:39:55 +02:00
Massimiliano Culpo
22c815f3d4 Do not halt concretization on unknown variants in externals (#45326)
* Do not halt concretization on unkwnown variants in externals
2024-09-21 12:39:55 +02:00
Massimiliano Culpo
4354288e44 Run minimization of weights only on known targets (#45269)
This prevents excessive output from clingo of the kind:

.../spack/lib/spack/spack/solver/concretize.lp:1640:5-11: info: tuple ignored:
  #sup@2
2024-09-21 12:39:55 +02:00
Massimiliano Culpo
ea2d43b4a6 Do not initialize previous store state in "use_store" (#45268)
The "use_store" context manager is used to swap the value
of a global variable (spack.store.STORE), while keeping
another global variable consistent (spack.config.CONFIG).

When doing that it tries to evaluate the previous value
of the store, if that was not done already. This is wrong,
since the configuration might be in an "intermediate" state
that was never meant to trigger side effects.

Remove that operation, and add a unit test to
prevent regressions.
2024-09-21 12:39:55 +02:00
Massimiliano Culpo
85e67d60a0 Add compatibility of sequoia with previous macOS versions (#45127)
* Add compatibility of sequoia with previous macOS versions

* Add compatibility of sequoia with previous macOS versions
2024-09-21 12:39:55 +02:00
Adam J. Stewart
bf6a9ff5ed Add support for macOS Sequoia (#45018) 2024-09-21 12:39:55 +02:00
Jordan Galby
1bdc30979d Fix regression in spec format string for indiviual variants (#46206)
Fix a regression in {variants.X} and {variants.X.value} spec format strings.
2024-09-21 12:39:55 +02:00
Harmen Stoppels
ef1eabe5b3 Add c to the list of languages (#45191) 2024-09-21 12:39:55 +02:00
Harmen Stoppels
43d673f915 Add pkg- prefix to builtin.mock a b c d ... (#45205) 2024-09-21 12:39:55 +02:00
Harmen Stoppels
8a9c501030 spec.py: fix __getitem__ looking outside of dag (#45090)
`Spec.__getitem__` queries dependent edges, which almost always point to
nodes outside the sub-dag considered. It should only ever look at edges
being traversed.
2024-09-21 12:39:55 +02:00
Harmen Stoppels
9f035ca030 Set version to v0.22.2.dev0 2024-09-21 12:39:55 +02:00
Harmen Stoppels
d66dce2d66 Set version to v0.22.1 2024-07-04 15:14:09 +02:00
Jordan Galby
ef2aa2f5f5 spack audit packages: Fix message (#45045)
Fix message formatting of the "virtual dependency cannot have variants" error.
2024-07-04 15:13:31 +02:00
Harmen Stoppels
41f5f6eaab iconv: require libiconv on linux (#45026)
otherwise it is still picked up from glibc as it is external
2024-07-04 15:07:05 +02:00
Massimiliano Culpo
cba347e0b7 Heuristic decays to default over time (#45023)
This modifies heuristic to decay to clingo default
over time. The hope is that this helps with specs
that have an optimal solution with a high penalty.

Let target and compiler heuristic decay too, do not
guess compiler
2024-07-04 15:07:05 +02:00
Harmen Stoppels
a3cef0f02e netlib-lapack: provide blas and lapack together (#44981)
If netlib-lapack is built with ~external-blas, it internally links
liblapack.so with libblas.so, meaning that whenever netlib-lapack is
used as a lapack provider, the package must also be a blas provider.

Conversely using netli-lapack as a blas provider does not imply that it
also must provide lapack, but nothing is lost disallowing that...
2024-07-01 16:56:31 +02:00
Harmen Stoppels
45fca040c3 Use composite stage also for develop specs (#44950) 2024-07-01 16:56:31 +02:00
Harmen Stoppels
eb2b5739b2 Remove DIYStage (#44949) 2024-07-01 16:56:31 +02:00
Massimiliano Culpo
d299e17d43 neoverse-v1: restore py-cinemasci (#44976)
Use a different tactic for determining conflicts.

Give more priority to setting False very old versions.
2024-07-01 16:56:31 +02:00
Massimiliano Culpo
d883883be0 Ensure parent runtime version >= child (#44834)
Fixes a bug where old gcc-runtime libraries would be loaded at runtime, but newer are required by dependencies, breaking the binaries.
2024-07-01 16:56:31 +02:00
Massimiliano Culpo
249dcb49e2 ASP-based solver: add a generic rule for propagation (#44870)
This adds a generic propagate/2 rule to propagate any
fact to children in the DAG.
2024-07-01 16:56:31 +02:00
Massimiliano Culpo
8628add66b Simplify and improve solver heuristic (#44893)
When we changed how to deal with errors in November,
we didn't realize that for an unconstrained choice
rule it is more important in the heuristic to guess
what is NOT in the answer set, since it will be the
majority of options.

Previously this was following automatically from what
was in the answer set, via `1 { ... } 1` cardinality
constraints.

Here we improve the heuristic and the solve time for specs.
2024-07-01 16:56:31 +02:00
Harmen Stoppels
aeccba8bc0 build_environment: fix ccache error handling (#44740) 2024-07-01 16:56:31 +02:00
Todd Gamblin
d94e8ab36f python: make every view a venv (#44382)
#40773 introduced python-venv, which improved build isolation and avoids issues with,
e.g., `ubuntu`'s system python modifying `sysconfig` to include a (very unwanted)
`local` directory within the default install layout.

This addresses a few cases where #40773 removed functionality, without harming the
default cases where we use `python-venv`.

Traditionally, *every* view with `python` in it was essentially a virtual environment,
because we would copy the `python` interpreter and `os.py` into every view when linking.
We now rely on `python-venv` to do that, but only when it's used (i.e. new builds) and
only for packages that have an `extends("python")` directive.

This again makes every view with `python` in it a virtual environment, but only
if we're not already using a package like `python-venv`. This uses a different
mechanism from before -- instead of using the `virtualenv` trick of copying `python`
into the prefix, we instead create a `pyvenv.cfg` like `venv` (the more modern way
to do it).

This fixes two things:
1. If you already had an environment before Spack `v0.22` that worked, it would
   stop working without a reconcretize and rebuild in `v0.22`, because we no longer
   copy the python interpreter on link. Adding `pyvenv.cfg` fixes this in a more
   modern way, so old views will keep working.

2. If you have an env that only includes python packages that use `depends_on("python")`
   instead of `extends("python")`, those packages will now be importable as before,
   though they won't have the same level of build isolation you'd get with `extends`
   and `python-venv`.

* views: avoid making client code deal with link functions

Users of views and ViewDescriptors shouldn't have to deal with link functions -- they
should just say what type of linking they want.

- [x] views take a link_type, not a link function
- [x] views work out the link function from the link type
- [x] view descriptors and commands now just tell the view what they want.

* python: simplify logic for avoiding pyvenv.cfg in copy views

Signed-off-by: Todd Gamblin <tgamblin@llnl.gov>
2024-07-01 16:56:31 +02:00
Massimiliano Culpo
e66c26871f Move unit tests into the same file, simplify main workflow 2024-07-01 16:56:31 +02:00
kwryankrattiger
2db4ff7061 Generate jobs should use x86_64_v3 runners only (#44582) 2024-07-01 16:56:31 +02:00
Tom Bradford
c248932a94 protobuf: fix 3.4:3.21 patch checksum (#44443) 2024-07-01 16:56:31 +02:00
dmagdavector
f15d302fc7 protobuf: update hash for patch needed when="@3.4:3.21" (#44210)
* protobuf: update hash for patch needed when="@3.4:3.21"

* Update var/spack/repos/builtin/packages/protobuf/package.py

Co-authored-by: Wouter Deconinck <wdconinc@gmail.com>

* Update var/spack/repos/builtin/packages/protobuf/package.py

Co-authored-by: Wouter Deconinck <wdconinc@gmail.com>

---------

Co-authored-by: Wouter Deconinck <wdconinc@gmail.com>
2024-07-01 16:56:31 +02:00
John W. Parent
74ef630241 Windows: Non config changes to support Gitlab CI (#43965)
* Quote python for shlex

* Remove python path quoting patch

* spack env: Allow `C` "protocol" for config_path

When running spack on windows, a path beginning with `C://...` is a valid path.

* Remove makefile from ci rebuild

* GPG use llnl.util.filesystem.getuid

* Cleanup process_command

* Remove unused lines

* Fix tyop in encode_path

* Double quote arguments

* Cleanup process_command

* Pass cdash args with =

* Escape parens in CMD script

* escape parens doesn't only apply to paths

* Install deps

* sfn prefix

* use sfn with libxml2

* Add hash to dep install

* WIP

* REview

* Changes missed in prior review commit

* Style

* Ensure we handle Windows paths with config scopes

* clarify docstring

* No more MAKE_COMMAND

* syntax cleanup

* Actually correct is_path_url

* Correct call

* raise on other errors

* url2path behaves differently on unix

* Ensure proper quoting

* actually prepend slash in slash_hash

---------

Co-authored-by: Ryan Krattiger <ryan.krattiger@kitware.com>
Co-authored-by: Mike VanDenburgh <michael.vandenburgh@kitware.com>
2024-07-01 16:56:31 +02:00
John W. Parent
a70ea11e69 Gitlab CI: Windows Configs (#43967)
Add support for Gitlab CI on Windows

This PR adds the config changes required to configure and execute
Gitlab pipelines running Windows builds on Windows runners using
the existing Gitlab CI infrastructure (and newly added Windows 
infrastructure).

* Adds support for generating child pipelines dispatched to Windows runners
* Refactors the relevant pre-scripts, scripts, and post scripts to be compatible with Windows
* Adds Windows config section describing Windows jobs
* Adds VTK as Windows build stack (to be expanded later)
* Modifies proj to build on Windows
* Refactors Windows rpath symlinking to avoid system libs and externals

---------

Co-authored-by: Ryan Krattiger <ryan.krattiger@kitware.com>
Co-authored-by: Mike VanDenburgh <michael.vandenburgh@kitware.com>
Co-authored-by: Todd Gamblin <tgamblin@llnl.gov>
Co-authored-by: Scott Wittenburg <scott.wittenburg@kitware.com>
2024-07-01 16:56:31 +02:00
John W. Parent
a79b1bd9af Buildcache/ensure symlinks proper prefix (#43851)
* archive: relative links only

Ensure all links written into tarfiles generated from Spack prefixes do not contain symlinks pointing outside the prefix

* binary_distribution: limit extraction to prefix

Ensure files extracted from spackballs are not links pointing outside of the prefix

* Ensure rpaths are properly set on Windows

* hard error on extraction of absolute links

* refactor for non link-modifying approach

* Restore tarball extraction to original impl

* use custom readlink

* cleanup symlink module

* make lstrip
2024-07-01 16:56:31 +02:00
John W. Parent
ac5d5485b9 Cdash reporting timeout (#44213)
* Add timeout to cdash reporter PUT request

Add cdash timeout everywhere
Correct mock responder api

* Style

* brief doc
2024-07-01 16:56:31 +02:00
John W. Parent
04258f9cce Prefer llnl.util.symlink.readlink to os.readlink (#44126)
Symlinks on Windows can use longpath prefixes (\\?\); these are fine
in the context of win32 API interactions but break numerous facets of
Spack behavior that rely on string parsing/matching (archiving,
binary distributions, tarball extraction, view regen, etc).

Spack's internal readlink method (llnl.util.symlink.readlink)
gracefully handles this by removing the prefix and otherwise behaving
exactly as os.readlink does, so we should prefer that in all cases.
2024-07-01 16:56:31 +02:00
Scott Wittenburg
1b14170bd1 gitlab ci: fix untouched spec pruning on windows (#44279)
Use correct path separator in get_all_package_diffs for all platforms.
Ensures correct package change computation on Windows when pruning unchanged specs in Gitlab CI
2024-07-01 16:56:31 +02:00
Massimiliano Culpo
a3bc9dbfe8 Make strong preferences even stronger (#44373)
Before this PR, if Spack could see a possibility to reuse a spec that
doesn't match a strong preference, it would do so. After the PR, a
strong preference would take precedence.
2024-07-01 16:56:31 +02:00
Greg Becker
e7c86259bd bugfix: external detection for compilers with os but not target (#44156)
avoid calling `spec.target` when None.

When an external compiler package has an `os` set but no `target` set, Spack
currently falls into a codepath that calls `spec.target` (which itself calls
`spec.architecture.target.Microarchitecture`) when `spec.architecture.target`
is None, throwing an error.

e.g.

```
packages:
  gcc:
    externals:
    - spec: gcc@12.3.1 os=rhel7
      prefix: /usr
```

---------

Co-authored-by: Todd Gamblin <tgamblin@llnl.gov>
2024-07-01 16:56:31 +02:00
Massimiliano Culpo
2605aeb072 ASP-based solver: fix reusing externals on linux (#44316)
We need to tell clingo the libc compatibility of external nodes
in buildcaches or stores, to allow reuse.
2024-07-01 16:56:31 +02:00
Massimiliano Culpo
94536d2b66 Enforce consistency of gl providers (#44307)
* glew: rework dependency on gl

This simplifies the package and ensures a single gl implementation is
pulled in. Before we were adding direct dependencies, and those are
not unified through the virtual.

* mesa-demos: rework dependency on gl

This simplifies the package and ensures a single gl implementation is
pulled in. Before we were adding direct dependencies, and those are
not unified through the virtual.

* mesa-glu: rework dependency on gl

This simplifies the package and ensures a single gl implementation is
pulled in. Before we were adding direct dependencies, and those are
not unified through the virtual.

* paraview: fix dependency on glew

* mesa: group dependency on when("+glx")

* Add missing dependency on libxml2

* paraview: remove the "osmesa" and "egl" variant

Instead, enforce consistency using the "gl" virtual that allows
only one provider.

* visit: remove osmesa variant

* Disable paraview in the aws-isc stacks

* data-vis-sdk: rework constrains to enforce front-ends

* e4s-power: remove redundant paraview

* Pipelines: update osmesa variants

* trilinos-catalyst-ioss-adapter: make gl a run dependency
2024-07-01 16:56:31 +02:00
Massimiliano Culpo
5e580fc82e Remove mesa18 and libosmesa (#44264)
* Remove mesa18 and libosmesa

mesa18 was introduced in #19528 as a way to maintain the old
autotools build of mesa separate from the new meson build.

We could add a second build system to mesa, but since mesa18 has
been deprecated for a long time, we'll just remove it.

libosmesa was used to multiplex the gl provider between mesa18
and mesa, and is thus unecessary. Remove it to reduce complexity
in the graphical stack.

* Remove references to mesa18 and libosmesa

* vtk: rework dependency on gl and osmesa

* memsurfer: rework dependency on vtk

* visit: minimal fix to avoid having both osmesa and glx
2024-07-01 16:56:31 +02:00
Harmen Stoppels
195bad8675 Prefer libiconv for iconv (#44335)
`glibc` and `musl` provide a basic implementation of `iconv` (`iconv`,
`iconv_open`, `iconv_close`), but in practice the installation may be
missing the character encoding methods to make them usable. On Fedora
for example, users need to

```yum install glibc-gconv-extra```

to get the character encodings that `gettext` requires during configure,
namely EUC-JP. Users may not have permissions to install the missing
parts of glibc.

Since Spack can install `libiconv`, it is simpler to use that by
default.
2024-07-01 16:56:31 +02:00
Harmen Stoppels
bd9f3f100a gcc: use -rpath {rpath_dir} not -rpath={rpath dir} (#44315)
to make macOS's linker happy.
2024-07-01 16:56:31 +02:00
Mosè Giordano
b5962613a0 suite-sparse: improve setting of the libs property (#44214)
on some distros it is in lib64/
2024-07-01 16:56:31 +02:00
Massimiliano Culpo
cbcfc7e10a Demote a warning to debug message, if C compiler is not there (#44182) 2024-07-01 16:56:31 +02:00
Massimiliano Culpo
579fadacd0 ASP-based solver: fix version optimization for roots (#44272)
This fixes a bug occurring when two root specs need to select
old versions, and these versions have the same penalty in the
optimization. This sometimes caused an older version to be
preferred to a more recent one.

The issue was the omission of `PackageNode` in the optimization
tuple.
2024-07-01 16:56:31 +02:00
Chris Green
b86d08b022 git: bump v2.39 to 2.45; deprecate unsafe versions (#44248) 2024-07-01 16:56:31 +02:00
Scott Wittenburg
02d62cf40f oci buildcache: handle pagination of tags (#43136)
This fixes an issue where ghcr, gitlab and possibly other container registries paginate tags by default, which violates the OCI spec v1.0, but is common practice (the spec was broken itself). After this commit, you can create build cache indices of > 100 specs on ghcr.

Co-authored-by: Harmen Stoppels <me@harmenstoppels.nl>
2024-07-01 16:56:31 +02:00
Harmen Stoppels
97369776f0 build_environment.py: deal with rpathing identical packages (#44219)
When multiple gcc-runtime packages exist in the same link sub-dag, only rpath
the latest.
2024-07-01 16:56:31 +02:00
Howard Pritchard
47af0159dc py-matplotlib: qualify when to do a post install (#44191)
* py-matplotlib: qualify when to do a post install

Older versions of py-matplotlib don't seem to have some of the
files that the post install step is trying to install.
Looks like the files first appeared in 3.6.0 and later.

Signed-off-by: Howard Pritchard <hppritcha@gmail.com>

* Change install paths for older matplotlib

---------

Signed-off-by: Howard Pritchard <hppritcha@gmail.com>
Co-authored-by: Adam J. Stewart <ajstewart426@gmail.com>
2024-07-01 16:56:31 +02:00
Alec Scott
db6ead6fc1 rust: fix v1.78.0 instructions (#44127) 2024-07-01 16:56:31 +02:00
Harmen Stoppels
b4aa2c3cab glibc: detect from "Free Software Foundation" not "gnu" (#44154)
which should be more generic
2024-07-01 16:56:31 +02:00
Harmen Stoppels
4108de1ce4 Set version to v0.22.1.dev0 2024-07-01 16:56:31 +02:00
Todd Gamblin
5fe93fee1e Update CHANGELOG.md for v0.22.0 2024-05-12 02:06:28 +02:00
Todd Gamblin
8207f11333 Bump version to v0.22 2024-05-11 17:54:12 +02:00
Todd Gamblin
5bb5d2696f changelog: add changes form 0.21.1 and 0.21.2 (#44136)
These changes were added to the release branch but did not make it onto `develop`.
2024-05-11 17:48:27 +02:00
Harmen Stoppels
55f37dffe5 oci: improve default_retry (#44132)
Apparently urllib can throw a range of different exceptions:

1. HTTPError
2. URLError with e.reason set to the actual exception
3. TimeoutError from getresponse, which is not wrapped
2024-05-11 15:44:40 +02:00
Harmen Stoppels
252a5bd71b PythonExtension: fix issue where package does not extend python (#44109) 2024-05-10 10:48:06 +02:00
Massimiliano Culpo
f55224f161 Fix filtering external specs (#44093)
When an include filter on externals is present, implicitly
include libcs.

Also, do not penalize deprecated versions if they come
from externals.
2024-05-09 20:48:43 +02:00
Massimiliano Culpo
189ae4b06e CI/Update macOS runners: macos-latest switched to macos-14 (#44094)
macos-latest switched to macos-14, so now we are running
two identical jobs.
2024-05-09 20:48:43 +02:00
Harmen Stoppels
5e9c702fa7 gcc: use -idirafter for libc headers (#44081)
GCC C++ headers like cstdlib use `#include_next <stdlib.h>` to wrap libc
headers. We're using `-isystem` for libc, which puts those headers too
early in the search path. `-idirafter` fixes this so `include_next`
works.
2024-05-08 20:45:39 +02:00
Harmen Stoppels
965bb4d3c0 gitlab ci: tutorial: add julia and vim (#44073) 2024-05-08 14:19:22 +02:00
Todd Gamblin
354f98c94a r: patch R-CVE-2024-27322 for r@3.5:4.3.3 (#44050)
Signed-off-by: Todd Gamblin <tgamblin@llnl.gov>
2024-05-08 14:19:22 +02:00
Tamara Dahlgren
5dce480154 Remove dead environment creation code (#44065) 2024-05-08 14:19:22 +02:00
Richarda Butler
f634d48b7c Include concrete environments with include_concrete (#33768)
Add the ability to include any number of (potentially nested) concrete environments, e.g.:

```yaml
   spack:
     specs: []
     concretizer:
         unify: true
     include_concrete:
     - /path/to/environment1
     - /path/to/environment2
```

or, from the CLI:

```console
   $ spack env create myenv
   $ spack -e myenv add python
   $ spack -e myenv concretize
   $ spack env create --include-concrete myenv included_env
```

The contents of included concrete environments' spack.lock files are
included in the environment's lock file at creation time. Any changes
to included concrete environments are only reflected after the environment
is re-concretized from the re-concretized included environments.

- [x] Concretize included envs
- [x] Save concrete specs in memory by hash
- [x] Add included envs to combined env's lock file
- [x] Add test
- [x] Update documentation

    Co-authored-by: Kayla Butler <<butler59@llnl.gov>
    Co-authored-by: Tamara Dahlgren <35777542+tldahlgren@users.noreply.github.co
m>
    Co-authored-by: Todd Gamblin <tgamblin@llnl.gov>
    Signed-off-by: Todd Gamblin <tgamblin@llnl.gov>
2024-05-08 14:19:22 +02:00
Massimiliano Culpo
4daee565ae Update the tutorial command to point to releases/v0.22 (#44056) 2024-05-08 14:19:22 +02:00
Massimiliano Culpo
8e4dbdc2d7 Bump removal version in deprecation messages (#44064) 2024-05-08 14:19:22 +02:00
Harmen Stoppels
4f6adc03cd gitlab: dont build paraview for neoverse v2 (#44060) 2024-05-08 14:19:22 +02:00
6021 changed files with 52361 additions and 90562 deletions

View File

@@ -5,7 +5,7 @@ coverage:
status:
project:
default:
threshold: 2.0%
threshold: 0.2%
ignore:
- lib/spack/spack/test/.*

View File

@@ -1,5 +1,4 @@
{
"name": "Ubuntu 20.04",
"image": "ghcr.io/spack/ubuntu20.04-runner-amd64-gcc-11.4:2023.08.01",
"postCreateCommand": "./.devcontainer/postCreateCommand.sh"
}

View File

@@ -1,5 +0,0 @@
{
"name": "Ubuntu 22.04",
"image": "ghcr.io/spack/ubuntu-22.04:v2024-05-07",
"postCreateCommand": "./.devcontainer/postCreateCommand.sh"
}

View File

@@ -5,10 +5,13 @@ updates:
directory: "/"
schedule:
interval: "daily"
# Requirements to run style checks and build documentation
# Requirements to build documentation
- package-ecosystem: "pip"
directories:
- "/.github/workflows/requirements/style/*"
- "/lib/spack/docs"
directory: "/lib/spack/docs"
schedule:
interval: "daily"
# Requirements to run style checks
- package-ecosystem: "pip"
directory: "/.github/workflows/style"
schedule:
interval: "daily"

View File

@@ -28,8 +28,8 @@ jobs:
run:
shell: ${{ matrix.system.shell }}
steps:
- uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
- uses: actions/setup-python@0b93645e9fea7318ecaed2b359559ac225c90a2b
- uses: actions/checkout@0ad4b8fadaa221de15dcec353f45205ec38ea70b
- uses: actions/setup-python@82c7e631bb3cdc910f68e0081d67478d79c6982d
with:
python-version: ${{inputs.python_version}}
- name: Install Python packages
@@ -40,35 +40,30 @@ jobs:
run: |
python -m pip install --upgrade pywin32
- name: Package audits (with coverage)
env:
COVERAGE_FILE: coverage/.coverage-audits-${{ matrix.system.os }}
if: ${{ inputs.with_coverage == 'true' && runner.os != 'Windows' }}
run: |
. share/spack/setup-env.sh
coverage run $(which spack) audit packages
coverage run $(which spack) audit configs
coverage run $(which spack) -d audit externals
coverage combine
coverage xml
- name: Package audits (without coverage)
if: ${{ inputs.with_coverage == 'false' && runner.os != 'Windows' }}
run: |
. share/spack/setup-env.sh
. share/spack/setup-env.sh
spack -d audit packages
spack -d audit configs
spack -d audit externals
- name: Package audits (without coverage)
if: ${{ runner.os == 'Windows' }}
run: |
. share/spack/setup-env.sh
. share/spack/setup-env.sh
spack -d audit packages
./share/spack/qa/validate_last_exit.ps1
spack -d audit configs
./share/spack/qa/validate_last_exit.ps1
spack -d audit externals
./share/spack/qa/validate_last_exit.ps1
- uses: actions/upload-artifact@b4b15b8c7c6ac21ea08fcf65892d2ee8f75cf882
if: ${{ inputs.with_coverage == 'true' && runner.os != 'Windows' }}
- uses: codecov/codecov-action@5ecb98a3c6b747ed38dc09f787459979aebb39be
if: ${{ inputs.with_coverage == 'true' }}
with:
name: coverage-audits-${{ matrix.system.os }}
path: coverage
include-hidden-files: true
flags: unittests,audits
token: ${{ secrets.CODECOV_TOKEN }}
verbose: true

View File

@@ -1,7 +1,7 @@
#!/bin/bash
set -e
source share/spack/setup-env.sh
$PYTHON bin/spack bootstrap disable github-actions-v0.5
$PYTHON bin/spack bootstrap disable github-actions-v0.4
$PYTHON bin/spack bootstrap disable spack-install
$PYTHON bin/spack $SPACK_FLAGS solve zlib
tree $BOOTSTRAP/store

View File

@@ -37,14 +37,14 @@ jobs:
make patch unzip which xz python3 python3-devel tree \
cmake bison
- name: Checkout
uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
uses: actions/checkout@0ad4b8fadaa221de15dcec353f45205ec38ea70b
with:
fetch-depth: 0
- name: Bootstrap clingo
run: |
source share/spack/setup-env.sh
spack bootstrap disable github-actions-v0.6
spack bootstrap disable github-actions-v0.5
spack bootstrap disable github-actions-v0.4
spack external find cmake bison
spack -d solve zlib
tree ~/.spack/bootstrap/store/
@@ -60,20 +60,20 @@ jobs:
run: |
brew install cmake bison tree
- name: Checkout
uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
uses: actions/checkout@0ad4b8fadaa221de15dcec353f45205ec38ea70b
with:
fetch-depth: 0
- uses: actions/setup-python@0b93645e9fea7318ecaed2b359559ac225c90a2b
- uses: actions/setup-python@82c7e631bb3cdc910f68e0081d67478d79c6982d
with:
python-version: "3.12"
- name: Bootstrap clingo
run: |
source share/spack/setup-env.sh
spack bootstrap disable github-actions-v0.6
spack bootstrap disable github-actions-v0.5
spack bootstrap disable github-actions-v0.4
spack external find --not-buildable cmake bison
spack -d solve zlib
tree $HOME/.spack/bootstrap/store/
tree ~/.spack/bootstrap/store/
gnupg-sources:
runs-on: ${{ matrix.runner }}
@@ -83,22 +83,24 @@ jobs:
steps:
- name: Setup macOS
if: ${{ matrix.runner != 'ubuntu-latest' }}
run: brew install tree gawk
- name: Remove system executables
run: |
while [ -n "$(command -v gpg gpg2 patchelf)" ]; do
sudo rm $(command -v gpg gpg2 patchelf)
done
brew install tree
# Remove GnuPG since we want to bootstrap it
sudo rm -rf /usr/local/bin/gpg
- name: Setup Ubuntu
if: ${{ matrix.runner == 'ubuntu-latest' }}
run: |
sudo rm -rf $(which gpg) $(which gpg2) $(which patchelf)
- name: Checkout
uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
uses: actions/checkout@0ad4b8fadaa221de15dcec353f45205ec38ea70b
with:
fetch-depth: 0
- name: Bootstrap GnuPG
run: |
source share/spack/setup-env.sh
spack solve zlib
spack bootstrap disable github-actions-v0.6
spack bootstrap disable github-actions-v0.5
spack bootstrap disable github-actions-v0.4
spack -d gpg list
tree ~/.spack/bootstrap/store/
@@ -110,17 +112,19 @@ jobs:
steps:
- name: Setup macOS
if: ${{ matrix.runner != 'ubuntu-latest' }}
run: brew install tree
- name: Remove system executables
run: |
while [ -n "$(command -v gpg gpg2 patchelf)" ]; do
sudo rm $(command -v gpg gpg2 patchelf)
done
brew install tree
# Remove GnuPG since we want to bootstrap it
sudo rm -rf /usr/local/bin/gpg
- name: Setup Ubuntu
if: ${{ matrix.runner == 'ubuntu-latest' }}
run: |
sudo rm -rf $(which gpg) $(which gpg2) $(which patchelf)
- name: Checkout
uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
uses: actions/checkout@0ad4b8fadaa221de15dcec353f45205ec38ea70b
with:
fetch-depth: 0
- uses: actions/setup-python@0b93645e9fea7318ecaed2b359559ac225c90a2b
- uses: actions/setup-python@82c7e631bb3cdc910f68e0081d67478d79c6982d
with:
python-version: |
3.8
@@ -128,16 +132,15 @@ jobs:
3.10
3.11
3.12
3.13
- name: Set bootstrap sources
run: |
source share/spack/setup-env.sh
spack bootstrap disable github-actions-v0.5
spack bootstrap disable github-actions-v0.4
spack bootstrap disable spack-install
- name: Bootstrap clingo
run: |
set -e
for ver in '3.8' '3.9' '3.10' '3.11' '3.12' '3.13'; do
for ver in '3.8' '3.9' '3.10' '3.11' '3.12' ; do
not_found=1
ver_dir="$(find $RUNNER_TOOL_CACHE/Python -wholename "*/${ver}.*/*/bin" | grep . || true)"
if [[ -d "$ver_dir" ]] ; then
@@ -147,7 +150,7 @@ jobs:
not_found=0
old_path="$PATH"
export PATH="$ver_dir:$PATH"
./bin/spack-tmpconfig -b ./.github/workflows/bin/bootstrap-test.sh
./bin/spack-tmpconfig -b ./.github/workflows/bootstrap-test.sh
export PATH="$old_path"
fi
fi
@@ -160,45 +163,5 @@ jobs:
run: |
source share/spack/setup-env.sh
spack -d gpg list
tree $HOME/.spack/bootstrap/store/
- name: Bootstrap File
run: |
source share/spack/setup-env.sh
spack -d python share/spack/qa/bootstrap-file.py
tree $HOME/.spack/bootstrap/store/
tree ~/.spack/bootstrap/store/
windows:
runs-on: "windows-latest"
steps:
- name: Checkout
uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
with:
fetch-depth: 0
- uses: actions/setup-python@0b93645e9fea7318ecaed2b359559ac225c90a2b
with:
python-version: "3.12"
- name: Setup Windows
run: |
Remove-Item -Path (Get-Command gpg).Path
Remove-Item -Path (Get-Command file).Path
- name: Bootstrap clingo
run: |
./share/spack/setup-env.ps1
spack bootstrap disable github-actions-v0.6
spack bootstrap disable github-actions-v0.5
spack external find --not-buildable cmake bison
spack -d solve zlib
./share/spack/qa/validate_last_exit.ps1
tree $env:userprofile/.spack/bootstrap/store/
- name: Bootstrap GnuPG
run: |
./share/spack/setup-env.ps1
spack -d gpg list
./share/spack/qa/validate_last_exit.ps1
tree $env:userprofile/.spack/bootstrap/store/
- name: Bootstrap File
run: |
./share/spack/setup-env.ps1
spack -d python share/spack/qa/bootstrap-file.py
./share/spack/qa/validate_last_exit.ps1
tree $env:userprofile/.spack/bootstrap/store/

View File

@@ -40,7 +40,8 @@ jobs:
# 1: Platforms to build for
# 2: Base image (e.g. ubuntu:22.04)
dockerfile: [[amazon-linux, 'linux/amd64,linux/arm64', 'amazonlinux:2'],
[centos-stream9, 'linux/amd64,linux/arm64,linux/ppc64le', 'centos:stream9'],
[centos7, 'linux/amd64,linux/arm64,linux/ppc64le', 'centos:7'],
[centos-stream, 'linux/amd64,linux/arm64,linux/ppc64le', 'centos:stream'],
[leap15, 'linux/amd64,linux/arm64,linux/ppc64le', 'opensuse/leap:15'],
[ubuntu-focal, 'linux/amd64,linux/arm64,linux/ppc64le', 'ubuntu:20.04'],
[ubuntu-jammy, 'linux/amd64,linux/arm64,linux/ppc64le', 'ubuntu:22.04'],
@@ -55,7 +56,7 @@ jobs:
if: github.repository == 'spack/spack'
steps:
- name: Checkout
uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
uses: actions/checkout@0ad4b8fadaa221de15dcec353f45205ec38ea70b
- uses: docker/metadata-action@8e5442c4ef9f78752691e2d8f8d19755c6f78e81
id: docker_meta
@@ -76,7 +77,7 @@ jobs:
env:
SPACK_YAML_OS: "${{ matrix.dockerfile[2] }}"
run: |
.github/workflows/bin/generate_spack_yaml_containerize.sh
.github/workflows/generate_spack_yaml_containerize.sh
. share/spack/setup-env.sh
mkdir -p dockerfiles/${{ matrix.dockerfile[0] }}
spack containerize --last-stage=bootstrap | tee dockerfiles/${{ matrix.dockerfile[0] }}/Dockerfile
@@ -87,19 +88,19 @@ jobs:
fi
- name: Upload Dockerfile
uses: actions/upload-artifact@b4b15b8c7c6ac21ea08fcf65892d2ee8f75cf882
uses: actions/upload-artifact@65462800fd760344b1a7b4382951275a0abb4808
with:
name: dockerfiles_${{ matrix.dockerfile[0] }}
path: dockerfiles
- name: Set up QEMU
uses: docker/setup-qemu-action@49b3bc8e6bdd4a60e6116a5414239cba5943d3cf
uses: docker/setup-qemu-action@68827325e0b33c7199eb31dd4e31fbe9023e06e3
- name: Set up Docker Buildx
uses: docker/setup-buildx-action@c47758b77c9736f4b2ef4073d4d51994fabfe349
uses: docker/setup-buildx-action@d70bba72b1f3fd22344832f00baa16ece964efeb
- name: Log in to GitHub Container Registry
uses: docker/login-action@9780b0c442fbb1117ed29e0efdff1e18412f7567
uses: docker/login-action@e92390c5fb421da1463c202d546fed0ec5c39f20
with:
registry: ghcr.io
username: ${{ github.actor }}
@@ -107,13 +108,13 @@ jobs:
- name: Log in to DockerHub
if: github.event_name != 'pull_request'
uses: docker/login-action@9780b0c442fbb1117ed29e0efdff1e18412f7567
uses: docker/login-action@e92390c5fb421da1463c202d546fed0ec5c39f20
with:
username: ${{ secrets.DOCKERHUB_USERNAME }}
password: ${{ secrets.DOCKERHUB_TOKEN }}
- name: Build & Deploy ${{ matrix.dockerfile[0] }}
uses: docker/build-push-action@4f58ea79222b3b9dc2c8bbdd6debcef730109a75
uses: docker/build-push-action@2cdde995de11925a030ce8070c3d77a52ffcf1c0
with:
context: dockerfiles/${{ matrix.dockerfile[0] }}
platforms: ${{ matrix.dockerfile[1] }}
@@ -126,7 +127,7 @@ jobs:
needs: deploy-images
steps:
- name: Merge Artifacts
uses: actions/upload-artifact/merge@b4b15b8c7c6ac21ea08fcf65892d2ee8f75cf882
uses: actions/upload-artifact/merge@65462800fd760344b1a7b4382951275a0abb4808
with:
name: dockerfiles
pattern: dockerfiles_*

View File

@@ -15,6 +15,18 @@ concurrency:
cancel-in-progress: true
jobs:
prechecks:
needs: [ changes ]
uses: ./.github/workflows/valid-style.yml
secrets: inherit
with:
with_coverage: ${{ needs.changes.outputs.core }}
all-prechecks:
needs: [ prechecks ]
runs-on: ubuntu-latest
steps:
- name: Success
run: "true"
# Check which files have been updated by the PR
changes:
runs-on: ubuntu-latest
@@ -24,7 +36,7 @@ jobs:
core: ${{ steps.filter.outputs.core }}
packages: ${{ steps.filter.outputs.packages }}
steps:
- uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
- uses: actions/checkout@0ad4b8fadaa221de15dcec353f45205ec38ea70b
if: ${{ github.event_name == 'push' }}
with:
fetch-depth: 0
@@ -41,13 +53,6 @@ jobs:
- 'var/spack/repos/builtin/packages/clingo/**'
- 'var/spack/repos/builtin/packages/python/**'
- 'var/spack/repos/builtin/packages/re2c/**'
- 'var/spack/repos/builtin/packages/gnupg/**'
- 'var/spack/repos/builtin/packages/libassuan/**'
- 'var/spack/repos/builtin/packages/libgcrypt/**'
- 'var/spack/repos/builtin/packages/libgpg-error/**'
- 'var/spack/repos/builtin/packages/libksba/**'
- 'var/spack/repos/builtin/packages/npth/**'
- 'var/spack/repos/builtin/packages/pinentry/**'
- 'lib/spack/**'
- 'share/spack/**'
- '.github/workflows/bootstrap.yml'
@@ -67,53 +72,14 @@ jobs:
needs: [ prechecks, changes ]
uses: ./.github/workflows/bootstrap.yml
secrets: inherit
unit-tests:
if: ${{ github.repository == 'spack/spack' && needs.changes.outputs.core == 'true' }}
needs: [ prechecks, changes ]
uses: ./.github/workflows/unit_tests.yaml
secrets: inherit
prechecks:
needs: [ changes ]
uses: ./.github/workflows/valid-style.yml
secrets: inherit
with:
with_coverage: ${{ needs.changes.outputs.core }}
all-prechecks:
needs: [ prechecks ]
if: ${{ always() }}
all:
needs: [ unit-tests, bootstrap ]
runs-on: ubuntu-latest
steps:
- name: Success
run: |
if [ "${{ needs.prechecks.result }}" == "failure" ] || [ "${{ needs.prechecks.result }}" == "canceled" ]; then
echo "Unit tests failed."
exit 1
else
exit 0
fi
coverage:
needs: [ unit-tests, prechecks ]
uses: ./.github/workflows/coverage.yml
secrets: inherit
all:
needs: [ unit-tests, coverage, bootstrap ]
if: ${{ always() }}
runs-on: ubuntu-latest
# See https://docs.github.com/en/actions/writing-workflows/choosing-what-your-workflow-does/accessing-contextual-information-about-workflow-runs#needs-context
steps:
- name: Status summary
run: |
if [ "${{ needs.unit-tests.result }}" == "failure" ] || [ "${{ needs.unit-tests.result }}" == "canceled" ]; then
echo "Unit tests failed."
exit 1
elif [ "${{ needs.bootstrap.result }}" == "failure" ] || [ "${{ needs.bootstrap.result }}" == "canceled" ]; then
echo "Bootstrap tests failed."
exit 1
else
exit 0
fi
run: "true"

View File

@@ -1,34 +0,0 @@
name: coverage
on:
workflow_call:
jobs:
# Upload coverage reports to codecov once as a single bundle
upload:
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
- uses: actions/setup-python@0b93645e9fea7318ecaed2b359559ac225c90a2b
with:
python-version: '3.11'
cache: 'pip'
- name: Install python dependencies
run: pip install -r .github/workflows/requirements/coverage/requirements.txt
- name: Download coverage artifact files
uses: actions/download-artifact@fa0a91b85d4f404e444e00e005971372dc801d16
with:
pattern: coverage-*
path: coverage
merge-multiple: true
- run: ls -la coverage
- run: coverage combine -a coverage/.coverage*
- run: coverage xml
- name: "Upload coverage report to CodeCov"
uses: codecov/codecov-action@5c47607acb93fed5485fdbf7232e8a31425f672a
with:
verbose: true

8
.github/workflows/install_spack.sh vendored Executable file
View File

@@ -0,0 +1,8 @@
#!/usr/bin/env sh
. share/spack/setup-env.sh
echo -e "config:\n build_jobs: 2" > etc/spack/config.yaml
spack config add "packages:all:target:[x86_64]"
spack compiler find
spack compiler info apple-clang
spack debug report
spack solve zlib

View File

@@ -14,10 +14,10 @@ jobs:
build-paraview-deps:
runs-on: windows-latest
steps:
- uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
- uses: actions/checkout@0ad4b8fadaa221de15dcec353f45205ec38ea70b
with:
fetch-depth: 0
- uses: actions/setup-python@0b93645e9fea7318ecaed2b359559ac225c90a2b
- uses: actions/setup-python@82c7e631bb3cdc910f68e0081d67478d79c6982d
with:
python-version: 3.9
- name: Install Python packages

View File

@@ -1 +0,0 @@
coverage==7.6.1

View File

@@ -1,7 +0,0 @@
black==24.10.0
clingo==5.7.1
flake8==7.1.1
isort==5.13.2
mypy==1.8.0
types-six==1.16.21.20241105
vermin==1.6.0

View File

@@ -0,0 +1,7 @@
black==24.4.2
clingo==5.7.1
flake8==7.0.0
isort==5.13.2
mypy==1.8.0
types-six==1.16.21.9
vermin==1.6.0

View File

@@ -16,34 +16,45 @@ jobs:
matrix:
os: [ubuntu-latest]
python-version: ['3.7', '3.8', '3.9', '3.10', '3.11', '3.12']
concretizer: ['clingo']
on_develop:
- ${{ github.ref == 'refs/heads/develop' }}
include:
- python-version: '3.11'
os: ubuntu-latest
concretizer: original
on_develop: ${{ github.ref == 'refs/heads/develop' }}
- python-version: '3.6'
os: ubuntu-20.04
concretizer: clingo
on_develop: ${{ github.ref == 'refs/heads/develop' }}
exclude:
- python-version: '3.7'
os: ubuntu-latest
concretizer: 'clingo'
on_develop: false
- python-version: '3.8'
os: ubuntu-latest
concretizer: 'clingo'
on_develop: false
- python-version: '3.9'
os: ubuntu-latest
concretizer: 'clingo'
on_develop: false
- python-version: '3.10'
os: ubuntu-latest
concretizer: 'clingo'
on_develop: false
- python-version: '3.11'
os: ubuntu-latest
concretizer: 'clingo'
on_develop: false
steps:
- uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
- uses: actions/checkout@0ad4b8fadaa221de15dcec353f45205ec38ea70b
with:
fetch-depth: 0
- uses: actions/setup-python@0b93645e9fea7318ecaed2b359559ac225c90a2b
- uses: actions/setup-python@82c7e631bb3cdc910f68e0081d67478d79c6982d
with:
python-version: ${{ matrix.python-version }}
- name: Install System packages
@@ -55,13 +66,13 @@ jobs:
cmake bison libbison-dev kcov
- name: Install Python packages
run: |
pip install --upgrade pip setuptools pytest pytest-xdist pytest-cov clingo
pip install --upgrade pip setuptools pytest pytest-xdist pytest-cov
pip install --upgrade flake8 "isort>=4.3.5" "mypy>=0.900" "click" "black"
- name: Setup git configuration
run: |
# Need this for the git tests to succeed.
git --version
. .github/workflows/bin/setup_git.sh
. .github/workflows/setup_git.sh
- name: Bootstrap clingo
if: ${{ matrix.concretizer == 'clingo' }}
env:
@@ -74,25 +85,25 @@ jobs:
- name: Run unit tests
env:
SPACK_PYTHON: python
SPACK_TEST_SOLVER: ${{ matrix.concretizer }}
SPACK_TEST_PARALLEL: 2
COVERAGE: true
COVERAGE_FILE: coverage/.coverage-${{ matrix.os }}-python${{ matrix.python-version }}
UNIT_TEST_COVERAGE: ${{ matrix.python-version == '3.11' }}
run: |
share/spack/qa/run-unit-tests
- uses: actions/upload-artifact@b4b15b8c7c6ac21ea08fcf65892d2ee8f75cf882
- uses: codecov/codecov-action@5ecb98a3c6b747ed38dc09f787459979aebb39be
with:
name: coverage-${{ matrix.os }}-python${{ matrix.python-version }}
path: coverage
include-hidden-files: true
flags: unittests,linux,${{ matrix.concretizer }}
token: ${{ secrets.CODECOV_TOKEN }}
verbose: true
# Test shell integration
shell:
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
- uses: actions/checkout@0ad4b8fadaa221de15dcec353f45205ec38ea70b
with:
fetch-depth: 0
- uses: actions/setup-python@0b93645e9fea7318ecaed2b359559ac225c90a2b
- uses: actions/setup-python@82c7e631bb3cdc910f68e0081d67478d79c6982d
with:
python-version: '3.11'
- name: Install System packages
@@ -107,17 +118,17 @@ jobs:
run: |
# Need this for the git tests to succeed.
git --version
. .github/workflows/bin/setup_git.sh
. .github/workflows/setup_git.sh
- name: Run shell tests
env:
COVERAGE: true
run: |
share/spack/qa/run-shell-tests
- uses: actions/upload-artifact@b4b15b8c7c6ac21ea08fcf65892d2ee8f75cf882
- uses: codecov/codecov-action@5ecb98a3c6b747ed38dc09f787459979aebb39be
with:
name: coverage-shell
path: coverage
include-hidden-files: true
flags: shelltests,linux
token: ${{ secrets.CODECOV_TOKEN }}
verbose: true
# Test RHEL8 UBI with platform Python. This job is run
# only on PRs modifying core Spack
@@ -130,13 +141,13 @@ jobs:
dnf install -y \
bzip2 curl file gcc-c++ gcc gcc-gfortran git gnupg2 gzip \
make patch tcl unzip which xz
- uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
- uses: actions/checkout@0ad4b8fadaa221de15dcec353f45205ec38ea70b
- name: Setup repo and non-root user
run: |
git --version
git config --global --add safe.directory /__w/spack/spack
git fetch --unshallow
. .github/workflows/bin/setup_git.sh
. .github/workflows/setup_git.sh
useradd spack-test
chown -R spack-test .
- name: Run unit tests
@@ -149,36 +160,36 @@ jobs:
clingo-cffi:
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
- uses: actions/checkout@0ad4b8fadaa221de15dcec353f45205ec38ea70b
with:
fetch-depth: 0
- uses: actions/setup-python@0b93645e9fea7318ecaed2b359559ac225c90a2b
- uses: actions/setup-python@82c7e631bb3cdc910f68e0081d67478d79c6982d
with:
python-version: '3.13'
python-version: '3.11'
- name: Install System packages
run: |
sudo apt-get -y update
sudo apt-get -y install coreutils gfortran graphviz gnupg2
sudo apt-get -y install coreutils cvs gfortran graphviz gnupg2 mercurial ninja-build kcov
- name: Install Python packages
run: |
pip install --upgrade pip setuptools pytest coverage[toml] pytest-cov clingo
pip install --upgrade pip setuptools pytest coverage[toml] pytest-cov clingo pytest-xdist
pip install --upgrade flake8 "isort>=4.3.5" "mypy>=0.900" "click" "black"
- name: Setup git configuration
run: |
# Need this for the git tests to succeed.
git --version
. .github/workflows/setup_git.sh
- name: Run unit tests (full suite with coverage)
env:
COVERAGE: true
COVERAGE_FILE: coverage/.coverage-clingo-cffi
SPACK_TEST_SOLVER: clingo
run: |
. share/spack/setup-env.sh
spack bootstrap disable spack-install
spack bootstrap disable github-actions-v0.5
spack bootstrap disable github-actions-v0.6
spack bootstrap status
spack unit-test --verbose --cov --cov-config=pyproject.toml --cov-report=xml:coverage.xml lib/spack/spack/test/concretization/core.py
- uses: actions/upload-artifact@b4b15b8c7c6ac21ea08fcf65892d2ee8f75cf882
share/spack/qa/run-unit-tests
- uses: codecov/codecov-action@5ecb98a3c6b747ed38dc09f787459979aebb39be
with:
name: coverage-clingo-cffi
path: coverage
include-hidden-files: true
flags: unittests,linux,clingo
token: ${{ secrets.CODECOV_TOKEN }}
verbose: true
# Run unit tests on MacOS
macos:
runs-on: ${{ matrix.os }}
@@ -187,10 +198,10 @@ jobs:
os: [macos-13, macos-14]
python-version: ["3.11"]
steps:
- uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
- uses: actions/checkout@0ad4b8fadaa221de15dcec353f45205ec38ea70b
with:
fetch-depth: 0
- uses: actions/setup-python@0b93645e9fea7318ecaed2b359559ac225c90a2b
- uses: actions/setup-python@82c7e631bb3cdc910f68e0081d67478d79c6982d
with:
python-version: ${{ matrix.python-version }}
- name: Install Python packages
@@ -199,24 +210,24 @@ jobs:
pip install --upgrade pytest coverage[toml] pytest-xdist pytest-cov
- name: Setup Homebrew packages
run: |
brew install dash fish gcc gnupg kcov
brew install dash fish gcc gnupg2 kcov
- name: Run unit tests
env:
SPACK_TEST_SOLVER: clingo
SPACK_TEST_PARALLEL: 4
COVERAGE_FILE: coverage/.coverage-${{ matrix.os }}-python${{ matrix.python-version }}
run: |
git --version
. .github/workflows/bin/setup_git.sh
. .github/workflows/setup_git.sh
. share/spack/setup-env.sh
$(which spack) bootstrap disable spack-install
$(which spack) solve zlib
common_args=(--dist loadfile --tx '4*popen//python=./bin/spack-tmpconfig python -u ./bin/spack python')
common_args=(--dist loadfile --tx '4*popen//python=./bin/spack-tmpconfig python -u ./bin/spack python' -x)
$(which spack) unit-test --verbose --cov --cov-config=pyproject.toml --cov-report=xml:coverage.xml "${common_args[@]}"
- uses: actions/upload-artifact@b4b15b8c7c6ac21ea08fcf65892d2ee8f75cf882
- uses: codecov/codecov-action@5ecb98a3c6b747ed38dc09f787459979aebb39be
with:
name: coverage-${{ matrix.os }}-python${{ matrix.python-version }}
path: coverage
include-hidden-files: true
flags: unittests,macos
token: ${{ secrets.CODECOV_TOKEN }}
verbose: true
# Run unit tests on Windows
windows:
defaults:
@@ -225,10 +236,10 @@ jobs:
powershell Invoke-Expression -Command "./share/spack/qa/windows_test_setup.ps1"; {0}
runs-on: windows-latest
steps:
- uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
- uses: actions/checkout@a5ac7e51b41094c92402da3b24376905380afc29
with:
fetch-depth: 0
- uses: actions/setup-python@0b93645e9fea7318ecaed2b359559ac225c90a2b
- uses: actions/setup-python@82c7e631bb3cdc910f68e0081d67478d79c6982d
with:
python-version: 3.9
- name: Install Python packages
@@ -236,15 +247,15 @@ jobs:
python -m pip install --upgrade pip pywin32 setuptools pytest-cov clingo
- name: Create local develop
run: |
./.github/workflows/bin/setup_git.ps1
./.github/workflows/setup_git.ps1
- name: Unit Test
env:
COVERAGE_FILE: coverage/.coverage-windows
run: |
spack unit-test --verbose --cov --cov-config=pyproject.toml
spack unit-test -x --verbose --cov --cov-config=pyproject.toml
./share/spack/qa/validate_last_exit.ps1
- uses: actions/upload-artifact@b4b15b8c7c6ac21ea08fcf65892d2ee8f75cf882
coverage combine -a
coverage xml
- uses: codecov/codecov-action@125fc84a9a348dbcf27191600683ec096ec9021c
with:
name: coverage-windows
path: coverage
include-hidden-files: true
flags: unittests,windows
token: ${{ secrets.CODECOV_TOKEN }}
verbose: true

View File

@@ -18,15 +18,15 @@ jobs:
validate:
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
- uses: actions/setup-python@0b93645e9fea7318ecaed2b359559ac225c90a2b
- uses: actions/checkout@0ad4b8fadaa221de15dcec353f45205ec38ea70b
- uses: actions/setup-python@82c7e631bb3cdc910f68e0081d67478d79c6982d
with:
python-version: '3.11'
cache: 'pip'
- name: Install Python Packages
run: |
pip install --upgrade pip setuptools
pip install -r .github/workflows/requirements/style/requirements.txt
pip install -r .github/workflows/style/requirements.txt
- name: vermin (Spack's Core)
run: vermin --backport importlib --backport argparse --violations --backport typing -t=3.6- -vvv lib/spack/spack/ lib/spack/llnl/ bin/
- name: vermin (Repositories)
@@ -35,22 +35,22 @@ jobs:
style:
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
- uses: actions/checkout@0ad4b8fadaa221de15dcec353f45205ec38ea70b
with:
fetch-depth: 0
- uses: actions/setup-python@0b93645e9fea7318ecaed2b359559ac225c90a2b
- uses: actions/setup-python@82c7e631bb3cdc910f68e0081d67478d79c6982d
with:
python-version: '3.11'
cache: 'pip'
- name: Install Python packages
run: |
pip install --upgrade pip setuptools
pip install -r .github/workflows/requirements/style/requirements.txt
pip install -r .github/workflows/style/requirements.txt
- name: Setup git configuration
run: |
# Need this for the git tests to succeed.
git --version
. .github/workflows/bin/setup_git.sh
. .github/workflows/setup_git.sh
- name: Run style tests
run: |
share/spack/qa/run-style-tests
@@ -70,13 +70,13 @@ jobs:
dnf install -y \
bzip2 curl file gcc-c++ gcc gcc-gfortran git gnupg2 gzip \
make patch tcl unzip which xz
- uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
- uses: actions/checkout@0ad4b8fadaa221de15dcec353f45205ec38ea70b
- name: Setup repo and non-root user
run: |
git --version
git config --global --add safe.directory /__w/spack/spack
git fetch --unshallow
. .github/workflows/bin/setup_git.sh
. .github/workflows/setup_git.sh
useradd spack-test
chown -R spack-test .
- name: Bootstrap Spack development environment
@@ -85,64 +85,5 @@ jobs:
source share/spack/setup-env.sh
spack debug report
spack -d bootstrap now --dev
spack -d style -t black
spack style -t black
spack unit-test -V
import-check:
runs-on: ubuntu-latest
steps:
- uses: julia-actions/setup-julia@v2
with:
version: '1.10'
- uses: julia-actions/cache@v2
# PR: use the base of the PR as the old commit
- name: Checkout PR base commit
if: github.event_name == 'pull_request'
uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
with:
ref: ${{ github.event.pull_request.base.sha }}
path: old
# not a PR: use the previous commit as the old commit
- name: Checkout previous commit
if: github.event_name != 'pull_request'
uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
with:
fetch-depth: 2
path: old
- name: Checkout previous commit
if: github.event_name != 'pull_request'
run: git -C old reset --hard HEAD^
- name: Checkout new commit
uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
with:
path: new
- name: Install circular import checker
uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
with:
repository: haampie/circular-import-fighter
ref: 9f60f51bc7134e0be73f27623f1b0357d1718427
path: circular-import-fighter
- name: Install dependencies
working-directory: circular-import-fighter
run: make -j dependencies
- name: Import cycles before
working-directory: circular-import-fighter
run: make SPACK_ROOT=../old && cp solution solution.old
- name: Import cycles after
working-directory: circular-import-fighter
run: make clean-graph && make SPACK_ROOT=../new && cp solution solution.new
- name: Compare import cycles
working-directory: circular-import-fighter
run: |
edges_before="$(grep -oP 'edges to delete: \K\d+' solution.old)"
edges_after="$(grep -oP 'edges to delete: \K\d+' solution.new)"
if [ "$edges_after" -gt "$edges_before" ]; then
printf '\033[1;31mImport check failed: %s imports need to be deleted, ' "$edges_after"
printf 'previously this was %s\033[0m\n' "$edges_before"
printf 'Compare \033[1;97m"Import cycles before"\033[0m and '
printf '\033[1;97m"Import cycles after"\033[0m to see problematic imports.\n'
exit 1
else
printf '\033[1;32mImport check passed: %s <= %s\033[0m\n' "$edges_after" "$edges_before"
fi

View File

@@ -14,26 +14,3 @@ sphinx:
python:
install:
- requirements: lib/spack/docs/requirements.txt
search:
ranking:
spack.html: -10
spack.*.html: -10
llnl.html: -10
llnl.*.html: -10
_modules/*: -10
command_index.html: -9
basic_usage.html: 5
configuration.html: 5
config_yaml.html: 5
packages_yaml.html: 5
build_settings.html: 5
environments.html: 5
containers.html: 5
mirrors.html: 5
module_file_support.html: 5
repositories.html: 5
binary_caches.html: 5
chain.html: 5
pipelines.html: 5
packaging_guide.html: 5

View File

@@ -60,6 +60,7 @@
- suite-sparse: improve setting of the `libs` property (#44214)
- netlib-lapack: provide blas and lapack together (#44981)
# v0.22.0 (2024-05-12)
`v0.22.0` is a major feature release.
@@ -380,15 +381,6 @@
* 344 committers to packages
* 45 committers to core
# v0.21.3 (2024-10-02)
## Bugfixes
- Forward compatibility with Spack 0.23 packages with language dependencies (#45205, #45191)
- Forward compatibility with `urllib` from Python 3.12.6+ (#46453, #46483)
- Bump `archspec` to 0.2.5-dev for better aarch64 and Windows support (#42854, #44005,
#45721, #46445)
- Support macOS Sequoia (#45018, #45127, #43862)
- CI and test maintenance (#42909, #42728, #46711, #41943, #43363)
# v0.21.2 (2024-03-01)
@@ -419,7 +411,7 @@
- spack graph: fix coloring with environments (#41240)
- spack info: sort variants in --variants-by-name (#41389)
- Spec.format: error on old style format strings (#41934)
- ASP-based solver:
- ASP-based solver:
- fix infinite recursion when computing concretization errors (#41061)
- don't error for type mismatch on preferences (#41138)
- don't emit spurious debug output (#41218)

View File

@@ -32,7 +32,7 @@
Spack is a multi-platform package manager that builds and installs
multiple versions and configurations of software. It works on Linux,
macOS, Windows, and many supercomputers. Spack is non-destructive: installing a
macOS, and many supercomputers. Spack is non-destructive: installing a
new version of a package does not break existing installations, so many
configurations of the same package can coexist.
@@ -46,18 +46,13 @@ See the
[Feature Overview](https://spack.readthedocs.io/en/latest/features.html)
for examples and highlights.
To install spack and your first package, make sure you have Python & Git.
To install spack and your first package, make sure you have Python.
Then:
$ git clone -c feature.manyFiles=true --depth=2 https://github.com/spack/spack.git
$ git clone -c feature.manyFiles=true https://github.com/spack/spack.git
$ cd spack/bin
$ ./spack install zlib
> [!TIP]
> `-c feature.manyFiles=true` improves git's performance on repositories with 1,000+ files.
>
> `--depth=2` prunes the git history to reduce the size of the Spack installation.
Documentation
----------------

View File

@@ -22,4 +22,4 @@
#
# This is compatible across platforms.
#
exec spack python "$@"
exec /usr/bin/env spack python "$@"

View File

@@ -188,27 +188,25 @@ if NOT "%_sp_args%"=="%_sp_args:--help=%" (
goto :end_switch
:case_load
if NOT defined _sp_args (
exit /B 0
)
:: If args contain --bat, or -h/--help: just execute.
if NOT "%_sp_args%"=="%_sp_args:--help=%" (
goto :default_case
) else if NOT "%_sp_args%"=="%_sp_args:-h=%" (
goto :default_case
) else if NOT "%_sp_args%"=="%_sp_args:--bat=%" (
goto :default_case
) else if NOT "%_sp_args%"=="%_sp_args:--list=%" (
goto :default_case
:: If args contain --sh, --csh, or -h/--help: just execute.
if defined _sp_args (
if NOT "%_sp_args%"=="%_sp_args:--help=%" (
goto :default_case
) else if NOT "%_sp_args%"=="%_sp_args:-h=%" (
goto :default_case
) else if NOT "%_sp_args%"=="%_sp_args:--bat=%" (
goto :default_case
)
)
for /f "tokens=* USEBACKQ" %%I in (
`python "%spack%" %_sp_flags% %_sp_subcommand% --bat %_sp_args%`
) do %%I
`python "%spack%" %_sp_flags% %_sp_subcommand% --bat %_sp_args%`) do %%I
goto :end_switch
:case_unload
goto :case_load
:default_case
python "%spack%" %_sp_flags% %_sp_subcommand% %_sp_args%
goto :end_switch

View File

@@ -1,11 +1,71 @@
@ECHO OFF
setlocal EnableDelayedExpansion
:: (c) 2021 Lawrence Livermore National Laboratory
:: To use this file independently of Spack's installer, execute this script in its directory, or add the
:: associated bin directory to your PATH. Invoke to launch Spack Shell.
::
:: source_dir/spack/bin/spack_cmd.bat
::
pushd %~dp0..
set SPACK_ROOT=%CD%
pushd %CD%\..
set spackinstdir=%CD%
popd
call "%~dp0..\share\spack\setup-env.bat"
pushd %SPACK_ROOT%
%comspec% /K
:: Check if Python is on the PATH
if not defined python_pf_ver (
(for /f "delims=" %%F in ('where python.exe') do (
set "python_pf_ver=%%F"
goto :found_python
) ) 2> NUL
)
:found_python
if not defined python_pf_ver (
:: If not, look for Python from the Spack installer
:get_builtin
(for /f "tokens=*" %%g in ('dir /b /a:d "!spackinstdir!\Python*"') do (
set "python_ver=%%g")) 2> NUL
if not defined python_ver (
echo Python was not found on your system.
echo Please install Python or add Python to your PATH.
) else (
set "py_path=!spackinstdir!\!python_ver!"
set "py_exe=!py_path!\python.exe"
)
goto :exitpoint
) else (
:: Python is already on the path
set "py_exe=!python_pf_ver!"
(for /F "tokens=* USEBACKQ" %%F in (
`"!py_exe!" --version`) do (set "output=%%F")) 2>NUL
if not "!output:Microsoft Store=!"=="!output!" goto :get_builtin
goto :exitpoint
)
:exitpoint
set "PATH=%SPACK_ROOT%\bin\;%PATH%"
if defined py_path (
set "PATH=%py_path%;%PATH%"
)
if defined py_exe (
"%py_exe%" "%SPACK_ROOT%\bin\haspywin.py"
)
set "EDITOR=notepad"
DOSKEY spacktivate=spack env activate $*
@echo **********************************************************************
@echo ** Spack Package Manager
@echo **********************************************************************
IF "%1"=="" GOTO CONTINUE
set
GOTO:EOF
:continue
set PROMPT=[spack] %PROMPT%
%comspec% /k

View File

@@ -9,15 +9,15 @@ bootstrap:
# may not be able to bootstrap all the software that Spack needs,
# depending on its type.
sources:
- name: github-actions-v0.6
metadata: $spack/share/spack/bootstrap/github-actions-v0.6
- name: github-actions-v0.5
- name: 'github-actions-v0.5'
metadata: $spack/share/spack/bootstrap/github-actions-v0.5
- name: spack-install
- name: 'github-actions-v0.4'
metadata: $spack/share/spack/bootstrap/github-actions-v0.4
- name: 'spack-install'
metadata: $spack/share/spack/bootstrap/spack-install
trusted:
# By default we trust bootstrapping from sources and from binaries
# produced on Github via the workflow
github-actions-v0.6: true
github-actions-v0.5: true
github-actions-v0.4: true
spack-install: true

View File

@@ -39,19 +39,11 @@ concretizer:
# Option to deal with possible duplicate nodes (i.e. different nodes from the same package) in the DAG.
duplicates:
# "none": allows a single node for any package in the DAG.
# "minimal": allows the duplication of 'build-tools' nodes only
# (e.g. py-setuptools, cmake etc.)
# "minimal": allows the duplication of 'build-tools' nodes only (e.g. py-setuptools, cmake etc.)
# "full" (experimental): allows separation of the entire build-tool stack (e.g. the entire "cmake" subDAG)
strategy: minimal
# Option to specify compatibility between operating systems for reuse of compilers and packages
# Specified as a key: [list] where the key is the os that is being targeted, and the list contains the OS's
# it can reuse. Note this is a directional compatibility so mutual compatibility between two OS's
# Option to specify compatiblity between operating systems for reuse of compilers and packages
# Specified as a key: [list] where the key is the os that is being targeted, and the list contains the OS's
# it can reuse. Note this is a directional compatibility so mutual compatibility between two OS's
# requires two entries i.e. os_compatible: {sonoma: [monterey], monterey: [sonoma]}
os_compatible: {}
# Option to specify whether to support splicing. Splicing allows for
# the relinking of concrete package dependencies in order to better
# reuse already built packages with ABI compatible dependencies
splice:
explicit: []
automatic: false

View File

@@ -19,7 +19,7 @@ config:
install_tree:
root: $spack/opt/spack
projections:
all: "{architecture.platform}/{architecture.target}/{name}-{version}-{hash}"
all: "{architecture}/{compiler.name}-{compiler.version}/{name}-{version}-{hash}"
# install_tree can include an optional padded length (int or boolean)
# default is False (do not pad)
# if padded_length is True, Spack will pad as close to the system max path
@@ -115,6 +115,12 @@ config:
suppress_gpg_warnings: false
# If set to true, Spack will attempt to build any compiler on the spec
# that is not already available. If set to False, Spack will only use
# compilers already configured in compilers.yaml
install_missing_compilers: false
# If set to true, Spack will always check checksums after downloading
# archives. If false, Spack skips the checksum step.
checksum: true
@@ -164,6 +170,23 @@ config:
# If set to true, Spack will use ccache to cache C compiles.
ccache: false
# The concretization algorithm to use in Spack. Options are:
#
# 'clingo': Uses a logic solver under the hood to solve DAGs with full
# backtracking and optimization for user preferences. Spack will
# try to bootstrap the logic solver, if not already available.
#
# 'original': Spack's original greedy, fixed-point concretizer. This
# algorithm can make decisions too early and will not backtrack
# sufficiently for many specs. This will soon be deprecated in
# favor of clingo.
#
# See `concretizer.yaml` for more settings you can fine-tune when
# using clingo.
concretizer: clingo
# How long to wait to lock the Spack installation database. This lock is used
# when Spack needs to manage its own package metadata and all operations are
# expected to complete within the default time limit. The timeout should

View File

@@ -0,0 +1,16 @@
# -------------------------------------------------------------------------
# This is the default configuration for Spack's module file generation.
#
# Settings here are versioned with Spack and are intended to provide
# sensible defaults out of the box. Spack maintainers should edit this
# file to keep it current.
#
# Users can override these settings by editing the following files.
#
# Per-spack-instance settings (overrides defaults):
# $SPACK_ROOT/etc/spack/modules.yaml
#
# Per-user settings (overrides default and site settings):
# ~/.spack/modules.yaml
# -------------------------------------------------------------------------
modules: {}

View File

@@ -15,11 +15,12 @@
# -------------------------------------------------------------------------
packages:
all:
compiler:
- apple-clang
- clang
- gcc
providers:
c: [apple-clang, llvm, gcc]
cxx: [apple-clang, llvm, gcc]
elf: [libelf]
fortran: [gcc]
fuse: [macfuse]
gl: [apple-gl]
glu: [apple-glu]

View File

@@ -0,0 +1,3 @@
packages:
iconv:
require: [libiconv]

View File

@@ -15,18 +15,16 @@
# -------------------------------------------------------------------------
packages:
all:
compiler: [gcc, clang, oneapi, xl, nag, fj, aocc]
providers:
awk: [gawk]
armci: [armcimpi]
blas: [openblas, amdblis]
c: [gcc, llvm, intel-oneapi-compilers, xl, aocc]
cxx: [gcc, llvm, intel-oneapi-compilers, xl, aocc]
D: [ldc]
daal: [intel-oneapi-daal]
elf: [elfutils]
fftw-api: [fftw, amdfftw]
flame: [libflame, amdlibflame]
fortran: [gcc, llvm]
fortran-rt: [gcc-runtime, intel-oneapi-runtime]
fuse: [libfuse]
gl: [glx, osmesa]
@@ -39,9 +37,9 @@ packages:
jpeg: [libjpeg-turbo, libjpeg]
lapack: [openblas, amdlibflame]
libc: [glibc, musl]
libgfortran: [gcc-runtime]
libgfortran: [ gcc-runtime ]
libglx: [mesa+glx]
libifcore: [intel-oneapi-runtime]
libifcore: [ intel-oneapi-runtime ]
libllvm: [llvm]
lua-lang: [lua, lua-luajit-openresty, lua-luajit]
luajit: [lua-luajit-openresty, lua-luajit]
@@ -63,7 +61,6 @@ packages:
tbb: [intel-tbb]
unwind: [libunwind]
uuid: [util-linux-uuid, libuuid]
wasi-sdk: [wasi-sdk-prebuilt]
xxd: [xxd-standalone, vim]
yacc: [bison, byacc]
ziglang: [zig]
@@ -71,13 +68,3 @@ packages:
permissions:
read: world
write: user
cray-mpich:
buildable: false
cray-mvapich2:
buildable: false
fujitsu-mpi:
buildable: false
hpcx-mpi:
buildable: false
spectrum-mpi:
buildable: false

View File

@@ -1,5 +1,6 @@
config:
locks: false
concretizer: clingo
build_stage::
- '$spack/.staging'
stage_name: '{name}-{version}-{hash:7}'

View File

@@ -15,8 +15,8 @@
# -------------------------------------------------------------------------
packages:
all:
compiler:
- msvc
providers:
c : [msvc]
cxx: [msvc]
mpi: [msmpi]
gl: [wgl]

View File

@@ -1175,17 +1175,6 @@ unspecified version, but packages can depend on other packages with
could depend on ``mpich@1.2:`` if it can only build with version
``1.2`` or higher of ``mpich``.
.. note:: Windows Spec Syntax Caveats
Windows has a few idiosyncrasies when it comes to the Spack spec syntax and the use of certain shells
Spack's spec dependency syntax uses the carat (``^``) character, however this is an escape string in CMD
so it must be escaped with an additional carat (i.e. ``^^``).
CMD also will attempt to interpret strings with ``=`` characters in them. Any spec including this symbol
must double quote the string.
Note: All of these issues are unique to CMD, they can be avoided by using Powershell.
For more context on these caveats see the related issues: `carat <https://github.com/spack/spack/issues/42833>`_ and `equals <https://github.com/spack/spack/issues/43348>`_
Below are more details about the specifiers that you can add to specs.
.. _version-specifier:
@@ -1359,10 +1348,6 @@ For example, for the ``stackstart`` variant:
mpileaks stackstart==4 # variant will be propagated to dependencies
mpileaks stackstart=4 # only mpileaks will have this variant value
Spack also allows variants to be propagated from a package that does
not have that variant.
^^^^^^^^^^^^^^
Compiler Flags
^^^^^^^^^^^^^^
@@ -1448,12 +1433,22 @@ the reserved keywords ``platform``, ``os`` and ``target``:
$ spack install libelf os=ubuntu18.04
$ spack install libelf target=broadwell
or together by using the reserved keyword ``arch``:
.. code-block:: console
$ spack install libelf arch=cray-CNL10-haswell
Normally users don't have to bother specifying the architecture if they
are installing software for their current host, as in that case the
values will be detected automatically. If you need fine-grained control
over which packages use which targets (or over *all* packages' default
target), see :ref:`package-preferences`.
.. admonition:: Cray machines
The situation is a little bit different for Cray machines and a detailed
explanation on how the architecture can be set on them can be found at :ref:`cray-support`
.. _support-for-microarchitectures:

View File

@@ -166,106 +166,3 @@ while `py-numpy` still needs an older version:
Up to Spack v0.20 ``duplicates:strategy:none`` was the default (and only) behavior. From Spack v0.21 the
default behavior is ``duplicates:strategy:minimal``.
--------
Splicing
--------
The ``splice`` key covers config attributes for splicing specs in the solver.
"Splicing" is a method for replacing a dependency with another spec
that provides the same package or virtual. There are two types of
splices, referring to different behaviors for shared dependencies
between the root spec and the new spec replacing a dependency:
"transitive" and "intransitive". A "transitive" splice is one that
resolves all conflicts by taking the dependency from the new node. An
"intransitive" splice is one that resolves all conflicts by taking the
dependency from the original root. From a theory perspective, hybrid
splices are possible but are not modeled by Spack.
All spliced specs retain a ``build_spec`` attribute that points to the
original Spec before any splice occurred. The ``build_spec`` for a
non-spliced spec is itself.
The figure below shows examples of transitive and intransitive splices:
.. figure:: images/splices.png
:align: center
The concretizer can be configured to explicitly splice particular
replacements for a target spec. Splicing will allow the user to make
use of generically built public binary caches, while swapping in
highly optimized local builds for performance critical components
and/or components that interact closely with the specific hardware
details of the system. The most prominent candidate for splicing is
MPI providers. MPI packages have relatively well-understood ABI
characteristics, and most High Performance Computing facilities deploy
highly optimized MPI packages tailored to their particular
hardware. The following config block configures Spack to replace
whatever MPI provider each spec was concretized to use with the
particular package of ``mpich`` with the hash that begins ``abcdef``.
.. code-block:: yaml
concretizer:
splice:
explicit:
- target: mpi
replacement: mpich/abcdef
transitive: false
.. warning::
When configuring an explicit splice, you as the user take on the
responsibility for ensuring ABI compatibility between the specs
matched by the target and the replacement you provide. If they are
not compatible, Spack will not warn you and your application will
fail to run.
The ``target`` field of an explicit splice can be any abstract
spec. The ``replacement`` field must be a spec that includes the hash
of a concrete spec, and the replacement must either be the same
package as the target, provide the virtual that is the target, or
provide a virtual that the target provides. The ``transitive`` field
is optional -- by default, splices will be transitive.
.. note::
With explicit splices configured, it is possible for Spack to
concretize to a spec that does not satisfy the input. For example,
with the config above ``hdf5 ^mvapich2`` will concretize to user
``mpich/abcdef`` instead of ``mvapich2`` as the MPI provider. Spack
will warn the user in this case, but will not fail the
concretization.
.. _automatic_splicing:
^^^^^^^^^^^^^^^^^^
Automatic Splicing
^^^^^^^^^^^^^^^^^^
The Spack solver can be configured to do automatic splicing for
ABI-compatible packages. Automatic splices are enabled in the concretizer
config section
.. code-block:: yaml
concretizer:
splice:
automatic: True
Packages can include ABI-compatibility information using the
``can_splice`` directive. See :ref:`the packaging
guide<abi_compatibility>` for instructions on specifying ABI
compatibility using the ``can_splice`` directive.
.. note::
The ``can_splice`` directive is experimental and may be changed in
future versions.
When automatic splicing is enabled, the concretizer will combine any
number of ABI-compatible specs if possible to reuse installed packages
and packages available from binary caches. The end result of these
specs is equivalent to a series of transitive/intransitive splices,
but the series may be non-obvious.

View File

@@ -147,15 +147,6 @@ example, the ``bash`` shell is used to run the ``autogen.sh`` script.
def autoreconf(self, spec, prefix):
which("bash")("autogen.sh")
If the ``package.py`` has build instructions in a separate
:ref:`builder class <multiple_build_systems>`, the signature for a phase changes slightly:
.. code-block:: python
class AutotoolsBuilder(AutotoolsBuilder):
def autoreconf(self, pkg, spec, prefix):
which("bash")("autogen.sh")
"""""""""""""""""""""""""""""""""""""""
patching configure or Makefile.in files
"""""""""""""""""""""""""""""""""""""""

View File

@@ -130,19 +130,14 @@ before or after a particular phase. For example, in ``perl``, we see:
@run_after("install")
def install_cpanm(self):
spec = self.spec
maker = make
cpan_dir = join_path("cpanm", "cpanm")
if sys.platform == "win32":
maker = nmake
cpan_dir = join_path(self.stage.source_path, cpan_dir)
cpan_dir = windows_sfn(cpan_dir)
if "+cpanm" in spec:
with working_dir(cpan_dir):
perl = spec["perl"].command
perl("Makefile.PL")
maker()
maker("install")
spec = self.spec
if spec.satisfies("+cpanm"):
with working_dir(join_path("cpanm", "cpanm")):
perl = spec["perl"].command
perl("Makefile.PL")
make()
make("install")
This extra step automatically installs ``cpanm`` in addition to the
base Perl installation.
@@ -181,14 +176,8 @@ In the ``perl`` package, we can see:
@run_after("build")
@on_package_attributes(run_tests=True)
def build_test(self):
if sys.platform == "win32":
win32_dir = os.path.join(self.stage.source_path, "win32")
win32_dir = windows_sfn(win32_dir)
with working_dir(win32_dir):
nmake("test", ignore_quotes=True)
else:
make("test")
def test(self):
make("test")
As you can guess, this runs ``make test`` *after* building the package,
if and only if testing is requested. Again, this is not specific to

View File

@@ -25,7 +25,7 @@ use Spack to build packages with the tools.
The Spack Python class ``IntelOneapiPackage`` is a base class that is
used by ``IntelOneapiCompilers``, ``IntelOneapiMkl``,
``IntelOneapiTbb`` and other classes to implement the oneAPI
packages. Search for ``oneAPI`` at `packages.spack.io <https://packages.spack.io>`_ for the full
packages. Search for ``oneAPI`` at `<packages.spack.io>`_ for the full
list of available oneAPI packages, or use::
spack list -d oneAPI

View File

@@ -49,14 +49,14 @@ following phases:
#. ``install`` - install the package
Package developers often add unit tests that can be invoked with
``scons test`` or ``scons check``. Spack provides a ``build_test`` method
``scons test`` or ``scons check``. Spack provides a ``test`` method
to handle this. Since we don't know which one the package developer
chose, the ``build_test`` method does nothing by default, but can be easily
chose, the ``test`` method does nothing by default, but can be easily
overridden like so:
.. code-block:: python
def build_test(self):
def test(self):
scons("check")

View File

@@ -11,8 +11,7 @@ Chaining Spack Installations (upstreams.yaml)
You can point your Spack installation to another installation to use any
packages that are installed there. To register the other Spack instance,
you can add it as an entry to ``upstreams.yaml`` at any of the
:ref:`configuration-scopes`:
you can add it as an entry to ``upstreams.yaml``:
.. code-block:: yaml
@@ -23,8 +22,7 @@ you can add it as an entry to ``upstreams.yaml`` at any of the
install_tree: /path/to/another/spack/opt/spack
``install_tree`` must point to the ``opt/spack`` directory inside of the
Spack base directory, or the location of the ``install_tree`` defined
in :ref:`config.yaml <config-yaml>`.
Spack base directory.
Once the upstream Spack instance has been added, ``spack find`` will
automatically check the upstream instance when querying installed packages,

View File

@@ -206,24 +206,17 @@ def setup(sphinx):
("py:class", "six.moves.urllib.parse.ParseResult"),
("py:class", "TextIO"),
("py:class", "hashlib._Hash"),
("py:class", "concurrent.futures._base.Executor"),
# Spack classes that are private and we don't want to expose
("py:class", "spack.provider_index._IndexBase"),
("py:class", "spack.repo._PrependFileLoader"),
("py:class", "spack.build_systems._checks.BuilderWithDefaults"),
("py:class", "spack.build_systems._checks.BaseBuilder"),
# Spack classes that intersphinx is unable to resolve
("py:class", "spack.version.StandardVersion"),
("py:class", "spack.spec.DependencySpec"),
("py:class", "spack.spec.ArchSpec"),
("py:class", "spack.spec.InstallStatus"),
("py:class", "spack.spec.SpecfileReaderBase"),
("py:class", "spack.install_test.Pb"),
("py:class", "spack.filesystem_view.SimpleFilesystemView"),
("py:class", "spack.traverse.EdgeAndDepth"),
("py:class", "archspec.cpu.microarchitecture.Microarchitecture"),
("py:class", "spack.compiler.CompilerCache"),
# TypeVar that is not handled correctly
("py:class", "llnl.util.lang.T"),
]
# The reST default role (used for this markup: `text`) to use for all documents.

View File

@@ -281,7 +281,7 @@ When spack queries for configuration parameters, it searches in
higher-precedence scopes first. So, settings in a higher-precedence file
can override those with the same key in a lower-precedence one. For
list-valued settings, Spack *prepends* higher-precedence settings to
lower-precedence settings. Completely ignoring lower-precedence configuration
lower-precedence settings. Completely ignoring higher-level configuration
options is supported with the ``::`` notation for keys (see
:ref:`config-overrides` below).
@@ -511,7 +511,6 @@ Spack understands over a dozen special variables. These are:
* ``$target_family``. The target family for the current host, as
detected by ArchSpec. E.g. ``x86_64`` or ``aarch64``.
* ``$date``: the current date in the format YYYY-MM-DD
* ``$spack_short_version``: the Spack version truncated to the first components.
Note that, as with shell variables, you can write these as ``$varname``

View File

@@ -203,9 +203,12 @@ The OS that are currently supported are summarized in the table below:
* - Ubuntu 24.04
- ``ubuntu:24.04``
- ``spack/ubuntu-noble``
* - CentOS Stream9
- ``quay.io/centos/centos:stream9``
- ``spack/centos-stream9``
* - CentOS 7
- ``centos:7``
- ``spack/centos7``
* - CentOS Stream
- ``quay.io/centos/centos:stream``
- ``spack/centos-stream``
* - openSUSE Leap
- ``opensuse/leap``
- ``spack/leap15``

View File

@@ -184,7 +184,7 @@ Style Tests
Spack uses `Flake8 <http://flake8.pycqa.org/en/latest/>`_ to test for
`PEP 8 <https://www.python.org/dev/peps/pep-0008/>`_ conformance and
`mypy <https://mypy.readthedocs.io/en/stable/>`_ for type checking. PEP 8 is
`mypy <https://mypy.readthedocs.io/en/stable/>` for type checking. PEP 8 is
a series of style guides for Python that provide suggestions for everything
from variable naming to indentation. In order to limit the number of PRs that
were mostly style changes, we decided to enforce PEP 8 conformance. Your PR
@@ -316,215 +316,6 @@ documentation tests to make sure there are no errors. Documentation changes can
in some obfuscated warning messages. If you don't understand what they mean, feel free
to ask when you submit your PR.
.. _spack-builders-and-pipelines:
^^^^^^^^^
GitLab CI
^^^^^^^^^
""""""""""""""""""
Build Cache Stacks
""""""""""""""""""
Spack welcomes the contribution of software stacks of interest to the community. These
stacks are used to test package recipes and generate publicly available build caches.
Spack uses GitLab CI for managing the orchestration of build jobs.
GitLab Entry Point
~~~~~~~~~~~~~~~~~~
Add stack entrypoint to the ``share/spack/gitlab/cloud_pipelines/.gitlab-ci.yml``. There
are two stages required for each new stack, the generation stage and the build stage.
The generate stage is defined using the job template ``.generate`` configured with
environment variables defining the name of the stack in ``SPACK_CI_STACK_NAME`` and the
platform (``SPACK_TARGET_PLATFORM``) and architecture (``SPACK_TARGET_ARCH``) configuration,
and the tags associated with the class of runners to build on.
.. note::
The ``SPACK_CI_STACK_NAME`` must match the name of the directory containing the
stacks ``spack.yaml``.
.. note::
The platform and architecture variables are specified in order to select the
correct configurations from the generic configurations used in Spack CI. The
configurations currently available are:
* ``.cray_rhel_zen4``
* ``.cray_sles_zen4``
* ``.darwin_aarch64``
* ``.darwin_x86_64``
* ``.linux_aarch64``
* ``.linux_icelake``
* ``.linux_neoverse_n1``
* ``.linux_neoverse_v1``
* ``.linux_neoverse_v2``
* ``.linux_power``
* ``.linux_skylake``
* ``.linux_x86_64``
* ``.linux_x86_64_v4``
New configurations can be added to accommodate new platforms and architectures.
The build stage is defined as a trigger job that consumes the GitLab CI pipeline generated in
the generate stage for this stack. Build stage jobs use the ``.build`` job template which
handles the basic configuration.
An example entry point for a new stack called ``my-super-cool-stack``
.. code-block:: yaml
.my-super-cool-stack:
extends: [ ".linux_x86_64_v3" ]
variables:
SPACK_CI_STACK_NAME: my-super-cool-stack
tags: [ "all", "tags", "your", "job", "needs"]
my-super-cool-stack-generate:
extends: [ ".generate", ".my-super-cool-stack" ]
image: my-super-cool-stack-image:0.0.1
my-super-cool-stack-build:
extends: [ ".build", ".my-super-cool-stack" ]
trigger:
include:
- artifact: jobs_scratch_dir/cloud-ci-pipeline.yml
job: my-super-cool-stack-generate
strategy: depend
needs:
- artifacts: True
job: my-super-cool-stack-generate
Stack Configuration
~~~~~~~~~~~~~~~~~~~
The stack configuration is a spack environment file with two additional sections added.
Stack configurations should be located in ``share/spack/gitlab/cloud_pipelines/stacks/<stack_name>/spack.yaml``.
The ``ci`` section is generally used to define stack specific mappings such as image or tags.
For more information on what can go into the ``ci`` section refer to the docs on pipelines.
The ``cdash`` section is used for defining where to upload the results of builds. Spack configures
most of the details for posting pipeline results to
`cdash.spack.io <https://cdash.spack.io/index.php?project=Spack+Testing>`_. The only
requirement in the stack configuration is to define a ``build-group`` that is unique,
this is usually the long name of the stack.
An example stack that builds ``zlib``.
.. code-block:: yaml
spack:
view: false
packages:
all:
require: ["%gcc", "target=x86_64_v3"]
specs:
- zlib
ci:
pipeline-gen
- build-job:
image: my-super-cool-stack-image:0.0.1
cdash:
build-group: My Super Cool Stack
.. note::
The ``image`` used in the ``*-generate`` job must match exactly the ``image`` used in the ``build-job``.
When the images do not match the build job may fail.
"""""""""""""""""""
Registering Runners
"""""""""""""""""""
Contributing computational resources to Spack's CI build farm is one way to help expand the
capabilities and offerings of the public Spack build caches. Currently, Spack utilizes linux runners
from AWS, Google, and the University of Oregon (UO).
Runners require three key peices:
* Runner Registration Token
* Accurate tags
* OIDC Authentication script
* GPG keys
Minimum GitLab Runner Version: ``16.1.0``
`Intallation instructions <https://docs.gitlab.com/runner/install/>`_
Registration Token
~~~~~~~~~~~~~~~~~~
The first step to contribute new runners is to open an issue in the `spack infrastructure <https://github.com/spack/spack-infrastructure/issues/new?assignees=&labels=runner-registration&projects=&template=runner_registration.yml>`_
project. This will be reported to the spack infrastructure team who will guide users through the process
of registering new runners for Spack CI.
The information needed to register a runner is the motivation for the new resources, a semi-detailed description of
the runner, and finallly the point of contact for maintaining the software on the runner.
The point of contact will then work with the infrastruture team to obtain runner registration token(s) for interacting with
with Spack's GitLab instance. Once the runner is active, this point of contact will also be responsible for updating the
GitLab runner software to keep pace with Spack's Gitlab.
Tagging
~~~~~~~
In the initial stages of runner registration it is important to **exclude** the special tag ``spack``. This will prevent
the new runner(s) from being picked up for production CI jobs while it is configured and evaluated. Once it is determined
that the runner is ready for production use the ``spack`` tag will be added.
Because gitlab has no concept of tag exclustion, runners that provide specialized resource also require specialized tags.
For example, a basic CPU only x86_64 runner may have a tag ``x86_64`` associated with it. However, a runner containing an
CUDA capable GPU may have the tag ``x86_64-cuda`` to denote that it should only be used for packages that will benefit from
a CUDA capable resource.
OIDC
~~~~
Spack runners use OIDC authentication for connecting to the appropriate AWS bucket
which is used for coordinating the communication of binaries between build jobs. In
order to configure OIDC authentication, Spack CI runners use a python script with minimal
dependencies. This script can be configured for runners as seen here using the ``pre_build_script``.
.. code-block:: toml
[[runners]]
pre_build_script = """
echo 'Executing Spack pre-build setup script'
for cmd in "${PY3:-}" python3 python; do
if command -v > /dev/null "$cmd"; then
export PY3="$(command -v "$cmd")"
break
fi
done
if [ -z "${PY3:-}" ]; then
echo "Unable to find python3 executable"
exit 1
fi
$PY3 -c "import urllib.request;urllib.request.urlretrieve('https://raw.githubusercontent.com/spack/spack-infrastructure/main/scripts/gitlab_runner_pre_build/pre_build.py', 'pre_build.py')"
$PY3 pre_build.py > envvars
. ./envvars
rm -f envvars
unset GITLAB_OIDC_TOKEN
"""
GPG Keys
~~~~~~~~
Runners that may be utilized for ``protected`` CI require the registration of an intermediate signing key that
can be used to sign packages. For more information on package signing read :ref:`key_architecture`.
--------
Coverage
--------

View File

@@ -181,6 +181,10 @@ Spec-related modules
:mod:`spack.parser`
Contains :class:`~spack.parser.SpecParser` and functions related to parsing specs.
:mod:`spack.concretize`
Contains :class:`~spack.concretize.Concretizer` implementation,
which allows site administrators to change Spack's :ref:`concretization-policies`.
:mod:`spack.version`
Implements a simple :class:`~spack.version.Version` class with simple
comparison semantics. Also implements :class:`~spack.version.VersionRange`
@@ -333,9 +337,13 @@ inserting them at different places in the spack code base. Whenever a hook
type triggers by way of a function call, we find all the hooks of that type,
and run them.
Spack defines hooks by way of a module in the ``lib/spack/spack/hooks`` directory.
This module has to be registered in ``__init__.py`` so that Spack is aware of it.
This section will cover the basic kind of hooks, and how to write them.
Spack defines hooks by way of a module at ``lib/spack/spack/hooks`` where we can define
types of hooks in the ``__init__.py``, and then python files in that folder
can use hook functions. The files are automatically parsed, so if you write
a new file for some integration (e.g., ``lib/spack/spack/hooks/myintegration.py``
you can then write hook functions in that file that will be automatically detected,
and run whenever your hook is called. This section will cover the basic kind
of hooks, and how to write them.
^^^^^^^^^^^^^^
Types of Hooks
@@ -708,27 +716,27 @@ Release branches
^^^^^^^^^^^^^^^^
There are currently two types of Spack releases: :ref:`major releases
<major-releases>` (``0.21.0``, ``0.22.0``, etc.) and :ref:`patch releases
<patch-releases>` (``0.22.1``, ``0.22.2``, ``0.22.3``, etc.). Here is a
<major-releases>` (``0.17.0``, ``0.18.0``, etc.) and :ref:`point releases
<point-releases>` (``0.17.1``, ``0.17.2``, ``0.17.3``, etc.). Here is a
diagram of how Spack release branches work::
o branch: develop (latest version, v0.23.0.dev0)
o branch: develop (latest version, v0.19.0.dev0)
|
o
| o branch: releases/v0.22, tag: v0.22.1
| o branch: releases/v0.18, tag: v0.18.1
o |
| o tag: v0.22.0
| o tag: v0.18.0
o |
| o
|/
o
|
o
| o branch: releases/v0.21, tag: v0.21.2
| o branch: releases/v0.17, tag: v0.17.2
o |
| o tag: v0.21.1
| o tag: v0.17.1
o |
| o tag: v0.21.0
| o tag: v0.17.0
o |
| o
|/
@@ -739,8 +747,8 @@ requests target ``develop``. The ``develop`` branch will report that its
version is that of the next **major** release with a ``.dev0`` suffix.
Each Spack release series also has a corresponding branch, e.g.
``releases/v0.22`` has ``v0.22.x`` versions of Spack, and
``releases/v0.21`` has ``v0.21.x`` versions. A major release is the first
``releases/v0.18`` has ``0.18.x`` versions of Spack, and
``releases/v0.17`` has ``0.17.x`` versions. A major release is the first
tagged version on a release branch. Minor releases are back-ported from
develop onto release branches. This is typically done by cherry-picking
bugfix commits off of ``develop``.
@@ -770,40 +778,27 @@ for more details.
Scheduling work for releases
^^^^^^^^^^^^^^^^^^^^^^^^^^^^
We schedule work for **major releases** through `milestones
<https://github.com/spack/spack/milestones>`_ and `GitHub Projects
<https://github.com/spack/spack/projects>`_, while **patch releases** use `labels
<https://github.com/spack/spack/labels>`_.
We schedule work for releases by creating `GitHub projects
<https://github.com/spack/spack/projects>`_. At any time, there may be
several open release projects. For example, below are two releases (from
some past version of the page linked above):
There is only one milestone open at a time. Its name corresponds to the next major version, for
example ``v0.23``. Important issues and pull requests should be assigned to this milestone by
core developers, so that they are not forgotten at the time of release. The milestone is closed
when the release is made, and a new milestone is created for the next major release.
.. image:: images/projects.png
Bug reports in GitHub issues are automatically labelled ``bug`` and ``triage``. Spack developers
assign one of the labels ``impact-low``, ``impact-medium`` or ``impact-high``. This will make the
issue appear in the `Triaged bugs <https://github.com/orgs/spack/projects/6>`_ project board.
Important issues should be assigned to the next milestone as well, so they appear at the top of
the project board.
This image shows one release in progress for ``0.15.1`` and another for
``0.16.0``. Each of these releases has a project board containing issues
and pull requests. GitHub shows a status bar with completed work in
green, work in progress in purple, and work not started yet in gray, so
it's fairly easy to see progress.
Spack's milestones are not firm commitments so we move work between releases frequently. If we
need to make a release and some tasks are not yet done, we will simply move them to the next major
release milestone, rather than delaying the release to complete them.
Spack's project boards are not firm commitments so we move work between
releases frequently. If we need to make a release and some tasks are not
yet done, we will simply move them to the next minor or major release, rather
than delaying the release to complete them.
^^^^^^^^^^^^^^^^^^^^^
Backporting bug fixes
^^^^^^^^^^^^^^^^^^^^^
For more on using GitHub project boards, see `GitHub's documentation
<https://docs.github.com/en/github/managing-your-work-on-github/about-project-boards>`_.
When a bug is fixed in the ``develop`` branch, it is often necessary to backport the fix to one
(or more) of the ``release/vX.Y`` branches. Only the release manager is responsible for doing
backports, but Spack maintainers are responsible for labelling pull requests (and issues if no bug
fix is available yet) with ``vX.Y.Z`` labels. The label should correspond to the next patch version
that the bug fix should be backported to.
Backports are done publicly by the release manager using a pull request named ``Backports vX.Y.Z``.
This pull request is opened from the ``backports/vX.Y.Z`` branch, targets the ``releases/vX.Y``
branch and contains a (growing) list of cherry-picked commits from the ``develop`` branch.
Typically there are one or two backport pull requests open at any given time.
.. _major-releases:
@@ -811,21 +806,25 @@ Typically there are one or two backport pull requests open at any given time.
Making major releases
^^^^^^^^^^^^^^^^^^^^^
Assuming all required work from the milestone is completed, the steps to make the major release
are:
Assuming a project board has already been created and all required work
completed, the steps to make the major release are:
#. `Create a new milestone <https://github.com/spack/spack/milestones>`_ for the next major
release.
#. Create two new project boards:
#. `Create a new label <https://github.com/spack/spack/labels>`_ for the next patch release.
* One for the next major release
* One for the next point release
#. Move any optional tasks that are not done to the next milestone.
#. Move any optional tasks that are not done to one of the new project boards.
In general, small bugfixes should go to the next point release. Major
features, refactors, and changes that could affect concretization should
go in the next major release.
#. Create a branch for the release, based on ``develop``:
.. code-block:: console
$ git checkout -b releases/v0.23 develop
$ git checkout -b releases/v0.15 develop
For a version ``vX.Y.Z``, the branch's name should be
``releases/vX.Y``. That is, you should create a ``releases/vX.Y``
@@ -861,8 +860,8 @@ are:
Create a pull request targeting the ``develop`` branch, bumping the major
version in ``lib/spack/spack/__init__.py`` with a ``dev0`` release segment.
For instance when you have just released ``v0.23.0``, set the version
to ``(0, 24, 0, 'dev0')`` on ``develop``.
For instance when you have just released ``v0.15.0``, set the version
to ``(0, 16, 0, 'dev0')`` on ``develop``.
#. Follow the steps in :ref:`publishing-releases`.
@@ -871,52 +870,82 @@ are:
#. Follow the steps in :ref:`announcing-releases`.
.. _patch-releases:
.. _point-releases:
^^^^^^^^^^^^^^^^^^^^^
Making patch releases
Making point releases
^^^^^^^^^^^^^^^^^^^^^
To make the patch release process both efficient and transparent, we use a *backports pull request*
which contains cherry-picked commits from the ``develop`` branch. The majority of the work is to
cherry-pick the bug fixes, which ideally should be done as soon as they land on ``develop``:
this ensures cherry-picking happens in order, and makes conflicts easier to resolve since the
changes are fresh in the mind of the developer.
Assuming a project board has already been created and all required work
completed, the steps to make the point release are:
The backports pull request is always titled ``Backports vX.Y.Z`` and is labelled ``backports``. It
is opened from a branch named ``backports/vX.Y.Z`` and targets the ``releases/vX.Y`` branch.
#. Create a new project board for the next point release.
Whenever a pull request labelled ``vX.Y.Z`` is merged, cherry-pick the associated squashed commit
on ``develop`` to the ``backports/vX.Y.Z`` branch. For pull requests that were rebased (or not
squashed), cherry-pick each associated commit individually. Never force push to the
``backports/vX.Y.Z`` branch.
#. Move any optional tasks that are not done to the next project board.
.. warning::
#. Check out the release branch (it should already exist).
Sometimes you may **still** get merge conflicts even if you have
cherry-picked all the commits in order. This generally means there
is some other intervening pull request that the one you're trying
to pick depends on. In these cases, you'll need to make a judgment
call regarding those pull requests. Consider the number of affected
files and/or the resulting differences.
For the ``X.Y.Z`` release, the release branch is called ``releases/vX.Y``.
For ``v0.15.1``, you would check out ``releases/v0.15``:
1. If the changes are small, you might just cherry-pick it.
.. code-block:: console
2. If the changes are large, then you may decide that this fix is not
worth including in a patch release, in which case you should remove
the label from the pull request. Remember that large, manual backports
are seldom the right choice for a patch release.
$ git checkout releases/v0.15
When all commits are cherry-picked in the ``backports/vX.Y.Z`` branch, make the patch
release as follows:
#. If a pull request to the release branch named ``Backports vX.Y.Z`` is not already
in the project, create it. This pull request ought to be created as early as
possible when working on a release project, so that we can build the release
commits incrementally, and identify potential conflicts at an early stage.
#. `Create a new label <https://github.com/spack/spack/labels>`_ ``vX.Y.{Z+1}`` for the next patch
release.
#. Cherry-pick each pull request in the ``Done`` column of the release
project board onto the ``Backports vX.Y.Z`` pull request.
#. Replace the label ``vX.Y.Z`` with ``vX.Y.{Z+1}`` for all PRs and issues that are not done.
This is **usually** fairly simple since we squash the commits from the
vast majority of pull requests. That means there is only one commit
per pull request to cherry-pick. For example, `this pull request
<https://github.com/spack/spack/pull/15777>`_ has three commits, but
they were squashed into a single commit on merge. You can see the
commit that was created here:
#. Manually push a single commit with commit message ``Set version to vX.Y.Z`` to the
``backports/vX.Y.Z`` branch, that both bumps the Spack version number and updates the changelog:
.. image:: images/pr-commit.png
You can easily cherry pick it like this (assuming you already have the
release branch checked out):
.. code-block:: console
$ git cherry-pick 7e46da7
For pull requests that were rebased (or not squashed), you'll need to
cherry-pick each associated commit individually.
.. warning::
It is important to cherry-pick commits in the order they happened,
otherwise you can get conflicts while cherry-picking. When
cherry-picking look at the merge date,
**not** the number of the pull request or the date it was opened.
Sometimes you may **still** get merge conflicts even if you have
cherry-picked all the commits in order. This generally means there
is some other intervening pull request that the one you're trying
to pick depends on. In these cases, you'll need to make a judgment
call regarding those pull requests. Consider the number of affected
files and or the resulting differences.
1. If the dependency changes are small, you might just cherry-pick it,
too. If you do this, add the task to the release board.
2. If the changes are large, then you may decide that this fix is not
worth including in a point release, in which case you should remove
the task from the release project.
3. You can always decide to manually back-port the fix to the release
branch if neither of the above options makes sense, but this can
require a lot of work. It's seldom the right choice.
#. When all the commits from the project board are cherry-picked into
the ``Backports vX.Y.Z`` pull request, you can push a commit to:
1. Bump the version in ``lib/spack/spack/__init__.py``.
2. Update ``CHANGELOG.md`` with a list of the changes.
@@ -925,22 +954,20 @@ release as follows:
release branch. See `the changelog from 0.14.1
<https://github.com/spack/spack/commit/ff0abb9838121522321df2a054d18e54b566b44a>`_.
#. Make sure CI passes on the **backports pull request**, including:
#. Merge the ``Backports vX.Y.Z`` PR with the **Rebase and merge** strategy. This
is needed to keep track in the release branch of all the commits that were
cherry-picked.
#. Make sure CI passes on the release branch, including:
* Regular unit tests
* Build tests
* The E4S pipeline at `gitlab.spack.io <https://gitlab.spack.io>`_
#. Merge the ``Backports vX.Y.Z`` PR with the **Rebase and merge** strategy. This
is needed to keep track in the release branch of all the commits that were
cherry-picked.
#. Make sure CI passes on the last commit of the **release branch**.
#. In the rare case you need to include additional commits in the patch release after the backports
PR is merged, it is best to delete the last commit ``Set version to vX.Y.Z`` from the release
branch with a single force push, open a new backports PR named ``Backports vX.Y.Z (2)``, and
repeat the process. Avoid repeated force pushes to the release branch.
If CI does not pass, you'll need to figure out why, and make changes
to the release branch until it does. You can make more commits, modify
or remove cherry-picked commits, or cherry-pick **more** from
``develop`` to make this happen.
#. Follow the steps in :ref:`publishing-releases`.
@@ -1015,31 +1042,25 @@ Updating `releases/latest`
If the new release is the **highest** Spack release yet, you should
also tag it as ``releases/latest``. For example, suppose the highest
release is currently ``0.22.3``:
release is currently ``0.15.3``:
* If you are releasing ``0.22.4`` or ``0.23.0``, then you should tag
it with ``releases/latest``, as these are higher than ``0.22.3``.
* If you are releasing ``0.15.4`` or ``0.16.0``, then you should tag
it with ``releases/latest``, as these are higher than ``0.15.3``.
* If you are making a new release of an **older** major version of
Spack, e.g. ``0.21.4``, then you should not tag it as
Spack, e.g. ``0.14.4``, then you should not tag it as
``releases/latest`` (as there are newer major versions).
To do so, first fetch the latest tag created on GitHub, since you may not have it locally:
To tag ``releases/latest``, do this:
.. code-block:: console
$ git fetch --force git@github.com:spack/spack vX.Y.Z
$ git checkout releases/vX.Y # vX.Y is the new release's branch
$ git tag --force releases/latest
$ git push --force --tags
Then tag ``vX.Y.Z`` as ``releases/latest`` and push the individual tag to GitHub.
.. code-block:: console
$ git tag --force releases/latest vX.Y.Z
$ git push --force git@github.com:spack/spack releases/latest
The ``--force`` argument to ``git tag`` makes ``git`` overwrite the existing ``releases/latest``
tag with the new one. Do **not** use the ``--tags`` flag when pushing, since this will push *all*
local tags.
The ``--force`` argument to ``git tag`` makes ``git`` overwrite the existing
``releases/latest`` tag with the new one.
.. _announcing-releases:

View File

@@ -5,56 +5,49 @@
.. _environments:
=====================================
Environments (spack.yaml, spack.lock)
=====================================
=========================
Environments (spack.yaml)
=========================
An environment is used to group a set of specs intended for some purpose
to be built, rebuilt, and deployed in a coherent fashion. Environments
define aspects of the installation of the software, such as:
An environment is used to group together a set of specs for the
purpose of building, rebuilding and deploying in a coherent fashion.
Environments provide a number of advantages over the *à la carte*
approach of building and loading individual Spack modules:
#. *which* specs to install;
#. *how* those specs are configured; and
#. *where* the concretized software will be installed.
Aggregating this information into an environment for processing has advantages
over the *à la carte* approach of building and loading individual Spack modules.
With environments, you concretize, install, or load (activate) all of the
specs with a single command. Concretization fully configures the specs
and dependencies of the environment in preparation for installing the
software. This is a more robust solution than ad-hoc installation scripts.
And you can share an environment or even re-use it on a different computer.
Environment definitions, especially *how* specs are configured, allow the
software to remain stable and repeatable even when Spack packages are upgraded. Changes are only picked up when the environment is explicitly re-concretized.
Defining *where* specs are installed supports a filesystem view of the
environment. Yet Spack maintains a single installation of the software that
can be re-used across multiple environments.
Activating an environment determines *when* all of the associated (and
installed) specs are loaded so limits the software loaded to those specs
actually needed by the environment. Spack can even generate a script to
load all modules related to an environment.
#. Environments separate the steps of (a) choosing what to
install, (b) concretizing, and (c) installing. This allows
Environments to remain stable and repeatable, even if Spack packages
are upgraded: specs are only re-concretized when the user
explicitly asks for it. It is even possible to reliably
transport environments between different computers running
different versions of Spack!
#. Environments allow several specs to be built at once; a more robust
solution than ad-hoc scripts making multiple calls to ``spack
install``.
#. An Environment that is built as a whole can be loaded as a whole
into the user environment. An Environment can be built to maintain
a filesystem view of its packages, and the environment can load
that view into the user environment at activation time. Spack can
also generate a script to load all modules related to an
environment.
Other packaging systems also provide environments that are similar in
some ways to Spack environments; for example, `Conda environments
<https://conda.io/docs/user-guide/tasks/manage-environments.html>`_ or
`Python Virtual Environments
<https://docs.python.org/3/tutorial/venv.html>`_. Spack environments
provide some distinctive features though:
provide some distinctive features:
#. A spec installed "in" an environment is no different from the same
spec installed anywhere else in Spack.
#. Spack environments may contain more than one spec of the same
spec installed anywhere else in Spack. Environments are assembled
simply by collecting together a set of specs.
#. Spack Environments may contain more than one spec of the same
package.
Spack uses a "manifest and lock" model similar to `Bundler gemfiles
<https://bundler.io/man/gemfile.5.html>`_ and other package managers.
The environment's user input file (or manifest), is named ``spack.yaml``.
The lock file, which contains the fully configured and concretized specs,
is named ``spack.lock``.
<https://bundler.io/man/gemfile.5.html>`_ and other package
managers. The user input file is named ``spack.yaml`` and the lock
file is named ``spack.lock``
.. _environments-using:
@@ -75,60 +68,55 @@ An environment is created by:
$ spack env create myenv
The directory ``$SPACK_ROOT/var/spack/environments/myenv`` is created
to manage the environment.
Spack then creates the directory ``var/spack/environments/myenv``.
.. note::
All managed environments by default are stored in the
``$SPACK_ROOT/var/spack/environments`` folder. This location can be changed
by setting the ``environments_root`` variable in ``config.yaml``.
All managed environments by default are stored in the ``var/spack/environments`` folder.
This location can be changed by setting the ``environments_root`` variable in ``config.yaml``.
Spack creates the file ``spack.yaml``, hidden directory ``.spack-env``, and
``spack.lock`` file under ``$SPACK_ROOT/var/spack/environments/myenv``. User
interaction occurs through the ``spack.yaml`` file and the Spack commands
that affect it. Metadata and, by default, the view are stored in the
``.spack-env`` directory. When the environment is concretized, Spack creates
the ``spack.lock`` file with the fully configured specs and dependencies for
In the ``var/spack/environments/myenv`` directory, Spack creates the
file ``spack.yaml`` and the hidden directory ``.spack-env``.
Spack stores metadata in the ``.spack-env`` directory. User
interaction will occur through the ``spack.yaml`` file and the Spack
commands that affect it. When the environment is concretized, Spack
will create a file ``spack.lock`` with the concrete information for
the environment.
The ``.spack-env`` subdirectory also contains:
In addition to being the default location for the view associated with
an Environment, the ``.spack-env`` directory also contains:
* ``repo/``: A subdirectory acting as the repo consisting of the Spack
packages used in the environment. It allows the environment to build
the same, in theory, even on different versions of Spack with different
* ``repo/``: A repo consisting of the Spack packages used in this
environment. This allows the environment to build the same, in
theory, even on different versions of Spack with different
packages!
* ``logs/``: A subdirectory containing the build logs for the packages
in this environment.
* ``logs/``: A directory containing the build logs for the packages
in this Environment.
Spack Environments can also be created from either the user input, or
manifest, file or the lockfile. Create an environment from a manifest using:
Spack Environments can also be created from either a manifest file
(usually but not necessarily named, ``spack.yaml``) or a lockfile.
To create an Environment from a manifest:
.. code-block:: console
$ spack env create myenv spack.yaml
The resulting environment is guaranteed to have the same root specs as
the original but may concretize differently in the presence of different
explicit or default configuration settings (e.g., a different version of
Spack or for a different user account).
Create an environment from a ``spack.lock`` file using:
To create an Environment from a ``spack.lock`` lockfile:
.. code-block:: console
$ spack env create myenv spack.lock
The resulting environment, when on the same or a compatible machine, is
guaranteed to initially have the same concrete specs as the original.
Either of these commands can also take a full path to the
initialization file.
.. note::
Environment creation also accepts a full path to the file.
If the path is not under the ``$SPACK_ROOT/var/spack/environments``
directory then the source is referred to as an
:ref:`independent environment <independent_environments>`.
A Spack Environment created from a ``spack.yaml`` manifest is
guaranteed to have the same root specs as the original Environment,
but may concretize differently. A Spack Environment created from a
``spack.lock`` lockfile is guaranteed to have the same concrete specs
as the original Environment. Either may obviously then differ as the
user modifies it.
^^^^^^^^^^^^^^^^^^^^^^^^^
Activating an Environment
@@ -141,7 +129,7 @@ To activate an environment, use the following command:
$ spack env activate myenv
By default, the ``spack env activate`` will load the view associated
with the environment into the user environment. The ``-v,
with the Environment into the user environment. The ``-v,
--with-view`` argument ensures this behavior, and the ``-V,
--without-view`` argument activates the environment without changing
the user environment variables.
@@ -154,11 +142,8 @@ user's prompt to begin with the environment name in brackets.
$ spack env activate -p myenv
[myenv] $ ...
The ``activate`` command can also be used to create a new environment, if it is
not already defined, by adding the ``--create`` flag. Managed and independent
environments can both be created using the same flags that `spack env create`
accepts. If an environment already exists then spack will simply activate it
and ignore the create-specific flags.
The ``activate`` command can also be used to create a new environment if it does not already
exist.
.. code-block:: console
@@ -183,50 +168,49 @@ or the shortcut alias
If the environment was activated with its view, deactivating the
environment will remove the view from the user environment.
.. _independent_environments:
^^^^^^^^^^^^^^^^^^^^^^
Anonymous Environments
^^^^^^^^^^^^^^^^^^^^^^
^^^^^^^^^^^^^^^^^^^^^^^^
Independent Environments
^^^^^^^^^^^^^^^^^^^^^^^^
Apart from managed environments, Spack also supports anonymous environments.
Independent environments can be located in any directory outside of Spack.
Anonymous environments can be placed in any directory of choice.
.. note::
When uninstalling packages, Spack asks the user to confirm the removal of packages
that are still used in a managed environment. This is not the case for independent
that are still used in a managed environment. This is not the case for anonymous
environments.
To create an independent environment, use one of the following commands:
To create an anonymous environment, use one of the following commands:
.. code-block:: console
$ spack env create --dir my_env
$ spack env create ./my_env
As a shorthand, you can also create an independent environment upon activation if it does not
As a shorthand, you can also create an anonymous environment upon activation if it does not
already exist:
.. code-block:: console
$ spack env activate --create ./my_env
For convenience, Spack can also place an independent environment in a temporary directory for you:
For convenience, Spack can also place an anonymous environment in a temporary directory for you:
.. code-block:: console
$ spack env activate --temp
^^^^^^^^^^^^^^^^^^^^^^^^^^
Environment-Aware Commands
^^^^^^^^^^^^^^^^^^^^^^^^^^
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Environment Sensitive Commands
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Spack commands are environment-aware. For example, the ``find``
command shows only the specs in the active environment if an
environment has been activated. Otherwise it shows all specs in
the Spack instance. The same rule applies to the ``install`` and
``uninstall`` commands.
Spack commands are environment sensitive. For example, the ``find``
command shows only the specs in the active Environment if an
Environment has been activated. Similarly, the ``install`` and
``uninstall`` commands act on the active environment.
.. code-block:: console
@@ -271,33 +255,32 @@ the Spack instance. The same rule applies to the ``install`` and
Note that when we installed the abstract spec ``zlib@1.2.8``, it was
presented as a root of the environment. All explicitly installed
packages will be listed as roots of the environment.
presented as a root of the Environment. All explicitly installed
packages will be listed as roots of the Environment.
All of the Spack commands that act on the list of installed specs are
environment-aware in this way, including ``install``,
``uninstall``, ``find``, ``extensions``, etcetera. In the
Environment-sensitive in this way, including ``install``,
``uninstall``, ``find``, ``extensions``, and more. In the
:ref:`environment-configuration` section we will discuss
environment-aware commands further.
Environment-sensitive commands further.
^^^^^^^^^^^^^^^^^^^^^
Adding Abstract Specs
^^^^^^^^^^^^^^^^^^^^^
An abstract spec is the user-specified spec before Spack applies
defaults or dependency information.
An abstract spec is the user-specified spec before Spack has applied
any defaults or dependency information.
Users can add abstract specs to an environment using the ``spack add``
command. The most important component of an environment is a list of
Users can add abstract specs to an Environment using the ``spack add``
command. The most important component of an Environment is a list of
abstract specs.
Adding a spec adds it as a root spec of the environment in the user
input file (``spack.yaml``). It does not affect the concrete specs
in the lock file (``spack.lock``) and it does not install the spec.
Adding a spec adds to the manifest (the ``spack.yaml`` file), which is
used to define the roots of the Environment, but does not affect the
concrete specs in the lockfile, nor does it install the spec.
The ``spack add`` command is environment-aware. It adds the spec to the
currently active environment. An error is generated if there isn't an
active environment. All environment-aware commands can also
The ``spack add`` command is environment aware. It adds to the
currently active environment. All environment aware commands can also
be called using the ``spack -e`` flag to specify the environment.
.. code-block:: console
@@ -317,11 +300,11 @@ or
Concretizing
^^^^^^^^^^^^
Once user specs have been added to an environment, they can be concretized.
There are three different modes of operation to concretize an environment,
explained in detail in :ref:`environments_concretization_config`.
Regardless of which mode of operation is chosen, the following
command will ensure all of the root specs are concretized according to the
Once some user specs have been added to an environment, they can be concretized.
There are at the moment three different modes of operation to concretize an environment,
which are explained in details in :ref:`environments_concretization_config`.
Regardless of which mode of operation has been chosen, the following
command will ensure all the root specs are concretized according to the
constraints that are prescribed in the configuration:
.. code-block:: console
@@ -330,15 +313,16 @@ constraints that are prescribed in the configuration:
In the case of specs that are not concretized together, the command
above will concretize only the specs that were added and not yet
concretized. Forcing a re-concretization of all of the specs can be done
by adding the ``-f`` option:
concretized. Forcing a re-concretization of all the specs can be done
instead with this command:
.. code-block:: console
[myenv]$ spack concretize -f
Without the option, Spack guarantees that already concretized specs are
unchanged in the environment.
When the ``-f`` flag is not used to reconcretize all specs, Spack
guarantees that already concretized specs are unchanged in the
environment.
The ``concretize`` command does not install any packages. For packages
that have already been installed outside of the environment, the
@@ -371,16 +355,16 @@ installed specs using the ``-c`` (``--concretized``) flag.
Installing an Environment
^^^^^^^^^^^^^^^^^^^^^^^^^
In addition to adding individual specs to an environment, one
can install the entire environment at once using the command
In addition to installing individual specs into an Environment, one
can install the entire Environment at once using the command
.. code-block:: console
[myenv]$ spack install
If the environment has been concretized, Spack will install the
concretized specs. Otherwise, ``spack install`` will concretize
the environment before installing the concretized specs.
If the Environment has been concretized, Spack will install the
concretized specs. Otherwise, ``spack install`` will first concretize
the Environment and then install the concretized specs.
.. note::
@@ -401,17 +385,17 @@ the environment before installing the concretized specs.
As it installs, ``spack install`` creates symbolic links in the
``logs/`` directory in the environment, allowing for easy inspection
``logs/`` directory in the Environment, allowing for easy inspection
of build logs related to that environment. The ``spack install``
command also stores a Spack repo containing the ``package.py`` file
used at install time for each package in the ``repos/`` directory in
the environment.
the Environment.
The ``--no-add`` option can be used in a concrete environment to tell
spack to install specs already present in the environment but not to
add any new root specs to the environment. For root specs provided
to ``spack install`` on the command line, ``--no-add`` is the default,
while for dependency specs, it is optional. In other
while for dependency specs on the other hand, it is optional. In other
words, if there is an unambiguous match in the active concrete environment
for a root spec provided to ``spack install`` on the command line, spack
does not require you to specify the ``--no-add`` option to prevent the spec
@@ -425,22 +409,12 @@ Developing Packages in a Spack Environment
The ``spack develop`` command allows one to develop Spack packages in
an environment. It requires a spec containing a concrete version, and
will configure Spack to install the package from local source.
If a version is not provided from the command line interface then spack
will automatically pick the highest version the package has defined.
This means any infinity versions (``develop``, ``main``, ``stable``) will be
preferred in this selection process.
By default, ``spack develop`` will also clone the package to a subdirectory in the
environment for the local source. This package will have a special variant ``dev_path``
will configure Spack to install the package from local source. By
default, it will also clone the package to a subdirectory in the
environment. This package will have a special variant ``dev_path``
set, and Spack will ensure the package and its dependents are rebuilt
any time the environment is installed if the package's local source
code has been modified. Spack's native implementation to check for modifications
is to check if ``mtime`` is newer than the installation.
A custom check can be created by overriding the ``detect_dev_src_change`` method
in your package class. This is particularly useful for projects using custom spack repo's
to drive development and want to optimize performance.
Spack ensures that all instances of a
code has been modified. Spack ensures that all instances of a
developed package in the environment are concretized to match the
version (and other constraints) passed as the spec argument to the
``spack develop`` command.
@@ -450,7 +424,7 @@ also be used as valid concrete versions (see :ref:`version-specifier`).
This means that for a package ``foo``, ``spack develop foo@git.main`` will clone
the ``main`` branch of the package, and ``spack install`` will install from
that git clone if ``foo`` is in the environment.
Further development on ``foo`` can be tested by re-installing the environment,
Further development on ``foo`` can be tested by reinstalling the environment,
and eventually committed and pushed to the upstream git repo.
If the package being developed supports out-of-source builds then users can use the
@@ -635,7 +609,7 @@ manipulate configuration inline in the ``spack.yaml`` file.
Inline configurations
^^^^^^^^^^^^^^^^^^^^^
Inline environment-scope configuration is done using the same yaml
Inline Environment-scope configuration is done using the same yaml
format as standard Spack configuration scopes, covered in the
:ref:`configuration` section. Each section is contained under a
top-level yaml object with it's name. For example, a ``spack.yaml``
@@ -660,7 +634,7 @@ Included configurations
Spack environments allow an ``include`` heading in their yaml
schema. This heading pulls in external configuration files and applies
them to the environment.
them to the Environment.
.. code-block:: yaml
@@ -673,9 +647,6 @@ them to the environment.
Environments can include files or URLs. File paths can be relative or
absolute. URLs include the path to the text for individual files or
can be the path to a directory containing configuration files.
Spack supports ``file``, ``http``, ``https`` and ``ftp`` protocols (or
schemes). Spack-specific, environment and user path variables may be
used in these paths. See :ref:`config-file-variables` for more information.
^^^^^^^^^^^^^^^^^^^^^^^^
Configuration precedence
@@ -690,7 +661,7 @@ have higher precedence, as the included configs are applied in reverse order.
Manually Editing the Specs List
-------------------------------
The list of abstract/root specs in the environment is maintained in
The list of abstract/root specs in the Environment is maintained in
the ``spack.yaml`` manifest under the heading ``specs``.
.. code-block:: yaml
@@ -798,7 +769,7 @@ evaluates to the cross-product of those specs. Spec matrices also
contain an ``excludes`` directive, which eliminates certain
combinations from the evaluated result.
The following two environment manifests are identical:
The following two Environment manifests are identical:
.. code-block:: yaml
@@ -873,7 +844,7 @@ files are identical.
In short files like the example, it may be easier to simply list the
included specs. However for more complicated examples involving many
packages across many toolchains, separately factored lists make
environments substantially more manageable.
Environments substantially more manageable.
Additionally, the ``-l`` option to the ``spack add`` command allows
one to add to named lists in the definitions section of the manifest
@@ -892,7 +863,7 @@ named list ``compilers`` is ``['%gcc', '%clang', '%intel']`` on
spack:
definitions:
- compilers: ['%gcc', '%clang']
- when: arch.satisfies('target=x86_64:')
- when: arch.satisfies('x86_64:')
compilers: ['%intel']
.. note::
@@ -960,89 +931,37 @@ This allows for a much-needed reduction in redundancy between packages
and constraints.
-----------------
Environment Views
-----------------
----------------
Filesystem Views
----------------
Spack Environments can have an associated filesystem view, which is a directory
with a more traditional structure ``<view>/bin``, ``<view>/lib``, ``<view>/include``
in which all files of the installed packages are linked.
By default a view is created for each environment, thanks to the ``view: true``
option in the ``spack.yaml`` manifest file:
.. code-block:: yaml
spack:
specs: [perl, python]
view: true
The view is created in a hidden directory ``.spack-env/view`` relative to the environment.
If you've used ``spack env activate``, you may have already interacted with this view. Spack
prepends its ``<view>/bin`` dir to ``PATH`` when the environment is activated, so that
you can directly run executables from all installed packages in the environment.
Views are highly customizable: you can control where they are put, modify their structure,
include and exclude specs, change how files are linked, and you can even generate multiple
views for a single environment.
Spack Environments can define filesystem views, which provide a direct access point
for software similar to the directory hierarchy that might exist under ``/usr/local``.
Filesystem views are updated every time the environment is written out to the lock
file ``spack.lock``, so the concrete environment and the view are always compatible.
The files of the view's installed packages are brought into the view by symbolic or
hard links, referencing the original Spack installation, or by copy.
.. _configuring_environment_views:
^^^^^^^^^^^^^^^^^^^^^^^^^^
Minimal view configuration
^^^^^^^^^^^^^^^^^^^^^^^^^^
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Configuration in ``spack.yaml``
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
The minimal configuration
.. code-block:: yaml
spack:
# ...
view: true
lets Spack generate a single view with default settings under the
``.spack-env/view`` directory of the environment.
Another short way to configure a view is to specify just where to put it:
.. code-block:: yaml
spack:
# ...
view: /path/to/view
Views can also be disabled by setting ``view: false``.
^^^^^^^^^^^^^^^^^^^^^^^^^^^
Advanced view configuration
^^^^^^^^^^^^^^^^^^^^^^^^^^^
One or more **view descriptors** can be defined under ``view``, keyed by a name.
The example from the previous section with ``view: /path/to/view`` is equivalent
to defining a view descriptor named ``default`` with a ``root`` attribute:
.. code-block:: yaml
spack:
# ...
view:
default: # name of the view
root: /path/to/view # view descriptor attribute
The ``default`` view descriptor name is special: when you ``spack env activate`` your
environment, this view will be used to update (among other things) your ``PATH``
variable.
View descriptors must contain the root of the view, and optionally projections,
``select`` and ``exclude`` lists and link information via ``link`` and
The Spack Environment manifest file has a top-level keyword
``view``. Each entry under that heading is a **view descriptor**, headed
by a name. Any number of views may be defined under the ``view`` heading.
The view descriptor contains the root of the view, and
optionally the projections for the view, ``select`` and
``exclude`` lists for the view and link information via ``link`` and
``link_type``.
As a more advanced example, in the following manifest
For example, in the following manifest
file snippet we define a view named ``mpis``, rooted at
``/path/to/view`` in which all projections use the package name,
version, and compiler name to determine the path for a given
package. This view selects all packages that depend on MPI, and
excludes those built with the GCC compiler at version 18.5.
excludes those built with the PGI compiler at version 18.5.
The root specs with their (transitive) link and run type dependencies
will be put in the view due to the ``link: all`` option,
and the files in the view will be symlinks to the spack install
@@ -1056,7 +975,7 @@ directories.
mpis:
root: /path/to/view
select: [^mpi]
exclude: ['%gcc@18.5']
exclude: ['%pgi@18.5']
projections:
all: '{name}/{version}-{compiler.name}'
link: all
@@ -1082,14 +1001,63 @@ of ``hardlink`` or ``copy``.
when the environment is not activated, and linked libraries will be located
*outside* of the view thanks to rpaths.
There are two shorthands for environments with a single view. If the
environment at ``/path/to/env`` has a single view, with a root at
``/path/to/env/.spack-env/view``, with default selection and exclusion
and the default projection, we can put ``view: True`` in the
environment manifest. Similarly, if the environment has a view with a
different root, but default selection, exclusion, and projections, the
manifest can say ``view: /path/to/view``. These views are
automatically named ``default``, so that
.. code-block:: yaml
spack:
# ...
view: True
is equivalent to
.. code-block:: yaml
spack:
# ...
view:
default:
root: .spack-env/view
and
.. code-block:: yaml
spack:
# ...
view: /path/to/view
is equivalent to
.. code-block:: yaml
spack:
# ...
view:
default:
root: /path/to/view
By default, Spack environments are configured with ``view: True`` in
the manifest. Environments can be configured without views using
``view: False``. For backwards compatibility reasons, environments
with no ``view`` key are treated the same as ``view: True``.
From the command line, the ``spack env create`` command takes an
argument ``--with-view [PATH]`` that sets the path for a single, default
view. If no path is specified, the default path is used (``view:
true``). The argument ``--without-view`` can be used to create an
True``). The argument ``--without-view`` can be used to create an
environment without any view configured.
The ``spack env view`` command can be used to change the manage views
of an environment. The subcommand ``spack env view enable`` will add a
of an Environment. The subcommand ``spack env view enable`` will add a
view named ``default`` to an environment. It takes an optional
argument to specify the path for the new default view. The subcommand
``spack env view disable`` will remove the view named ``default`` from
@@ -1151,18 +1119,11 @@ the projection under ``all`` before reaching those entries.
Activating environment views
^^^^^^^^^^^^^^^^^^^^^^^^^^^^
The ``spack env activate <env>`` has two effects:
1. It activates the environment so that further Spack commands such
as ``spack install`` will run in the context of the environment.
2. It activates the view so that environment variables such as
``PATH`` are updated to include the view.
Without further arguments, the ``default`` view of the environment is
activated. If a view with a different name has to be activated,
``spack env activate --with-view <name> <env>`` can be
used instead. You can also activate the environment without modifying
further environment variables using ``--without-view``.
The ``spack env activate`` command will put the default view for the
environment into the user's path, in addition to activating the
environment for Spack commands. The arguments ``-v,--with-view`` and
``-V,--without-view`` can be used to tune this behavior. The default
behavior is to activate with the environment view if there is one.
The environment variables affected by the ``spack env activate``
command and the paths that are used to update them are determined by
@@ -1185,8 +1146,8 @@ relevant variable if the path exists. For this reason, it is not
recommended to use non-default projections with the default view of an
environment.
The ``spack env deactivate`` command will remove the active view of
the Spack environment from the user's environment variables.
The ``spack env deactivate`` command will remove the default view of
the environment from the user's path.
.. _env-generate-depfile:
@@ -1257,7 +1218,7 @@ gets installed and is available for use in the ``env`` target.
$(SPACK) -e . env depfile -o $@ --make-prefix spack
env: spack/env
$(info environment installed!)
$(info Environment installed!)
clean:
rm -rf spack.lock env.mk spack/
@@ -1345,7 +1306,7 @@ index once every package is pushed. Note how this target uses the generated
example/push/%: example/install/%
@mkdir -p $(dir $@)
$(info About to push $(SPEC) to a buildcache)
$(SPACK) -e . buildcache push --only=package $(BUILDCACHE_DIR) /$(HASH)
$(SPACK) -e . buildcache push --allow-root --only=package $(BUILDCACHE_DIR) /$(HASH)
@touch $@
push: $(addprefix example/push/,$(example/SPACK_PACKAGE_IDS))

View File

@@ -35,7 +35,7 @@ A build matrix showing which packages are working on which systems is shown belo
.. code-block:: console
apt update
apt install bzip2 ca-certificates file g++ gcc gfortran git gzip lsb-release patch python3 tar unzip xz-utils zstd
apt install build-essential ca-certificates coreutils curl environment-modules gfortran git gpg lsb-release python3 python3-distutils python3-venv unzip zip
.. tab-item:: RHEL
@@ -43,14 +43,14 @@ A build matrix showing which packages are working on which systems is shown belo
dnf install epel-release
dnf group install "Development Tools"
dnf install gcc-gfortran redhat-lsb-core python3 unzip
dnf install curl findutils gcc-gfortran gnupg2 hostname iproute redhat-lsb-core python3 python3-pip python3-setuptools unzip python3-boto3
.. tab-item:: macOS Brew
.. code-block:: console
brew update
brew install gcc git zip
brew install curl gcc git gnupg zip
------------
Installation
@@ -61,15 +61,10 @@ Getting Spack is easy. You can clone it from the `github repository
.. code-block:: console
$ git clone -c feature.manyFiles=true --depth=2 https://github.com/spack/spack.git
$ git clone -c feature.manyFiles=true https://github.com/spack/spack.git
This will create a directory called ``spack``.
.. note::
``-c feature.manyFiles=true`` improves git's performance on repositories with 1,000+ files.
``--depth=2`` prunes the git history to reduce the size of the Spack installation.
.. _shell-support:
^^^^^^^^^^^^^
@@ -283,6 +278,10 @@ compilers`` or ``spack compiler list``:
intel@14.0.1 intel@13.0.1 intel@12.1.2 intel@10.1
-- clang -------------------------------------------------------
clang@3.4 clang@3.3 clang@3.2 clang@3.1
-- pgi ---------------------------------------------------------
pgi@14.3-0 pgi@13.2-0 pgi@12.1-0 pgi@10.9-0 pgi@8.0-1
pgi@13.10-0 pgi@13.1-1 pgi@11.10-0 pgi@10.2-0 pgi@7.1-3
pgi@13.6-0 pgi@12.8-0 pgi@11.1-0 pgi@9.0-4 pgi@7.0-6
Any of these compilers can be used to build Spack packages. More on
how this is done is in :ref:`sec-specs`.
@@ -802,6 +801,65 @@ flags to the ``icc`` command:
spec: intel@15.0.24.4.9.3
^^^
PGI
^^^
PGI comes with two sets of compilers for C++ and Fortran,
distinguishable by their names. "Old" compilers:
.. code-block:: yaml
cc: /soft/pgi/15.10/linux86-64/15.10/bin/pgcc
cxx: /soft/pgi/15.10/linux86-64/15.10/bin/pgCC
f77: /soft/pgi/15.10/linux86-64/15.10/bin/pgf77
fc: /soft/pgi/15.10/linux86-64/15.10/bin/pgf90
"New" compilers:
.. code-block:: yaml
cc: /soft/pgi/15.10/linux86-64/15.10/bin/pgcc
cxx: /soft/pgi/15.10/linux86-64/15.10/bin/pgc++
f77: /soft/pgi/15.10/linux86-64/15.10/bin/pgfortran
fc: /soft/pgi/15.10/linux86-64/15.10/bin/pgfortran
Older installations of PGI contains just the old compilers; whereas
newer installations contain the old and the new. The new compiler is
considered preferable, as some packages
(``hdf``) will not build with the old compiler.
When auto-detecting a PGI compiler, there are cases where Spack will
find the old compilers, when you really want it to find the new
compilers. It is best to check this ``compilers.yaml``; and if the old
compilers are being used, change ``pgf77`` and ``pgf90`` to
``pgfortran``.
Other issues:
* There are reports that some packages will not build with PGI,
including ``libpciaccess`` and ``openssl``. A workaround is to
build these packages with another compiler and then use them as
dependencies for PGI-build packages. For example:
.. code-block:: console
$ spack install openmpi%pgi ^libpciaccess%gcc
* PGI requires a license to use; see :ref:`licensed-compilers` for more
information on installation.
.. note::
It is believed the problem with HDF 4 is that everything is
compiled with the ``F77`` compiler, but at some point some Fortran
90 code slipped in there. So compilers that can handle both FORTRAN
77 and Fortran 90 (``gfortran``, ``pgfortran``, etc) are fine. But
compilers specific to one or the other (``pgf77``, ``pgf90``) won't
work.
^^^
NAG
^^^
@@ -1306,6 +1364,187 @@ This will write the private key to the file `dinosaur.priv`.
or for help on an issue or the Spack slack.
.. _cray-support:
-------------
Spack on Cray
-------------
Spack differs slightly when used on a Cray system. The architecture spec
can differentiate between the front-end and back-end processor and operating system.
For example, on Edison at NERSC, the back-end target processor
is "Ivy Bridge", so you can specify to use the back-end this way:
.. code-block:: console
$ spack install zlib target=ivybridge
You can also use the operating system to build against the back-end:
.. code-block:: console
$ spack install zlib os=CNL10
Notice that the name includes both the operating system name and the major
version number concatenated together.
Alternatively, if you want to build something for the front-end,
you can specify the front-end target processor. The processor for a login node
on Edison is "Sandy bridge" so we specify on the command line like so:
.. code-block:: console
$ spack install zlib target=sandybridge
And the front-end operating system is:
.. code-block:: console
$ spack install zlib os=SuSE11
^^^^^^^^^^^^^^^^^^^^^^^
Cray compiler detection
^^^^^^^^^^^^^^^^^^^^^^^
Spack can detect compilers using two methods. For the front-end, we treat
everything the same. The difference lies in back-end compiler detection.
Back-end compiler detection is made via the Tcl module avail command.
Once it detects the compiler it writes the appropriate PrgEnv and compiler
module name to compilers.yaml and sets the paths to each compiler with Cray\'s
compiler wrapper names (i.e. cc, CC, ftn). During build time, Spack will load
the correct PrgEnv and compiler module and will call appropriate wrapper.
The compilers.yaml config file will also differ. There is a
modules section that is filled with the compiler's Programming Environment
and module name. On other systems, this field is empty []:
.. code-block:: yaml
- compiler:
modules:
- PrgEnv-intel
- intel/15.0.109
As mentioned earlier, the compiler paths will look different on a Cray system.
Since most compilers are invoked using cc, CC and ftn, the paths for each
compiler are replaced with their respective Cray compiler wrapper names:
.. code-block:: yaml
paths:
cc: cc
cxx: CC
f77: ftn
fc: ftn
As opposed to an explicit path to the compiler executable. This allows Spack
to call the Cray compiler wrappers during build time.
For more on compiler configuration, check out :ref:`compiler-config`.
Spack sets the default Cray link type to dynamic, to better match other
other platforms. Individual packages can enable static linking (which is the
default outside of Spack on cray systems) using the ``-static`` flag.
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Setting defaults and using Cray modules
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
If you want to use default compilers for each PrgEnv and also be able
to load cray external modules, you will need to set up a ``packages.yaml``.
Here's an example of an external configuration for cray modules:
.. code-block:: yaml
packages:
mpich:
externals:
- spec: "mpich@7.3.1%gcc@5.2.0 arch=cray_xc-haswell-CNL10"
modules:
- cray-mpich
- spec: "mpich@7.3.1%intel@16.0.0.109 arch=cray_xc-haswell-CNL10"
modules:
- cray-mpich
all:
providers:
mpi: [mpich]
This tells Spack that for whatever package that depends on mpi, load the
cray-mpich module into the environment. You can then be able to use whatever
environment variables, libraries, etc, that are brought into the environment
via module load.
.. note::
For Cray-provided packages, it is best to use ``modules:`` instead of ``prefix:``
in ``packages.yaml``, because the Cray Programming Environment heavily relies on
modules (e.g., loading the ``cray-mpich`` module adds MPI libraries to the
compiler wrapper link line).
You can set the default compiler that Spack can use for each compiler type.
If you want to use the Cray defaults, then set them under ``all:`` in packages.yaml.
In the compiler field, set the compiler specs in your order of preference.
Whenever you build with that compiler type, Spack will concretize to that version.
Here is an example of a full packages.yaml used at NERSC
.. code-block:: yaml
packages:
mpich:
externals:
- spec: "mpich@7.3.1%gcc@5.2.0 arch=cray_xc-CNL10-ivybridge"
modules:
- cray-mpich
- spec: "mpich@7.3.1%intel@16.0.0.109 arch=cray_xc-SuSE11-ivybridge"
modules:
- cray-mpich
buildable: False
netcdf:
externals:
- spec: "netcdf@4.3.3.1%gcc@5.2.0 arch=cray_xc-CNL10-ivybridge"
modules:
- cray-netcdf
- spec: "netcdf@4.3.3.1%intel@16.0.0.109 arch=cray_xc-CNL10-ivybridge"
modules:
- cray-netcdf
buildable: False
hdf5:
externals:
- spec: "hdf5@1.8.14%gcc@5.2.0 arch=cray_xc-CNL10-ivybridge"
modules:
- cray-hdf5
- spec: "hdf5@1.8.14%intel@16.0.0.109 arch=cray_xc-CNL10-ivybridge"
modules:
- cray-hdf5
buildable: False
all:
compiler: [gcc@5.2.0, intel@16.0.0.109]
providers:
mpi: [mpich]
Here we tell spack that whenever we want to build with gcc use version 5.2.0 or
if we want to build with intel compilers, use version 16.0.0.109. We add a spec
for each compiler type for each cray modules. This ensures that for each
compiler on our system we can use that external module.
For more on external packages check out the section :ref:`sec-external-packages`.
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Using Linux containers on Cray machines
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Spack uses environment variables particular to the Cray programming
environment to determine which systems are Cray platforms. These
environment variables may be propagated into containers that are not
using the Cray programming environment.
To ensure that Spack does not autodetect the Cray programming
environment, unset the environment variable ``MODULEPATH``. This
will cause Spack to treat a linux container on a Cray system as a base
linux distro.
.. _windows_support:
----------------
@@ -1417,14 +1656,16 @@ in a Windows CMD prompt.
Step 3: Run and configure Spack
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
On Windows, Spack supports both primary native shells, Powershell and the traditional command prompt.
To use Spack, pick your favorite shell, and run ``bin\spack_cmd.bat`` or ``share/spack/setup-env.ps1``
(you may need to Run as Administrator) from the top-level spack
directory. This will provide a Spack enabled shell. If you receive a warning message that Python is not in your ``PATH``
To use Spack, run ``bin\spack_cmd.bat`` (you may need to Run as Administrator) from the top-level spack
directory. This will provide a Windows command prompt with an environment properly set up with Spack
and its prerequisites. If you receive a warning message that Python is not in your ``PATH``
(which may happen if you installed Python from the website and not the Windows Store) add the location
of the Python executable to your ``PATH`` now. You can permanently add Python to your ``PATH`` variable
by using the ``Edit the system environment variables`` utility in Windows Control Panel.
.. note::
Alternatively, Powershell can be used in place of CMD
To configure Spack, first run the following command inside the Spack console:
.. code-block:: console
@@ -1489,7 +1730,7 @@ and not tabs, so ensure that this is the case when editing one directly.
.. note:: Cygwin
The use of Cygwin is not officially supported by Spack and is not tested.
However Spack will not prevent this, so use if choosing to use Spack
However Spack will not throw an error, so use if choosing to use Spack
with Cygwin, know that no functionality is garunteed.
^^^^^^^^^^^^^^^^^
@@ -1503,12 +1744,21 @@ Spack console via:
spack install cpuinfo
If in the previous step, you did not have CMake or Ninja installed, running the command above should install both packages
If in the previous step, you did not have CMake or Ninja installed, running the command above should bootstrap both packages
.. note:: Spec Syntax Caveats
Windows has a few idiosyncrasies when it comes to the Spack spec syntax and the use of certain shells
See the Spack spec syntax doc for more information
"""""""""""""""""""""""""""
Windows Compatible Packages
"""""""""""""""""""""""""""
Not all spack packages currently have Windows support. Some are inherently incompatible with the
platform, and others simply have yet to be ported. To view the current set of packages with Windows
support, the list command should be used via `spack list -t windows`. If there's a package you'd like
to install on Windows but is not in that list, feel free to reach out to request the port or contribute
the port yourself.
.. note::
This is by no means a comprehensive list, some packages may have ports that were not tagged
while others may just work out of the box on Windows and have not been tagged as such.
^^^^^^^^^^^^^^
For developers
@@ -1518,3 +1768,6 @@ The intent is to provide a Windows installer that will automatically set up
Python, Git, and Spack, instead of requiring the user to do so manually.
Instructions for creating the installer are at
https://github.com/spack/spack/blob/develop/lib/spack/spack/cmd/installer/README.md
Alternatively a pre-built copy of the Windows installer is available as an artifact of Spack's Windows CI
available at each run of the CI on develop or any PR.

Binary file not shown.

After

Width:  |  Height:  |  Size: 44 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 68 KiB

Binary file not shown.

Before

Width:  |  Height:  |  Size: 358 KiB

View File

@@ -12,6 +12,10 @@
Spack
===================
.. epigraph::
`These are docs for the Spack package manager. For sphere packing, see` `pyspack <https://pyspack.readthedocs.io>`_.
Spack is a package management tool designed to support multiple
versions and configurations of software on a wide variety of platforms
and environments. It was designed for large supercomputing centers,
@@ -35,15 +39,10 @@ package:
.. code-block:: console
$ git clone -c feature.manyFiles=true --depth=2 https://github.com/spack/spack.git
$ git clone -c feature.manyFiles=true https://github.com/spack/spack.git
$ cd spack/bin
$ ./spack install libelf
.. note::
``-c feature.manyFiles=true`` improves git's performance on repositories with 1,000+ files.
``--depth=2`` prunes the git history to reduce the size of the Spack installation.
If you're new to spack and want to start using it, see :doc:`getting_started`,
or refer to the full manual below.

View File

@@ -457,11 +457,11 @@ For instance, the following config options,
tcl:
all:
suffixes:
^python@3: 'python{^python.version}'
^python@3.12: 'python-3.12'
^openblas: 'openblas'
will add a ``python-3.12.1`` version string to any packages compiled with
Python matching the spec, ``python@3``. This is useful to know which
will add a ``python-3.12`` version string to any packages compiled with
Python matching the spec, ``python@3.12``. This is useful to know which
version of Python a set of Python extensions is associated with. Likewise, the
``openblas`` string is attached to any program that has openblas in the spec,
most likely via the ``+blas`` variant specification.

File diff suppressed because it is too large Load Diff

View File

@@ -59,7 +59,7 @@ Functional Example
------------------
The simplest fully functional standalone example of a working pipeline can be
examined live at this example `project <https://gitlab.com/spack/pipeline-quickstart>`_
examined live at this example `project <https://gitlab.com/scott.wittenburg/spack-pipeline-demo>`_
on gitlab.com.
Here's the ``.gitlab-ci.yml`` file from that example that builds and runs the
@@ -67,46 +67,39 @@ pipeline:
.. code-block:: yaml
stages: [ "generate", "build" ]
stages: [generate, build]
variables:
SPACK_REPOSITORY: "https://github.com/spack/spack.git"
SPACK_REF: "develop-2024-10-06"
SPACK_USER_CONFIG_PATH: ${CI_PROJECT_DIR}
SPACK_BACKTRACE: 1
SPACK_REPO: https://github.com/scottwittenburg/spack.git
SPACK_REF: pipelines-reproducible-builds
generate-pipeline:
tags:
- saas-linux-small-amd64
stage: generate
tags:
- docker
image:
name: ghcr.io/spack/ubuntu20.04-runner-x86_64:2023-01-01
script:
- git clone ${SPACK_REPOSITORY}
- cd spack && git checkout ${SPACK_REF} && cd ../
name: ghcr.io/scottwittenburg/ecpe4s-ubuntu18.04-runner-x86_64:2020-09-01
entrypoint: [""]
before_script:
- git clone ${SPACK_REPO}
- pushd spack && git checkout ${SPACK_REF} && popd
- . "./spack/share/spack/setup-env.sh"
- spack --version
script:
- spack env activate --without-view .
- spack -d -v --color=always
ci generate
--check-index-only
- spack -d ci generate
--artifacts-root "${CI_PROJECT_DIR}/jobs_scratch_dir"
--output-file "${CI_PROJECT_DIR}/jobs_scratch_dir/cloud-ci-pipeline.yml"
--output-file "${CI_PROJECT_DIR}/jobs_scratch_dir/pipeline.yml"
artifacts:
paths:
- "${CI_PROJECT_DIR}/jobs_scratch_dir"
build-pipeline:
build-jobs:
stage: build
trigger:
include:
- artifact: jobs_scratch_dir/cloud-ci-pipeline.yml
- artifact: "jobs_scratch_dir/pipeline.yml"
job: generate-pipeline
strategy: depend
needs:
- artifacts: True
job: generate-pipeline
The key thing to note above is that there are two jobs: The first job to run,
``generate-pipeline``, runs the ``spack ci generate`` command to generate a
@@ -121,93 +114,82 @@ And here's the spack environment built by the pipeline represented as a
spack:
view: false
concretizer:
unify: true
reuse: false
unify: false
definitions:
- pkgs:
- zlib
- bzip2 ~debug
- compiler:
- '%gcc'
- bzip2
- arch:
- '%gcc@7.5.0 arch=linux-ubuntu18.04-x86_64'
specs:
- matrix:
- - $pkgs
- - $compiler
- - $arch
mirrors: { "mirror": "s3://spack-public/mirror" }
ci:
target: gitlab
enable-artifacts-buildcache: True
rebuild-index: False
pipeline-gen:
- any-job:
tags:
- saas-linux-small-amd64
image:
name: ghcr.io/spack/ubuntu20.04-runner-x86_64:2023-01-01
before_script:
- git clone ${SPACK_REPOSITORY}
- cd spack && git checkout ${SPACK_REF} && cd ../
- . "./spack/share/spack/setup-env.sh"
- spack --version
- export SPACK_USER_CONFIG_PATH=${CI_PROJECT_DIR}
- spack config blame mirrors
- git clone ${SPACK_REPO}
- pushd spack && git checkout ${SPACK_CHECKOUT_VERSION} && popd
- . "./spack/share/spack/setup-env.sh"
- build-job:
tags: [docker]
image:
name: ghcr.io/scottwittenburg/ecpe4s-ubuntu18.04-runner-x86_64:2020-09-01
entrypoint: [""]
The elements of this file important to spack ci pipelines are described in more
detail below, but there are a couple of things to note about the above working
example:
.. note::
The use of ``reuse: false`` in spack environments used for pipelines is
almost always what you want, as without it your pipelines will not rebuild
packages even if package hashes have changed. This is due to the concretizer
strongly preferring known hashes when ``reuse: true``.
There is no ``script`` attribute specified for here. The reason for this is
Spack CI will automatically generate reasonable default scripts. More
detail on what is in these scripts can be found below.
The ``ci`` section in the above environment file contains the bare minimum
configuration required for ``spack ci generate`` to create a working pipeline.
The ``target: gitlab`` tells spack that the desired pipeline output is for
gitlab. However, this isn't strictly required, as currently gitlab is the
only possible output format for pipelines. The ``pipeline-gen`` section
contains the key information needed to specify attributes for the generated
jobs. Notice that it contains a list which has only a single element in
this case. In real pipelines it will almost certainly have more elements,
and in those cases, order is important: spack starts at the bottom of the
list and works upwards when applying attributes.
Also notice the ``before_script`` section. It is required when using any of the
default scripts to source the ``setup-env.sh`` script in order to inform
the default scripts where to find the ``spack`` executable.
But in this simple case, we use only the special key ``any-job`` to
indicate that spack should apply the specified attributes (``tags``, ``image``,
and ``before_script``) to any job it generates. This includes jobs for
building/pushing all packages, a ``rebuild-index`` job at the end of the
pipeline, as well as any ``noop`` jobs that might be needed by gitlab when
no rebuilds are required.
Normally ``enable-artifacts-buildcache`` is not recommended in production as it
results in large binary artifacts getting transferred back and forth between
gitlab and the runners. But in this example on gitlab.com where there is no
shared, persistent file system, and where no secrets are stored for giving
permission to write to an S3 bucket, ``enabled-buildcache-artifacts`` is the only
way to propagate binaries from jobs to their dependents.
Something to note is that in this simple case, we rely on spack to
generate a reasonable script for the package build jobs (it just creates
a script that invokes ``spack ci rebuild``).
Also, it is usually a good idea to let the pipeline generate a final "rebuild the
buildcache index" job, so that subsequent pipeline generation can quickly determine
which specs are up to date and which need to be rebuilt (it's a good idea for other
reasons as well, but those are out of scope for this discussion). In this case we
have disabled it (using ``rebuild-index: False``) because the index would only be
generated in the artifacts mirror anyway, and consequently would not be available
during subsequent pipeline runs.
Another thing to note is the use of the ``SPACK_USER_CONFIG_DIR`` environment
variable in any generated jobs. The purpose of this is to make spack
aware of one final file in the example, the one that contains the mirror
configuration. This file, ``mirrors.yaml`` looks like this:
.. note::
With the addition of reproducible builds (#22887) a previously working
pipeline will require some changes:
.. code-block:: yaml
* In the build-jobs, the environment location changed.
This will typically show as a ``KeyError`` in the failing job. Be sure to
point to ``${SPACK_CONCRETE_ENV_DIR}``.
mirrors:
buildcache-destination:
url: oci://registry.gitlab.com/spack/pipeline-quickstart
binary: true
access_pair:
id_variable: CI_REGISTRY_USER
secret_variable: CI_REGISTRY_PASSWORD
* When using ``include`` in your environment, be sure to make the included
files available in the build jobs. This means adding those files to the
artifact directory. Those files will also be missing in the reproducibility
artifact.
Note the name of the mirror is ``buildcache-destination``, which is required
as of Spack 0.23 (see below for more information). The mirror url simply
points to the container registry associated with the project, while
``id_variable`` and ``secret_variable`` refer to to environment variables
containing the access credentials for the mirror.
When spack builds packages for this example project, they will be pushed to
the project container registry, where they will be available for subsequent
jobs to install as dependencies, or for other pipelines to use to build runnable
container images.
* Because the location of the environment changed, including files with
relative path may have to be adapted to work both in the project context
(generation job) and in the concrete env dir context (build job).
-----------------------------------
Spack commands supporting pipelines
@@ -271,6 +253,17 @@ can easily happen if it is not updated frequently, this behavior ensures that
spack has a way to know for certain about the status of any concrete spec on
the remote mirror, but can slow down pipeline generation significantly.
The ``--optimize`` argument is experimental and runs the generated pipeline
document through a series of optimization passes designed to reduce the size
of the generated file.
The ``--dependencies`` is also experimental and disables what in Gitlab is
referred to as DAG scheduling, internally using the ``dependencies`` keyword
rather than ``needs`` to list dependency jobs. The drawback of using this option
is that before any job can begin, all jobs in previous stages must first
complete. The benefit is that Gitlab allows more dependencies to be listed
when using ``dependencies`` instead of ``needs``.
The optional ``--output-file`` argument should be an absolute path (including
file name) to the generated pipeline, and if not given, the default is
``./.gitlab-ci.yml``.
@@ -435,6 +428,15 @@ configuration with a ``script`` attribute. Specifying a signing job without a sc
does not create a signing job and the job configuration attributes will be ignored.
Signing jobs are always assigned the runner tags ``aws``, ``protected``, and ``notary``.
^^^^^^^^^^^^^^^^^
Cleanup (cleanup)
^^^^^^^^^^^^^^^^^
When using ``temporary-storage-url-prefix`` the cleanup job will destroy the mirror
created for the associated Gitlab pipeline. Cleanup jobs do not allow modifying the
script, but do expect that the spack command is in the path and require a
``before_script`` to be specified that sources the ``setup-env.sh`` script.
.. _noop_jobs:
^^^^^^^^^^^^
@@ -601,77 +603,6 @@ the attributes will be merged starting from the bottom match going up to the top
In the case that no match is found in a submapping section, no additional attributes will be applied.
^^^^^^^^^^^^^^^^^^^^^^^^
Dynamic Mapping Sections
^^^^^^^^^^^^^^^^^^^^^^^^
For large scale CI where cost optimization is required, dynamic mapping allows for the use of real-time
mapping schemes served by a web service. This type of mapping does not support the ``-remove`` type
behavior, but it does follow the rest of the merge rules for configurations.
The dynamic mapping service needs to implement a single REST API interface for getting
requests ``GET <URL>[:PORT][/PATH]?spec=<pkg_name@pkg_version +variant1+variant2%compiler@compiler_version>``.
example request.
.. code-block::
https://my-dyn-mapping.spack.io/allocation?spec=zlib-ng@2.1.6 +compat+opt+shared+pic+new_strategies arch=linux-ubuntu20.04-x86_64_v3%gcc@12.0.0
With an example response the updates kubernetes request variables, overrides the max retries for gitlab,
and prepends a note about the modifications made by the my-dyn-mapping.spack.io service.
.. code-block::
200 OK
{
"variables":
{
"KUBERNETES_CPU_REQUEST": "500m",
"KUBERNETES_MEMORY_REQUEST": "2G",
},
"retry": { "max:": "1"}
"script+:":
[
"echo \"Job modified by my-dyn-mapping.spack.io\""
]
}
The ci.yaml configuration section takes the URL endpoint as well as a number of options to configure how responses are handled.
It is possible to specify a list of allowed and ignored configuration attributes under ``allow`` and ``ignore``
respectively. It is also possible to configure required attributes under ``required`` section.
Options to configure the client timeout and SSL verification using the ``timeout`` and ``verify_ssl`` options.
By default, the ``timeout`` is set to the option in ``config:timeout`` and ``veryify_ssl`` is set the the option in ``config::verify_ssl``.
Passing header parameters to the request can be achieved through the ``header`` section. The values of the variables passed to the
header may be environment variables that are expanded at runtime, such as a private token configured on the runner.
Here is an example configuration pointing to ``my-dyn-mapping.spack.io/allocation``.
.. code-block:: yaml
ci:
- dynamic-mapping:
endpoint: my-dyn-mapping.spack.io/allocation
timeout: 10
verify_ssl: True
header:
PRIVATE_TOKEN: ${MY_PRIVATE_TOKEN}
MY_CONFIG: "fuzz_allocation:false"
allow:
- variables
ignore:
- script
require: []
^^^^^^^^^^^^^
Bootstrapping
^^^^^^^^^^^^^
@@ -743,13 +674,26 @@ build the package.
When including a bootstrapping phase as in the example above, the result is that
the bootstrapped compiler packages will be pushed to the binary mirror (and the
local artifacts mirror) before the actual release specs are built.
local artifacts mirror) before the actual release specs are built. In this case,
the jobs corresponding to subsequent release specs are configured to
``install_missing_compilers``, so that if spack is asked to install a package
with a compiler it doesn't know about, it can be quickly installed from the
binary mirror first.
Since bootstrapping compilers is optional, those items can be left out of the
environment/stack file, and in that case no bootstrapping will be done (only the
specs will be staged for building) and the runners will be expected to already
have all needed compilers installed and configured for spack to use.
^^^^^^^^^^^^^^^^^^^
Pipeline Buildcache
^^^^^^^^^^^^^^^^^^^
The ``enable-artifacts-buildcache`` key
takes a boolean and determines whether the pipeline uses artifacts to store and
pass along the buildcaches from one stage to the next (the default if you don't
provide this option is ``False``).
^^^^^^^^^^^^^^^^
Broken Specs URL
^^^^^^^^^^^^^^^^

View File

@@ -476,3 +476,9 @@ implemented using Python's built-in `sys.path
:py:mod:`spack.repo` module implements a custom `Python importer
<https://docs.python.org/2/library/imp.html>`_.
.. warning::
The mechanism for extending packages is not yet extensively tested,
and extending packages across repositories imposes inter-repo
dependencies, which may be hard to manage. Use this feature at your
own risk, but let us know if you have a use case for it.

View File

@@ -1,13 +1,13 @@
sphinx==8.1.3
sphinx==7.2.6
sphinxcontrib-programoutput==0.17
sphinx_design==0.6.1
sphinx-rtd-theme==3.0.2
python-levenshtein==0.26.1
docutils==0.21.2
pygments==2.18.0
urllib3==2.2.3
pytest==8.3.3
sphinx_design==0.5.0
sphinx-rtd-theme==2.0.0
python-levenshtein==0.25.1
docutils==0.20.1
pygments==2.17.2
urllib3==2.2.1
pytest==8.2.0
isort==5.13.2
black==24.10.0
flake8==7.1.1
mypy==1.11.1
black==24.4.2
flake8==7.0.0
mypy==1.10.0

349
lib/spack/env/cc vendored
View File

@@ -41,6 +41,10 @@ SPACK_ENV_PATH
SPACK_DEBUG_LOG_DIR
SPACK_DEBUG_LOG_ID
SPACK_COMPILER_SPEC
SPACK_CC_RPATH_ARG
SPACK_CXX_RPATH_ARG
SPACK_F77_RPATH_ARG
SPACK_FC_RPATH_ARG
SPACK_LINKER_ARG
SPACK_SHORT_SPEC
SPACK_SYSTEM_DIRS
@@ -97,9 +101,10 @@ setsep() {
esac
}
# prepend LISTNAME ELEMENT
# prepend LISTNAME ELEMENT [SEP]
#
# Prepend ELEMENT to the list stored in the variable LISTNAME.
# Prepend ELEMENT to the list stored in the variable LISTNAME,
# assuming the list is separated by SEP.
# Handles empty lists and single-element lists.
prepend() {
varname="$1"
@@ -169,46 +174,6 @@ preextend() {
unset IFS
}
execute() {
# dump the full command if the caller supplies SPACK_TEST_COMMAND=dump-args
if [ -n "${SPACK_TEST_COMMAND=}" ]; then
case "$SPACK_TEST_COMMAND" in
dump-args)
IFS="$lsep"
for arg in $full_command_list; do
echo "$arg"
done
unset IFS
exit
;;
dump-env-*)
var=${SPACK_TEST_COMMAND#dump-env-}
eval "printf '%s\n' \"\$0: \$var: \$$var\""
;;
*)
die "Unknown test command: '$SPACK_TEST_COMMAND'"
;;
esac
fi
#
# Write the input and output commands to debug logs if it's asked for.
#
if [ "$SPACK_DEBUG" = TRUE ]; then
input_log="$SPACK_DEBUG_LOG_DIR/spack-cc-$SPACK_DEBUG_LOG_ID.in.log"
output_log="$SPACK_DEBUG_LOG_DIR/spack-cc-$SPACK_DEBUG_LOG_ID.out.log"
echo "[$mode] $command $input_command" >> "$input_log"
IFS="$lsep"
echo "[$mode] "$full_command_list >> "$output_log"
unset IFS
fi
# Execute the full command, preserving spaces with IFS set
# to the alarm bell separator.
IFS="$lsep"; exec $full_command_list
exit
}
# Fail with a clear message if the input contains any bell characters.
if eval "[ \"\${*#*${lsep}}\" != \"\$*\" ]"; then
die "Compiler command line contains our separator ('${lsep}'). Cannot parse."
@@ -219,7 +184,6 @@ for param in $params; do
if eval "test -z \"\${${param}:-}\""; then
die "Spack compiler must be run from Spack! Input '$param' is missing."
fi
# FIXME (compiler as nodes) add checks on whether `SPACK_XX_RPATH` is set if `SPACK_XX` is set
done
# eval this because SPACK_MANAGED_DIRS and SPACK_SYSTEM_DIRS are inputs we don't wanna loop over.
@@ -234,36 +198,6 @@ esac
}
"
# path_list functions. Path_lists have 3 parts: spack_store_<list>, <list> and system_<list>,
# which are used to prioritize paths when assembling the final command line.
# init_path_lists LISTNAME
# Set <LISTNAME>, spack_store_<LISTNAME>, and system_<LISTNAME> to "".
init_path_lists() {
eval "spack_store_$1=\"\""
eval "$1=\"\""
eval "system_$1=\"\""
}
# assign_path_lists LISTNAME1 LISTNAME2
# Copy contents of LISTNAME2 into LISTNAME1, for each path_list prefix.
assign_path_lists() {
eval "spack_store_$1=\"\${spack_store_$2}\""
eval "$1=\"\${$2}\""
eval "system_$1=\"\${system_$2}\""
}
# append_path_lists LISTNAME ELT
# Append the provided ELT to the appropriate list, based on the result of path_order().
append_path_lists() {
path_order "$2"
case $? in
0) eval "append spack_store_$1 \"\$2\"" ;;
1) eval "append $1 \"\$2\"" ;;
2) eval "append system_$1 \"\$2\"" ;;
esac
}
# Check if optional parameters are defined
# If we aren't asking for debug flags, don't add them
if [ -z "${SPACK_ADD_DEBUG_FLAGS:-}" ]; then
@@ -297,17 +231,12 @@ fi
# ld link
# ccld compile & link
# Note. SPACK_ALWAYS_XFLAGS are applied for all compiler invocations,
# including version checks (SPACK_XFLAGS variants are not applied
# for version checks).
command="${0##*/}"
comp="CC"
vcheck_flags=""
case "$command" in
cpp)
mode=cpp
debug_flags="-g"
vcheck_flags="${SPACK_ALWAYS_CPPFLAGS}"
;;
cc|c89|c99|gcc|clang|armclang|icc|icx|pgcc|nvc|xlc|xlc_r|fcc|amdclang|cl.exe|craycc)
command="$SPACK_CC"
@@ -315,7 +244,6 @@ case "$command" in
comp="CC"
lang_flags=C
debug_flags="-g"
vcheck_flags="${SPACK_ALWAYS_CFLAGS}"
;;
c++|CC|g++|clang++|armclang++|icpc|icpx|pgc++|nvc++|xlc++|xlc++_r|FCC|amdclang++|crayCC)
command="$SPACK_CXX"
@@ -323,7 +251,6 @@ case "$command" in
comp="CXX"
lang_flags=CXX
debug_flags="-g"
vcheck_flags="${SPACK_ALWAYS_CXXFLAGS}"
;;
ftn|f90|fc|f95|gfortran|flang|armflang|ifort|ifx|pgfortran|nvfortran|xlf90|xlf90_r|nagfor|frt|amdflang|crayftn)
command="$SPACK_FC"
@@ -331,7 +258,6 @@ case "$command" in
comp="FC"
lang_flags=F
debug_flags="-g"
vcheck_flags="${SPACK_ALWAYS_FFLAGS}"
;;
f77|xlf|xlf_r|pgf77)
command="$SPACK_F77"
@@ -339,7 +265,6 @@ case "$command" in
comp="F77"
lang_flags=F
debug_flags="-g"
vcheck_flags="${SPACK_ALWAYS_FFLAGS}"
;;
ld|ld.gold|ld.lld)
mode=ld
@@ -400,7 +325,7 @@ dtags_to_strip="${SPACK_DTAGS_TO_STRIP}"
linker_arg="${SPACK_LINKER_ARG}"
# Set up rpath variable according to language.
rpath="ERROR: RPATH ARG WAS NOT SET, MAYBE THE PACKAGE DOES NOT DEPEND ON ${comp}?"
rpath="ERROR: RPATH ARG WAS NOT SET"
eval "rpath=\${SPACK_${comp}_RPATH_ARG:?${rpath}}"
# Dump the mode and exit if the command is dump-mode.
@@ -440,11 +365,7 @@ unset IFS
export PATH="$new_dirs"
if [ "$mode" = vcheck ]; then
full_command_list="$command"
args="$@"
extend full_command_list vcheck_flags
extend full_command_list args
execute
exec "${command}" "$@"
fi
# Darwin's linker has a -r argument that merges object files together.
@@ -496,7 +417,12 @@ input_command="$*"
parse_Wl() {
while [ $# -ne 0 ]; do
if [ "$wl_expect_rpath" = yes ]; then
append_path_lists return_rpath_dirs_list "$1"
path_order "$1"
case $? in
0) append return_spack_store_rpath_dirs_list "$1" ;;
1) append return_rpath_dirs_list "$1" ;;
2) append return_system_rpath_dirs_list "$1" ;;
esac
wl_expect_rpath=no
else
case "$1" in
@@ -505,14 +431,24 @@ parse_Wl() {
if [ -z "$arg" ]; then
shift; continue
fi
append_path_lists return_rpath_dirs_list "$arg"
path_order "$arg"
case $? in
0) append return_spack_store_rpath_dirs_list "$arg" ;;
1) append return_rpath_dirs_list "$arg" ;;
2) append return_system_rpath_dirs_list "$arg" ;;
esac
;;
--rpath=*)
arg="${1#--rpath=}"
if [ -z "$arg" ]; then
shift; continue
fi
append_path_lists return_rpath_dirs_list "$arg"
path_order "$arg"
case $? in
0) append return_spack_store_rpath_dirs_list "$arg" ;;
1) append return_rpath_dirs_list "$arg" ;;
2) append return_system_rpath_dirs_list "$arg" ;;
esac
;;
-rpath|--rpath)
wl_expect_rpath=yes
@@ -520,7 +456,8 @@ parse_Wl() {
"$dtags_to_strip")
;;
-Wl)
# Nested -Wl,-Wl means we're in NAG compiler territory. We don't support it.
# Nested -Wl,-Wl means we're in NAG compiler territory, we don't support
# it.
return 1
;;
*)
@@ -539,10 +476,21 @@ categorize_arguments() {
return_other_args_list=""
return_isystem_was_used=""
init_path_lists return_isystem_include_dirs_list
init_path_lists return_include_dirs_list
init_path_lists return_lib_dirs_list
init_path_lists return_rpath_dirs_list
return_isystem_spack_store_include_dirs_list=""
return_isystem_system_include_dirs_list=""
return_isystem_include_dirs_list=""
return_spack_store_include_dirs_list=""
return_system_include_dirs_list=""
return_include_dirs_list=""
return_spack_store_lib_dirs_list=""
return_system_lib_dirs_list=""
return_lib_dirs_list=""
return_spack_store_rpath_dirs_list=""
return_system_rpath_dirs_list=""
return_rpath_dirs_list=""
# Global state for keeping track of -Wl,-rpath -Wl,/path
wl_expect_rpath=no
@@ -608,17 +556,32 @@ categorize_arguments() {
arg="${1#-isystem}"
return_isystem_was_used=true
if [ -z "$arg" ]; then shift; arg="$1"; fi
append_path_lists return_isystem_include_dirs_list "$arg"
path_order "$arg"
case $? in
0) append return_isystem_spack_store_include_dirs_list "$arg" ;;
1) append return_isystem_include_dirs_list "$arg" ;;
2) append return_isystem_system_include_dirs_list "$arg" ;;
esac
;;
-I*)
arg="${1#-I}"
if [ -z "$arg" ]; then shift; arg="$1"; fi
append_path_lists return_include_dirs_list "$arg"
path_order "$arg"
case $? in
0) append return_spack_store_include_dirs_list "$arg" ;;
1) append return_include_dirs_list "$arg" ;;
2) append return_system_include_dirs_list "$arg" ;;
esac
;;
-L*)
arg="${1#-L}"
if [ -z "$arg" ]; then shift; arg="$1"; fi
append_path_lists return_lib_dirs_list "$arg"
path_order "$arg"
case $? in
0) append return_spack_store_lib_dirs_list "$arg" ;;
1) append return_lib_dirs_list "$arg" ;;
2) append return_system_lib_dirs_list "$arg" ;;
esac
;;
-l*)
# -loopopt=0 is generated erroneously in autoconf <= 2.69,
@@ -651,17 +614,32 @@ categorize_arguments() {
break
elif [ "$xlinker_expect_rpath" = yes ]; then
# Register the path of -Xlinker -rpath <other args> -Xlinker <path>
append_path_lists return_rpath_dirs_list "$1"
path_order "$1"
case $? in
0) append return_spack_store_rpath_dirs_list "$1" ;;
1) append return_rpath_dirs_list "$1" ;;
2) append return_system_rpath_dirs_list "$1" ;;
esac
xlinker_expect_rpath=no
else
case "$1" in
-rpath=*)
arg="${1#-rpath=}"
append_path_lists return_rpath_dirs_list "$arg"
path_order "$arg"
case $? in
0) append return_spack_store_rpath_dirs_list "$arg" ;;
1) append return_rpath_dirs_list "$arg" ;;
2) append return_system_rpath_dirs_list "$arg" ;;
esac
;;
--rpath=*)
arg="${1#--rpath=}"
append_path_lists return_rpath_dirs_list "$arg"
path_order "$arg"
case $? in
0) append return_spack_store_rpath_dirs_list "$arg" ;;
1) append return_rpath_dirs_list "$arg" ;;
2) append return_system_rpath_dirs_list "$arg" ;;
esac
;;
-rpath|--rpath)
xlinker_expect_rpath=yes
@@ -678,36 +656,7 @@ categorize_arguments() {
"$dtags_to_strip")
;;
*)
# if mode is not ld, we can just add to other args
if [ "$mode" != "ld" ]; then
append return_other_args_list "$1"
shift
continue
fi
# if we're in linker mode, we need to parse raw RPATH args
case "$1" in
-rpath=*)
arg="${1#-rpath=}"
append_path_lists return_rpath_dirs_list "$arg"
;;
--rpath=*)
arg="${1#--rpath=}"
append_path_lists return_rpath_dirs_list "$arg"
;;
-rpath|--rpath)
if [ $# -eq 1 ]; then
# -rpath without value: let the linker raise an error.
append return_other_args_list "$1"
break
fi
shift
append_path_lists return_rpath_dirs_list "$1"
;;
*)
append return_other_args_list "$1"
;;
esac
append return_other_args_list "$1"
;;
esac
shift
@@ -729,10 +678,21 @@ categorize_arguments() {
categorize_arguments "$@"
assign_path_lists isystem_include_dirs_list return_isystem_include_dirs_list
assign_path_lists include_dirs_list return_include_dirs_list
assign_path_lists lib_dirs_list return_lib_dirs_list
assign_path_lists rpath_dirs_list return_rpath_dirs_list
spack_store_include_dirs_list="$return_spack_store_include_dirs_list"
system_include_dirs_list="$return_system_include_dirs_list"
include_dirs_list="$return_include_dirs_list"
spack_store_lib_dirs_list="$return_spack_store_lib_dirs_list"
system_lib_dirs_list="$return_system_lib_dirs_list"
lib_dirs_list="$return_lib_dirs_list"
spack_store_rpath_dirs_list="$return_spack_store_rpath_dirs_list"
system_rpath_dirs_list="$return_system_rpath_dirs_list"
rpath_dirs_list="$return_rpath_dirs_list"
isystem_spack_store_include_dirs_list="$return_isystem_spack_store_include_dirs_list"
isystem_system_include_dirs_list="$return_isystem_system_include_dirs_list"
isystem_include_dirs_list="$return_isystem_include_dirs_list"
isystem_was_used="$return_isystem_was_used"
other_args_list="$return_other_args_list"
@@ -762,7 +722,6 @@ case "$mode" in
cc|ccld)
case $lang_flags in
F)
extend spack_flags_list SPACK_ALWAYS_FFLAGS
extend spack_flags_list SPACK_FFLAGS
;;
esac
@@ -772,7 +731,6 @@ esac
# C preprocessor flags come before any C/CXX flags
case "$mode" in
cpp|as|cc|ccld)
extend spack_flags_list SPACK_ALWAYS_CPPFLAGS
extend spack_flags_list SPACK_CPPFLAGS
;;
esac
@@ -783,19 +741,15 @@ case "$mode" in
cc|ccld)
case $lang_flags in
C)
extend spack_flags_list SPACK_ALWAYS_CFLAGS
extend spack_flags_list SPACK_CFLAGS
preextend flags_list SPACK_TARGET_ARGS_CC
;;
CXX)
extend spack_flags_list SPACK_ALWAYS_CXXFLAGS
extend spack_flags_list SPACK_CXXFLAGS
preextend flags_list SPACK_TARGET_ARGS_CXX
;;
F)
preextend flags_list SPACK_TARGET_ARGS_FORTRAN
;;
esac
# prepend target args
preextend flags_list SPACK_TARGET_ARGS
;;
esac
@@ -810,10 +764,21 @@ IFS="$lsep"
categorize_arguments $spack_flags_list
unset IFS
assign_path_lists spack_flags_isystem_include_dirs_list return_isystem_include_dirs_list
assign_path_lists spack_flags_include_dirs_list return_include_dirs_list
assign_path_lists spack_flags_lib_dirs_list return_lib_dirs_list
assign_path_lists spack_flags_rpath_dirs_list return_rpath_dirs_list
spack_flags_isystem_spack_store_include_dirs_list="$return_isystem_spack_store_include_dirs_list"
spack_flags_isystem_system_include_dirs_list="$return_isystem_system_include_dirs_list"
spack_flags_isystem_include_dirs_list="$return_isystem_include_dirs_list"
spack_flags_spack_store_include_dirs_list="$return_spack_store_include_dirs_list"
spack_flags_system_include_dirs_list="$return_system_include_dirs_list"
spack_flags_include_dirs_list="$return_include_dirs_list"
spack_flags_spack_store_lib_dirs_list="$return_spack_store_lib_dirs_list"
spack_flags_system_lib_dirs_list="$return_system_lib_dirs_list"
spack_flags_lib_dirs_list="$return_lib_dirs_list"
spack_flags_spack_store_rpath_dirs_list="$return_spack_store_rpath_dirs_list"
spack_flags_system_rpath_dirs_list="$return_system_rpath_dirs_list"
spack_flags_rpath_dirs_list="$return_rpath_dirs_list"
spack_flags_isystem_was_used="$return_isystem_was_used"
spack_flags_other_args_list="$return_other_args_list"
@@ -872,7 +837,7 @@ esac
case "$mode" in
cpp|cc|as|ccld)
if [ "$spack_flags_isystem_was_used" = "true" ] || [ "$isystem_was_used" = "true" ]; then
extend spack_store_isystem_include_dirs_list SPACK_STORE_INCLUDE_DIRS
extend isystem_spack_store_include_dirs_list SPACK_STORE_INCLUDE_DIRS
extend isystem_include_dirs_list SPACK_INCLUDE_DIRS
else
extend spack_store_include_dirs_list SPACK_STORE_INCLUDE_DIRS
@@ -888,63 +853,64 @@ args_list="$flags_list"
# Include search paths partitioned by (in store, non-sytem, system)
# NOTE: adding ${lsep} to the prefix here turns every added element into two
extend args_list spack_store_spack_flags_include_dirs_list -I
extend args_list spack_flags_spack_store_include_dirs_list -I
extend args_list spack_store_include_dirs_list -I
extend args_list spack_flags_include_dirs_list -I
extend args_list include_dirs_list -I
extend args_list spack_store_spack_flags_isystem_include_dirs_list "-isystem${lsep}"
extend args_list spack_store_isystem_include_dirs_list "-isystem${lsep}"
extend args_list spack_flags_isystem_spack_store_include_dirs_list "-isystem${lsep}"
extend args_list isystem_spack_store_include_dirs_list "-isystem${lsep}"
extend args_list spack_flags_isystem_include_dirs_list "-isystem${lsep}"
extend args_list isystem_include_dirs_list "-isystem${lsep}"
extend args_list system_spack_flags_include_dirs_list -I
extend args_list spack_flags_system_include_dirs_list -I
extend args_list system_include_dirs_list -I
extend args_list system_spack_flags_isystem_include_dirs_list "-isystem${lsep}"
extend args_list system_isystem_include_dirs_list "-isystem${lsep}"
extend args_list spack_flags_isystem_system_include_dirs_list "-isystem${lsep}"
extend args_list isystem_system_include_dirs_list "-isystem${lsep}"
# Library search paths partitioned by (in store, non-sytem, system)
extend args_list spack_store_spack_flags_lib_dirs_list "-L"
extend args_list spack_flags_spack_store_lib_dirs_list "-L"
extend args_list spack_store_lib_dirs_list "-L"
extend args_list spack_flags_lib_dirs_list "-L"
extend args_list lib_dirs_list "-L"
extend args_list system_spack_flags_lib_dirs_list "-L"
extend args_list spack_flags_system_lib_dirs_list "-L"
extend args_list system_lib_dirs_list "-L"
# RPATHs arguments
rpath_prefix=""
case "$mode" in
ccld)
if [ -n "$dtags_to_add" ] ; then
append args_list "$linker_arg$dtags_to_add"
fi
rpath_prefix="$rpath"
extend args_list spack_flags_spack_store_rpath_dirs_list "$rpath"
extend args_list spack_store_rpath_dirs_list "$rpath"
extend args_list spack_flags_rpath_dirs_list "$rpath"
extend args_list rpath_dirs_list "$rpath"
extend args_list spack_flags_system_rpath_dirs_list "$rpath"
extend args_list system_rpath_dirs_list "$rpath"
;;
ld)
if [ -n "$dtags_to_add" ] ; then
append args_list "$dtags_to_add"
fi
rpath_prefix="-rpath${lsep}"
extend args_list spack_flags_spack_store_rpath_dirs_list "-rpath${lsep}"
extend args_list spack_store_rpath_dirs_list "-rpath${lsep}"
extend args_list spack_flags_rpath_dirs_list "-rpath${lsep}"
extend args_list rpath_dirs_list "-rpath${lsep}"
extend args_list spack_flags_system_rpath_dirs_list "-rpath${lsep}"
extend args_list system_rpath_dirs_list "-rpath${lsep}"
;;
esac
# if mode is ccld or ld, extend RPATH lists with the prefix determined above
if [ -n "$rpath_prefix" ]; then
extend args_list spack_store_spack_flags_rpath_dirs_list "$rpath_prefix"
extend args_list spack_store_rpath_dirs_list "$rpath_prefix"
extend args_list spack_flags_rpath_dirs_list "$rpath_prefix"
extend args_list rpath_dirs_list "$rpath_prefix"
extend args_list system_spack_flags_rpath_dirs_list "$rpath_prefix"
extend args_list system_rpath_dirs_list "$rpath_prefix"
fi
# Other arguments from the input command
extend args_list other_args_list
extend args_list spack_flags_other_args_list
@@ -967,4 +933,39 @@ if [ -n "$SPACK_CCACHE_BINARY" ]; then
esac
fi
execute
# dump the full command if the caller supplies SPACK_TEST_COMMAND=dump-args
if [ -n "${SPACK_TEST_COMMAND=}" ]; then
case "$SPACK_TEST_COMMAND" in
dump-args)
IFS="$lsep"
for arg in $full_command_list; do
echo "$arg"
done
unset IFS
exit
;;
dump-env-*)
var=${SPACK_TEST_COMMAND#dump-env-}
eval "printf '%s\n' \"\$0: \$var: \$$var\""
;;
*)
die "Unknown test command: '$SPACK_TEST_COMMAND'"
;;
esac
fi
#
# Write the input and output commands to debug logs if it's asked for.
#
if [ "$SPACK_DEBUG" = TRUE ]; then
input_log="$SPACK_DEBUG_LOG_DIR/spack-cc-$SPACK_DEBUG_LOG_ID.in.log"
output_log="$SPACK_DEBUG_LOG_DIR/spack-cc-$SPACK_DEBUG_LOG_ID.out.log"
echo "[$mode] $command $input_command" >> "$input_log"
IFS="$lsep"
echo "[$mode] "$full_command_list >> "$output_log"
unset IFS
fi
# Execute the full command, preserving spaces with IFS set
# to the alarm bell separator.
IFS="$lsep"; exec $full_command_list

View File

@@ -18,7 +18,7 @@
* Homepage: https://pypi.python.org/pypi/archspec
* Usage: Labeling, comparison and detection of microarchitectures
* Version: 0.2.5 (commit 38ce485258ffc4fc6dd6688f8dc90cb269478c47)
* Version: 0.2.5-dev (commit cbb1fd5eb397a70d466e5160b393b87b0dbcc78f)
astunparse
----------------

View File

@@ -1265,29 +1265,27 @@ def _distro_release_info(self) -> Dict[str, str]:
match = _DISTRO_RELEASE_BASENAME_PATTERN.match(basename)
else:
try:
with os.scandir(self.etc_dir) as it:
etc_files = [
p.path for p in it
if p.is_file() and p.name not in _DISTRO_RELEASE_IGNORE_BASENAMES
]
basenames = [
basename
for basename in os.listdir(self.etc_dir)
if basename not in _DISTRO_RELEASE_IGNORE_BASENAMES
and os.path.isfile(os.path.join(self.etc_dir, basename))
]
# We sort for repeatability in cases where there are multiple
# distro specific files; e.g. CentOS, Oracle, Enterprise all
# containing `redhat-release` on top of their own.
etc_files.sort()
basenames.sort()
except OSError:
# This may occur when /etc is not readable but we can't be
# sure about the *-release files. Check common entries of
# /etc for information. If they turn out to not be there the
# error is handled in `_parse_distro_release_file()`.
etc_files = [
os.path.join(self.etc_dir, basename)
for basename in _DISTRO_RELEASE_BASENAMES
]
for filepath in etc_files:
match = _DISTRO_RELEASE_BASENAME_PATTERN.match(os.path.basename(filepath))
basenames = _DISTRO_RELEASE_BASENAMES
for basename in basenames:
match = _DISTRO_RELEASE_BASENAME_PATTERN.match(basename)
if match is None:
continue
filepath = os.path.join(self.etc_dir, basename)
distro_info = self._parse_distro_release_file(filepath)
# The name is always present if the pattern matches.
if "name" not in distro_info:

View File

@@ -231,6 +231,96 @@ def is_host_name(instance):
return True
try:
# The built-in `idna` codec only implements RFC 3890, so we go elsewhere.
import idna
except ImportError:
pass
else:
@_checks_drafts(draft7="idn-hostname", raises=idna.IDNAError)
def is_idn_host_name(instance):
if not isinstance(instance, str_types):
return True
idna.encode(instance)
return True
try:
import rfc3987
except ImportError:
try:
from rfc3986_validator import validate_rfc3986
except ImportError:
pass
else:
@_checks_drafts(name="uri")
def is_uri(instance):
if not isinstance(instance, str_types):
return True
return validate_rfc3986(instance, rule="URI")
@_checks_drafts(
draft6="uri-reference",
draft7="uri-reference",
raises=ValueError,
)
def is_uri_reference(instance):
if not isinstance(instance, str_types):
return True
return validate_rfc3986(instance, rule="URI_reference")
else:
@_checks_drafts(draft7="iri", raises=ValueError)
def is_iri(instance):
if not isinstance(instance, str_types):
return True
return rfc3987.parse(instance, rule="IRI")
@_checks_drafts(draft7="iri-reference", raises=ValueError)
def is_iri_reference(instance):
if not isinstance(instance, str_types):
return True
return rfc3987.parse(instance, rule="IRI_reference")
@_checks_drafts(name="uri", raises=ValueError)
def is_uri(instance):
if not isinstance(instance, str_types):
return True
return rfc3987.parse(instance, rule="URI")
@_checks_drafts(
draft6="uri-reference",
draft7="uri-reference",
raises=ValueError,
)
def is_uri_reference(instance):
if not isinstance(instance, str_types):
return True
return rfc3987.parse(instance, rule="URI_reference")
try:
from strict_rfc3339 import validate_rfc3339
except ImportError:
try:
from rfc3339_validator import validate_rfc3339
except ImportError:
validate_rfc3339 = None
if validate_rfc3339:
@_checks_drafts(name="date-time")
def is_datetime(instance):
if not isinstance(instance, str_types):
return True
return validate_rfc3339(instance)
@_checks_drafts(draft7="time")
def is_time(instance):
if not isinstance(instance, str_types):
return True
return is_datetime("1970-01-01T" + instance)
@_checks_drafts(name="regex", raises=re.error)
def is_regex(instance):
if not isinstance(instance, str_types):
@@ -250,3 +340,86 @@ def is_draft3_time(instance):
if not isinstance(instance, str_types):
return True
return datetime.datetime.strptime(instance, "%H:%M:%S")
try:
import webcolors
except ImportError:
pass
else:
def is_css_color_code(instance):
return webcolors.normalize_hex(instance)
@_checks_drafts(draft3="color", raises=(ValueError, TypeError))
def is_css21_color(instance):
if (
not isinstance(instance, str_types) or
instance.lower() in webcolors.css21_names_to_hex
):
return True
return is_css_color_code(instance)
def is_css3_color(instance):
if instance.lower() in webcolors.css3_names_to_hex:
return True
return is_css_color_code(instance)
try:
import jsonpointer
except ImportError:
pass
else:
@_checks_drafts(
draft6="json-pointer",
draft7="json-pointer",
raises=jsonpointer.JsonPointerException,
)
def is_json_pointer(instance):
if not isinstance(instance, str_types):
return True
return jsonpointer.JsonPointer(instance)
# TODO: I don't want to maintain this, so it
# needs to go either into jsonpointer (pending
# https://github.com/stefankoegl/python-json-pointer/issues/34) or
# into a new external library.
@_checks_drafts(
draft7="relative-json-pointer",
raises=jsonpointer.JsonPointerException,
)
def is_relative_json_pointer(instance):
# Definition taken from:
# https://tools.ietf.org/html/draft-handrews-relative-json-pointer-01#section-3
if not isinstance(instance, str_types):
return True
non_negative_integer, rest = [], ""
for i, character in enumerate(instance):
if character.isdigit():
non_negative_integer.append(character)
continue
if not non_negative_integer:
return False
rest = instance[i:]
break
return (rest == "#") or jsonpointer.JsonPointer(rest)
try:
import uritemplate.exceptions
except ImportError:
pass
else:
@_checks_drafts(
draft6="uri-template",
draft7="uri-template",
raises=uritemplate.exceptions.InvalidTemplate,
)
def is_uri_template(
instance,
template_validator=uritemplate.Validator().force_balanced_braces(),
):
template = uritemplate.URITemplate(instance)
return template_validator.validate(template)

View File

@@ -81,13 +81,8 @@ def __init__(self, name, parents, vendor, features, compilers, generation=0, cpu
self.generation = generation
# Only relevant for AArch64
self.cpu_part = cpu_part
# Cache the "ancestor" computation
# Cache the ancestor computation
self._ancestors = None
# Cache the "generic" computation
self._generic = None
# Cache the "family" computation
self._family = None
@property
def ancestors(self):
@@ -120,9 +115,6 @@ def __eq__(self, other):
and self.cpu_part == other.cpu_part
)
def __hash__(self):
return hash(self.name)
@coerce_target_names
def __ne__(self, other):
return not self == other
@@ -179,22 +171,18 @@ def __contains__(self, feature):
@property
def family(self):
"""Returns the architecture family a given target belongs to"""
if self._family is None:
roots = [x for x in [self] + self.ancestors if not x.ancestors]
msg = "a target is expected to belong to just one architecture family"
msg += f"[found {', '.join(str(x) for x in roots)}]"
assert len(roots) == 1, msg
self._family = roots.pop()
roots = [x for x in [self] + self.ancestors if not x.ancestors]
msg = "a target is expected to belong to just one architecture family"
msg += f"[found {', '.join(str(x) for x in roots)}]"
assert len(roots) == 1, msg
return self._family
return roots.pop()
@property
def generic(self):
"""Returns the best generic architecture that is compatible with self"""
if self._generic is None:
generics = [x for x in [self] + self.ancestors if x.vendor == "generic"]
self._generic = max(generics, key=lambda x: len(x.ancestors))
return self._generic
generics = [x for x in [self] + self.ancestors if x.vendor == "generic"]
return max(generics, key=lambda x: len(x.ancestors))
def to_dict(self):
"""Returns a dictionary representation of this object."""

View File

@@ -1482,6 +1482,7 @@
"cldemote",
"movdir64b",
"movdiri",
"pdcm",
"serialize",
"waitpkg"
],
@@ -2236,84 +2237,6 @@
]
}
},
"zen5": {
"from": ["zen4"],
"vendor": "AuthenticAMD",
"features": [
"abm",
"aes",
"avx",
"avx2",
"avx512_bf16",
"avx512_bitalg",
"avx512bw",
"avx512cd",
"avx512dq",
"avx512f",
"avx512ifma",
"avx512vbmi",
"avx512_vbmi2",
"avx512vl",
"avx512_vnni",
"avx512_vp2intersect",
"avx512_vpopcntdq",
"avx_vnni",
"bmi1",
"bmi2",
"clflushopt",
"clwb",
"clzero",
"cppc",
"cx16",
"f16c",
"flush_l1d",
"fma",
"fsgsbase",
"gfni",
"ibrs_enhanced",
"mmx",
"movbe",
"movdir64b",
"movdiri",
"pclmulqdq",
"popcnt",
"rdseed",
"sse",
"sse2",
"sse4_1",
"sse4_2",
"sse4a",
"ssse3",
"tsc_adjust",
"vaes",
"vpclmulqdq",
"xsavec",
"xsaveopt"
],
"compilers": {
"gcc": [
{
"versions": "14.1:",
"name": "znver5",
"flags": "-march={name} -mtune={name}"
}
],
"aocc": [
{
"versions": "5.0:",
"name": "znver5",
"flags": "-march={name} -mtune={name}"
}
],
"clang": [
{
"versions": "19.1:",
"name": "znver5",
"flags": "-march={name} -mtune={name}"
}
]
}
},
"ppc64": {
"from": [],
"vendor": "generic",

View File

@@ -1,45 +0,0 @@
diff --git a/lib/spack/external/_vendoring/distro/distro.py b/lib/spack/external/_vendoring/distro/distro.py
index 89e1868047..50c3b18d4d 100644
--- a/lib/spack/external/_vendoring/distro/distro.py
+++ b/lib/spack/external/_vendoring/distro/distro.py
@@ -1265,27 +1265,29 @@ def _distro_release_info(self) -> Dict[str, str]:
match = _DISTRO_RELEASE_BASENAME_PATTERN.match(basename)
else:
try:
- basenames = [
- basename
- for basename in os.listdir(self.etc_dir)
- if basename not in _DISTRO_RELEASE_IGNORE_BASENAMES
- and os.path.isfile(os.path.join(self.etc_dir, basename))
- ]
+ with os.scandir(self.etc_dir) as it:
+ etc_files = [
+ p.path for p in it
+ if p.is_file() and p.name not in _DISTRO_RELEASE_IGNORE_BASENAMES
+ ]
# We sort for repeatability in cases where there are multiple
# distro specific files; e.g. CentOS, Oracle, Enterprise all
# containing `redhat-release` on top of their own.
- basenames.sort()
+ etc_files.sort()
except OSError:
# This may occur when /etc is not readable but we can't be
# sure about the *-release files. Check common entries of
# /etc for information. If they turn out to not be there the
# error is handled in `_parse_distro_release_file()`.
- basenames = _DISTRO_RELEASE_BASENAMES
- for basename in basenames:
- match = _DISTRO_RELEASE_BASENAME_PATTERN.match(basename)
+ etc_files = [
+ os.path.join(self.etc_dir, basename)
+ for basename in _DISTRO_RELEASE_BASENAMES
+ ]
+
+ for filepath in etc_files:
+ match = _DISTRO_RELEASE_BASENAME_PATTERN.match(os.path.basename(filepath))
if match is None:
continue
- filepath = os.path.join(self.etc_dir, basename)
distro_info = self._parse_distro_release_file(filepath)
# The name is always present if the pattern matches.
if "name" not in distro_info:

View File

@@ -13,191 +13,3 @@ index 6b630cdfbb..1791fe7fbf 100644
-__version__ = metadata.version("jsonschema")
+
+__version__ = "3.2.0"
diff --git a/lib/spack/external/_vendoring/jsonschema/_format.py b/lib/spack/external/_vendoring/jsonschema/_format.py
index 281a7cfcff..29061e3661 100644
--- a/lib/spack/external/_vendoring/jsonschema/_format.py
+++ b/lib/spack/external/_vendoring/jsonschema/_format.py
@@ -231,96 +231,6 @@ def is_host_name(instance):
return True
-try:
- # The built-in `idna` codec only implements RFC 3890, so we go elsewhere.
- import idna
-except ImportError:
- pass
-else:
- @_checks_drafts(draft7="idn-hostname", raises=idna.IDNAError)
- def is_idn_host_name(instance):
- if not isinstance(instance, str_types):
- return True
- idna.encode(instance)
- return True
-
-
-try:
- import rfc3987
-except ImportError:
- try:
- from rfc3986_validator import validate_rfc3986
- except ImportError:
- pass
- else:
- @_checks_drafts(name="uri")
- def is_uri(instance):
- if not isinstance(instance, str_types):
- return True
- return validate_rfc3986(instance, rule="URI")
-
- @_checks_drafts(
- draft6="uri-reference",
- draft7="uri-reference",
- raises=ValueError,
- )
- def is_uri_reference(instance):
- if not isinstance(instance, str_types):
- return True
- return validate_rfc3986(instance, rule="URI_reference")
-
-else:
- @_checks_drafts(draft7="iri", raises=ValueError)
- def is_iri(instance):
- if not isinstance(instance, str_types):
- return True
- return rfc3987.parse(instance, rule="IRI")
-
- @_checks_drafts(draft7="iri-reference", raises=ValueError)
- def is_iri_reference(instance):
- if not isinstance(instance, str_types):
- return True
- return rfc3987.parse(instance, rule="IRI_reference")
-
- @_checks_drafts(name="uri", raises=ValueError)
- def is_uri(instance):
- if not isinstance(instance, str_types):
- return True
- return rfc3987.parse(instance, rule="URI")
-
- @_checks_drafts(
- draft6="uri-reference",
- draft7="uri-reference",
- raises=ValueError,
- )
- def is_uri_reference(instance):
- if not isinstance(instance, str_types):
- return True
- return rfc3987.parse(instance, rule="URI_reference")
-
-
-try:
- from strict_rfc3339 import validate_rfc3339
-except ImportError:
- try:
- from rfc3339_validator import validate_rfc3339
- except ImportError:
- validate_rfc3339 = None
-
-if validate_rfc3339:
- @_checks_drafts(name="date-time")
- def is_datetime(instance):
- if not isinstance(instance, str_types):
- return True
- return validate_rfc3339(instance)
-
- @_checks_drafts(draft7="time")
- def is_time(instance):
- if not isinstance(instance, str_types):
- return True
- return is_datetime("1970-01-01T" + instance)
-
-
@_checks_drafts(name="regex", raises=re.error)
def is_regex(instance):
if not isinstance(instance, str_types):
@@ -340,86 +250,3 @@ def is_draft3_time(instance):
if not isinstance(instance, str_types):
return True
return datetime.datetime.strptime(instance, "%H:%M:%S")
-
-
-try:
- import webcolors
-except ImportError:
- pass
-else:
- def is_css_color_code(instance):
- return webcolors.normalize_hex(instance)
-
- @_checks_drafts(draft3="color", raises=(ValueError, TypeError))
- def is_css21_color(instance):
- if (
- not isinstance(instance, str_types) or
- instance.lower() in webcolors.css21_names_to_hex
- ):
- return True
- return is_css_color_code(instance)
-
- def is_css3_color(instance):
- if instance.lower() in webcolors.css3_names_to_hex:
- return True
- return is_css_color_code(instance)
-
-
-try:
- import jsonpointer
-except ImportError:
- pass
-else:
- @_checks_drafts(
- draft6="json-pointer",
- draft7="json-pointer",
- raises=jsonpointer.JsonPointerException,
- )
- def is_json_pointer(instance):
- if not isinstance(instance, str_types):
- return True
- return jsonpointer.JsonPointer(instance)
-
- # TODO: I don't want to maintain this, so it
- # needs to go either into jsonpointer (pending
- # https://github.com/stefankoegl/python-json-pointer/issues/34) or
- # into a new external library.
- @_checks_drafts(
- draft7="relative-json-pointer",
- raises=jsonpointer.JsonPointerException,
- )
- def is_relative_json_pointer(instance):
- # Definition taken from:
- # https://tools.ietf.org/html/draft-handrews-relative-json-pointer-01#section-3
- if not isinstance(instance, str_types):
- return True
- non_negative_integer, rest = [], ""
- for i, character in enumerate(instance):
- if character.isdigit():
- non_negative_integer.append(character)
- continue
-
- if not non_negative_integer:
- return False
-
- rest = instance[i:]
- break
- return (rest == "#") or jsonpointer.JsonPointer(rest)
-
-
-try:
- import uritemplate.exceptions
-except ImportError:
- pass
-else:
- @_checks_drafts(
- draft6="uri-template",
- draft7="uri-template",
- raises=uritemplate.exceptions.InvalidTemplate,
- )
- def is_uri_template(
- instance,
- template_validator=uritemplate.Validator().force_balanced_braces(),
- ):
- template = uritemplate.URITemplate(instance)
- return template_validator.validate(template)

View File

@@ -41,20 +41,6 @@ def comma_and(sequence: List[str]) -> str:
return comma_list(sequence, "and")
def ordinal(number: int) -> str:
"""Return the ordinal representation (1st, 2nd, 3rd, etc.) for the provided number.
Args:
number: int to convert to ordinal number
Returns: number's corresponding ordinal
"""
idx = (number % 10) << 1
tens = number % 100 // 10
suffix = "th" if tens == 1 or idx > 6 else "thstndrd"[idx : idx + 2]
return f"{number}{suffix}"
def quote(sequence: List[str], q: str = "'") -> List[str]:
"""Quotes each item in the input list with the quote character passed as second argument."""
return [f"{q}{e}{q}" for e in sequence]

View File

@@ -20,25 +20,15 @@
import tempfile
from contextlib import contextmanager
from itertools import accumulate
from typing import (
Callable,
Deque,
Dict,
Iterable,
List,
Match,
Optional,
Sequence,
Set,
Tuple,
Union,
)
from typing import Callable, Iterable, List, Match, Optional, Tuple, Union
import llnl.util.symlink
from llnl.util import tty
from llnl.util.lang import dedupe, fnmatch_translate_multiple, memoized
from llnl.util.lang import dedupe, memoized
from llnl.util.symlink import islink, readlink, resolve_link_target_relative_to_the_link, symlink
from spack.util.executable import Executable, which
from ..path import path_to_os_path, system_path_filter
if sys.platform != "win32":
@@ -59,11 +49,11 @@
"copy_mode",
"filter_file",
"find",
"find_first",
"find_headers",
"find_all_headers",
"find_libraries",
"find_system_libraries",
"fix_darwin_install_name",
"force_remove",
"force_symlink",
"getuid",
@@ -97,8 +87,6 @@
"visit_directory_tree",
]
Path = Union[str, pathlib.Path]
if sys.version_info < (3, 7, 4):
# monkeypatch shutil.copystat to fix PermissionError when copying read-only
# files on Lustre when using Python < 3.7.4
@@ -260,6 +248,42 @@ def path_contains_subdirectory(path, root):
return norm_path.startswith(norm_root)
@memoized
def file_command(*args):
"""Creates entry point to `file` system command with provided arguments"""
file_cmd = which("file", required=True)
for arg in args:
file_cmd.add_default_arg(arg)
return file_cmd
@memoized
def _get_mime_type():
"""Generate method to call `file` system command to aquire mime type
for a specified path
"""
if sys.platform == "win32":
# -h option (no-dereference) does not exist in Windows
return file_command("-b", "--mime-type")
else:
return file_command("-b", "-h", "--mime-type")
def mime_type(filename):
"""Returns the mime type and subtype of a file.
Args:
filename: file to be analyzed
Returns:
Tuple containing the MIME type and subtype
"""
output = _get_mime_type()(filename, output=str, error=str).strip()
tty.debug("==> " + output)
type, _, subtype = output.partition("/")
return type, subtype
#: This generates the library filenames that may appear on any OS.
library_extensions = ["a", "la", "so", "tbd", "dylib"]
@@ -742,6 +766,7 @@ def copy_tree(
src: str,
dest: str,
symlinks: bool = True,
allow_broken_symlinks: bool = sys.platform != "win32",
ignore: Optional[Callable[[str], bool]] = None,
_permissions: bool = False,
):
@@ -764,6 +789,8 @@ def copy_tree(
src (str): the directory to copy
dest (str): the destination directory
symlinks (bool): whether or not to preserve symlinks
allow_broken_symlinks (bool): whether or not to allow broken (dangling) symlinks,
On Windows, setting this to True will raise an exception. Defaults to true on unix.
ignore (typing.Callable): function indicating which files to ignore
_permissions (bool): for internal use only
@@ -771,6 +798,8 @@ def copy_tree(
IOError: if *src* does not match any files or directories
ValueError: if *src* is a parent directory of *dest*
"""
if allow_broken_symlinks and sys.platform == "win32":
raise llnl.util.symlink.SymlinkError("Cannot allow broken symlinks on Windows!")
if _permissions:
tty.debug("Installing {0} to {1}".format(src, dest))
else:
@@ -843,14 +872,16 @@ def escaped_path(path):
copy_mode(s, d)
for target, d, s in links:
symlink(target, d)
symlink(target, d, allow_broken_symlinks=allow_broken_symlinks)
if _permissions:
set_install_permissions(d)
copy_mode(s, d)
@system_path_filter
def install_tree(src, dest, symlinks=True, ignore=None):
def install_tree(
src, dest, symlinks=True, ignore=None, allow_broken_symlinks=sys.platform != "win32"
):
"""Recursively install an entire directory tree rooted at *src*.
Same as :py:func:`copy_tree` with the addition of setting proper
@@ -861,12 +892,21 @@ def install_tree(src, dest, symlinks=True, ignore=None):
dest (str): the destination directory
symlinks (bool): whether or not to preserve symlinks
ignore (typing.Callable): function indicating which files to ignore
allow_broken_symlinks (bool): whether or not to allow broken (dangling) symlinks,
On Windows, setting this to True will raise an exception.
Raises:
IOError: if *src* does not match any files or directories
ValueError: if *src* is a parent directory of *dest*
"""
copy_tree(src, dest, symlinks=symlinks, ignore=ignore, _permissions=True)
copy_tree(
src,
dest,
symlinks=symlinks,
allow_broken_symlinks=allow_broken_symlinks,
ignore=ignore,
_permissions=True,
)
@system_path_filter
@@ -1600,12 +1640,6 @@ def remove_linked_tree(path):
shutil.rmtree(os.path.realpath(path), **kwargs)
os.unlink(path)
else:
if sys.platform == "win32":
# Adding this prefix allows shutil to remove long paths on windows
# https://learn.microsoft.com/en-us/windows/win32/fileio/maximum-file-path-limitation?tabs=registry
long_path_pfx = "\\\\?\\"
if not path.startswith(long_path_pfx):
path = long_path_pfx + path
shutil.rmtree(path, **kwargs)
@@ -1655,6 +1689,41 @@ def safe_remove(*files_or_dirs):
raise
@system_path_filter
def fix_darwin_install_name(path):
"""Fix install name of dynamic libraries on Darwin to have full path.
There are two parts of this task:
1. Use ``install_name('-id', ...)`` to change install name of a single lib
2. Use ``install_name('-change', ...)`` to change the cross linking between
libs. The function assumes that all libraries are in one folder and
currently won't follow subfolders.
Parameters:
path (str): directory in which .dylib files are located
"""
libs = glob.glob(join_path(path, "*.dylib"))
for lib in libs:
# fix install name first:
install_name_tool = Executable("install_name_tool")
install_name_tool("-id", lib, lib)
otool = Executable("otool")
long_deps = otool("-L", lib, output=str).split("\n")
deps = [dep.partition(" ")[0][1::] for dep in long_deps[2:-1]]
# fix all dependencies:
for dep in deps:
for loc in libs:
# We really want to check for either
# dep == os.path.basename(loc) or
# dep == join_path(builddir, os.path.basename(loc)),
# but we don't know builddir (nor how symbolic links look
# in builddir). We thus only compare the basenames.
if os.path.basename(dep) == os.path.basename(loc):
install_name_tool("-change", dep, loc, lib)
break
def find_first(root: str, files: Union[Iterable[str], str], bfs_depth: int = 2) -> Optional[str]:
"""Find the first file matching a pattern.
@@ -1687,203 +1756,105 @@ def find_first(root: str, files: Union[Iterable[str], str], bfs_depth: int = 2)
return FindFirstFile(root, *files, bfs_depth=bfs_depth).find()
def find(
root: Union[Path, Sequence[Path]],
files: Union[str, Sequence[str]],
recursive: bool = True,
max_depth: Optional[int] = None,
) -> List[str]:
"""Finds all files matching the patterns from ``files`` starting from ``root``. This function
returns a deterministic result for the same input and directory structure when run multiple
times. Symlinked directories are followed, and unique directories are searched only once. Each
matching file is returned only once at lowest depth in case multiple paths exist due to
symlinked directories.
def find(root, files, recursive=True):
"""Search for ``files`` starting from the ``root`` directory.
Like GNU/BSD find but written entirely in Python.
Examples:
.. code-block:: console
$ find /usr -name python
is equivalent to:
>>> find('/usr', 'python')
.. code-block:: console
$ find /usr/local/bin -maxdepth 1 -name python
is equivalent to:
>>> find('/usr/local/bin', 'python', recursive=False)
Accepts any glob characters accepted by fnmatch:
========== ====================================
Pattern Meaning
========== ====================================
``*`` matches one or more characters
``*`` matches everything
``?`` matches any single character
``[seq]`` matches any character in ``seq``
``[!seq]`` matches any character not in ``seq``
========== ====================================
Examples:
>>> find("/usr", "*.txt", recursive=True, max_depth=2)
finds all files with the extension ``.txt`` in the directory ``/usr`` and subdirectories up to
depth 2.
>>> find(["/usr", "/var"], ["*.txt", "*.log"], recursive=True)
finds all files with the extension ``.txt`` or ``.log`` in the directories ``/usr`` and
``/var`` at any depth.
>>> find("/usr", "GL/*.h", recursive=True)
finds all header files in a directory GL at any depth in the directory ``/usr``.
Parameters:
root: One or more root directories to start searching from
files: One or more filename patterns to search for
recursive: if False search only root, if True descends from roots. Defaults to True.
max_depth: if set, don't search below this depth. Cannot be set if recursive is False
root (str): The root directory to start searching from
files (str or collections.abc.Sequence): Library name(s) to search for
recursive (bool): if False search only root folder,
if True descends top-down from the root. Defaults to True.
Returns a list of absolute, matching file paths.
Returns:
list: The files that have been found
"""
if isinstance(root, (str, pathlib.Path)):
root = [root]
elif not isinstance(root, collections.abc.Sequence):
raise TypeError(f"'root' arg must be a path or a sequence of paths, not '{type(root)}']")
if isinstance(files, str):
files = [files]
elif not isinstance(files, collections.abc.Sequence):
raise TypeError(f"'files' arg must be str or a sequence of str, not '{type(files)}']")
# If recursive is false, max_depth can only be None or 0
if max_depth and not recursive:
raise ValueError(f"max_depth ({max_depth}) cannot be set if recursive is False")
if recursive:
tty.debug(f"Find (recursive): {root} {str(files)}")
result = _find_recursive(root, files)
else:
tty.debug(f"Find (not recursive): {root} {str(files)}")
result = _find_non_recursive(root, files)
tty.debug(f"Find (max depth = {max_depth}): {root} {files}")
if not recursive:
max_depth = 0
elif max_depth is None:
max_depth = sys.maxsize
result = _find_max_depth(root, files, max_depth)
tty.debug(f"Find complete: {root} {files}")
tty.debug(f"Find complete: {root} {str(files)}")
return result
def _log_file_access_issue(e: OSError, path: str) -> None:
errno_name = errno.errorcode.get(e.errno, "UNKNOWN")
tty.debug(f"find must skip {path}: {errno_name} {e}")
@system_path_filter
def _find_recursive(root, search_files):
# The variable here is **on purpose** a defaultdict. The idea is that
# we want to poke the filesystem as little as possible, but still maintain
# stability in the order of the answer. Thus we are recording each library
# found in a key, and reconstructing the stable order later.
found_files = collections.defaultdict(list)
# Make the path absolute to have os.walk also return an absolute path
root = os.path.abspath(root)
for path, _, list_files in os.walk(root):
for search_file in search_files:
matches = glob.glob(os.path.join(path, search_file))
matches = [os.path.join(path, x) for x in matches]
found_files[search_file].extend(matches)
answer = []
for search_file in search_files:
answer.extend(found_files[search_file])
return answer
def _file_id(s: os.stat_result) -> Tuple[int, int]:
# Note: on windows, st_ino is the file index and st_dev is the volume serial number. See
# https://github.com/python/cpython/blob/3.9/Python/fileutils.c
return (s.st_ino, s.st_dev)
@system_path_filter
def _find_non_recursive(root, search_files):
# The variable here is **on purpose** a defaultdict as os.list_dir
# can return files in any order (does not preserve stability)
found_files = collections.defaultdict(list)
# Make the path absolute to have absolute path returned
root = os.path.abspath(root)
def _dedupe_files(paths: List[str]) -> List[str]:
"""Deduplicate files by inode and device, dropping files that cannot be accessed."""
unique_files: List[str] = []
# tuple of (inode, device) for each file without following symlinks
visited: Set[Tuple[int, int]] = set()
for path in paths:
try:
stat_info = os.lstat(path)
except OSError as e:
_log_file_access_issue(e, path)
continue
file_id = _file_id(stat_info)
if file_id not in visited:
unique_files.append(path)
visited.add(file_id)
return unique_files
for search_file in search_files:
matches = glob.glob(os.path.join(root, search_file))
matches = [os.path.join(root, x) for x in matches]
found_files[search_file].extend(matches)
answer = []
for search_file in search_files:
answer.extend(found_files[search_file])
def _find_max_depth(
roots: Sequence[Path], globs: Sequence[str], max_depth: int = sys.maxsize
) -> List[str]:
"""See ``find`` for the public API."""
# We optimize for the common case of simple filename only patterns: a single, combined regex
# is used. For complex patterns that include path components, we use a slower glob call from
# every directory we visit within max_depth.
filename_only_patterns = {
f"pattern_{i}": os.path.normcase(x) for i, x in enumerate(globs) if "/" not in x
}
complex_patterns = {f"pattern_{i}": x for i, x in enumerate(globs) if "/" in x}
regex = re.compile(fnmatch_translate_multiple(filename_only_patterns))
# Ordered dictionary that keeps track of what pattern found which files
matched_paths: Dict[str, List[str]] = {f"pattern_{i}": [] for i, _ in enumerate(globs)}
# Ensure returned paths are always absolute
roots = [os.path.abspath(r) for r in roots]
# Breadth-first search queue. Each element is a tuple of (depth, dir)
dir_queue: Deque[Tuple[int, str]] = collections.deque()
# Set of visited directories. Each element is a tuple of (inode, device)
visited_dirs: Set[Tuple[int, int]] = set()
for root in roots:
try:
stat_root = os.stat(root)
except OSError as e:
_log_file_access_issue(e, root)
continue
dir_id = _file_id(stat_root)
if dir_id not in visited_dirs:
dir_queue.appendleft((0, root))
visited_dirs.add(dir_id)
while dir_queue:
depth, curr_dir = dir_queue.pop()
try:
dir_iter = os.scandir(curr_dir)
except OSError as e:
_log_file_access_issue(e, curr_dir)
continue
# Use glob.glob for complex patterns.
for pattern_name, pattern in complex_patterns.items():
matched_paths[pattern_name].extend(
path for path in glob.glob(os.path.join(curr_dir, pattern))
)
# List of subdirectories by path and (inode, device) tuple
subdirs: List[Tuple[str, Tuple[int, int]]] = []
with dir_iter:
for dir_entry in dir_iter:
# Match filename only patterns
if filename_only_patterns:
m = regex.match(os.path.normcase(dir_entry.name))
if m:
for pattern_name in filename_only_patterns:
if m.group(pattern_name):
matched_paths[pattern_name].append(dir_entry.path)
break
# Collect subdirectories
if depth >= max_depth:
continue
try:
if not dir_entry.is_dir(follow_symlinks=True):
continue
if sys.platform == "win32":
# Note: st_ino/st_dev on DirEntry.stat are not set on Windows, so we have
# to call os.stat
stat_info = os.stat(dir_entry.path, follow_symlinks=True)
else:
stat_info = dir_entry.stat(follow_symlinks=True)
except OSError as e:
# Possible permission issue, or a symlink that cannot be resolved (ELOOP).
_log_file_access_issue(e, dir_entry.path)
continue
subdirs.append((dir_entry.path, _file_id(stat_info)))
# Enqueue subdirectories in a deterministic order
if subdirs:
subdirs.sort(key=lambda s: os.path.basename(s[0]))
for subdir, subdir_id in subdirs:
if subdir_id not in visited_dirs:
dir_queue.appendleft((depth + 1, subdir))
visited_dirs.add(subdir_id)
# Sort the matched paths for deterministic output
for paths in matched_paths.values():
paths.sort()
all_matching_paths = [path for paths in matched_paths.values() for path in paths]
# We only dedupe files if we have any complex patterns, since only they can match the same file
# multiple times
return _dedupe_files(all_matching_paths) if complex_patterns else all_matching_paths
return answer
# Utilities for libraries and headers
@@ -2322,9 +2293,7 @@ def find_system_libraries(libraries, shared=True):
return libraries_found
def find_libraries(
libraries, root, shared=True, recursive=False, runtime=True, max_depth: Optional[int] = None
):
def find_libraries(libraries, root, shared=True, recursive=False, runtime=True):
"""Returns an iterable of full paths to libraries found in a root dir.
Accepts any glob characters accepted by fnmatch:
@@ -2345,8 +2314,6 @@ def find_libraries(
otherwise for static. Defaults to True.
recursive (bool): if False search only root folder,
if True descends top-down from the root. Defaults to False.
max_depth (int): if set, don't search below this depth. Cannot be set
if recursive is False
runtime (bool): Windows only option, no-op elsewhere. If true,
search for runtime shared libs (.DLL), otherwise, search
for .Lib files. If shared is false, this has no meaning.
@@ -2355,7 +2322,6 @@ def find_libraries(
Returns:
LibraryList: The libraries that have been found
"""
if isinstance(libraries, str):
libraries = [libraries]
elif not isinstance(libraries, collections.abc.Sequence):
@@ -2388,10 +2354,8 @@ def find_libraries(
libraries = ["{0}.{1}".format(lib, suffix) for lib in libraries for suffix in suffixes]
if not recursive:
if max_depth:
raise ValueError(f"max_depth ({max_depth}) cannot be set if recursive is False")
# If not recursive, look for the libraries directly in root
return LibraryList(find(root, libraries, recursive=False))
return LibraryList(find(root, libraries, False))
# To speedup the search for external packages configured e.g. in /usr,
# perform first non-recursive search in root/lib then in root/lib64 and
@@ -2409,7 +2373,7 @@ def find_libraries(
if found_libs:
break
else:
found_libs = find(root, libraries, recursive=True, max_depth=max_depth)
found_libs = find(root, libraries, True)
return LibraryList(found_libs)

View File

@@ -5,20 +5,18 @@
import collections.abc
import contextlib
import fnmatch
import functools
import inspect
import itertools
import os
import re
import sys
import traceback
import typing
import warnings
from datetime import datetime, timedelta
from typing import Callable, Dict, Iterable, List, Tuple, TypeVar
from typing import Any, Callable, Iterable, List, Tuple
# Ignore emacs backups when listing modules
ignore_modules = r"^\.#|~$"
ignore_modules = [r"^\.#", "~$"]
def index_by(objects, *funcs):
@@ -73,7 +71,7 @@ def index_by(objects, *funcs):
if isinstance(f, str):
f = lambda x: getattr(x, funcs[0])
elif isinstance(f, tuple):
f = lambda x: tuple(getattr(x, p, None) for p in funcs[0])
f = lambda x: tuple(getattr(x, p) for p in funcs[0])
result = {}
for o in objects:
@@ -86,6 +84,20 @@ def index_by(objects, *funcs):
return result
def caller_locals():
"""This will return the locals of the *parent* of the caller.
This allows a function to insert variables into its caller's
scope. Yes, this is some black magic, and yes it's useful
for implementing things like depends_on and provides.
"""
# Passing zero here skips line context for speed.
stack = inspect.stack(0)
try:
return stack[2][0].f_locals
finally:
del stack
def attr_setdefault(obj, name, value):
"""Like dict.setdefault, but for objects."""
if not hasattr(obj, name):
@@ -93,6 +105,15 @@ def attr_setdefault(obj, name, value):
return getattr(obj, name)
def has_method(cls, name):
for base in inspect.getmro(cls):
if base is object:
continue
if name in base.__dict__:
return True
return False
def union_dicts(*dicts):
"""Use update() to combine all dicts into one.
@@ -157,22 +178,19 @@ def list_modules(directory, **kwargs):
order."""
list_directories = kwargs.setdefault("directories", True)
ignore = re.compile(ignore_modules)
for name in os.listdir(directory):
if name == "__init__.py":
continue
with os.scandir(directory) as it:
for entry in it:
if entry.name == "__init__.py" or entry.name == "__pycache__":
continue
path = os.path.join(directory, name)
if list_directories and os.path.isdir(path):
init_py = os.path.join(path, "__init__.py")
if os.path.isfile(init_py):
yield name
if (
list_directories
and entry.is_dir()
and os.path.isfile(os.path.join(entry.path, "__init__.py"))
):
yield entry.name
elif entry.name.endswith(".py") and entry.is_file() and not ignore.search(entry.name):
yield entry.name[:-3] # strip .py
elif name.endswith(".py"):
if not any(re.search(pattern, name) for pattern in ignore_modules):
yield re.sub(".py$", "", name)
def decorator_with_or_without_args(decorator):
@@ -219,8 +237,8 @@ def setter(name, value):
value.__name__ = name
setattr(cls, name, value)
if not hasattr(cls, "_cmp_key"):
raise TypeError(f"'{cls.__name__}' doesn't define _cmp_key().")
if not has_method(cls, "_cmp_key"):
raise TypeError("'%s' doesn't define _cmp_key()." % cls.__name__)
setter("__eq__", lambda s, o: (s is o) or (o is not None and s._cmp_key() == o._cmp_key()))
setter("__lt__", lambda s, o: o is not None and s._cmp_key() < o._cmp_key())
@@ -370,8 +388,8 @@ def cd_fun():
TypeError: If the class does not have a ``_cmp_iter`` method
"""
if not hasattr(cls, "_cmp_iter"):
raise TypeError(f"'{cls.__name__}' doesn't define _cmp_iter().")
if not has_method(cls, "_cmp_iter"):
raise TypeError("'%s' doesn't define _cmp_iter()." % cls.__name__)
# comparison operators are implemented in terms of lazy_eq and lazy_lt
def eq(self, other):
@@ -846,32 +864,20 @@ def uniq(sequence):
return uniq_list
def elide_list(line_list: List[str], max_num: int = 10) -> List[str]:
def elide_list(line_list, max_num=10):
"""Takes a long list and limits it to a smaller number of elements,
replacing intervening elements with '...'. For example::
elide_list(["1", "2", "3", "4", "5", "6"], 4)
elide_list([1,2,3,4,5,6], 4)
gives::
["1", "2", "3", "...", "6"]
[1, 2, 3, '...', 6]
"""
if len(line_list) > max_num:
return [*line_list[: max_num - 1], "...", line_list[-1]]
return line_list
if sys.version_info >= (3, 9):
PatternStr = re.Pattern[str]
else:
PatternStr = typing.Pattern[str]
def fnmatch_translate_multiple(named_patterns: Dict[str, str]) -> str:
"""Similar to ``fnmatch.translate``, but takes an ordered dictionary where keys are pattern
names, and values are filename patterns. The output is a regex that matches any of the
patterns in order, and named capture groups are used to identify which pattern matched."""
return "|".join(f"(?P<{n}>{fnmatch.translate(p)})" for n, p in named_patterns.items())
return line_list[: max_num - 1] + ["..."] + line_list[-1:]
else:
return line_list
@contextlib.contextmanager
@@ -886,12 +892,18 @@ class UnhashableArguments(TypeError):
"""Raise when an @memoized function receives unhashable arg or kwarg values."""
T = TypeVar("T")
def enum(**kwargs):
"""Return an enum-like class.
Args:
**kwargs: explicit dictionary of enums
"""
return type("Enum", (object,), kwargs)
def stable_partition(
input_iterable: Iterable[T], predicate_fn: Callable[[T], bool]
) -> Tuple[List[T], List[T]]:
input_iterable: Iterable, predicate_fn: Callable[[Any], bool]
) -> Tuple[List[Any], List[Any]]:
"""Partition the input iterable according to a custom predicate.
Args:
@@ -903,13 +915,12 @@ def stable_partition(
Tuple of the list of elements evaluating to True, and
list of elements evaluating to False.
"""
true_items: List[T] = []
false_items: List[T] = []
true_items, false_items = [], []
for item in input_iterable:
if predicate_fn(item):
true_items.append(item)
else:
false_items.append(item)
continue
false_items.append(item)
return true_items, false_items
@@ -921,21 +932,6 @@ def ensure_last(lst, *elements):
lst.append(lst.pop(lst.index(elt)))
class Const:
"""Class level constant, raises when trying to set the attribute"""
__slots__ = ["value"]
def __init__(self, value):
self.value = value
def __get__(self, instance, owner):
return self.value
def __set__(self, instance, value):
raise TypeError(f"Const value does not support assignment [value={self.value}]")
class TypedMutableSequence(collections.abc.MutableSequence):
"""Base class that behaves like a list, just with a different type.
@@ -995,8 +991,11 @@ def _receive_forwarded(self, context: str, exc: Exception, tb: List[str]):
def grouped_message(self, with_tracebacks: bool = True) -> str:
"""Print out an error message coalescing all the forwarded errors."""
each_exception_message = [
"\n\t{0} raised {1}: {2}\n{3}".format(
context, exc.__class__.__name__, exc, f"\n{''.join(tb)}" if with_tracebacks else ""
"{0} raised {1}: {2}{3}".format(
context,
exc.__class__.__name__,
exc,
"\n{0}".format("".join(tb)) if with_tracebacks else "",
)
for context, exc, tb in self.exceptions
]
@@ -1037,42 +1036,3 @@ def __init__(self, callback):
def __get__(self, instance, owner):
return self.callback(owner)
class DeprecatedProperty:
"""Data descriptor to error or warn when a deprecated property is accessed.
Derived classes must define a factory method to return an adaptor for the deprecated
property, if the descriptor is not set to error.
"""
__slots__ = ["name"]
#: 0 - Nothing
#: 1 - Warning
#: 2 - Error
error_lvl = 0
def __init__(self, name: str) -> None:
self.name = name
def __get__(self, instance, owner):
if instance is None:
return self
if self.error_lvl == 1:
warnings.warn(
f"accessing the '{self.name}' property of '{instance}', which is deprecated"
)
elif self.error_lvl == 2:
raise AttributeError(f"cannot access the '{self.name}' attribute of '{instance}'")
return self.factory(instance, owner)
def __set__(self, instance, value):
raise TypeError(
f"the deprecated property '{self.name}' of '{instance}' does not support assignment"
)
def factory(self, instance, owner):
raise NotImplementedError("must be implemented by derived classes")

View File

@@ -8,7 +8,6 @@
import subprocess
import sys
import tempfile
from typing import Union
from llnl.util import lang, tty
@@ -17,66 +16,92 @@
if sys.platform == "win32":
from win32file import CreateHardLink
is_windows = sys.platform == "win32"
def _windows_symlink(
src: str, dst: str, target_is_directory: bool = False, *, dir_fd: Union[int, None] = None
):
"""On Windows with System Administrator privileges this will be a normal symbolic link via
os.symlink. On Windows without privledges the link will be a junction for a directory and a
hardlink for a file. On Windows the various link types are:
Symbolic Link: A link to a file or directory on the same or different volume (drive letter) or
even to a remote file or directory (using UNC in its path). Need System Administrator
privileges to make these.
def symlink(source_path: str, link_path: str, allow_broken_symlinks: bool = not is_windows):
"""
Create a link.
Hard Link: A link to a file on the same volume (drive letter) only. Every file (file's data)
has at least 1 hard link (file's name). But when this method creates a new hard link there will
be 2. Deleting all hard links effectively deletes the file. Don't need System Administrator
privileges.
On non-Windows and Windows with System Administrator
privleges this will be a normal symbolic link via
os.symlink.
Junction: A link to a directory on the same or different volume (drive letter) but not to a
remote directory. Don't need System Administrator privileges."""
source_path = os.path.normpath(src)
On Windows without privledges the link will be a
junction for a directory and a hardlink for a file.
On Windows the various link types are:
Symbolic Link: A link to a file or directory on the
same or different volume (drive letter) or even to
a remote file or directory (using UNC in its path).
Need System Administrator privileges to make these.
Hard Link: A link to a file on the same volume (drive
letter) only. Every file (file's data) has at least 1
hard link (file's name). But when this method creates
a new hard link there will be 2. Deleting all hard
links effectively deletes the file. Don't need System
Administrator privileges.
Junction: A link to a directory on the same or different
volume (drive letter) but not to a remote directory. Don't
need System Administrator privileges.
Parameters:
source_path (str): The real file or directory that the link points to.
Must be absolute OR relative to the link.
link_path (str): The path where the link will exist.
allow_broken_symlinks (bool): On Linux or Mac, don't raise an exception if the source_path
doesn't exist. This will still raise an exception on Windows.
"""
source_path = os.path.normpath(source_path)
win_source_path = source_path
link_path = os.path.normpath(dst)
link_path = os.path.normpath(link_path)
# Perform basic checks to make sure symlinking will succeed
if os.path.lexists(link_path):
raise AlreadyExistsError(f"Link path ({link_path}) already exists. Cannot create link.")
# Never allow broken links on Windows.
if sys.platform == "win32" and allow_broken_symlinks:
raise ValueError("allow_broken_symlinks parameter cannot be True on Windows.")
if not os.path.exists(source_path):
if os.path.isabs(source_path):
# An absolute source path that does not exist will result in a broken link.
raise SymlinkError(
f"Source path ({source_path}) is absolute but does not exist. Resulting "
f"link would be broken so not making link."
if not allow_broken_symlinks:
# Perform basic checks to make sure symlinking will succeed
if os.path.lexists(link_path):
raise AlreadyExistsError(
f"Link path ({link_path}) already exists. Cannot create link."
)
else:
# os.symlink can create a link when the given source path is relative to
# the link path. Emulate this behavior and check to see if the source exists
# relative to the link path ahead of link creation to prevent broken
# links from being made.
link_parent_dir = os.path.dirname(link_path)
relative_path = os.path.join(link_parent_dir, source_path)
if os.path.exists(relative_path):
# In order to work on windows, the source path needs to be modified to be
# relative because hardlink/junction dont resolve relative paths the same
# way as os.symlink. This is ignored on other operating systems.
win_source_path = relative_path
else:
if not os.path.exists(source_path):
if os.path.isabs(source_path) and not allow_broken_symlinks:
# An absolute source path that does not exist will result in a broken link.
raise SymlinkError(
f"The source path ({source_path}) is not relative to the link path "
f"({link_path}). Resulting link would be broken so not making link."
f"Source path ({source_path}) is absolute but does not exist. Resulting "
f"link would be broken so not making link."
)
else:
# os.symlink can create a link when the given source path is relative to
# the link path. Emulate this behavior and check to see if the source exists
# relative to the link path ahead of link creation to prevent broken
# links from being made.
link_parent_dir = os.path.dirname(link_path)
relative_path = os.path.join(link_parent_dir, source_path)
if os.path.exists(relative_path):
# In order to work on windows, the source path needs to be modified to be
# relative because hardlink/junction dont resolve relative paths the same
# way as os.symlink. This is ignored on other operating systems.
win_source_path = relative_path
elif not allow_broken_symlinks:
raise SymlinkError(
f"The source path ({source_path}) is not relative to the link path "
f"({link_path}). Resulting link would be broken so not making link."
)
# Create the symlink
if not _windows_can_symlink():
if sys.platform == "win32" and not _windows_can_symlink():
_windows_create_link(win_source_path, link_path)
else:
os.symlink(source_path, link_path, target_is_directory=os.path.isdir(source_path))
def _windows_islink(path: str) -> bool:
def islink(path: str) -> bool:
"""Override os.islink to give correct answer for spack logic.
For Non-Windows: a link can be determined with the os.path.islink method.
@@ -244,7 +269,7 @@ def _windows_create_hard_link(path: str, link: str):
CreateHardLink(link, path)
def _windows_readlink(path: str, *, dir_fd=None):
def readlink(path: str, *, dir_fd=None):
"""Spack utility to override of os.readlink method to work cross platform"""
if _windows_is_hardlink(path):
return _windows_read_hard_link(path)
@@ -313,16 +338,6 @@ def resolve_link_target_relative_to_the_link(link):
return os.path.join(link_dir, target)
if sys.platform == "win32":
symlink = _windows_symlink
readlink = _windows_readlink
islink = _windows_islink
else:
symlink = os.symlink
readlink = os.readlink
islink = os.path.islink
class SymlinkError(RuntimeError):
"""Exception class for errors raised while creating symlinks,
junctions and hard links

View File

@@ -263,9 +263,7 @@ def match_to_ansi(match):
f"Incomplete color format: '{match.group(0)}' in '{match.string}'"
)
color_number = colors.get(color_code, "")
semi = ";" if color_number else ""
ansi_code = _escape(f"{styles[style]}{semi}{color_number}", color, enclose, zsh)
ansi_code = _escape(f"{styles[style]};{colors.get(color_code, '')}", color, enclose, zsh)
if text:
return f"{ansi_code}{text}{_escape(0, color, enclose, zsh)}"
else:

View File

@@ -18,10 +18,9 @@
import threading
import traceback
from contextlib import contextmanager
from multiprocessing.connection import Connection
from threading import Thread
from types import ModuleType
from typing import Callable, Optional
from typing import Optional
import llnl.util.tty as tty
@@ -34,23 +33,8 @@
pass
esc, bell, lbracket, bslash, newline = r"\x1b", r"\x07", r"\[", r"\\", r"\n"
# Ansi Control Sequence Introducers (CSI) are a well-defined format
# Standard ECMA-48: Control Functions for Character-Imaging I/O Devices, section 5.4
# https://www.ecma-international.org/wp-content/uploads/ECMA-48_5th_edition_june_1991.pdf
csi_pre = f"{esc}{lbracket}"
csi_param, csi_inter, csi_post = r"[0-?]", r"[ -/]", r"[@-~]"
ansi_csi = f"{csi_pre}{csi_param}*{csi_inter}*{csi_post}"
# General ansi escape sequences have well-defined prefixes,
# but content and suffixes are less reliable.
# Conservatively assume they end with either "<ESC>\" or "<BELL>",
# with no intervening "<ESC>"/"<BELL>" keys or newlines
esc_pre = f"{esc}[@-_]"
esc_content = f"[^{esc}{bell}{newline}]"
esc_post = f"(?:{esc}{bslash}|{bell})"
ansi_esc = f"{esc_pre}{esc_content}*{esc_post}"
# Use this to strip escape sequences
_escape = re.compile(f"{ansi_csi}|{ansi_esc}")
_escape = re.compile(r"\x1b[^m]*m|\x1b\[?1034h|\x1b\][0-9]+;[^\x07]*\x07")
# control characters for enabling/disabling echo
#
@@ -345,6 +329,49 @@ def close(self):
self.file.close()
class MultiProcessFd:
"""Return an object which stores a file descriptor and can be passed as an
argument to a function run with ``multiprocessing.Process``, such that
the file descriptor is available in the subprocess."""
def __init__(self, fd):
self._connection = None
self._fd = None
if sys.version_info >= (3, 8):
self._connection = multiprocessing.connection.Connection(fd)
else:
self._fd = fd
@property
def fd(self):
if self._connection:
return self._connection._handle
else:
return self._fd
def close(self):
if self._connection:
self._connection.close()
else:
os.close(self._fd)
def close_connection_and_file(multiprocess_fd, file):
# MultiprocessFd is intended to transmit a FD
# to a child process, this FD is then opened to a Python File object
# (using fdopen). In >= 3.8, MultiprocessFd encapsulates a
# multiprocessing.connection.Connection; Connection closes the FD
# when it is deleted, and prints a warning about duplicate closure if
# it is not explicitly closed. In < 3.8, MultiprocessFd encapsulates a
# simple FD; closing the FD here appears to conflict with
# closure of the File object (in < 3.8 that is). Therefore this needs
# to choose whether to close the File or the Connection.
if sys.version_info >= (3, 8):
multiprocess_fd.close()
else:
file.close()
@contextmanager
def replace_environment(env):
"""Replace the current environment (`os.environ`) with `env`.
@@ -502,20 +529,22 @@ def __enter__(self):
# forcing debug output.
self._saved_debug = tty._debug
# Pipe for redirecting output to logger
read_fd, self.write_fd = multiprocessing.Pipe(duplex=False)
# OS-level pipe for redirecting output to logger
read_fd, write_fd = os.pipe()
# Pipe for communication back from the daemon
read_multiprocess_fd = MultiProcessFd(read_fd)
# Multiprocessing pipe for communication back from the daemon
# Currently only used to save echo value between uses
self.parent_pipe, child_pipe = multiprocessing.Pipe(duplex=False)
self.parent_pipe, child_pipe = multiprocessing.Pipe()
# Sets a daemon that writes to file what it reads from a pipe
try:
# need to pass this b/c multiprocessing closes stdin in child.
input_fd = None
input_multiprocess_fd = None
try:
if sys.stdin.isatty():
input_fd = Connection(os.dup(sys.stdin.fileno()))
input_multiprocess_fd = MultiProcessFd(os.dup(sys.stdin.fileno()))
except BaseException:
# just don't forward input if this fails
pass
@@ -524,9 +553,9 @@ def __enter__(self):
self.process = multiprocessing.Process(
target=_writer_daemon,
args=(
input_fd,
read_fd,
self.write_fd,
input_multiprocess_fd,
read_multiprocess_fd,
write_fd,
self.echo,
self.log_file,
child_pipe,
@@ -537,9 +566,9 @@ def __enter__(self):
self.process.start()
finally:
if input_fd:
input_fd.close()
read_fd.close()
if input_multiprocess_fd:
input_multiprocess_fd.close()
read_multiprocess_fd.close()
# Flush immediately before redirecting so that anything buffered
# goes to the original stream
@@ -557,9 +586,9 @@ def __enter__(self):
self._saved_stderr = os.dup(sys.stderr.fileno())
# redirect to the pipe we created above
os.dup2(self.write_fd.fileno(), sys.stdout.fileno())
os.dup2(self.write_fd.fileno(), sys.stderr.fileno())
self.write_fd.close()
os.dup2(write_fd, sys.stdout.fileno())
os.dup2(write_fd, sys.stderr.fileno())
os.close(write_fd)
else:
# Handle I/O the Python way. This won't redirect lower-level
@@ -572,7 +601,7 @@ def __enter__(self):
self._saved_stderr = sys.stderr
# create a file object for the pipe; redirect to it.
pipe_fd_out = os.fdopen(self.write_fd.fileno(), "w", closefd=False)
pipe_fd_out = os.fdopen(write_fd, "w")
sys.stdout = pipe_fd_out
sys.stderr = pipe_fd_out
@@ -608,7 +637,6 @@ def __exit__(self, exc_type, exc_val, exc_tb):
else:
sys.stdout = self._saved_stdout
sys.stderr = self._saved_stderr
self.write_fd.close()
# print log contents in parent if needed.
if self.log_file.write_in_parent:
@@ -822,14 +850,14 @@ def force_echo(self):
def _writer_daemon(
stdin_fd: Optional[Connection],
read_fd: Connection,
write_fd: Connection,
echo: bool,
log_file_wrapper: FileWrapper,
control_fd: Connection,
filter_fn: Optional[Callable[[str], str]],
) -> None:
stdin_multiprocess_fd,
read_multiprocess_fd,
write_fd,
echo,
log_file_wrapper,
control_pipe,
filter_fn,
):
"""Daemon used by ``log_output`` to write to a log file and to ``stdout``.
The daemon receives output from the parent process and writes it both
@@ -866,37 +894,43 @@ def _writer_daemon(
``StringIO`` in the parent. This is mainly for testing.
Arguments:
stdin_fd: optional input from the terminal
read_fd: pipe for reading from parent's redirected stdout
echo: initial echo setting -- controlled by user and preserved across multiple writer
daemons
log_file_wrapper: file to log all output
control_pipe: multiprocessing pipe on which to send control information to the parent
filter_fn: optional function to filter each line of output
stdin_multiprocess_fd (int): input from the terminal
read_multiprocess_fd (int): pipe for reading from parent's redirected
stdout
echo (bool): initial echo setting -- controlled by user and
preserved across multiple writer daemons
log_file_wrapper (FileWrapper): file to log all output
control_pipe (Pipe): multiprocessing pipe on which to send control
information to the parent
filter_fn (callable, optional): function to filter each line of output
"""
# This process depends on closing all instances of write_pipe to terminate the reading loop
write_fd.close()
# If this process was forked, then it will inherit file descriptors from
# the parent process. This process depends on closing all instances of
# write_fd to terminate the reading loop, so we close the file descriptor
# here. Forking is the process spawning method everywhere except Mac OS
# for Python >= 3.8 and on Windows
if sys.version_info < (3, 8) or sys.platform != "darwin":
os.close(write_fd)
# 1. Use line buffering (3rd param = 1) since Python 3 has a bug
# that prevents unbuffered text I/O.
# 2. Python 3.x before 3.7 does not open with UTF-8 encoding by default
# 3. closefd=False because Connection has "ownership"
read_file = os.fdopen(read_fd.fileno(), "r", 1, encoding="utf-8", closefd=False)
in_pipe = os.fdopen(read_multiprocess_fd.fd, "r", 1, encoding="utf-8")
if stdin_fd:
stdin_file = os.fdopen(stdin_fd.fileno(), closefd=False)
if stdin_multiprocess_fd:
stdin = os.fdopen(stdin_multiprocess_fd.fd)
else:
stdin_file = None
stdin = None
# list of streams to select from
istreams = [read_file, stdin_file] if stdin_file else [read_file]
istreams = [in_pipe, stdin] if stdin else [in_pipe]
force_echo = False # parent can force echo for certain output
log_file = log_file_wrapper.unwrap()
try:
with keyboard_input(stdin_file) as kb:
with keyboard_input(stdin) as kb:
while True:
# fix the terminal settings if we recently came to
# the foreground
@@ -909,12 +943,12 @@ def _writer_daemon(
# Allow user to toggle echo with 'v' key.
# Currently ignores other chars.
# only read stdin if we're in the foreground
if stdin_file and stdin_file in rlist and not _is_background_tty(stdin_file):
if stdin in rlist and not _is_background_tty(stdin):
# it's possible to be backgrounded between the above
# check and the read, so we ignore SIGTTIN here.
with ignore_signal(signal.SIGTTIN):
try:
if stdin_file.read(1) == "v":
if stdin.read(1) == "v":
echo = not echo
except IOError as e:
# If SIGTTIN is ignored, the system gives EIO
@@ -923,13 +957,13 @@ def _writer_daemon(
if e.errno != errno.EIO:
raise
if read_file in rlist:
if in_pipe in rlist:
line_count = 0
try:
while line_count < 100:
# Handle output from the calling process.
try:
line = _retry(read_file.readline)()
line = _retry(in_pipe.readline)()
except UnicodeDecodeError:
# installs like --test=root gpgme produce non-UTF8 logs
line = "<line lost: output was not encoded as UTF-8>\n"
@@ -958,7 +992,7 @@ def _writer_daemon(
if xoff in controls:
force_echo = False
if not _input_available(read_file):
if not _input_available(in_pipe):
break
finally:
if line_count > 0:
@@ -973,14 +1007,14 @@ def _writer_daemon(
finally:
# send written data back to parent if we used a StringIO
if isinstance(log_file, io.StringIO):
control_fd.send(log_file.getvalue())
control_pipe.send(log_file.getvalue())
log_file_wrapper.close()
read_fd.close()
if stdin_fd:
stdin_fd.close()
close_connection_and_file(read_multiprocess_fd, in_pipe)
if stdin_multiprocess_fd:
close_connection_and_file(stdin_multiprocess_fd, stdin)
# send echo value back to the parent so it can be preserved.
control_fd.send(echo)
control_pipe.send(echo)
def _retry(function):

View File

@@ -3,15 +3,8 @@
#
# SPDX-License-Identifier: (Apache-2.0 OR MIT)
import os
import re
from typing import Optional
import spack.paths
import spack.util.git
#: PEP440 canonical <major>.<minor>.<micro>.<devN> string
__version__ = "1.0.0-alpha.1"
__version__ = "0.22.2"
spack_version = __version__
@@ -26,58 +19,4 @@ def __try_int(v):
spack_version_info = tuple([__try_int(v) for v in __version__.split(".")])
def get_spack_commit() -> Optional[str]:
"""Get the Spack git commit sha.
Returns:
(str or None) the commit sha if available, otherwise None
"""
git_path = os.path.join(spack.paths.prefix, ".git")
if not os.path.exists(git_path):
return None
git = spack.util.git.git()
if not git:
return None
rev = git(
"-C",
spack.paths.prefix,
"rev-parse",
"HEAD",
output=str,
error=os.devnull,
fail_on_error=False,
)
if git.returncode != 0:
return None
match = re.match(r"[a-f\d]{7,}$", rev)
return match.group(0) if match else None
def get_version() -> str:
"""Get a descriptive version of this instance of Spack.
Outputs '<PEP440 version> (<git commit sha>)'.
The commit sha is only added when available.
"""
commit = get_spack_commit()
if commit:
return f"{spack_version} ({commit})"
return spack_version
def get_short_version() -> str:
"""Short Spack version."""
return f"{spack_version_info[0]}.{spack_version_info[1]}"
__all__ = [
"spack_version_info",
"spack_version",
"get_version",
"get_spack_commit",
"get_short_version",
]
__all__ = ["spack_version_info", "spack_version"]

131
lib/spack/spack/abi.py Normal file
View File

@@ -0,0 +1,131 @@
# Copyright 2013-2024 Lawrence Livermore National Security, LLC and other
# Spack Project Developers. See the top-level COPYRIGHT file for details.
#
# SPDX-License-Identifier: (Apache-2.0 OR MIT)
import os
from llnl.util.lang import memoized
import spack.spec
import spack.version
from spack.compilers.clang import Clang
from spack.util.executable import Executable, ProcessError
class ABI:
"""This class provides methods to test ABI compatibility between specs.
The current implementation is rather rough and could be improved."""
def architecture_compatible(
self, target: spack.spec.Spec, constraint: spack.spec.Spec
) -> bool:
"""Return true if architecture of target spec is ABI compatible
to the architecture of constraint spec. If either the target
or constraint specs have no architecture, target is also defined
as architecture ABI compatible to constraint."""
return (
not target.architecture
or not constraint.architecture
or target.architecture.intersects(constraint.architecture)
)
@memoized
def _gcc_get_libstdcxx_version(self, version):
"""Returns gcc ABI compatibility info by getting the library version of
a compiler's libstdc++ or libgcc_s"""
from spack.build_environment import dso_suffix
spec = spack.spec.CompilerSpec("gcc", version)
compilers = spack.compilers.compilers_for_spec(spec)
if not compilers:
return None
compiler = compilers[0]
rungcc = None
libname = None
output = None
if compiler.cxx:
rungcc = Executable(compiler.cxx)
libname = "libstdc++." + dso_suffix
elif compiler.cc:
rungcc = Executable(compiler.cc)
libname = "libgcc_s." + dso_suffix
else:
return None
try:
# Some gcc's are actually clang and don't respond properly to
# --print-file-name (they just print the filename, not the
# full path). Ignore these and expect them to be handled as clang.
if Clang.default_version(rungcc.exe[0]) != "unknown":
return None
output = rungcc("--print-file-name=%s" % libname, output=str)
except ProcessError:
return None
if not output:
return None
libpath = os.path.realpath(output.strip())
if not libpath:
return None
return os.path.basename(libpath)
@memoized
def _gcc_compiler_compare(self, pversion, cversion):
"""Returns true iff the gcc version pversion and cversion
are ABI compatible."""
plib = self._gcc_get_libstdcxx_version(pversion)
clib = self._gcc_get_libstdcxx_version(cversion)
if not plib or not clib:
return False
return plib == clib
def _intel_compiler_compare(
self, pversion: spack.version.ClosedOpenRange, cversion: spack.version.ClosedOpenRange
) -> bool:
"""Returns true iff the intel version pversion and cversion
are ABI compatible"""
# Test major and minor versions. Ignore build version.
pv = pversion.lo
cv = cversion.lo
return pv.up_to(2) == cv.up_to(2)
def compiler_compatible(
self, parent: spack.spec.Spec, child: spack.spec.Spec, loose: bool = False
) -> bool:
"""Return true if compilers for parent and child are ABI compatible."""
if not parent.compiler or not child.compiler:
return True
if parent.compiler.name != child.compiler.name:
# Different compiler families are assumed ABI incompatible
return False
if loose:
return True
# TODO: Can we move the specialized ABI matching stuff
# TODO: into compiler classes?
for pversion in parent.compiler.versions:
for cversion in child.compiler.versions:
# For a few compilers use specialized comparisons.
# Otherwise match on version match.
if pversion.intersects(cversion):
return True
elif parent.compiler.name == "gcc" and self._gcc_compiler_compare(
pversion, cversion
):
return True
elif parent.compiler.name == "intel" and self._intel_compiler_compare(
pversion, cversion
):
return True
return False
def compatible(
self, target: spack.spec.Spec, constraint: spack.spec.Spec, loose: bool = False
) -> bool:
"""Returns true if target spec is ABI compatible to constraint spec"""
return self.architecture_compatible(target, constraint) and self.compiler_compatible(
target, constraint, loose=loose
)

View File

@@ -42,21 +42,15 @@ def _search_duplicate_compilers(error_cls):
import inspect
import io
import itertools
import os
import pathlib
import pickle
import re
import warnings
from typing import Iterable, List, Set, Tuple
from urllib.request import urlopen
import llnl.util.lang
from llnl.string import plural
import spack.builder
import spack.config
import spack.deptypes
import spack.fetch_strategy
import spack.patch
import spack.repo
import spack.spec
@@ -79,9 +73,7 @@ def __init__(self, summary, details):
self.details = tuple(details)
def __str__(self):
if self.details:
return f"{self.summary}\n" + "\n".join(f" {detail}" for detail in self.details)
return self.summary
return self.summary + "\n" + "\n".join([" " + detail for detail in self.details])
def __eq__(self, other):
if self.summary != other.summary or self.details != other.details:
@@ -218,11 +210,6 @@ def _search_duplicate_compilers(error_cls):
group="configs", tag="CFG-PACKAGES", description="Sanity checks on packages.yaml", kwargs=()
)
#: Sanity checks on packages.yaml
config_repos = AuditClass(
group="configs", tag="CFG-REPOS", description="Sanity checks on repositories", kwargs=()
)
@config_packages
def _search_duplicate_specs_in_externals(error_cls):
@@ -265,6 +252,40 @@ def _search_duplicate_specs_in_externals(error_cls):
return errors
@config_packages
def _deprecated_preferences(error_cls):
"""Search package preferences deprecated in v0.21 (and slated for removal in v0.23)"""
# TODO (v0.23): remove this audit as the attributes will not be allowed in config
errors = []
packages_yaml = spack.config.CONFIG.get_config("packages")
def make_error(attribute_name, config_data, summary):
s = io.StringIO()
s.write("Occurring in the following file:\n")
dict_view = syaml.syaml_dict((k, v) for k, v in config_data.items() if k == attribute_name)
syaml.dump_config(dict_view, stream=s, blame=True)
return error_cls(summary=summary, details=[s.getvalue()])
if "all" in packages_yaml and "version" in packages_yaml["all"]:
summary = "Using the deprecated 'version' attribute under 'packages:all'"
errors.append(make_error("version", packages_yaml["all"], summary))
for package_name in packages_yaml:
if package_name == "all":
continue
package_conf = packages_yaml[package_name]
for attribute in ("compiler", "providers", "target"):
if attribute not in package_conf:
continue
summary = (
f"Using the deprecated '{attribute}' attribute " f"under 'packages:{package_name}'"
)
errors.append(make_error(attribute, package_conf, summary))
return errors
@config_packages
def _avoid_mismatched_variants(error_cls):
"""Warns if variant preferences have mismatched types or names."""
@@ -285,7 +306,7 @@ def _avoid_mismatched_variants(error_cls):
pkg_cls = spack.repo.PATH.get_pkg_class(pkg_name)
for variant in current_spec.variants.values():
# Variant does not exist at all
if variant.name not in pkg_cls.variant_names():
if variant.name not in pkg_cls.variants:
summary = (
f"Setting a preference for the '{pkg_name}' package to the "
f"non-existing variant '{variant.name}'"
@@ -294,8 +315,9 @@ def _avoid_mismatched_variants(error_cls):
continue
# Variant cannot accept this value
s = spack.spec.Spec(pkg_name)
try:
spack.variant.prevalidate_variant_value(pkg_cls, variant, strict=True)
s.update_variant_validate(variant.name, variant.value)
except Exception:
summary = (
f"Setting the variant '{variant.name}' of the '{pkg_name}' package "
@@ -329,43 +351,6 @@ def _wrongly_named_spec(error_cls):
return errors
@config_packages
def _ensure_all_virtual_packages_have_default_providers(error_cls):
"""All virtual packages must have a default provider explicitly set."""
configuration = spack.config.create()
defaults = configuration.get("packages", scope="defaults")
default_providers = defaults["all"]["providers"]
virtuals = spack.repo.PATH.provider_index.providers
default_providers_filename = configuration.scopes["defaults"].get_section_filename("packages")
return [
error_cls(f"'{virtual}' must have a default provider in {default_providers_filename}", [])
for virtual in virtuals
if virtual not in default_providers
]
@config_repos
def _ensure_no_folders_without_package_py(error_cls):
"""Check that we don't leave any folder without a package.py in repos"""
errors = []
for repository in spack.repo.PATH.repos:
missing = []
for entry in os.scandir(repository.packages_path):
if not entry.is_dir():
continue
package_py = pathlib.Path(entry.path) / spack.repo.package_file_name
if not package_py.exists():
missing.append(entry.path)
if missing:
summary = (
f"The '{repository.namespace}' repository misses a package.py file"
f" in the following folders"
)
errors.append(error_cls(summary=summary, details=[f"{x}" for x in missing]))
return errors
def _make_config_error(config_data, summary, error_cls):
s = io.StringIO()
s.write("Occurring in the following file:\n")
@@ -389,14 +374,6 @@ def _make_config_error(config_data, summary, error_cls):
)
package_deprecated_attributes = AuditClass(
group="packages",
tag="PKG-DEPRECATED-ATTRIBUTES",
description="Sanity checks to preclude use of deprecated package attributes",
kwargs=("pkgs",),
)
package_properties = AuditClass(
group="packages",
tag="PKG-PROPERTIES",
@@ -415,23 +392,22 @@ def _make_config_error(config_data, summary, error_cls):
)
@package_properties
@package_directives
def _check_build_test_callbacks(pkgs, error_cls):
"""Ensure stand-alone test methods are not included in build-time callbacks.
Test methods are for checking the installed software as stand-alone tests.
They could also be called during the post-install phase of a build.
"""
"""Ensure stand-alone test method is not included in build-time callbacks"""
errors = []
for pkg_name in pkgs:
pkg_cls = spack.repo.PATH.get_pkg_class(pkg_name)
test_callbacks = getattr(pkg_cls, "build_time_test_callbacks", None)
has_test_method = test_callbacks and any([m.startswith("test_") for m in test_callbacks])
# TODO (post-34236): "test*"->"test_*" once remove deprecated methods
# TODO (post-34236): "test"->"test_" once remove deprecated methods
has_test_method = test_callbacks and any([m.startswith("test") for m in test_callbacks])
if has_test_method:
msg = f"Package {pkg_name} includes stand-alone test methods in build-time checks."
callbacks = ", ".join(test_callbacks)
instr = f"Remove the following from 'build_time_test_callbacks': {callbacks}"
msg = '{0} package contains "test*" method(s) in ' "build_time_test_callbacks"
instr = 'Remove all methods whose names start with "test" from: [{0}]'.format(
", ".join(test_callbacks)
)
errors.append(error_cls(msg.format(pkg_name), [instr]))
return errors
@@ -445,10 +421,6 @@ def _check_patch_urls(pkgs, error_cls):
r"^https?://(?:patch-diff\.)?github(?:usercontent)?\.com/"
r".+/.+/(?:commit|pull)/[a-fA-F0-9]+\.(?:patch|diff)"
)
github_pull_commits_re = (
r"^https?://(?:patch-diff\.)?github(?:usercontent)?\.com/"
r".+/.+/pull/\d+/commits/[a-fA-F0-9]+\.(?:patch|diff)"
)
# Only .diff URLs have stable/full hashes:
# https://forum.gitlab.com/t/patches-with-full-index/29313
gitlab_patch_url_re = (
@@ -464,24 +436,14 @@ def _check_patch_urls(pkgs, error_cls):
if not isinstance(patch, spack.patch.UrlPatch):
continue
if re.match(github_pull_commits_re, patch.url):
url = re.sub(r"/pull/\d+/commits/", r"/commit/", patch.url)
url = re.sub(r"^(.*)(?<!full_index=1)$", r"\1?full_index=1", url)
errors.append(
error_cls(
f"patch URL in package {pkg_cls.name} "
+ "must not be a pull request commit; "
+ f"instead use {url}",
[patch.url],
)
)
elif re.match(github_patch_url_re, patch.url):
if re.match(github_patch_url_re, patch.url):
full_index_arg = "?full_index=1"
if not patch.url.endswith(full_index_arg):
errors.append(
error_cls(
f"patch URL in package {pkg_cls.name} "
+ f"must end with {full_index_arg}",
"patch URL in package {0} must end with {1}".format(
pkg_cls.name, full_index_arg
),
[patch.url],
)
)
@@ -489,7 +451,9 @@ def _check_patch_urls(pkgs, error_cls):
if not patch.url.endswith(".diff"):
errors.append(
error_cls(
f"patch URL in package {pkg_cls.name} must end with .diff",
"patch URL in package {0} must end with .diff".format(
pkg_cls.name
),
[patch.url],
)
)
@@ -506,7 +470,7 @@ def _search_for_reserved_attributes_names_in_packages(pkgs, error_cls):
name_definitions = collections.defaultdict(list)
pkg_cls = spack.repo.PATH.get_pkg_class(pkg_name)
for cls_item in pkg_cls.__mro__:
for cls_item in inspect.getmro(pkg_cls):
for name in RESERVED_NAMES:
current_value = cls_item.__dict__.get(name)
if current_value is None:
@@ -529,58 +493,13 @@ def _search_for_reserved_attributes_names_in_packages(pkgs, error_cls):
return errors
@package_deprecated_attributes
def _search_for_deprecated_package_methods(pkgs, error_cls):
"""Ensure the package doesn't define or use deprecated methods"""
DEPRECATED_METHOD = (("test", "a name starting with 'test_'"),)
DEPRECATED_USE = (
("self.cache_extra_test_sources(", "cache_extra_test_sources(self, ..)"),
("self.install_test_root(", "install_test_root(self, ..)"),
("self.run_test(", "test_part(self, ..)"),
)
errors = []
for pkg_name in pkgs:
pkg_cls = spack.repo.PATH.get_pkg_class(pkg_name)
methods = inspect.getmembers(pkg_cls, predicate=lambda x: inspect.isfunction(x))
method_errors = collections.defaultdict(list)
for name, function in methods:
for deprecated_name, alternate in DEPRECATED_METHOD:
if name == deprecated_name:
msg = f"Rename '{deprecated_name}' method to {alternate} instead."
method_errors[name].append(msg)
source = inspect.getsource(function)
for deprecated_name, alternate in DEPRECATED_USE:
if deprecated_name in source:
msg = f"Change '{deprecated_name}' to '{alternate}' in '{name}' method."
method_errors[name].append(msg)
num_methods = len(method_errors)
if num_methods > 0:
methods = plural(num_methods, "method", show_n=False)
error_msg = (
f"Package '{pkg_name}' implements or uses unsupported deprecated {methods}."
)
instr = [f"Make changes to '{pkg_cls.__module__}':"]
for name in sorted(method_errors):
instr.extend([f" {msg}" for msg in method_errors[name]])
errors.append(error_cls(error_msg, instr))
return errors
@package_properties
def _ensure_all_package_names_are_lowercase(pkgs, error_cls):
"""Ensure package names are lowercase and consistent"""
reserved_names = ("all",)
badname_regex, errors = re.compile(r"[_A-Z]"), []
for pkg_name in pkgs:
if pkg_name in reserved_names:
error_msg = f"The name '{pkg_name}' is reserved, and cannot be used for packages"
errors.append(error_cls(error_msg, []))
if badname_regex.search(pkg_name):
error_msg = f"Package name '{pkg_name}' should be lowercase and must not contain '_'"
error_msg = "Package name '{}' is either lowercase or conatine '_'".format(pkg_name)
errors.append(error_cls(error_msg, []))
return errors
@@ -719,17 +638,12 @@ def _ensure_env_methods_are_ported_to_builders(pkgs, error_cls):
errors = []
for pkg_name in pkgs:
pkg_cls = spack.repo.PATH.get_pkg_class(pkg_name)
# values are either ConditionalValue objects or the values themselves
build_system_names = set(
v.value if isinstance(v, spack.variant.ConditionalValue) else v
for _, variant in pkg_cls.variant_definitions("build_system")
for v in variant.values
)
builder_cls_names = [spack.builder.BUILDER_CLS[x].__name__ for x in build_system_names]
buildsystem_variant, _ = pkg_cls.variants["build_system"]
buildsystem_names = [getattr(x, "value", x) for x in buildsystem_variant.values]
builder_cls_names = [spack.builder.BUILDER_CLS[x].__name__ for x in buildsystem_names]
module = pkg_cls.module
has_builders_in_package_py = any(
spack.builder.get_builder_class(pkg_cls, name) for name in builder_cls_names
getattr(module, name, False) for name in builder_cls_names
)
if not has_builders_in_package_py:
continue
@@ -745,171 +659,6 @@ def _ensure_env_methods_are_ported_to_builders(pkgs, error_cls):
return errors
class DeprecatedMagicGlobals(ast.NodeVisitor):
def __init__(self, magic_globals: Iterable[str]):
super().__init__()
self.magic_globals: Set[str] = set(magic_globals)
# State to track whether we're in a class function
self.depth: int = 0
self.in_function: bool = False
self.path = (ast.Module, ast.ClassDef, ast.FunctionDef)
# Defined locals in the current function (heuristically at least)
self.locals: Set[str] = set()
# List of (name, lineno) tuples for references to magic globals
self.references_to_globals: List[Tuple[str, int]] = []
def descend_in_function_def(self, node: ast.AST) -> None:
if not isinstance(node, self.path[self.depth]):
return
self.depth += 1
if self.depth == len(self.path):
self.in_function = True
super().generic_visit(node)
if self.depth == len(self.path):
self.in_function = False
self.locals.clear()
self.depth -= 1
def generic_visit(self, node: ast.AST) -> None:
# Recurse into function definitions
if self.depth < len(self.path):
return self.descend_in_function_def(node)
elif not self.in_function:
return
elif isinstance(node, ast.Global):
for name in node.names:
if name in self.magic_globals:
self.references_to_globals.append((name, node.lineno))
elif isinstance(node, ast.Assign):
# visit the rhs before lhs
super().visit(node.value)
for target in node.targets:
super().visit(target)
elif isinstance(node, ast.Name) and node.id in self.magic_globals:
if isinstance(node.ctx, ast.Load) and node.id not in self.locals:
self.references_to_globals.append((node.id, node.lineno))
elif isinstance(node.ctx, ast.Store):
self.locals.add(node.id)
else:
super().generic_visit(node)
@package_properties
def _uses_deprecated_globals(pkgs, error_cls):
"""Ensure that packages do not use deprecated globals"""
errors = []
for pkg_name in pkgs:
# some packages scheduled to be removed in v0.23 are not worth fixing.
pkg_cls = spack.repo.PATH.get_pkg_class(pkg_name)
if all(v.get("deprecated", False) for v in pkg_cls.versions.values()):
continue
file = spack.repo.PATH.filename_for_package_name(pkg_name)
tree = ast.parse(open(file).read())
visitor = DeprecatedMagicGlobals(("std_cmake_args", "std_meson_args", "std_pip_args"))
visitor.visit(tree)
if visitor.references_to_globals:
errors.append(
error_cls(
f"Package '{pkg_name}' uses deprecated globals",
[
f"{file}:{line} references '{name}'"
for name, line in visitor.references_to_globals
],
)
)
return errors
@package_properties
def _ensure_test_docstring(pkgs, error_cls):
"""Ensure stand-alone test methods have a docstring.
The docstring of a test method is implicitly used as the description of
the corresponding test part during test results reporting.
"""
doc_regex = r'\s+("""[^"]+""")'
errors = []
for pkg_name in pkgs:
pkg_cls = spack.repo.PATH.get_pkg_class(pkg_name)
methods = inspect.getmembers(pkg_cls, predicate=lambda x: inspect.isfunction(x))
method_names = []
for name, test_fn in methods:
if not name.startswith("test_"):
continue
# Ensure the test method has a docstring
source = inspect.getsource(test_fn)
match = re.search(doc_regex, source)
if match is None or len(match.group(0).replace('"', "").strip()) == 0:
method_names.append(name)
num_methods = len(method_names)
if num_methods > 0:
methods = plural(num_methods, "method", show_n=False)
docstrings = plural(num_methods, "docstring", show_n=False)
msg = f"Package {pkg_name} has test {methods} with empty or missing {docstrings}."
names = ", ".join(method_names)
instr = [
"Docstrings are used as descriptions in test outputs.",
f"Add a concise summary to the following {methods} in '{pkg_cls.__module__}':",
f"{names}",
]
errors.append(error_cls(msg, instr))
return errors
@package_properties
def _ensure_test_implemented(pkgs, error_cls):
"""Ensure stand-alone test methods are implemented.
The test method is also required to be non-empty.
"""
def skip(line):
ln = line.strip()
return ln.startswith("#") or "pass" in ln
doc_regex = r'\s+("""[^"]+""")'
errors = []
for pkg_name in pkgs:
pkg_cls = spack.repo.PATH.get_pkg_class(pkg_name)
methods = inspect.getmembers(pkg_cls, predicate=lambda x: inspect.isfunction(x))
method_names = []
for name, test_fn in methods:
if not name.startswith("test_"):
continue
source = inspect.getsource(test_fn)
# Attempt to ensure the test method is implemented.
impl = re.sub(doc_regex, r"", source).splitlines()[1:]
lines = [ln.strip() for ln in impl if not skip(ln)]
if not lines:
method_names.append(name)
num_methods = len(method_names)
if num_methods > 0:
methods = plural(num_methods, "method", show_n=False)
msg = f"Package {pkg_name} has empty or missing test {methods}."
names = ", ".join(method_names)
instr = [
f"Implement or remove the following {methods} from '{pkg_cls.__module__}': {names}"
]
errors.append(error_cls(msg, instr))
return errors
@package_https_directives
def _linting_package_file(pkgs, error_cls):
"""Check for correctness of links"""
@@ -1015,14 +764,7 @@ def _issues_in_depends_on_directive(pkgs, error_cls):
# depends_on('foo+bar ^fee+baz')
#
# but we'd like to have two dependencies listed instead.
nested_dependencies = dep.spec.edges_to_dependencies()
# Filter out pure build dependencies, like:
#
# depends_on('foo+bar%gcc')
#
nested_dependencies = [
x for x in nested_dependencies if x.depflag != spack.deptypes.BUILD
]
nested_dependencies = dep.spec.dependencies()
if nested_dependencies:
summary = f"{pkg_name}: nested dependency declaration '{dep.spec}'"
ndir = len(nested_dependencies) + 1
@@ -1083,22 +825,20 @@ def check_virtual_with_variants(spec, msg):
# check variants
dependency_variants = dep.spec.variants
for name, variant in dependency_variants.items():
for name, value in dependency_variants.items():
try:
spack.variant.prevalidate_variant_value(
dependency_pkg_cls, variant, dep.spec, strict=True
)
v, _ = dependency_pkg_cls.variants[name]
v.validate_or_raise(value, pkg_cls=dependency_pkg_cls)
except Exception as e:
summary = (
f"{pkg_name}: wrong variant used for dependency in 'depends_on()'"
)
error_msg = str(e)
if isinstance(e, KeyError):
error_msg = (
f"variant {str(e).strip()} does not exist in package {dep_name}"
f" in package '{dep_name}'"
)
error_msg += f" in package '{dep_name}'"
errors.append(
error_cls(summary=summary, details=[error_msg, f"in {filename}"])
@@ -1110,38 +850,39 @@ def check_virtual_with_variants(spec, msg):
@package_directives
def _ensure_variant_defaults_are_parsable(pkgs, error_cls):
"""Ensures that variant defaults are present and parsable from cli"""
def check_variant(pkg_cls, variant, vname):
# bool is a subclass of int in python. Permitting a default that is an instance
# of 'int' means both foo=false and foo=0 are accepted. Other falsish values are
# not allowed, since they can't be parsed from CLI ('foo=')
default_is_parsable = isinstance(variant.default, int) or variant.default
if not default_is_parsable:
msg = f"Variant '{vname}' of package '{pkg_cls.name}' has an unparsable default value"
return [error_cls(msg, [])]
try:
vspec = variant.make_default()
except spack.variant.MultipleValuesInExclusiveVariantError:
msg = f"Can't create default value for variant '{vname}' in package '{pkg_cls.name}'"
return [error_cls(msg, [])]
try:
variant.validate_or_raise(vspec, pkg_cls.name)
except spack.variant.InvalidVariantValueError:
msg = "Default value of variant '{vname}' in package '{pkg.name}' is invalid"
question = "Is it among the allowed values?"
return [error_cls(msg, [question])]
return []
errors = []
for pkg_name in pkgs:
pkg_cls = spack.repo.PATH.get_pkg_class(pkg_name)
for vname in pkg_cls.variant_names():
for _, variant_def in pkg_cls.variant_definitions(vname):
errors.extend(check_variant(pkg_cls, variant_def, vname))
for variant_name, entry in pkg_cls.variants.items():
variant, _ = entry
default_is_parsable = (
# Permitting a default that is an instance on 'int' permits
# to have foo=false or foo=0. Other falsish values are
# not allowed, since they can't be parsed from cli ('foo=')
isinstance(variant.default, int)
or variant.default
)
if not default_is_parsable:
error_msg = "Variant '{}' of package '{}' has a bad default value"
errors.append(error_cls(error_msg.format(variant_name, pkg_name), []))
continue
try:
vspec = variant.make_default()
except spack.variant.MultipleValuesInExclusiveVariantError:
error_msg = "Cannot create a default value for the variant '{}' in package '{}'"
errors.append(error_cls(error_msg.format(variant_name, pkg_name), []))
continue
try:
variant.validate_or_raise(vspec, pkg_cls=pkg_cls)
except spack.variant.InvalidVariantValueError:
error_msg = (
"The default value of the variant '{}' in package '{}' failed validation"
)
question = "Is it among the allowed values?"
errors.append(error_cls(error_msg.format(variant_name, pkg_name), [question]))
return errors
@@ -1151,11 +892,11 @@ def _ensure_variants_have_descriptions(pkgs, error_cls):
errors = []
for pkg_name in pkgs:
pkg_cls = spack.repo.PATH.get_pkg_class(pkg_name)
for name in pkg_cls.variant_names():
for when, variant in pkg_cls.variant_definitions(name):
if not variant.description:
msg = f"Variant '{name}' in package '{pkg_name}' is missing a description"
errors.append(error_cls(msg, []))
for variant_name, entry in pkg_cls.variants.items():
variant, _ = entry
if not variant.description:
error_msg = "Variant '{}' in package '{}' is missing a description"
errors.append(error_cls(error_msg.format(variant_name, pkg_name), []))
return errors
@@ -1212,26 +953,29 @@ def _version_constraints_are_satisfiable_by_some_version_in_repo(pkgs, error_cls
def _analyze_variants_in_directive(pkg, constraint, directive, error_cls):
variant_exceptions = (
spack.variant.InconsistentValidationError,
spack.variant.MultipleValuesInExclusiveVariantError,
spack.variant.InvalidVariantValueError,
KeyError,
)
errors = []
variant_names = pkg.variant_names()
summary = f"{pkg.name}: wrong variant in '{directive}' directive"
filename = spack.repo.PATH.filename_for_package_name(pkg.name)
for name, v in constraint.variants.items():
if name not in variant_names:
msg = f"variant {name} does not exist in {pkg.name}"
errors.append(error_cls(summary=summary, details=[msg, f"in {filename}"]))
continue
try:
spack.variant.prevalidate_variant_value(pkg, v, constraint, strict=True)
except (
spack.variant.InconsistentValidationError,
spack.variant.MultipleValuesInExclusiveVariantError,
spack.variant.InvalidVariantValueError,
) as e:
msg = str(e).strip()
errors.append(error_cls(summary=summary, details=[msg, f"in {filename}"]))
variant, _ = pkg.variants[name]
variant.validate_or_raise(v, pkg_cls=pkg)
except variant_exceptions as e:
summary = pkg.name + ': wrong variant in "{0}" directive'
summary = summary.format(directive)
filename = spack.repo.PATH.filename_for_package_name(pkg.name)
error_msg = str(e).strip()
if isinstance(e, KeyError):
error_msg = "the variant {0} does not exist".format(error_msg)
err = error_cls(summary=summary, details=[error_msg, "in " + filename])
errors.append(err)
return errors
@@ -1269,10 +1013,9 @@ def _extracts_errors(triggers, summary):
for dname in dnames
)
for when, variants_by_name in pkg_cls.variants.items():
for vname, variant in variants_by_name.items():
summary = f"{pkg_name}: wrong 'when=' condition for the '{vname}' variant"
errors.extend(_extracts_errors([when], summary))
for vname, (variant, triggers) in pkg_cls.variants.items():
summary = f"{pkg_name}: wrong 'when=' condition for the '{vname}' variant"
errors.extend(_extracts_errors(triggers, summary))
for when, providers, details in _error_items(pkg_cls.provided):
errors.extend(

File diff suppressed because it is too large Load Diff

View File

@@ -9,7 +9,6 @@
all_core_root_specs,
ensure_clingo_importable_or_raise,
ensure_core_dependencies,
ensure_file_in_path_or_raise,
ensure_gpg_in_path_or_raise,
ensure_patchelf_in_path_or_raise,
)
@@ -20,7 +19,6 @@
"is_bootstrapping",
"ensure_bootstrap_configuration",
"ensure_core_dependencies",
"ensure_file_in_path_or_raise",
"ensure_gpg_in_path_or_raise",
"ensure_clingo_importable_or_raise",
"ensure_patchelf_in_path_or_raise",

View File

@@ -4,8 +4,6 @@
# SPDX-License-Identifier: (Apache-2.0 OR MIT)
"""Common basic functions used through the spack.bootstrap package"""
import fnmatch
import glob
import importlib
import os.path
import re
import sys
@@ -30,7 +28,7 @@
def _python_import(module: str) -> bool:
try:
importlib.import_module(module)
__import__(module)
except ImportError:
return False
return True
@@ -61,19 +59,10 @@ def _try_import_from_store(
python, *_ = candidate_spec.dependencies("python-venv")
else:
python, *_ = candidate_spec.dependencies("python")
# if python is installed, ask it for the layout
if python.installed:
module_paths = [
os.path.join(candidate_spec.prefix, python.package.purelib),
os.path.join(candidate_spec.prefix, python.package.platlib),
]
# otherwise search for the site-packages directory
# (clingo from binaries with truncated python-venv runtime)
else:
module_paths = glob.glob(
os.path.join(candidate_spec.prefix, "lib", "python*", "site-packages")
)
module_paths = [
os.path.join(candidate_spec.prefix, python.package.purelib),
os.path.join(candidate_spec.prefix, python.package.platlib),
]
path_before = list(sys.path)
# NOTE: try module_paths first and last, last allows an existing version in path
@@ -224,19 +213,15 @@ def _root_spec(spec_str: str) -> str:
Args:
spec_str: spec to be bootstrapped. Must be without compiler and target.
"""
# Add a compiler and platform requirement to the root spec.
# Add a compiler requirement to the root spec.
platform = str(spack.platforms.host())
# FIXME (compiler as nodes): recover the compiler for source bootstrapping
# if platform == "darwin":
# spec_str += " %apple-clang"
if platform == "windows":
spec_str += " %msvc"
# elif platform == "linux":
# spec_str += " %gcc"
if platform == "darwin":
spec_str += " %apple-clang"
elif platform == "linux":
spec_str += " %gcc"
elif platform == "freebsd":
spec_str += " %clang"
spec_str += f" platform={platform}"
target = archspec.cpu.host().family
spec_str += f" target={target}"

View File

@@ -1,167 +0,0 @@
# Copyright 2013-2024 Lawrence Livermore National Security, LLC and other
# Spack Project Developers. See the top-level COPYRIGHT file for details.
#
# SPDX-License-Identifier: (Apache-2.0 OR MIT)
"""Bootstrap concrete specs for clingo
Spack uses clingo to concretize specs. When clingo itself needs to be bootstrapped from sources,
we need to rely on another mechanism to get a concrete spec that fits the current host.
This module contains the logic to get a concrete spec for clingo, starting from a prototype
JSON file for a similar platform.
"""
import pathlib
import sys
from typing import Dict, Optional, Tuple
import archspec.cpu
import spack.compilers.config
import spack.compilers.libraries
import spack.config
import spack.platforms
import spack.spec
import spack.traverse
from .config import spec_for_current_python
class ClingoBootstrapConcretizer:
def __init__(self, configuration):
self.host_platform = spack.platforms.host()
self.host_os = self.host_platform.operating_system("frontend")
self.host_target = archspec.cpu.host().family
self.host_architecture = spack.spec.ArchSpec.frontend_arch()
self.host_architecture.target = str(self.host_target)
self.host_compiler = self._valid_compiler_or_raise()
self.host_python = self.python_external_spec()
if str(self.host_platform) == "linux":
self.host_libc = self.libc_external_spec()
self.external_cmake, self.external_bison = self._externals_from_yaml(configuration)
def _valid_compiler_or_raise(self):
if str(self.host_platform) == "linux":
compiler_name = "gcc"
elif str(self.host_platform) == "darwin":
compiler_name = "apple-clang"
elif str(self.host_platform) == "windows":
compiler_name = "msvc"
elif str(self.host_platform) == "freebsd":
compiler_name = "llvm"
else:
raise RuntimeError(f"Cannot bootstrap clingo from sources on {self.host_platform}")
candidates = [
x
for x in spack.compilers.config.CompilerFactory.from_packages_yaml(spack.config.CONFIG)
if x.name == compiler_name
]
if not candidates:
raise RuntimeError(
f"Cannot find any version of {compiler_name} to bootstrap clingo from sources"
)
candidates.sort(key=lambda x: x.version, reverse=True)
best = candidates[0]
# Get compilers for bootstrapping from the 'builtin' repository
best.namespace = "builtin"
# If the compiler does not support C++ 14, fail with a legible error message
try:
_ = best.package.standard_flag(language="cxx", standard="14")
except RuntimeError as e:
raise RuntimeError(
"cannot find a compiler supporting C++ 14 [needed to bootstrap clingo]"
) from e
return candidates[0]
def _externals_from_yaml(
self, configuration: "spack.config.Configuration"
) -> Tuple[Optional["spack.spec.Spec"], Optional["spack.spec.Spec"]]:
packages_yaml = configuration.get("packages")
requirements = {"cmake": "@3.20:", "bison": "@2.5:"}
selected: Dict[str, Optional["spack.spec.Spec"]] = {"cmake": None, "bison": None}
for pkg_name in ["cmake", "bison"]:
if pkg_name not in packages_yaml:
continue
candidates = packages_yaml[pkg_name].get("externals", [])
for candidate in candidates:
s = spack.spec.Spec(candidate["spec"], external_path=candidate["prefix"])
if not s.satisfies(requirements[pkg_name]):
continue
if not s.intersects(f"arch={self.host_architecture}"):
continue
selected[pkg_name] = self._external_spec(s)
break
return selected["cmake"], selected["bison"]
def prototype_path(self) -> pathlib.Path:
"""Path to a prototype concrete specfile for clingo"""
parent_dir = pathlib.Path(__file__).parent
result = parent_dir / "prototypes" / f"clingo-{self.host_platform}-{self.host_target}.json"
if str(self.host_platform) == "linux":
# Using aarch64 as a fallback, since it has gnuconfig (x86_64 doesn't have it)
if not result.exists():
result = parent_dir / "prototypes" / f"clingo-{self.host_platform}-aarch64.json"
elif str(self.host_platform) == "freebsd":
result = parent_dir / "prototypes" / f"clingo-{self.host_platform}-amd64.json"
elif not result.exists():
raise RuntimeError(f"Cannot bootstrap clingo from sources on {self.host_platform}")
return result
def concretize(self) -> "spack.spec.Spec":
# Read the prototype and mark it NOT concrete
s = spack.spec.Spec.from_specfile(str(self.prototype_path()))
s._mark_concrete(False)
# Tweak it to conform to the host architecture
for node in s.traverse():
node.architecture.os = str(self.host_os)
node.architecture = self.host_architecture
if node.name == "gcc-runtime":
node.versions = self.host_compiler.versions
for edge in spack.traverse.traverse_edges([s], cover="edges"):
if edge.spec.name == "python":
edge.spec = self.host_python
if edge.spec.name == "bison" and self.external_bison:
edge.spec = self.external_bison
if edge.spec.name == "cmake" and self.external_cmake:
edge.spec = self.external_cmake
if edge.spec.name == self.host_compiler.name:
edge.spec = self.host_compiler
if "libc" in edge.virtuals:
edge.spec = self.host_libc
s._finalize_concretization()
# Work around the fact that the installer calls Spec.dependents() and
# we modified edges inconsistently
return s.copy()
def python_external_spec(self) -> "spack.spec.Spec":
"""Python external spec corresponding to the current running interpreter"""
result = spack.spec.Spec(spec_for_current_python(), external_path=sys.exec_prefix)
return self._external_spec(result)
def libc_external_spec(self) -> "spack.spec.Spec":
detector = spack.compilers.libraries.CompilerPropertyDetector(self.host_compiler)
result = detector.default_libc()
return self._external_spec(result)
def _external_spec(self, initial_spec) -> "spack.spec.Spec":
initial_spec.namespace = "builtin"
initial_spec.architecture = self.host_architecture
for flag_type in spack.spec.FlagMap.valid_compiler_flags():
initial_spec.compiler_flags[flag_type] = []
return spack.spec.parse_with_version_concrete(initial_spec)

View File

@@ -11,10 +11,9 @@
from llnl.util import tty
import spack.compilers.config
import spack.compilers
import spack.config
import spack.environment
import spack.modules
import spack.paths
import spack.platforms
import spack.repo
@@ -130,10 +129,10 @@ def _bootstrap_config_scopes() -> Sequence["spack.config.ConfigScope"]:
configuration_paths = (spack.config.CONFIGURATION_DEFAULTS_PATH, ("bootstrap", _config_path()))
for name, path in configuration_paths:
platform = spack.platforms.host().name
platform_scope = spack.config.DirectoryConfigScope(
f"{name}/{platform}", os.path.join(path, platform)
platform_scope = spack.config.ConfigScope(
"/".join([name, platform]), os.path.join(path, platform)
)
generic_scope = spack.config.DirectoryConfigScope(name, path)
generic_scope = spack.config.ConfigScope(name, path)
config_scopes.extend([generic_scope, platform_scope])
msg = "[BOOTSTRAP CONFIG SCOPE] name={0}, path={1}"
tty.debug(msg.format(generic_scope.name, generic_scope.path))
@@ -143,8 +142,12 @@ def _bootstrap_config_scopes() -> Sequence["spack.config.ConfigScope"]:
def _add_compilers_if_missing() -> None:
arch = spack.spec.ArchSpec.frontend_arch()
if not spack.compilers.config.compilers_for_arch(arch):
spack.compilers.config.find_compilers()
if not spack.compilers.compilers_for_arch(arch):
new_compilers = spack.compilers.find_new_compilers(
mixed_toolchain=sys.platform == "darwin"
)
if new_compilers:
spack.compilers.add_compilers_to_config(new_compilers)
@contextlib.contextmanager
@@ -153,7 +156,7 @@ def _ensure_bootstrap_configuration() -> Generator:
bootstrap_store_path = store_path()
user_configuration = _read_and_sanitize_configuration()
with spack.environment.no_active_environment():
with spack.platforms.use_platform(
with spack.platforms.prevent_cray_detection(), spack.platforms.use_platform(
spack.platforms.real_host()
), spack.repo.use_repositories(spack.paths.packages_path):
# Default configuration scopes excluding command line

View File

@@ -37,19 +37,23 @@
import spack.binary_distribution
import spack.config
import spack.detection
import spack.mirror
import spack.environment
import spack.modules
import spack.paths
import spack.platforms
import spack.platforms.linux
import spack.repo
import spack.spec
import spack.store
import spack.user_environment
import spack.util.environment
import spack.util.executable
import spack.util.path
import spack.util.spack_yaml
import spack.util.url
import spack.version
from spack.installer import PackageInstaller
from ._common import _executables_in_store, _python_import, _root_spec, _try_import_from_store
from .clingo import ClingoBootstrapConcretizer
from .config import spack_python_interpreter, spec_for_current_python
#: Name of the file containing metadata about the bootstrapping source
@@ -91,7 +95,12 @@ def __init__(self, conf: ConfigDictionary) -> None:
self.metadata_dir = spack.util.path.canonicalize_path(conf["metadata"])
# Promote (relative) paths to file urls
self.url = spack.mirror.Mirror(conf["info"]["url"]).fetch_url
url = conf["info"]["url"]
if spack.util.url.is_path_instead_of_url(url):
if not os.path.isabs(url):
url = os.path.join(self.metadata_dir, url)
url = spack.util.url.path_to_file_url(url)
self.url = url
@property
def mirror_scope(self) -> spack.config.InternalConfigScope:
@@ -170,15 +179,7 @@ def _install_by_hash(
query = spack.binary_distribution.BinaryCacheQuery(all_architectures=True)
for match in spack.store.find([f"/{pkg_hash}"], multiple=False, query_fn=query):
spack.binary_distribution.install_root_node(
# allow_missing is true since when bootstrapping clingo we truncate runtime
# deps such as gcc-runtime, since we link libstdc++ statically, and the other
# further runtime deps are loaded by the Python interpreter. This just silences
# warnings about missing dependencies.
match,
unsigned=True,
force=True,
sha256=pkg_sha256,
allow_missing=True,
match, unsigned=True, force=True, sha256=pkg_sha256
)
def _install_and_test(
@@ -267,13 +268,15 @@ def try_import(self, module: str, abstract_spec_str: str) -> bool:
# Try to build and install from sources
with spack_python_interpreter():
# Add hint to use frontend operating system on Cray
concrete_spec = spack.spec.Spec(abstract_spec_str + " ^" + spec_for_current_python())
if module == "clingo":
bootstrapper = ClingoBootstrapConcretizer(configuration=spack.config.CONFIG)
concrete_spec = bootstrapper.concretize()
else:
concrete_spec = spack.spec.Spec(
abstract_spec_str + " ^" + spec_for_current_python()
# TODO: remove when the old concretizer is deprecated # pylint: disable=fixme
concrete_spec._old_concretize( # pylint: disable=protected-access
deprecation_warning=False
)
else:
concrete_spec.concretize()
msg = "[BOOTSTRAP MODULE {0}] Try installing '{1}' from sources"
@@ -281,12 +284,7 @@ def try_import(self, module: str, abstract_spec_str: str) -> bool:
# Install the spec that should make the module importable
with spack.config.override(self.mirror_scope):
PackageInstaller(
[concrete_spec.package],
fail_fast=True,
package_use_cache=False,
dependencies_use_cache=False,
).install()
concrete_spec.package.do_install(fail_fast=True)
if _try_import_from_store(module, query_spec=concrete_spec, query_info=info):
self.last_search = info
@@ -305,11 +303,18 @@ def try_search_path(self, executables: Tuple[str], abstract_spec_str: str) -> bo
# might reduce compilation time by a fair amount
_add_externals_if_missing()
concrete_spec = spack.spec.Spec(abstract_spec_str).concretized()
concrete_spec = spack.spec.Spec(abstract_spec_str)
if concrete_spec.name == "patchelf":
concrete_spec._old_concretize( # pylint: disable=protected-access
deprecation_warning=False
)
else:
concrete_spec.concretize()
msg = "[BOOTSTRAP] Try installing '{0}' from sources"
tty.debug(msg.format(abstract_spec_str))
with spack.config.override(self.mirror_scope):
PackageInstaller([concrete_spec.package], fail_fast=True).install()
concrete_spec.package.do_install()
if _executables_in_store(executables, concrete_spec, query_info=info):
self.last_search = info
return True
@@ -322,10 +327,11 @@ def create_bootstrapper(conf: ConfigDictionary):
return _bootstrap_methods[btype](conf)
def source_is_enabled(conf: ConfigDictionary):
def source_is_enabled_or_raise(conf: ConfigDictionary):
"""Raise ValueError if the source is not enabled for bootstrapping"""
trusted, name = spack.config.get("bootstrap:trusted"), conf["name"]
return trusted.get(name, False)
if not trusted.get(name, False):
raise ValueError("source is not trusted")
def ensure_module_importable_or_raise(module: str, abstract_spec: Optional[str] = None):
@@ -355,10 +361,8 @@ def ensure_module_importable_or_raise(module: str, abstract_spec: Optional[str]
exception_handler = GroupedExceptionHandler()
for current_config in bootstrapping_sources():
if not source_is_enabled(current_config):
continue
with exception_handler.forward(current_config["name"], Exception):
source_is_enabled_or_raise(current_config)
current_bootstrapper = create_bootstrapper(current_config)
if current_bootstrapper.try_import(module, abstract_spec):
return
@@ -370,7 +374,11 @@ def ensure_module_importable_or_raise(module: str, abstract_spec: Optional[str]
msg = f'cannot bootstrap the "{module}" Python module '
if abstract_spec:
msg += f'from spec "{abstract_spec}" '
msg += exception_handler.grouped_message(with_tracebacks=tty.is_debug())
if tty.is_debug():
msg += exception_handler.grouped_message(with_tracebacks=True)
else:
msg += exception_handler.grouped_message(with_tracebacks=False)
msg += "\nRun `spack --debug ...` for more detailed errors"
raise ImportError(msg)
@@ -408,9 +416,8 @@ def ensure_executables_in_path_or_raise(
exception_handler = GroupedExceptionHandler()
for current_config in bootstrapping_sources():
if not source_is_enabled(current_config):
continue
with exception_handler.forward(current_config["name"], Exception):
source_is_enabled_or_raise(current_config)
current_bootstrapper = create_bootstrapper(current_config)
if current_bootstrapper.try_search_path(executables, abstract_spec):
# Additional environment variables needed
@@ -473,8 +480,7 @@ def ensure_clingo_importable_or_raise() -> None:
def gnupg_root_spec() -> str:
"""Return the root spec used to bootstrap GnuPG"""
root_spec_name = "win-gpg" if IS_WINDOWS else "gnupg"
return _root_spec(f"{root_spec_name}@2.3:")
return _root_spec("gnupg@2.3:")
def ensure_gpg_in_path_or_raise() -> None:
@@ -484,19 +490,6 @@ def ensure_gpg_in_path_or_raise() -> None:
)
def file_root_spec() -> str:
"""Return the root spec used to bootstrap file"""
root_spec_name = "win-file" if IS_WINDOWS else "file"
return _root_spec(root_spec_name)
def ensure_file_in_path_or_raise() -> None:
"""Ensure file is in the PATH or raise"""
return ensure_executables_in_path_or_raise(
executables=["file"], abstract_spec=file_root_spec()
)
def patchelf_root_spec() -> str:
"""Return the root spec used to bootstrap patchelf"""
# 0.13.1 is the last version not to require C++17.
@@ -580,15 +573,14 @@ def ensure_core_dependencies() -> None:
"""Ensure the presence of all the core dependencies."""
if sys.platform.lower() == "linux":
ensure_patchelf_in_path_or_raise()
elif sys.platform == "win32":
ensure_file_in_path_or_raise()
ensure_gpg_in_path_or_raise()
if not IS_WINDOWS:
ensure_gpg_in_path_or_raise()
ensure_clingo_importable_or_raise()
def all_core_root_specs() -> List[str]:
"""Return a list of all the core root specs that may be used to bootstrap Spack"""
return [clingo_root_spec(), gnupg_root_spec(), patchelf_root_spec(), file_root_spec()]
return [clingo_root_spec(), gnupg_root_spec(), patchelf_root_spec()]
def bootstrapping_sources(scope: Optional[str] = None):
@@ -605,10 +597,7 @@ def bootstrapping_sources(scope: Optional[str] = None):
current = copy.copy(entry)
metadata_dir = spack.util.path.canonicalize_path(entry["metadata"])
metadata_yaml = os.path.join(metadata_dir, METADATA_YAML_FILENAME)
try:
with open(metadata_yaml, encoding="utf-8") as stream:
current.update(spack.util.spack_yaml.load(stream))
list_of_sources.append(current)
except OSError:
pass
with open(metadata_yaml, encoding="utf-8") as stream:
current.update(spack.util.spack_yaml.load(stream))
list_of_sources.append(current)
return list_of_sources

View File

@@ -14,9 +14,9 @@
from llnl.util import tty
import spack.environment
import spack.spec
import spack.tengine
import spack.util.path
import spack.util.cpus
import spack.util.executable
from ._common import _root_spec
from .config import root_path, spec_for_current_python, store_path

File diff suppressed because one or more lines are too long

File diff suppressed because one or more lines are too long

File diff suppressed because one or more lines are too long

File diff suppressed because one or more lines are too long

File diff suppressed because one or more lines are too long

File diff suppressed because one or more lines are too long

File diff suppressed because one or more lines are too long

View File

@@ -88,7 +88,7 @@ def _core_requirements() -> List[RequiredResponseType]:
def _buildcache_requirements() -> List[RequiredResponseType]:
_buildcache_exes = {
"file": _missing("file", "required to analyze files for buildcaches", system_only=False),
"file": _missing("file", "required to analyze files for buildcaches"),
("gpg2", "gpg"): _missing("gpg2", "required to sign/verify buildcaches", False),
}
if platform.system().lower() == "darwin":
@@ -124,7 +124,7 @@ def _development_requirements() -> List[RequiredResponseType]:
# Ensure we trigger environment modifications if we have an environment
if BootstrapEnvironment.spack_yaml().exists():
with BootstrapEnvironment() as env:
env.load()
env.update_syspath_and_environ()
return [
_required_executable(

View File

@@ -43,10 +43,7 @@
from collections import defaultdict
from enum import Flag, auto
from itertools import chain
from multiprocessing.connection import Connection
from typing import Callable, Dict, List, Optional, Set, Tuple
import archspec.cpu
from typing import Dict, List, Set, Tuple
import llnl.util.tty as tty
from llnl.string import plural
@@ -54,29 +51,36 @@
from llnl.util.lang import dedupe, stable_partition
from llnl.util.symlink import symlink
from llnl.util.tty.color import cescape, colorize
from llnl.util.tty.log import MultiProcessFd
import spack.build_systems.cmake
import spack.build_systems.meson
import spack.build_systems.python
import spack.builder
import spack.compilers.libraries
import spack.compilers
import spack.config
import spack.deptypes as dt
import spack.error
import spack.multimethod
import spack.main
import spack.package_base
import spack.paths
import spack.platforms
import spack.repo
import spack.schema.environment
import spack.spec
import spack.stage
import spack.store
import spack.subprocess_context
import spack.user_environment
import spack.util.executable
import spack.util.path
import spack.util.pattern
from spack import traverse
from spack.context import Context
from spack.error import InstallError, NoHeadersError, NoLibrariesError
from spack.error import NoHeadersError, NoLibrariesError
from spack.install_test import spack_install_test_log
from spack.installer import InstallError
from spack.util.cpus import determine_number_of_jobs
from spack.util.environment import (
SYSTEM_DIR_CASE_ENTRY,
EnvironmentModifications,
@@ -88,7 +92,7 @@
)
from spack.util.executable import Executable
from spack.util.log_parse import make_log_context, parse_log_events
from spack.util.module_cmd import load_module
from spack.util.module_cmd import load_module, module, path_from_modules
#
# This can be set by the user to globally disable parallel builds.
@@ -187,6 +191,14 @@ def __call__(self, *args, **kwargs):
return super().__call__(*args, **kwargs)
def _on_cray():
host_platform = spack.platforms.host()
host_os = host_platform.operating_system("default_os")
on_cray = str(host_platform) == "cray"
using_cnl = re.match(r"cnl\d+", str(host_os))
return on_cray, using_cnl
def clean_environment():
# Stuff in here sanitizes the build environment to eliminate
# anything the user has set that may interfere. We apply it immediately
@@ -230,6 +242,17 @@ def clean_environment():
if varname.endswith("_ROOT") and varname != "SPACK_ROOT":
env.unset(varname)
# On Cray "cluster" systems, unset CRAY_LD_LIBRARY_PATH to avoid
# interference with Spack dependencies.
# CNL requires these variables to be set (or at least some of them,
# depending on the CNL version).
on_cray, using_cnl = _on_cray()
if on_cray and not using_cnl:
env.unset("CRAY_LD_LIBRARY_PATH")
for varname in os.environ.keys():
if "PKGCONF" in varname:
env.unset(varname)
# Unset the following variables because they can affect installation of
# Autotools and CMake packages.
build_system_vars = [
@@ -295,10 +318,62 @@ def _add_werror_handling(keep_werror, env):
env.set("SPACK_COMPILER_FLAGS_REPLACE", " ".join(["|".join(item) for item in replace_flags]))
def set_wrapper_environment_variables_for_flags(pkg, env):
def set_compiler_environment_variables(pkg, env):
assert pkg.spec.concrete
compiler = pkg.compiler
spec = pkg.spec
# Make sure the executables for this compiler exist
compiler.verify_executables()
# Set compiler variables used by CMake and autotools
assert all(key in compiler.link_paths for key in ("cc", "cxx", "f77", "fc"))
# Populate an object with the list of environment modifications
# and return it
# TODO : add additional kwargs for better diagnostics, like requestor,
# ttyout, ttyerr, etc.
link_dir = spack.paths.build_env_path
# Set SPACK compiler variables so that our wrapper knows what to
# call. If there is no compiler configured then use a default
# wrapper which will emit an error if it is used.
if compiler.cc:
env.set("SPACK_CC", compiler.cc)
env.set("CC", os.path.join(link_dir, compiler.link_paths["cc"]))
else:
env.set("CC", os.path.join(link_dir, "cc"))
if compiler.cxx:
env.set("SPACK_CXX", compiler.cxx)
env.set("CXX", os.path.join(link_dir, compiler.link_paths["cxx"]))
else:
env.set("CC", os.path.join(link_dir, "c++"))
if compiler.f77:
env.set("SPACK_F77", compiler.f77)
env.set("F77", os.path.join(link_dir, compiler.link_paths["f77"]))
else:
env.set("F77", os.path.join(link_dir, "f77"))
if compiler.fc:
env.set("SPACK_FC", compiler.fc)
env.set("FC", os.path.join(link_dir, compiler.link_paths["fc"]))
else:
env.set("FC", os.path.join(link_dir, "fc"))
# Set SPACK compiler rpath flags so that our wrapper knows what to use
env.set("SPACK_CC_RPATH_ARG", compiler.cc_rpath_arg)
env.set("SPACK_CXX_RPATH_ARG", compiler.cxx_rpath_arg)
env.set("SPACK_F77_RPATH_ARG", compiler.f77_rpath_arg)
env.set("SPACK_FC_RPATH_ARG", compiler.fc_rpath_arg)
env.set("SPACK_LINKER_ARG", compiler.linker_arg)
# Check whether we want to force RPATH or RUNPATH
if spack.config.get("config:shared_linking:type") == "rpath":
env.set("SPACK_DTAGS_TO_STRIP", compiler.enable_new_dtags)
env.set("SPACK_DTAGS_TO_ADD", compiler.disable_new_dtags)
else:
env.set("SPACK_DTAGS_TO_STRIP", compiler.disable_new_dtags)
env.set("SPACK_DTAGS_TO_ADD", compiler.enable_new_dtags)
if pkg.keep_werror is not None:
keep_werror = pkg.keep_werror
else:
@@ -306,6 +381,14 @@ def set_wrapper_environment_variables_for_flags(pkg, env):
_add_werror_handling(keep_werror, env)
# Set the target parameters that the compiler will add
# Don't set on cray platform because the targeting module handles this
if spec.satisfies("platform=cray"):
isa_arg = ""
else:
isa_arg = spec.architecture.target.optimization_flags(compiler)
env.set("SPACK_TARGET_ARGS", isa_arg)
# Trap spack-tracked compiler flags as appropriate.
# env_flags are easy to accidentally override.
inject_flags = {}
@@ -339,27 +422,15 @@ def set_wrapper_environment_variables_for_flags(pkg, env):
env.set(flag.upper(), " ".join(f for f in env_flags[flag]))
pkg.flags_to_build_system_args(build_system_flags)
env.set("SPACK_COMPILER_SPEC", str(spec.compiler))
env.set("SPACK_SYSTEM_DIRS", SYSTEM_DIR_CASE_ENTRY)
# FIXME (compiler as nodes): recover this one in the correct packages
# compiler.setup_custom_environment(pkg, env)
compiler.setup_custom_environment(pkg, env)
return env
def optimization_flags(compiler, target):
# Try to check if the current compiler comes with a version number or
# has an unexpected suffix. If so, treat it as a compiler with a
# custom spec.
version_number, _ = archspec.cpu.version_components(compiler.version.dotted_numeric_string)
try:
result = target.optimization_flags(compiler.name, version_number)
except (ValueError, archspec.cpu.UnsupportedMicroarchitecture):
result = ""
return result
def set_wrapper_variables(pkg, env):
"""Set environment variables used by the Spack compiler wrapper (which have the prefix
`SPACK_`) and also add the compiler wrappers to PATH.
@@ -368,15 +439,46 @@ def set_wrapper_variables(pkg, env):
this function computes these options in a manner that is intended to match the DAG traversal
order in `SetupContext`. TODO: this is not the case yet, we're using post order, SetupContext
is using topo order."""
# Set compiler flags injected from the spec
set_wrapper_environment_variables_for_flags(pkg, env)
# Set environment variables if specified for
# the given compiler
compiler = pkg.compiler
env.extend(spack.schema.environment.parse(compiler.environment))
if compiler.extra_rpaths:
extra_rpaths = ":".join(compiler.extra_rpaths)
env.set("SPACK_COMPILER_EXTRA_RPATHS", extra_rpaths)
# Add spack build environment path with compiler wrappers first in
# the path. We add the compiler wrapper path, which includes default
# wrappers (cc, c++, f77, f90), AND a subdirectory containing
# compiler-specific symlinks. The latter ensures that builds that
# are sensitive to the *name* of the compiler see the right name when
# we're building with the wrappers.
#
# Conflicts on case-insensitive systems (like "CC" and "cc") are
# handled by putting one in the <build_env_path>/case-insensitive
# directory. Add that to the path too.
env_paths = []
compiler_specific = os.path.join(
spack.paths.build_env_path, os.path.dirname(pkg.compiler.link_paths["cc"])
)
for item in [spack.paths.build_env_path, compiler_specific]:
env_paths.append(item)
ci = os.path.join(item, "case-insensitive")
if os.path.isdir(ci):
env_paths.append(ci)
tty.debug("Adding compiler bin/ paths: " + " ".join(env_paths))
for item in env_paths:
env.prepend_path("PATH", item)
env.set_path(SPACK_ENV_PATH, env_paths)
# Working directory for the spack command itself, for debug logs.
if spack.config.get("config:debug"):
env.set(SPACK_DEBUG, "TRUE")
env.set(SPACK_SHORT_SPEC, pkg.spec.short_spec)
env.set(SPACK_DEBUG_LOG_ID, pkg.spec.format("{name}-{hash:7}"))
env.set(SPACK_DEBUG_LOG_DIR, spack.paths.spack_working_dir)
env.set(SPACK_DEBUG_LOG_DIR, spack.main.spack_working_dir)
if spack.config.get("config:ccache"):
# Enable ccache in the compiler wrapper
@@ -386,65 +488,70 @@ def set_wrapper_variables(pkg, env):
env.set("CCACHE_DISABLE", "1")
# Gather information about various types of dependencies
rpath_hashes = set(s.dag_hash() for s in get_rpath_deps(pkg))
link_deps = pkg.spec.traverse(root=False, order="topo", deptype=dt.LINK)
external_link_deps, nonexternal_link_deps = stable_partition(link_deps, lambda d: d.external)
link_deps = set(pkg.spec.traverse(root=False, deptype=("link")))
rpath_deps = get_rpath_deps(pkg)
link_dirs = []
include_dirs = []
rpath_dirs = []
for dep in chain(external_link_deps, nonexternal_link_deps):
# TODO: is_system_path is wrong, but even if we knew default -L, -I flags from the compiler
# and default search dirs from the dynamic linker, it's not obvious how to avoid a possibly
# expensive search in `query.libs.directories` and `query.headers.directories`, which is
# what this branch is trying to avoid.
if is_system_path(dep.prefix):
continue
# TODO: as of Spack 0.22, multiple instances of the same package may occur among the link
# deps, so keying by name is wrong. In practice it is not problematic: we obtain the same
# gcc-runtime / glibc here, and repeatedly add the same dirs that are later deduped.
query = pkg.spec[dep.name]
dep_link_dirs = []
try:
# Locating libraries can be time consuming, so log start and finish.
tty.debug(f"Collecting libraries for {dep.name}")
dep_link_dirs.extend(query.libs.directories)
tty.debug(f"Libraries for {dep.name} have been collected.")
except NoLibrariesError:
tty.debug(f"No libraries found for {dep.name}")
def _prepend_all(list_to_modify, items_to_add):
# Update the original list (creating a new list would be faster but
# may not be convenient)
for item in reversed(list(items_to_add)):
list_to_modify.insert(0, item)
for default_lib_dir in ("lib", "lib64"):
default_lib_prefix = os.path.join(dep.prefix, default_lib_dir)
if os.path.isdir(default_lib_prefix):
dep_link_dirs.append(default_lib_prefix)
def update_compiler_args_for_dep(dep):
if dep in link_deps and (not is_system_path(dep.prefix)):
query = pkg.spec[dep.name]
dep_link_dirs = list()
try:
# In some circumstances (particularly for externals) finding
# libraries packages can be time consuming, so indicate that
# we are performing this operation (and also report when it
# finishes).
tty.debug("Collecting libraries for {0}".format(dep.name))
dep_link_dirs.extend(query.libs.directories)
tty.debug("Libraries for {0} have been collected.".format(dep.name))
except NoLibrariesError:
tty.debug("No libraries found for {0}".format(dep.name))
link_dirs[:0] = dep_link_dirs
if dep.dag_hash() in rpath_hashes:
rpath_dirs[:0] = dep_link_dirs
for default_lib_dir in ["lib", "lib64"]:
default_lib_prefix = os.path.join(dep.prefix, default_lib_dir)
if os.path.isdir(default_lib_prefix):
dep_link_dirs.append(default_lib_prefix)
try:
tty.debug(f"Collecting headers for {dep.name}")
include_dirs[:0] = query.headers.directories
tty.debug(f"Headers for {dep.name} have been collected.")
except NoHeadersError:
tty.debug(f"No headers found for {dep.name}")
_prepend_all(link_dirs, dep_link_dirs)
if dep in rpath_deps:
_prepend_all(rpath_dirs, dep_link_dirs)
# The top-level package is heuristically rpath'ed.
for libdir in ("lib64", "lib"):
try:
_prepend_all(include_dirs, query.headers.directories)
except NoHeadersError:
tty.debug("No headers found for {0}".format(dep.name))
for dspec in pkg.spec.traverse(root=False, order="post"):
if dspec.external:
update_compiler_args_for_dep(dspec)
# Just above, we prepended entries for -L/-rpath for externals. We
# now do this for non-external packages so that Spack-built packages
# are searched first for libraries etc.
for dspec in pkg.spec.traverse(root=False, order="post"):
if not dspec.external:
update_compiler_args_for_dep(dspec)
# The top-level package is always RPATHed. It hasn't been installed yet
# so the RPATHs are added unconditionally (e.g. even though lib64/ may
# not be created for the install).
for libdir in ["lib64", "lib"]:
lib_path = os.path.join(pkg.prefix, libdir)
rpath_dirs.insert(0, lib_path)
# TODO: filter_system_paths is again wrong (and probably unnecessary due to the is_system_path
# branch above). link_dirs should be filtered with entries from _parse_link_paths.
link_dirs = list(dedupe(filter_system_paths(link_dirs)))
include_dirs = list(dedupe(filter_system_paths(include_dirs)))
rpath_dirs = list(dedupe(filter_system_paths(rpath_dirs)))
default_dynamic_linker_filter = spack.compilers.libraries.dynamic_linker_filter_for(pkg.spec)
if default_dynamic_linker_filter:
rpath_dirs = default_dynamic_linker_filter(rpath_dirs)
# Spack managed directories include the stage, store and upstream stores. We extend this with
# their real paths to make it more robust (e.g. /tmp vs /private/tmp on macOS).
spack_managed_dirs: Set[str] = {
@@ -473,12 +580,14 @@ def set_package_py_globals(pkg, context: Context = Context.BUILD):
"""
module = ModuleChangePropagator(pkg)
jobs = spack.config.determine_number_of_jobs(parallel=pkg.parallel)
module.make_jobs = jobs
if context == Context.BUILD:
module.std_cmake_args = spack.build_systems.cmake.CMakeBuilder.std_args(pkg)
module.std_meson_args = spack.build_systems.meson.MesonBuilder.std_args(pkg)
module.std_pip_args = spack.build_systems.python.PythonPipBuilder.std_args(pkg)
jobs = determine_number_of_jobs(parallel=pkg.parallel)
module.make_jobs = jobs
# TODO: make these build deps that can be installed if not found.
module.make = MakeExecutable("make", jobs)
module.gmake = MakeExecutable("gmake", jobs)
@@ -499,19 +608,22 @@ def set_package_py_globals(pkg, context: Context = Context.BUILD):
# Put spack compiler paths in module scope. (Some packages use it
# in setup_run_environment etc, so don't put it context == build)
link_dir = spack.paths.build_env_path
pkg_compiler = None
try:
pkg_compiler = pkg.compiler
except spack.compilers.NoCompilerForSpecError as e:
tty.debug(f"cannot set 'spack_cc': {str(e)}")
# FIXME (compiler as nodes): make this more general, and not tied to three languages
# Maybe add a callback?
global_names = {
"c": ("spack_cc",),
"cxx": ("spack_cxx",),
"fortran": ("spack_fc", "spack_f77"),
}
for language in ("c", "cxx", "fortran"):
spec = pkg.spec.dependencies(virtuals=[language])
value = None if not spec else os.path.join(link_dir, spec[0].package.link_paths[language])
for name in global_names[language]:
setattr(module, name, value)
if pkg_compiler is not None:
module.spack_cc = os.path.join(link_dir, pkg_compiler.link_paths["cc"])
module.spack_cxx = os.path.join(link_dir, pkg_compiler.link_paths["cxx"])
module.spack_f77 = os.path.join(link_dir, pkg_compiler.link_paths["f77"])
module.spack_fc = os.path.join(link_dir, pkg_compiler.link_paths["fc"])
else:
module.spack_cc = None
module.spack_cxx = None
module.spack_f77 = None
module.spack_fc = None
# Useful directories within the prefix are encapsulated in
# a Prefix object.
@@ -643,6 +755,19 @@ def get_rpath_deps(pkg: spack.package_base.PackageBase) -> List[spack.spec.Spec]
return _get_rpath_deps_from_spec(pkg.spec, pkg.transitive_rpaths)
def get_rpaths(pkg):
"""Get a list of all the rpaths for a package."""
rpaths = [pkg.prefix.lib, pkg.prefix.lib64]
deps = get_rpath_deps(pkg)
rpaths.extend(d.prefix.lib for d in deps if os.path.isdir(d.prefix.lib))
rpaths.extend(d.prefix.lib64 for d in deps if os.path.isdir(d.prefix.lib64))
# Second module is our compiler mod name. We use that to get rpaths from
# module show output.
if pkg.compiler.modules and len(pkg.compiler.modules) > 1:
rpaths.append(path_from_modules([pkg.compiler.modules[1]]))
return list(dedupe(filter_system_paths(rpaths)))
def load_external_modules(pkg):
"""Traverse a package's spec DAG and load any external modules.
@@ -678,11 +803,13 @@ def setup_package(pkg, dirty, context: Context = Context.BUILD):
context == Context.TEST and pkg.test_requires_compiler
)
if need_compiler:
set_compiler_environment_variables(pkg, env_mods)
set_wrapper_variables(pkg, env_mods)
# Platform specific setup goes before package specific setup. This is for setting
# defaults like MACOSX_DEPLOYMENT_TARGET on macOS.
platform = spack.platforms.by_name(pkg.spec.architecture.platform)
target = platform.target(pkg.spec.architecture.target)
platform.setup_platform_environment(pkg, env_mods)
tty.debug("setup_package: grabbing modifications from dependencies")
@@ -702,8 +829,28 @@ def setup_package(pkg, dirty, context: Context = Context.BUILD):
# Load modules on an already clean environment, just before applying Spack's
# own environment modifications. This ensures Spack controls CC/CXX/... variables.
if need_compiler:
tty.debug("setup_package: loading compiler modules")
for mod in pkg.compiler.modules:
load_module(mod)
# kludge to handle cray mpich and libsci being automatically loaded by
# PrgEnv modules on cray platform. Module unload does no damage when
# unnecessary
on_cray, _ = _on_cray()
if on_cray and not dirty:
for mod in ["cray-mpich", "cray-libsci"]:
module("unload", mod)
if target and target.module_name:
load_module(target.module_name)
load_external_modules(pkg)
implicit_rpaths = pkg.compiler.implicit_rpaths()
if implicit_rpaths:
env_mods.set("SPACK_COMPILER_IMPLICIT_RPATHS", ":".join(implicit_rpaths))
# Make sure nothing's strange about the Spack environment.
validate(env_mods, tty.warn)
env_mods.apply_modifications()
@@ -893,12 +1040,6 @@ def set_all_package_py_globals(self):
# This includes runtime dependencies, also runtime deps of direct build deps.
set_package_py_globals(pkg, context=Context.RUN)
# Looping over the set of packages a second time
# ensures all globals are loaded into the module space prior to
# any package setup. This guarantees package setup methods have
# access to expected module level definitions such as "spack_cc"
for dspec, flag in chain(self.external, self.nonexternal):
pkg = dspec.package
for spec in dspec.dependents():
# Note: some specs have dependents that are unreachable from the root, so avoid
# setting globals for those.
@@ -908,15 +1049,6 @@ def set_all_package_py_globals(self):
pkg.setup_dependent_package(dependent_module, spec)
dependent_module.propagate_changes_to_mro()
if self.context == Context.BUILD:
pkg = self.specs[0].package
module = ModuleChangePropagator(pkg)
# std_cmake_args is not sufficiently static to be defined
# in set_package_py_globals and is deprecated so its handled
# here as a special case
module.std_cmake_args = spack.build_systems.cmake.CMakeBuilder.std_args(pkg)
module.propagate_changes_to_mro()
def get_env_modifications(self) -> EnvironmentModifications:
"""Returns the environment variable modifications for the given input specs and context.
Environment modifications include:
@@ -986,61 +1118,45 @@ def _make_runnable(self, dep: spack.spec.Spec, env: EnvironmentModifications):
env.prepend_path("PATH", bin_dir)
def get_cmake_prefix_path(pkg):
# Note that unlike modifications_from_dependencies, this does not include
# any edits to CMAKE_PREFIX_PATH defined in custom
# setup_dependent_build_environment implementations of dependency packages
build_deps = set(pkg.spec.dependencies(deptype=("build", "test")))
link_deps = set(pkg.spec.traverse(root=False, deptype=("link")))
build_link_deps = build_deps | link_deps
spack_built = []
externals = []
# modifications_from_dependencies updates CMAKE_PREFIX_PATH by first
# prepending all externals and then all non-externals
for dspec in pkg.spec.traverse(root=False, order="post"):
if dspec in build_link_deps:
if dspec.external:
externals.insert(0, dspec)
else:
spack_built.insert(0, dspec)
ordered_build_link_deps = spack_built + externals
cmake_prefix_path_entries = []
for spec in ordered_build_link_deps:
cmake_prefix_path_entries.extend(spec.package.cmake_prefix_paths)
return filter_system_paths(cmake_prefix_path_entries)
def _setup_pkg_and_run(
serialized_pkg: "spack.subprocess_context.PackageInstallContext",
function: Callable,
kwargs: Dict,
write_pipe: Connection,
input_pipe: Optional[Connection],
jsfd1: Optional[Connection],
jsfd2: Optional[Connection],
serialized_pkg, function, kwargs, write_pipe, input_multiprocess_fd, jsfd1, jsfd2
):
"""Main entry point in the child process for Spack builds.
``_setup_pkg_and_run`` is called by the child process created in
``start_build_process()``, and its main job is to run ``function()`` on behalf of
some Spack installation (see :ref:`spack.installer.PackageInstaller._install_task`).
The child process is passed a ``write_pipe``, on which it's expected to send one of
the following:
* ``StopPhase``: error raised by a build process indicating it's stopping at a
particular build phase.
* ``BaseException``: any exception raised by a child build process, which will be
wrapped in ``ChildError`` (which adds a bunch of debug info and log context) and
raised in the parent.
* The return value of ``function()``, which can be anything (except an exception).
This is returned to the caller.
Note: ``jsfd1`` and ``jsfd2`` are passed solely to ensure that the child process
does not close these file descriptors. Some ``multiprocessing`` backends will close
them automatically in the child if they are not passed at process creation time.
Arguments:
serialized_pkg: Spack package install context object (serialized form of the
package that we'll build in the child process).
function: function to call in the child process; serialized_pkg is passed to
this as the first argument.
kwargs: additional keyword arguments to pass to ``function()``.
write_pipe: multiprocessing ``Connection`` to the parent process, to which the
child *must* send a result (or an error) back to parent on.
input_multiprocess_fd: stdin from the parent (not passed currently on Windows)
jsfd1: gmake Jobserver file descriptor 1.
jsfd2: gmake Jobserver file descriptor 2.
"""
context: str = kwargs.get("context", "build")
try:
# We are in the child process. Python sets sys.stdin to open(os.devnull) to prevent our
# process and its parent from simultaneously reading from the original stdin. But, we
# assume that the parent process is not going to read from it till we are done with the
# child, so we undo Python's precaution. closefd=False since Connection has ownership.
if input_pipe is not None:
sys.stdin = os.fdopen(input_pipe.fileno(), closefd=False)
# We are in the child process. Python sets sys.stdin to
# open(os.devnull) to prevent our process and its parent from
# simultaneously reading from the original stdin. But, we assume
# that the parent process is not going to read from it till we
# are done with the child, so we undo Python's precaution.
if input_multiprocess_fd is not None:
sys.stdin = os.fdopen(input_multiprocess_fd.fd)
pkg = serialized_pkg.restore()
@@ -1052,18 +1168,17 @@ def _setup_pkg_and_run(
return_value = function(pkg, kwargs)
write_pipe.send(return_value)
except spack.error.StopPhase as e:
except StopPhase as e:
# Do not create a full ChildError from this, it's not an error
# it's a control statement.
write_pipe.send(e)
except BaseException as e:
except BaseException:
# catch ANYTHING that goes wrong in the child process
exc_type, exc, tb = sys.exc_info()
# Need to unwind the traceback in the child because traceback
# objects can't be sent to the parent.
exc_type = type(e)
tb = e.__traceback__
tb_string = "".join(traceback.format_exception(exc_type, e, tb))
tb_string = traceback.format_exc()
# build up some context from the offending package so we can
# show that, too.
@@ -1080,8 +1195,8 @@ def _setup_pkg_and_run(
elif context == "test":
logfile = os.path.join(pkg.test_suite.stage, pkg.test_suite.test_log_name(pkg.spec))
error_msg = str(e)
if isinstance(e, (spack.multimethod.NoSuchMethodError, AttributeError)):
error_msg = str(exc)
if isinstance(exc, (spack.multimethod.NoSuchMethodError, AttributeError)):
process = "test the installation" if context == "test" else "build from sources"
error_msg = (
"The '{}' package cannot find an attribute while trying to {}. "
@@ -1091,7 +1206,7 @@ def _setup_pkg_and_run(
"More information at https://spack.readthedocs.io/en/latest/packaging_guide.html#installation-procedure"
).format(pkg.name, process, context)
error_msg = colorize("@*R{{{}}}".format(error_msg))
error_msg = "{}\n\n{}".format(str(e), error_msg)
error_msg = "{}\n\n{}".format(str(exc), error_msg)
# make a pickleable exception to send to parent.
msg = "%s: %s" % (exc_type.__name__, error_msg)
@@ -1109,8 +1224,8 @@ def _setup_pkg_and_run(
finally:
write_pipe.close()
if input_pipe is not None:
input_pipe.close()
if input_multiprocess_fd is not None:
input_multiprocess_fd.close()
def start_build_process(pkg, function, kwargs):
@@ -1137,9 +1252,23 @@ def child_fun():
If something goes wrong, the child process catches the error and
passes it to the parent wrapped in a ChildError. The parent is
expected to handle (or re-raise) the ChildError.
This uses `multiprocessing.Process` to create the child process. The
mechanism used to create the process differs on different operating
systems and for different versions of Python. In some cases "fork"
is used (i.e. the "fork" system call) and some cases it starts an
entirely new Python interpreter process (in the docs this is referred
to as the "spawn" start method). Breaking it down by OS:
- Linux always uses fork.
- Mac OS uses fork before Python 3.8 and "spawn" for 3.8 and after.
- Windows always uses the "spawn" start method.
For more information on `multiprocessing` child process creation
mechanisms, see https://docs.python.org/3/library/multiprocessing.html#contexts-and-start-methods
"""
read_pipe, write_pipe = multiprocessing.Pipe(duplex=False)
input_fd = None
input_multiprocess_fd = None
jobserver_fd1 = None
jobserver_fd2 = None
@@ -1148,13 +1277,14 @@ def child_fun():
try:
# Forward sys.stdin when appropriate, to allow toggling verbosity
if sys.platform != "win32" and sys.stdin.isatty() and hasattr(sys.stdin, "fileno"):
input_fd = Connection(os.dup(sys.stdin.fileno()))
input_fd = os.dup(sys.stdin.fileno())
input_multiprocess_fd = MultiProcessFd(input_fd)
mflags = os.environ.get("MAKEFLAGS", False)
if mflags:
m = re.search(r"--jobserver-[^=]*=(\d),(\d)", mflags)
if m:
jobserver_fd1 = Connection(int(m.group(1)))
jobserver_fd2 = Connection(int(m.group(2)))
jobserver_fd1 = MultiProcessFd(int(m.group(1)))
jobserver_fd2 = MultiProcessFd(int(m.group(2)))
p = multiprocessing.Process(
target=_setup_pkg_and_run,
@@ -1163,7 +1293,7 @@ def child_fun():
function,
kwargs,
write_pipe,
input_fd,
input_multiprocess_fd,
jobserver_fd1,
jobserver_fd2,
),
@@ -1183,8 +1313,8 @@ def child_fun():
finally:
# Close the input stream in the parent process
if input_fd is not None:
input_fd.close()
if input_multiprocess_fd is not None:
input_multiprocess_fd.close()
def exitcode_msg(p):
typ = "exit" if p.exitcode >= 0 else "signal"
@@ -1199,7 +1329,7 @@ def exitcode_msg(p):
p.join()
# If returns a StopPhase, raise it
if isinstance(child_result, spack.error.StopPhase):
if isinstance(child_result, StopPhase):
# do not print
raise child_result
@@ -1222,7 +1352,7 @@ def exitcode_msg(p):
return child_result
CONTEXT_BASES = (spack.package_base.PackageBase, spack.builder.Builder)
CONTEXT_BASES = (spack.package_base.PackageBase, spack.build_systems._checks.BaseBuilder)
def get_package_context(traceback, context=3):
@@ -1375,7 +1505,7 @@ def long_message(self):
out.write(" {0}\n".format(self.log_name))
# Also output the test log path IF it exists
if self.context != "test" and have_log:
if self.context != "test":
test_log = join_path(os.path.dirname(self.log_name), spack_install_test_log)
if os.path.isfile(test_log):
out.write("\nSee test log for details:\n")
@@ -1408,6 +1538,17 @@ def _make_child_error(msg, module, name, traceback, log, log_type, context):
return ChildError(msg, module, name, traceback, log, log_type, context)
class StopPhase(spack.error.SpackError):
"""Pickle-able exception to control stopped builds."""
def __reduce__(self):
return _make_stop_phase, (self.message, self.long_message)
def _make_stop_phase(msg, long_msg):
return StopPhase(msg, long_msg)
def write_log_summary(out, log_type, log, last=None):
errors, warnings = parse_log_events(log)
nerr = len(errors)
@@ -1441,21 +1582,21 @@ class ModuleChangePropagator:
_PROTECTED_NAMES = ("package", "current_module", "modules_in_mro", "_set_attributes")
def __init__(self, package: spack.package_base.PackageBase) -> None:
def __init__(self, package):
self._set_self_attributes("package", package)
self._set_self_attributes("current_module", package.module)
#: Modules for the classes in the MRO up to PackageBase
modules_in_mro = []
for cls in package.__class__.__mro__:
module = getattr(cls, "module", None)
for cls in inspect.getmro(type(package)):
module = cls.module
if module is None or module is spack.package_base:
break
if module is self.current_module:
if module == self.current_module:
continue
if module == spack.package_base:
break
modules_in_mro.append(module)
self._set_self_attributes("modules_in_mro", modules_in_mro)
self._set_self_attributes("_set_attributes", {})

View File

@@ -8,8 +8,7 @@
import llnl.util.lang
import spack.builder
import spack.error
import spack.phase_callbacks
import spack.installer
import spack.relocate
import spack.spec
import spack.store
@@ -35,7 +34,7 @@ def check_paths(path_list, filetype, predicate):
if not predicate(abs_path):
msg = "Install failed for {0}. No such {1} in prefix: {2}"
msg = msg.format(pkg.name, filetype, path)
raise spack.error.InstallError(msg)
raise spack.installer.InstallError(msg)
check_paths(pkg.sanity_check_is_file, "file", os.path.isfile)
check_paths(pkg.sanity_check_is_dir, "directory", os.path.isdir)
@@ -43,7 +42,7 @@ def check_paths(path_list, filetype, predicate):
ignore_file = llnl.util.lang.match_predicate(spack.store.STORE.layout.hidden_file_regexes)
if all(map(ignore_file, os.listdir(pkg.prefix))):
msg = "Install failed for {0}. Nothing was installed!"
raise spack.error.InstallError(msg.format(pkg.name))
raise spack.installer.InstallError(msg.format(pkg.name))
def apply_macos_rpath_fixups(builder: spack.builder.Builder):
@@ -64,7 +63,7 @@ def apply_macos_rpath_fixups(builder: spack.builder.Builder):
def ensure_build_dependencies_or_raise(
spec: spack.spec.Spec, dependencies: List[str], error_msg: str
spec: spack.spec.Spec, dependencies: List[spack.spec.Spec], error_msg: str
):
"""Ensure that some build dependencies are present in the concrete spec.
@@ -72,7 +71,7 @@ def ensure_build_dependencies_or_raise(
Args:
spec: concrete spec to be checked.
dependencies: list of package names of required build dependencies
dependencies: list of abstract specs to be satisfied
error_msg: brief error message to be prepended to a longer description
Raises:
@@ -128,8 +127,8 @@ def execute_install_time_tests(builder: spack.builder.Builder):
builder.pkg.tester.phase_tests(builder, "install", builder.install_time_test_callbacks)
class BuilderWithDefaults(spack.builder.Builder):
"""Base class for all specific builders with common callbacks registered."""
class BaseBuilder(spack.builder.Builder):
"""Base class for builders to register common checks"""
# Check that self.prefix is there after installation
spack.phase_callbacks.run_after("install")(sanity_check_prefix)
spack.builder.run_after("install")(sanity_check_prefix)

View File

@@ -2,11 +2,10 @@
# Spack Project Developers. See the top-level COPYRIGHT file for details.
#
# SPDX-License-Identifier: (Apache-2.0 OR MIT)
import os
import llnl.util.filesystem as fs
import spack.directives
import spack.package_base
import spack.util.executable
from .autotools import AutotoolsBuilder, AutotoolsPackage
@@ -47,12 +46,18 @@ class AspellDictPackage(AutotoolsPackage):
#: Override the default autotools builder
AutotoolsBuilder = AspellBuilder
def patch(self):
def view_destination(self, view):
aspell_spec = self.spec["aspell"]
if view.get_projection_for_spec(aspell_spec) != aspell_spec.prefix:
raise spack.package_base.ExtensionError(
"aspell does not support non-global extensions"
)
aspell = aspell_spec.command
dictdir = aspell("dump", "config", "dict-dir", output=str).strip()
datadir = aspell("dump", "config", "data-dir", output=str).strip()
dictdir = os.path.relpath(dictdir, aspell_spec.prefix)
datadir = os.path.relpath(datadir, aspell_spec.prefix)
fs.filter_file(r"^dictdir=.*$", f"dictdir=/{dictdir}", "configure")
fs.filter_file(r"^datadir=.*$", f"datadir=/{datadir}", "configure")
return aspell("dump", "config", "dict-dir", output=str).strip()
def view_source(self):
return self.prefix.lib
def patch(self):
fs.filter_file(r"^dictdir=.*$", "dictdir=/lib", "configure")
fs.filter_file(r"^datadir=.*$", "datadir=/lib", "configure")

View File

@@ -2,23 +2,19 @@
# Spack Project Developers. See the top-level COPYRIGHT file for details.
#
# SPDX-License-Identifier: (Apache-2.0 OR MIT)
import inspect
import os
import os.path
import stat
import subprocess
from typing import Callable, List, Optional, Set, Tuple, Union
from typing import List
import llnl.util.filesystem as fs
import llnl.util.tty as tty
import spack.build_environment
import spack.builder
import spack.compilers.libraries
import spack.error
import spack.package_base
import spack.phase_callbacks
import spack.spec
import spack.util.prefix
from spack.directives import build_system, conflicts, depends_on
from spack.multimethod import when
from spack.operating_systems.mac_os import macos_version
@@ -26,7 +22,7 @@
from spack.version import Version
from ._checks import (
BuilderWithDefaults,
BaseBuilder,
apply_macos_rpath_fixups,
ensure_build_dependencies_or_raise,
execute_build_time_tests,
@@ -73,14 +69,14 @@ def flags_to_build_system_args(self, flags):
# Legacy methods (used by too many packages to change them,
# need to forward to the builder)
def enable_or_disable(self, *args, **kwargs):
return spack.builder.create(self).enable_or_disable(*args, **kwargs)
return self.builder.enable_or_disable(*args, **kwargs)
def with_or_without(self, *args, **kwargs):
return spack.builder.create(self).with_or_without(*args, **kwargs)
return self.builder.with_or_without(*args, **kwargs)
@spack.builder.builder("autotools")
class AutotoolsBuilder(BuilderWithDefaults):
class AutotoolsBuilder(BaseBuilder):
"""The autotools builder encodes the default way of installing software built
with autotools. It has four phases that can be overridden, if need be:
@@ -161,7 +157,7 @@ class AutotoolsBuilder(BuilderWithDefaults):
install_libtool_archives = False
@property
def patch_config_files(self) -> bool:
def patch_config_files(self):
"""Whether to update old ``config.guess`` and ``config.sub`` files
distributed with the tarball.
@@ -181,7 +177,7 @@ def patch_config_files(self) -> bool:
)
@property
def _removed_la_files_log(self) -> str:
def _removed_la_files_log(self):
"""File containing the list of removed libtool archives"""
build_dir = self.build_directory
if not os.path.isabs(self.build_directory):
@@ -189,15 +185,15 @@ def _removed_la_files_log(self) -> str:
return os.path.join(build_dir, "removed_la_files.txt")
@property
def archive_files(self) -> List[str]:
def archive_files(self):
"""Files to archive for packages based on autotools"""
files = [os.path.join(self.build_directory, "config.log")]
if not self.install_libtool_archives:
files.append(self._removed_la_files_log)
return files
@spack.phase_callbacks.run_after("autoreconf")
def _do_patch_config_files(self) -> None:
@spack.builder.run_after("autoreconf")
def _do_patch_config_files(self):
"""Some packages ship with older config.guess/config.sub files and need to
have these updated when installed on a newer architecture.
@@ -253,7 +249,7 @@ def runs_ok(script_abs_path):
# An external gnuconfig may not not have a prefix.
if gnuconfig_dir is None:
raise spack.error.InstallError(
raise spack.build_environment.InstallError(
"Spack could not find substitutes for GNU config files because no "
"prefix is available for the `gnuconfig` package. Make sure you set a "
"prefix path instead of modules for external `gnuconfig`."
@@ -273,7 +269,7 @@ def runs_ok(script_abs_path):
msg += (
" or the `gnuconfig` package prefix is misconfigured as" " an external package"
)
raise spack.error.InstallError(msg)
raise spack.build_environment.InstallError(msg)
# Filter working substitutes
candidates = [f for f in candidates if runs_ok(f)]
@@ -298,7 +294,9 @@ def runs_ok(script_abs_path):
and set the prefix to the directory containing the `config.guess` and
`config.sub` files.
"""
raise spack.error.InstallError(msg.format(", ".join(to_be_found), self.pkg.name))
raise spack.build_environment.InstallError(
msg.format(", ".join(to_be_found), self.name)
)
# Copy the good files over the bad ones
for abs_path in to_be_patched:
@@ -308,8 +306,8 @@ def runs_ok(script_abs_path):
fs.copy(substitutes[name], abs_path)
os.chmod(abs_path, mode)
@spack.phase_callbacks.run_before("configure")
def _patch_usr_bin_file(self) -> None:
@spack.builder.run_before("configure")
def _patch_usr_bin_file(self):
"""On NixOS file is not available in /usr/bin/file. Patch configure
scripts to use file from path."""
@@ -320,8 +318,8 @@ def _patch_usr_bin_file(self) -> None:
with fs.keep_modification_time(*x.filenames):
x.filter(regex="/usr/bin/file", repl="file", string=True)
@spack.phase_callbacks.run_before("configure")
def _set_autotools_environment_variables(self) -> None:
@spack.builder.run_before("configure")
def _set_autotools_environment_variables(self):
"""Many autotools builds use a version of mknod.m4 that fails when
running as root unless FORCE_UNSAFE_CONFIGURE is set to 1.
@@ -334,8 +332,8 @@ def _set_autotools_environment_variables(self) -> None:
"""
os.environ["FORCE_UNSAFE_CONFIGURE"] = "1"
@spack.phase_callbacks.run_before("configure")
def _do_patch_libtool_configure(self) -> None:
@spack.builder.run_before("configure")
def _do_patch_libtool_configure(self):
"""Patch bugs that propagate from libtool macros into "configure" and
further into "libtool". Note that patches that can be fixed by patching
"libtool" directly should be implemented in the _do_patch_libtool method
@@ -362,8 +360,8 @@ def _do_patch_libtool_configure(self) -> None:
# Support Libtool 2.4.2 and older:
x.filter(regex=r'^(\s*test \$p = "-R")(; then\s*)$', repl=r'\1 || test x-l = x"$p"\2')
@spack.phase_callbacks.run_after("configure")
def _do_patch_libtool(self) -> None:
@spack.builder.run_after("configure")
def _do_patch_libtool(self):
"""If configure generates a "libtool" script that does not correctly
detect the compiler (and patch_libtool is set), patch in the correct
values for libtool variables.
@@ -397,44 +395,33 @@ def _do_patch_libtool(self) -> None:
markers[tag] = "LIBTOOL TAG CONFIG: {0}".format(tag.upper())
# Replace empty linker flag prefixes:
if self.spec.satisfies("%nag"):
if self.pkg.compiler.name == "nag":
# Nag is mixed with gcc and g++, which are recognized correctly.
# Therefore, we change only Fortran values:
nag_pkg = self.spec["fortran"].package
for tag in ["fc", "f77"]:
marker = markers[tag]
x.filter(
regex='^wl=""$',
repl=f'wl="{nag_pkg.linker_arg}"',
start_at=f"# ### BEGIN {marker}",
stop_at=f"# ### END {marker}",
repl='wl="{0}"'.format(self.pkg.compiler.linker_arg),
start_at="# ### BEGIN {0}".format(marker),
stop_at="# ### END {0}".format(marker),
)
else:
compiler_spec = spack.compilers.libraries.compiler_spec(self.spec)
if compiler_spec:
x.filter(regex='^wl=""$', repl='wl="{0}"'.format(compiler_spec.package.linker_arg))
x.filter(regex='^wl=""$', repl='wl="{0}"'.format(self.pkg.compiler.linker_arg))
# Replace empty PIC flag values:
for compiler, marker in markers.items():
if compiler == "cc":
language = "c"
elif compiler == "cxx":
language = "cxx"
else:
language = "fortran"
if language not in self.spec:
continue
for cc, marker in markers.items():
x.filter(
regex='^pic_flag=""$',
repl=f'pic_flag="{self.spec[language].package.pic_flag}"',
start_at=f"# ### BEGIN {marker}",
stop_at=f"# ### END {marker}",
repl='pic_flag="{0}"'.format(
getattr(self.pkg.compiler, "{0}_pic_flag".format(cc))
),
start_at="# ### BEGIN {0}".format(marker),
stop_at="# ### END {0}".format(marker),
)
# Other compiler-specific patches:
if self.spec.satisfies("%fj"):
if self.pkg.compiler.name == "fj":
x.filter(regex="-nostdlib", repl="", string=True)
rehead = r"/\S*/"
for o in [
@@ -447,7 +434,7 @@ def _do_patch_libtool(self) -> None:
r"crtendS\.o",
]:
x.filter(regex=(rehead + o), repl="")
elif self.spec.satisfies("%nag"):
elif self.pkg.compiler.name == "nag":
for tag in ["fc", "f77"]:
marker = markers[tag]
start_at = "# ### BEGIN {0}".format(marker)
@@ -522,64 +509,27 @@ def _do_patch_libtool(self) -> None:
)
@property
def configure_directory(self) -> str:
def configure_directory(self):
"""Return the directory where 'configure' resides."""
return self.pkg.stage.source_path
@property
def configure_abs_path(self) -> str:
def configure_abs_path(self):
# Absolute path to configure
configure_abs_path = os.path.join(os.path.abspath(self.configure_directory), "configure")
return configure_abs_path
@property
def build_directory(self) -> str:
def build_directory(self):
"""Override to provide another place to build the package"""
return self.configure_directory
@spack.phase_callbacks.run_before("autoreconf")
def delete_configure_to_force_update(self) -> None:
@spack.builder.run_before("autoreconf")
def delete_configure_to_force_update(self):
if self.force_autoreconf:
fs.force_remove(self.configure_abs_path)
@property
def autoreconf_search_path_args(self) -> List[str]:
"""Search path includes for autoreconf. Add an -I flag for all `aclocal` dirs
of build deps, skips the default path of automake, move external include
flags to the back, since they might pull in unrelated m4 files shadowing
spack dependencies."""
return _autoreconf_search_path_args(self.spec)
@spack.phase_callbacks.run_after("autoreconf")
def set_configure_or_die(self) -> None:
"""Ensure the presence of a "configure" script, or raise. If the "configure"
is found, a module level attribute is set.
Raises:
RuntimeError: if the "configure" script is not found
"""
# Check if the "configure" script is there. If not raise a RuntimeError.
if not os.path.exists(self.configure_abs_path):
msg = "configure script not found in {0}"
raise RuntimeError(msg.format(self.configure_directory))
# Monkey-patch the configure script in the corresponding module
globals_for_pkg = spack.build_environment.ModuleChangePropagator(self.pkg)
globals_for_pkg.configure = Executable(self.configure_abs_path)
globals_for_pkg.propagate_changes_to_mro()
def configure_args(self) -> List[str]:
"""Return the list of all the arguments that must be passed to configure,
except ``--prefix`` which will be pre-pended to the list.
"""
return []
def autoreconf(
self,
pkg: spack.package_base.PackageBase,
spec: spack.spec.Spec,
prefix: spack.util.prefix.Prefix,
) -> None:
def autoreconf(self, pkg, spec, prefix):
"""Not needed usually, configure should be already there"""
# If configure exists nothing needs to be done
@@ -599,19 +549,45 @@ def autoreconf(
tty.warn("* a custom AUTORECONF phase in the package *")
tty.warn("*********************************************************")
with fs.working_dir(self.configure_directory):
m = inspect.getmodule(self.pkg)
# This line is what is needed most of the time
# --install, --verbose, --force
autoreconf_args = ["-ivf"]
autoreconf_args += self.autoreconf_search_path_args
autoreconf_args += self.autoreconf_extra_args
self.pkg.module.autoreconf(*autoreconf_args)
m.autoreconf(*autoreconf_args)
def configure(
self,
pkg: spack.package_base.PackageBase,
spec: spack.spec.Spec,
prefix: spack.util.prefix.Prefix,
) -> None:
@property
def autoreconf_search_path_args(self):
"""Search path includes for autoreconf. Add an -I flag for all `aclocal` dirs
of build deps, skips the default path of automake, move external include
flags to the back, since they might pull in unrelated m4 files shadowing
spack dependencies."""
return _autoreconf_search_path_args(self.spec)
@spack.builder.run_after("autoreconf")
def set_configure_or_die(self):
"""Ensure the presence of a "configure" script, or raise. If the "configure"
is found, a module level attribute is set.
Raises:
RuntimeError: if the "configure" script is not found
"""
# Check if the "configure" script is there. If not raise a RuntimeError.
if not os.path.exists(self.configure_abs_path):
msg = "configure script not found in {0}"
raise RuntimeError(msg.format(self.configure_directory))
# Monkey-patch the configure script in the corresponding module
inspect.getmodule(self.pkg).configure = Executable(self.configure_abs_path)
def configure_args(self):
"""Return the list of all the arguments that must be passed to configure,
except ``--prefix`` which will be pre-pended to the list.
"""
return []
def configure(self, pkg, spec, prefix):
"""Run "configure", with the arguments specified by the builder and an
appropriately set prefix.
"""
@@ -620,64 +596,51 @@ def configure(
options += self.configure_args()
with fs.working_dir(self.build_directory, create=True):
pkg.module.configure(*options)
inspect.getmodule(self.pkg).configure(*options)
def build(
self,
pkg: spack.package_base.PackageBase,
spec: spack.spec.Spec,
prefix: spack.util.prefix.Prefix,
) -> None:
def build(self, pkg, spec, prefix):
"""Run "make" on the build targets specified by the builder."""
# See https://autotools.io/automake/silent.html
params = ["V=1"]
params += self.build_targets
with fs.working_dir(self.build_directory):
pkg.module.make(*params)
inspect.getmodule(self.pkg).make(*params)
def install(
self,
pkg: spack.package_base.PackageBase,
spec: spack.spec.Spec,
prefix: spack.util.prefix.Prefix,
) -> None:
def install(self, pkg, spec, prefix):
"""Run "make" on the install targets specified by the builder."""
with fs.working_dir(self.build_directory):
pkg.module.make(*self.install_targets)
inspect.getmodule(self.pkg).make(*self.install_targets)
spack.phase_callbacks.run_after("build")(execute_build_time_tests)
spack.builder.run_after("build")(execute_build_time_tests)
def check(self) -> None:
def check(self):
"""Run "make" on the ``test`` and ``check`` targets, if found."""
with fs.working_dir(self.build_directory):
self.pkg._if_make_target_execute("test")
self.pkg._if_make_target_execute("check")
def _activate_or_not(
self,
name: str,
activation_word: str,
deactivation_word: str,
activation_value: Optional[Union[Callable, str]] = None,
variant=None,
) -> List[str]:
self, name, activation_word, deactivation_word, activation_value=None, variant=None
):
"""This function contain the current implementation details of
:meth:`~spack.build_systems.autotools.AutotoolsBuilder.with_or_without` and
:meth:`~spack.build_systems.autotools.AutotoolsBuilder.enable_or_disable`.
Args:
name: name of the option that is being activated or not
activation_word: the default activation word ('with' in the case of
``with_or_without``)
deactivation_word: the default deactivation word ('without' in the case of
``with_or_without``)
activation_value: callable that accepts a single value. This value is either one of the
allowed values for a multi-valued variant or the name of a bool-valued variant.
name (str): name of the option that is being activated or not
activation_word (str): the default activation word ('with' in the
case of ``with_or_without``)
deactivation_word (str): the default deactivation word ('without'
in the case of ``with_or_without``)
activation_value (typing.Callable): callable that accepts a single
value. This value is either one of the allowed values for a
multi-valued variant or the name of a bool-valued variant.
Returns the parameter to be used when the value is activated.
The special value "prefix" can also be assigned and will return
The special value 'prefix' can also be assigned and will return
``spec[name].prefix`` as activation parameter.
variant: name of the variant that is being processed (if different from option name)
variant (str): name of the variant that is being processed
(if different from option name)
Examples:
@@ -685,19 +648,19 @@ def _activate_or_not(
.. code-block:: python
variant("foo", values=("x", "y"), description=")
variant("bar", default=True, description=")
variant("ba_z", default=True, description=")
variant('foo', values=('x', 'y'), description='')
variant('bar', default=True, description='')
variant('ba_z', default=True, description='')
calling this function like:
.. code-block:: python
_activate_or_not(
"foo", "with", "without", activation_value="prefix"
'foo', 'with', 'without', activation_value='prefix'
)
_activate_or_not("bar", "with", "without")
_activate_or_not("ba-z", "with", "without", variant="ba_z")
_activate_or_not('bar', 'with', 'without')
_activate_or_not('ba-z', 'with', 'without', variant='ba_z')
will generate the following configuration options:
@@ -717,16 +680,17 @@ def _activate_or_not(
Raises:
KeyError: if name is not among known variants
"""
spec: spack.spec.Spec = self.pkg.spec
args: List[str] = []
spec = self.pkg.spec
args = []
if activation_value == "prefix":
activation_value = lambda x: spec[x].prefix
variant = variant or name
# Defensively look that the name passed as argument is among variants
if not self.pkg.has_variant(variant):
# Defensively look that the name passed as argument is among
# variants
if variant not in self.pkg.variants:
msg = '"{0}" is not a variant of "{1}"'
raise KeyError(msg.format(variant, self.pkg.name))
@@ -735,24 +699,34 @@ def _activate_or_not(
# Create a list of pairs. Each pair includes a configuration
# option and whether or not that option is activated
vdef = self.pkg.get_variant(variant)
if set(vdef.values) == set((True, False)): # type: ignore
variant_desc, _ = self.pkg.variants[variant]
if set(variant_desc.values) == set((True, False)):
# BoolValuedVariant carry information about a single option.
# Nonetheless, for uniformity of treatment we'll package them
# in an iterable of one element.
options = [(name, f"+{variant}" in spec)]
condition = "+{name}".format(name=variant)
options = [(name, condition in spec)]
else:
condition = "{variant}={value}"
# "feature_values" is used to track values which correspond to
# features which can be enabled or disabled as understood by the
# package's build system. It excludes values which have special
# meanings and do not correspond to features (e.g. "none")
feature_values = getattr(vdef.values, "feature_values", None) or vdef.values
options = [(v, f"{variant}={v}" in spec) for v in feature_values] # type: ignore
feature_values = (
getattr(variant_desc.values, "feature_values", None) or variant_desc.values
)
options = [
(value, condition.format(variant=variant, value=value) in spec)
for value in feature_values
]
# For each allowed value in the list of values
for option_value, activated in options:
# Search for an override in the package for this value
override_name = f"{activation_word}_or_{deactivation_word}_{option_value}"
override_name = "{0}_or_{1}_{2}".format(
activation_word, deactivation_word, option_value
)
line_generator = getattr(self, override_name, None) or getattr(
self.pkg, override_name, None
)
@@ -761,24 +735,19 @@ def _activate_or_not(
def _default_generator(is_activated):
if is_activated:
line = f"--{activation_word}-{option_value}"
line = "--{0}-{1}".format(activation_word, option_value)
if activation_value is not None and activation_value(
option_value
): # NOQA=ignore=E501
line = f"{line}={activation_value(option_value)}"
line += "={0}".format(activation_value(option_value))
return line
return f"--{deactivation_word}-{option_value}"
return "--{0}-{1}".format(deactivation_word, option_value)
line_generator = _default_generator
args.append(line_generator(activated))
return args
def with_or_without(
self,
name: str,
activation_value: Optional[Union[Callable, str]] = None,
variant: Optional[str] = None,
) -> List[str]:
def with_or_without(self, name, activation_value=None, variant=None):
"""Inspects a variant and returns the arguments that activate
or deactivate the selected feature(s) for the configure options.
@@ -793,11 +762,12 @@ def with_or_without(
``variant=value`` is in the spec.
Args:
name: name of a valid multi-valued variant
activation_value: callable that accepts a single value and returns the parameter to be
used leading to an entry of the type ``--with-{name}={parameter}``.
name (str): name of a valid multi-valued variant
activation_value (typing.Callable): callable that accepts a single
value and returns the parameter to be used leading to an entry
of the type ``--with-{name}={parameter}``.
The special value "prefix" can also be assigned and will return
The special value 'prefix' can also be assigned and will return
``spec[name].prefix`` as activation parameter.
Returns:
@@ -805,22 +775,18 @@ def with_or_without(
"""
return self._activate_or_not(name, "with", "without", activation_value, variant)
def enable_or_disable(
self,
name: str,
activation_value: Optional[Union[Callable, str]] = None,
variant: Optional[str] = None,
) -> List[str]:
def enable_or_disable(self, name, activation_value=None, variant=None):
"""Same as
:meth:`~spack.build_systems.autotools.AutotoolsBuilder.with_or_without`
but substitute ``with`` with ``enable`` and ``without`` with ``disable``.
Args:
name: name of a valid multi-valued variant
activation_value: if present accepts a single value and returns the parameter to be
used leading to an entry of the type ``--enable-{name}={parameter}``
name (str): name of a valid multi-valued variant
activation_value (typing.Callable): if present accepts a single value
and returns the parameter to be used leading to an entry of the
type ``--enable-{name}={parameter}``
The special value "prefix" can also be assigned and will return
The special value 'prefix' can also be assigned and will return
``spec[name].prefix`` as activation parameter.
Returns:
@@ -828,15 +794,15 @@ def enable_or_disable(
"""
return self._activate_or_not(name, "enable", "disable", activation_value, variant)
spack.phase_callbacks.run_after("install")(execute_install_time_tests)
spack.builder.run_after("install")(execute_install_time_tests)
def installcheck(self) -> None:
def installcheck(self):
"""Run "make" on the ``installcheck`` target, if found."""
with fs.working_dir(self.build_directory):
self.pkg._if_make_target_execute("installcheck")
@spack.phase_callbacks.run_after("install")
def remove_libtool_archives(self) -> None:
@spack.builder.run_after("install")
def remove_libtool_archives(self):
"""Remove all .la files in prefix sub-folders if the package sets
``install_libtool_archives`` to be False.
"""
@@ -858,13 +824,12 @@ def setup_build_environment(self, env):
env.set("MACOSX_DEPLOYMENT_TARGET", "10.16")
# On macOS, force rpaths for shared library IDs and remove duplicate rpaths
spack.phase_callbacks.run_after("install", when="platform=darwin")(apply_macos_rpath_fixups)
spack.builder.run_after("install", when="platform=darwin")(apply_macos_rpath_fixups)
def _autoreconf_search_path_args(spec: spack.spec.Spec) -> List[str]:
dirs_seen: Set[Tuple[int, int]] = set()
flags_spack: List[str] = []
flags_external: List[str] = []
def _autoreconf_search_path_args(spec):
dirs_seen = set()
flags_spack, flags_external = [], []
# We don't want to add an include flag for automake's default search path.
for automake in spec.dependencies(name="automake", deptype="build"):

View File

@@ -10,8 +10,8 @@
import llnl.util.filesystem as fs
import llnl.util.tty as tty
import spack.phase_callbacks
from spack.directives import depends_on
import spack.build_environment
import spack.builder
from .cmake import CMakeBuilder, CMakePackage
@@ -69,7 +69,12 @@ class CachedCMakeBuilder(CMakeBuilder):
@property
def cache_name(self):
return f"{self.pkg.name}-{self.spec.architecture.platform}-{self.spec.dag_hash()}.cmake"
return "{0}-{1}-{2}@{3}.cmake".format(
self.pkg.name,
self.pkg.spec.architecture,
self.pkg.spec.compiler.name,
self.pkg.spec.compiler.version,
)
@property
def cache_path(self):
@@ -84,7 +89,7 @@ def define_cmake_cache_from_variant(self, cmake_var, variant=None, comment=""):
if variant is None:
variant = cmake_var.lower()
if not self.pkg.has_variant(variant):
if variant not in self.pkg.variants:
raise KeyError('"{0}" is not a variant of "{1}"'.format(variant, self.pkg.name))
if variant not in self.pkg.spec.variants:
@@ -112,9 +117,7 @@ def initconfig_compiler_entries(self):
# Fortran compiler is optional
if "FC" in os.environ:
spack_fc_entry = cmake_cache_path("CMAKE_Fortran_COMPILER", os.environ["FC"])
system_fc_entry = cmake_cache_path(
"CMAKE_Fortran_COMPILER", self.spec["fortran"].package.fortran
)
system_fc_entry = cmake_cache_path("CMAKE_Fortran_COMPILER", self.pkg.compiler.fc)
else:
spack_fc_entry = "# No Fortran compiler defined in spec"
system_fc_entry = "# No Fortran compiler defined in spec"
@@ -130,8 +133,8 @@ def initconfig_compiler_entries(self):
" " + cmake_cache_path("CMAKE_CXX_COMPILER", os.environ["CXX"]),
" " + spack_fc_entry,
"else()\n",
" " + cmake_cache_path("CMAKE_C_COMPILER", self.spec["c"].package.cc),
" " + cmake_cache_path("CMAKE_CXX_COMPILER", self.spec["cxx"].package.cxx),
" " + cmake_cache_path("CMAKE_C_COMPILER", self.pkg.compiler.cc),
" " + cmake_cache_path("CMAKE_CXX_COMPILER", self.pkg.compiler.cxx),
" " + system_fc_entry,
"endif()\n",
]
@@ -159,9 +162,7 @@ def initconfig_compiler_entries(self):
ld_flags = " ".join(flags["ldflags"])
ld_format_string = "CMAKE_{0}_LINKER_FLAGS"
# CMake has separate linker arguments for types of builds.
# 'ldflags' should not be used with CMAKE_STATIC_LINKER_FLAGS which
# is used by the archiver, so don't include "STATIC" in this loop:
for ld_type in ["EXE", "MODULE", "SHARED"]:
for ld_type in ["EXE", "MODULE", "SHARED", "STATIC"]:
ld_string = ld_format_string.format(ld_type)
entries.append(cmake_cache_string(ld_string, ld_flags))
@@ -190,10 +191,7 @@ def initconfig_mpi_entries(self):
entries.append(cmake_cache_path("MPI_C_COMPILER", spec["mpi"].mpicc))
entries.append(cmake_cache_path("MPI_CXX_COMPILER", spec["mpi"].mpicxx))
# not all MPIs have Fortran wrappers
if hasattr(spec["mpi"], "mpifc"):
entries.append(cmake_cache_path("MPI_Fortran_COMPILER", spec["mpi"].mpifc))
entries.append(cmake_cache_path("MPI_Fortran_COMPILER", spec["mpi"].mpifc))
# Check for slurm
using_slurm = False
@@ -297,6 +295,18 @@ def initconfig_hardware_entries(self):
def std_initconfig_entries(self):
cmake_prefix_path_env = os.environ["CMAKE_PREFIX_PATH"]
cmake_prefix_path = cmake_prefix_path_env.replace(os.pathsep, ";")
cmake_rpaths_env = spack.build_environment.get_rpaths(self.pkg)
cmake_rpaths_path = ";".join(cmake_rpaths_env)
complete_rpath_list = cmake_rpaths_path
if "SPACK_COMPILER_EXTRA_RPATHS" in os.environ:
spack_extra_rpaths_env = os.environ["SPACK_COMPILER_EXTRA_RPATHS"]
spack_extra_rpaths_path = spack_extra_rpaths_env.replace(os.pathsep, ";")
complete_rpath_list = "{0};{1}".format(complete_rpath_list, spack_extra_rpaths_path)
if "SPACK_COMPILER_IMPLICIT_RPATHS" in os.environ:
spack_implicit_rpaths_env = os.environ["SPACK_COMPILER_IMPLICIT_RPATHS"]
spack_implicit_rpaths_path = spack_implicit_rpaths_env.replace(os.pathsep, ";")
complete_rpath_list = "{0};{1}".format(complete_rpath_list, spack_implicit_rpaths_path)
return [
"#------------------{0}".format("-" * 60),
@@ -306,6 +316,8 @@ def std_initconfig_entries(self):
"#------------------{0}\n".format("-" * 60),
cmake_cache_string("CMAKE_PREFIX_PATH", cmake_prefix_path),
cmake_cache_string("CMAKE_INSTALL_RPATH_USE_LINK_PATH", "ON"),
cmake_cache_string("CMAKE_BUILD_RPATH", complete_rpath_list),
cmake_cache_string("CMAKE_INSTALL_RPATH", complete_rpath_list),
self.define_cmake_cache_from_variant("CMAKE_BUILD_TYPE", "build_type"),
]
@@ -333,7 +345,7 @@ def std_cmake_args(self):
args.extend(["-C", self.cache_path])
return args
@spack.phase_callbacks.run_after("install")
@spack.builder.run_after("install")
def install_cmake_cache(self):
fs.mkdirp(self.pkg.spec.prefix.share.cmake)
fs.install(self.cache_path, self.pkg.spec.prefix.share.cmake)
@@ -350,10 +362,6 @@ class CachedCMakePackage(CMakePackage):
CMakeBuilder = CachedCMakeBuilder
# These dependencies are assumed in the builder
depends_on("c", type="build")
depends_on("cxx", type="build")
def flag_handler(self, name, flags):
if name in ("cflags", "cxxflags", "cppflags", "fflags"):
return None, None, None # handled in the cmake cache

View File

@@ -3,15 +3,16 @@
#
# SPDX-License-Identifier: (Apache-2.0 OR MIT)
import inspect
import llnl.util.filesystem as fs
import spack.builder
import spack.package_base
import spack.phase_callbacks
from spack.directives import build_system, depends_on
from spack.multimethod import when
from ._checks import BuilderWithDefaults, execute_install_time_tests
from ._checks import BaseBuilder, execute_install_time_tests
class CargoPackage(spack.package_base.PackageBase):
@@ -28,7 +29,7 @@ class CargoPackage(spack.package_base.PackageBase):
@spack.builder.builder("cargo")
class CargoBuilder(BuilderWithDefaults):
class CargoBuilder(BaseBuilder):
"""The Cargo builder encodes the most common way of building software with
a rust Cargo.toml file. It has two phases that can be overridden, if need be:
@@ -71,16 +72,18 @@ def check_args(self):
def build(self, pkg, spec, prefix):
"""Runs ``cargo install`` in the source directory"""
with fs.working_dir(self.build_directory):
pkg.module.cargo("install", "--root", "out", "--path", ".", *self.build_args)
inspect.getmodule(pkg).cargo(
"install", "--root", "out", "--path", ".", *self.build_args
)
def install(self, pkg, spec, prefix):
"""Copy build files into package prefix."""
with fs.working_dir(self.build_directory):
fs.install_tree("out", prefix)
spack.phase_callbacks.run_after("install")(execute_install_time_tests)
spack.builder.run_after("install")(execute_install_time_tests)
def check(self):
"""Run "cargo test"."""
with fs.working_dir(self.build_directory):
self.pkg.module.cargo("test", *self.check_args)
inspect.getmodule(self.pkg).cargo("test", *self.check_args)

View File

@@ -3,29 +3,24 @@
#
# SPDX-License-Identifier: (Apache-2.0 OR MIT)
import collections.abc
import inspect
import os
import pathlib
import platform
import re
import sys
from itertools import chain
from typing import Any, List, Optional, Set, Tuple
from typing import List, Optional, Tuple
import llnl.util.filesystem as fs
from llnl.util.lang import stable_partition
import spack.build_environment
import spack.builder
import spack.deptypes as dt
import spack.error
import spack.package_base
import spack.phase_callbacks
import spack.spec
import spack.util.prefix
from spack.directives import build_system, conflicts, depends_on, variant
from spack.multimethod import when
from spack.util.environment import filter_system_paths
from ._checks import BuilderWithDefaults, execute_build_time_tests
from ._checks import BaseBuilder, execute_build_time_tests
# Regex to extract the primary generator from the CMake generator
# string.
@@ -51,9 +46,9 @@ def _maybe_set_python_hints(pkg: spack.package_base.PackageBase, args: List[str]
python_executable = pkg.spec["python"].command.path
args.extend(
[
define("PYTHON_EXECUTABLE", python_executable),
define("Python_EXECUTABLE", python_executable),
define("Python3_EXECUTABLE", python_executable),
CMakeBuilder.define("PYTHON_EXECUTABLE", python_executable),
CMakeBuilder.define("Python_EXECUTABLE", python_executable),
CMakeBuilder.define("Python3_EXECUTABLE", python_executable),
]
)
@@ -88,7 +83,7 @@ def _conditional_cmake_defaults(pkg: spack.package_base.PackageBase, args: List[
ipo = False
if cmake.satisfies("@3.9:"):
args.append(define("CMAKE_INTERPROCEDURAL_OPTIMIZATION", ipo))
args.append(CMakeBuilder.define("CMAKE_INTERPROCEDURAL_OPTIMIZATION", ipo))
# Disable Package Registry: export(PACKAGE) may put files in the user's home directory, and
# find_package may search there. This is not what we want.
@@ -96,36 +91,25 @@ def _conditional_cmake_defaults(pkg: spack.package_base.PackageBase, args: List[
# Do not populate CMake User Package Registry
if cmake.satisfies("@3.15:"):
# see https://cmake.org/cmake/help/latest/policy/CMP0090.html
args.append(define("CMAKE_POLICY_DEFAULT_CMP0090", "NEW"))
args.append(CMakeBuilder.define("CMAKE_POLICY_DEFAULT_CMP0090", "NEW"))
elif cmake.satisfies("@3.1:"):
# see https://cmake.org/cmake/help/latest/variable/CMAKE_EXPORT_NO_PACKAGE_REGISTRY.html
args.append(define("CMAKE_EXPORT_NO_PACKAGE_REGISTRY", True))
args.append(CMakeBuilder.define("CMAKE_EXPORT_NO_PACKAGE_REGISTRY", True))
# Do not use CMake User/System Package Registry
# https://cmake.org/cmake/help/latest/manual/cmake-packages.7.html#disabling-the-package-registry
if cmake.satisfies("@3.16:"):
args.append(define("CMAKE_FIND_USE_PACKAGE_REGISTRY", False))
args.append(CMakeBuilder.define("CMAKE_FIND_USE_PACKAGE_REGISTRY", False))
elif cmake.satisfies("@3.1:3.15"):
args.append(define("CMAKE_FIND_PACKAGE_NO_PACKAGE_REGISTRY", False))
args.append(define("CMAKE_FIND_PACKAGE_NO_SYSTEM_PACKAGE_REGISTRY", False))
args.append(CMakeBuilder.define("CMAKE_FIND_PACKAGE_NO_PACKAGE_REGISTRY", False))
args.append(CMakeBuilder.define("CMAKE_FIND_PACKAGE_NO_SYSTEM_PACKAGE_REGISTRY", False))
# Export a compilation database if supported.
if _supports_compilation_databases(pkg):
args.append(define("CMAKE_EXPORT_COMPILE_COMMANDS", True))
# Enable MACOSX_RPATH by default when cmake_minimum_required < 3
# https://cmake.org/cmake/help/latest/policy/CMP0042.html
if pkg.spec.satisfies("platform=darwin") and cmake.satisfies("@3:"):
args.append(define("CMAKE_POLICY_DEFAULT_CMP0042", "NEW"))
# Disable find package's config mode for versions of Boost that
# didn't provide it. See https://github.com/spack/spack/issues/20169
# and https://cmake.org/cmake/help/latest/module/FindBoost.html
if pkg.spec.satisfies("^boost@:1.69.0"):
args.append(define("Boost_NO_BOOST_CMAKE", True))
args.append(CMakeBuilder.define("CMAKE_EXPORT_COMPILE_COMMANDS", True))
def generator(*names: str, default: Optional[str] = None) -> None:
def generator(*names: str, default: Optional[str] = None):
"""The build system generator to use.
See ``cmake --help`` for a list of valid generators.
@@ -157,30 +141,11 @@ def _values(x):
default=default,
values=_values,
description="the build system generator to use",
when="build_system=cmake",
)
for x in not_used:
conflicts(f"generator={x}")
def get_cmake_prefix_path(pkg: spack.package_base.PackageBase) -> List[str]:
"""Obtain the CMAKE_PREFIX_PATH entries for a package, based on the cmake_prefix_path package
attribute of direct build/test and transitive link dependencies."""
# Add direct build/test deps
selected: Set[str] = {s.dag_hash() for s in pkg.spec.dependencies(deptype=dt.BUILD | dt.TEST)}
# Add transitive link deps
selected.update(s.dag_hash() for s in pkg.spec.traverse(root=False, deptype=dt.LINK))
# Separate out externals so they do not shadow Spack prefixes
externals, spack_built = stable_partition(
(s for s in pkg.spec.traverse(root=False, order="topo") if s.dag_hash() in selected),
lambda x: x.external,
)
return filter_system_paths(
path for spec in chain(spack_built, externals) for path in spec.package.cmake_prefix_paths
)
class CMakePackage(spack.package_base.PackageBase):
"""Specialized class for packages built using CMake
@@ -272,15 +237,15 @@ def flags_to_build_system_args(self, flags):
# Legacy methods (used by too many packages to change them,
# need to forward to the builder)
def define(self, cmake_var: str, value: Any) -> str:
return define(cmake_var, value)
def define(self, *args, **kwargs):
return self.builder.define(*args, **kwargs)
def define_from_variant(self, cmake_var: str, variant: Optional[str] = None) -> str:
return define_from_variant(self, cmake_var, variant)
def define_from_variant(self, *args, **kwargs):
return self.builder.define_from_variant(*args, **kwargs)
@spack.builder.builder("cmake")
class CMakeBuilder(BuilderWithDefaults):
class CMakeBuilder(BaseBuilder):
"""The cmake builder encodes the default way of building software with CMake. IT
has three phases that can be overridden:
@@ -330,15 +295,15 @@ class CMakeBuilder(BuilderWithDefaults):
build_time_test_callbacks = ["check"]
@property
def archive_files(self) -> List[str]:
def archive_files(self):
"""Files to archive for packages based on CMake"""
files = [os.path.join(self.build_directory, "CMakeCache.txt")]
if _supports_compilation_databases(self.pkg):
if _supports_compilation_databases(self):
files.append(os.path.join(self.build_directory, "compile_commands.json"))
return files
@property
def root_cmakelists_dir(self) -> str:
def root_cmakelists_dir(self):
"""The relative path to the directory containing CMakeLists.txt
This path is relative to the root of the extracted tarball,
@@ -347,17 +312,16 @@ def root_cmakelists_dir(self) -> str:
return self.pkg.stage.source_path
@property
def generator(self) -> str:
def generator(self):
if self.spec.satisfies("generator=make"):
return "Unix Makefiles"
if self.spec.satisfies("generator=ninja"):
return "Ninja"
raise ValueError(
f'{self.spec.format()} has an unsupported value for the "generator" variant'
)
msg = f'{self.spec.format()} has an unsupported value for the "generator" variant'
raise ValueError(msg)
@property
def std_cmake_args(self) -> List[str]:
def std_cmake_args(self):
"""Standard cmake arguments provided as a property for
convenience of package writers
"""
@@ -366,9 +330,7 @@ def std_cmake_args(self) -> List[str]:
return args
@staticmethod
def std_args(
pkg: spack.package_base.PackageBase, generator: Optional[str] = None
) -> List[str]:
def std_args(pkg, generator=None):
"""Computes the standard cmake arguments for a generic package"""
default_generator = "Ninja" if sys.platform == "win32" else "Unix Makefiles"
generator = generator or default_generator
@@ -378,27 +340,18 @@ def std_args(
msg = "Invalid CMake generator: '{0}'\n".format(generator)
msg += "CMakePackage currently supports the following "
msg += "primary generators: '{0}'".format("', '".join(valid_primary_generators))
raise spack.error.InstallError(msg)
raise spack.package_base.InstallError(msg)
try:
build_type = pkg.spec.variants["build_type"].value
except KeyError:
build_type = "RelWithDebInfo"
define = CMakeBuilder.define
args = [
"-G",
generator,
define("CMAKE_INSTALL_PREFIX", pathlib.Path(pkg.prefix).as_posix()),
define("CMAKE_INSTALL_RPATH_USE_LINK_PATH", True),
# only include the install prefix lib dirs; rpaths for deps are added by USE_LINK_PATH
define(
"CMAKE_INSTALL_RPATH",
[
pathlib.Path(pkg.prefix, "lib").as_posix(),
pathlib.Path(pkg.prefix, "lib64").as_posix(),
],
),
define("CMAKE_PREFIX_PATH", get_cmake_prefix_path(pkg)),
define("CMAKE_BUILD_TYPE", build_type),
]
@@ -413,34 +366,164 @@ def std_args(
_conditional_cmake_defaults(pkg, args)
_maybe_set_python_hints(pkg, args)
# Set up CMake rpath
args.extend(
[
define("CMAKE_INSTALL_RPATH_USE_LINK_PATH", True),
define("CMAKE_INSTALL_RPATH", spack.build_environment.get_rpaths(pkg)),
define("CMAKE_PREFIX_PATH", spack.build_environment.get_cmake_prefix_path(pkg)),
]
)
return args
@staticmethod
def define_cuda_architectures(pkg: spack.package_base.PackageBase) -> str:
return define_cuda_architectures(pkg)
def define_cuda_architectures(pkg):
"""Returns the str ``-DCMAKE_CUDA_ARCHITECTURES:STRING=(expanded cuda_arch)``.
``cuda_arch`` is variant composed of a list of target CUDA architectures and
it is declared in the cuda package.
This method is no-op for cmake<3.18 and when ``cuda_arch`` variant is not set.
"""
cmake_flag = str()
if "cuda_arch" in pkg.spec.variants and pkg.spec.satisfies("^cmake@3.18:"):
cmake_flag = CMakeBuilder.define(
"CMAKE_CUDA_ARCHITECTURES", pkg.spec.variants["cuda_arch"].value
)
return cmake_flag
@staticmethod
def define_hip_architectures(pkg: spack.package_base.PackageBase) -> str:
return define_hip_architectures(pkg)
def define_hip_architectures(pkg):
"""Returns the str ``-DCMAKE_HIP_ARCHITECTURES:STRING=(expanded amdgpu_target)``.
``amdgpu_target`` is variant composed of a list of the target HIP
architectures and it is declared in the rocm package.
This method is no-op for cmake<3.18 and when ``amdgpu_target`` variant is
not set.
"""
cmake_flag = str()
if "amdgpu_target" in pkg.spec.variants and pkg.spec.satisfies("^cmake@3.21:"):
cmake_flag = CMakeBuilder.define(
"CMAKE_HIP_ARCHITECTURES", pkg.spec.variants["amdgpu_target"].value
)
return cmake_flag
@staticmethod
def define(cmake_var: str, value: Any) -> str:
return define(cmake_var, value)
def define(cmake_var, value):
"""Return a CMake command line argument that defines a variable.
def define_from_variant(self, cmake_var: str, variant: Optional[str] = None) -> str:
return define_from_variant(self.pkg, cmake_var, variant)
The resulting argument will convert boolean values to OFF/ON
and lists/tuples to CMake semicolon-separated string lists. All other
values will be interpreted as strings.
Examples:
.. code-block:: python
[define('BUILD_SHARED_LIBS', True),
define('CMAKE_CXX_STANDARD', 14),
define('swr', ['avx', 'avx2'])]
will generate the following configuration options:
.. code-block:: console
["-DBUILD_SHARED_LIBS:BOOL=ON",
"-DCMAKE_CXX_STANDARD:STRING=14",
"-DSWR:STRING=avx;avx2]
"""
# Create a list of pairs. Each pair includes a configuration
# option and whether or not that option is activated
if isinstance(value, bool):
kind = "BOOL"
value = "ON" if value else "OFF"
else:
kind = "STRING"
if isinstance(value, collections.abc.Sequence) and not isinstance(value, str):
value = ";".join(str(v) for v in value)
else:
value = str(value)
return "".join(["-D", cmake_var, ":", kind, "=", value])
def define_from_variant(self, cmake_var, variant=None):
"""Return a CMake command line argument from the given variant's value.
The optional ``variant`` argument defaults to the lower-case transform
of ``cmake_var``.
This utility function is similar to
:meth:`~spack.build_systems.autotools.AutotoolsBuilder.with_or_without`.
Examples:
Given a package with:
.. code-block:: python
variant('cxxstd', default='11', values=('11', '14'),
multi=False, description='')
variant('shared', default=True, description='')
variant('swr', values=any_combination_of('avx', 'avx2'),
description='')
calling this function like:
.. code-block:: python
[self.define_from_variant('BUILD_SHARED_LIBS', 'shared'),
self.define_from_variant('CMAKE_CXX_STANDARD', 'cxxstd'),
self.define_from_variant('SWR')]
will generate the following configuration options:
.. code-block:: console
["-DBUILD_SHARED_LIBS:BOOL=ON",
"-DCMAKE_CXX_STANDARD:STRING=14",
"-DSWR:STRING=avx;avx2]
for ``<spec-name> cxxstd=14 +shared swr=avx,avx2``
Note: if the provided variant is conditional, and the condition is not met,
this function returns an empty string. CMake discards empty strings
provided on the command line.
"""
if variant is None:
variant = cmake_var.lower()
if variant not in self.pkg.variants:
raise KeyError('"{0}" is not a variant of "{1}"'.format(variant, self.pkg.name))
if variant not in self.pkg.spec.variants:
return ""
value = self.pkg.spec.variants[variant].value
if isinstance(value, (tuple, list)):
# Sort multi-valued variants for reproducibility
value = sorted(value)
return self.define(cmake_var, value)
@property
def build_dirname(self) -> str:
def build_dirname(self):
"""Directory name to use when building the package."""
return f"spack-build-{self.pkg.spec.dag_hash(7)}"
return "spack-build-%s" % self.pkg.spec.dag_hash(7)
@property
def build_directory(self) -> str:
def build_directory(self):
"""Full-path to the directory to use when building the package."""
return os.path.join(self.pkg.stage.path, self.build_dirname)
def cmake_args(self) -> List[str]:
def cmake_args(self):
"""List of all the arguments that must be passed to cmake, except:
* CMAKE_INSTALL_PREFIX
@@ -450,56 +533,34 @@ def cmake_args(self) -> List[str]:
"""
return []
def cmake(
self,
pkg: spack.package_base.PackageBase,
spec: spack.spec.Spec,
prefix: spack.util.prefix.Prefix,
) -> None:
def cmake(self, pkg, spec, prefix):
"""Runs ``cmake`` in the build directory"""
# skip cmake phase if it is an incremental develop build
if spec.is_develop and os.path.isfile(
os.path.join(self.build_directory, "CMakeCache.txt")
):
return
options = self.std_cmake_args
options += self.cmake_args()
options.append(os.path.abspath(self.root_cmakelists_dir))
with fs.working_dir(self.build_directory, create=True):
pkg.module.cmake(*options)
inspect.getmodule(self.pkg).cmake(*options)
def build(
self,
pkg: spack.package_base.PackageBase,
spec: spack.spec.Spec,
prefix: spack.util.prefix.Prefix,
) -> None:
def build(self, pkg, spec, prefix):
"""Make the build targets"""
with fs.working_dir(self.build_directory):
if self.generator == "Unix Makefiles":
pkg.module.make(*self.build_targets)
inspect.getmodule(self.pkg).make(*self.build_targets)
elif self.generator == "Ninja":
self.build_targets.append("-v")
pkg.module.ninja(*self.build_targets)
inspect.getmodule(self.pkg).ninja(*self.build_targets)
def install(
self,
pkg: spack.package_base.PackageBase,
spec: spack.spec.Spec,
prefix: spack.util.prefix.Prefix,
) -> None:
def install(self, pkg, spec, prefix):
"""Make the install targets"""
with fs.working_dir(self.build_directory):
if self.generator == "Unix Makefiles":
pkg.module.make(*self.install_targets)
inspect.getmodule(self.pkg).make(*self.install_targets)
elif self.generator == "Ninja":
pkg.module.ninja(*self.install_targets)
inspect.getmodule(self.pkg).ninja(*self.install_targets)
spack.phase_callbacks.run_after("build")(execute_build_time_tests)
spack.builder.run_after("build")(execute_build_time_tests)
def check(self) -> None:
def check(self):
"""Search the CMake-generated files for the targets ``test`` and ``check``,
and runs them if found.
"""
@@ -510,133 +571,3 @@ def check(self) -> None:
elif self.generator == "Ninja":
self.pkg._if_ninja_target_execute("test", jobs_env="CTEST_PARALLEL_LEVEL")
self.pkg._if_ninja_target_execute("check")
def define(cmake_var: str, value: Any) -> str:
"""Return a CMake command line argument that defines a variable.
The resulting argument will convert boolean values to OFF/ON and lists/tuples to CMake
semicolon-separated string lists. All other values will be interpreted as strings.
Examples:
.. code-block:: python
[define("BUILD_SHARED_LIBS", True),
define("CMAKE_CXX_STANDARD", 14),
define("swr", ["avx", "avx2"])]
will generate the following configuration options:
.. code-block:: console
["-DBUILD_SHARED_LIBS:BOOL=ON",
"-DCMAKE_CXX_STANDARD:STRING=14",
"-DSWR:STRING=avx;avx2]
"""
# Create a list of pairs. Each pair includes a configuration
# option and whether or not that option is activated
if isinstance(value, bool):
kind = "BOOL"
value = "ON" if value else "OFF"
else:
kind = "STRING"
if isinstance(value, collections.abc.Sequence) and not isinstance(value, str):
value = ";".join(str(v) for v in value)
else:
value = str(value)
return "".join(["-D", cmake_var, ":", kind, "=", value])
def define_from_variant(
pkg: spack.package_base.PackageBase, cmake_var: str, variant: Optional[str] = None
) -> str:
"""Return a CMake command line argument from the given variant's value.
The optional ``variant`` argument defaults to the lower-case transform
of ``cmake_var``.
Examples:
Given a package with:
.. code-block:: python
variant("cxxstd", default="11", values=("11", "14"),
multi=False, description="")
variant("shared", default=True, description="")
variant("swr", values=any_combination_of("avx", "avx2"),
description="")
calling this function like:
.. code-block:: python
[
self.define_from_variant("BUILD_SHARED_LIBS", "shared"),
self.define_from_variant("CMAKE_CXX_STANDARD", "cxxstd"),
self.define_from_variant("SWR"),
]
will generate the following configuration options:
.. code-block:: console
[
"-DBUILD_SHARED_LIBS:BOOL=ON",
"-DCMAKE_CXX_STANDARD:STRING=14",
"-DSWR:STRING=avx;avx2",
]
for ``<spec-name> cxxstd=14 +shared swr=avx,avx2``
Note: if the provided variant is conditional, and the condition is not met, this function
returns an empty string. CMake discards empty strings provided on the command line.
"""
if variant is None:
variant = cmake_var.lower()
if not pkg.has_variant(variant):
raise KeyError('"{0}" is not a variant of "{1}"'.format(variant, pkg.name))
if variant not in pkg.spec.variants:
return ""
value = pkg.spec.variants[variant].value
if isinstance(value, (tuple, list)):
# Sort multi-valued variants for reproducibility
value = sorted(value)
return define(cmake_var, value)
def define_hip_architectures(pkg: spack.package_base.PackageBase) -> str:
"""Returns the str ``-DCMAKE_HIP_ARCHITECTURES:STRING=(expanded amdgpu_target)``.
``amdgpu_target`` is variant composed of a list of the target HIP
architectures and it is declared in the rocm package.
This method is no-op for cmake<3.18 and when ``amdgpu_target`` variant is
not set.
"""
if "amdgpu_target" in pkg.spec.variants and pkg.spec.satisfies("^cmake@3.21:"):
return define("CMAKE_HIP_ARCHITECTURES", pkg.spec.variants["amdgpu_target"].value)
return ""
def define_cuda_architectures(pkg: spack.package_base.PackageBase) -> str:
"""Returns the str ``-DCMAKE_CUDA_ARCHITECTURES:STRING=(expanded cuda_arch)``.
``cuda_arch`` is variant composed of a list of target CUDA architectures and
it is declared in the cuda package.
This method is no-op for cmake<3.18 and when ``cuda_arch`` variant is not set.
"""
if "cuda_arch" in pkg.spec.variants and pkg.spec.satisfies("^cmake@3.18:"):
return define("CMAKE_CUDA_ARCHITECTURES", pkg.spec.variants["cuda_arch"].value)
return ""

View File

@@ -5,21 +5,15 @@
import itertools
import os
import pathlib
import platform
import re
import sys
from typing import Dict, List, Optional, Sequence, Tuple, Union
import archspec.cpu
from typing import Dict, List, Sequence, Tuple, Union
import llnl.util.tty as tty
from llnl.util.lang import classproperty, memoized
from llnl.util.lang import classproperty
import spack.compilers.libraries
import spack.config
import spack.compiler
import spack.package_base
import spack.paths
import spack.util.executable
# Local "type" for type hints
Path = Union[str, pathlib.Path]
@@ -49,9 +43,6 @@ class CompilerPackage(spack.package_base.PackageBase):
#: Static definition of languages supported by this class
compiler_languages: Sequence[str] = ["c", "cxx", "fortran"]
#: Relative path to compiler wrappers
link_paths: Dict[str, str] = {}
def __init__(self, spec: "spack.spec.Spec"):
super().__init__(spec)
msg = f"Supported languages for {spec} are not a subset of possible supported languages"
@@ -86,14 +77,14 @@ def executables(cls) -> Sequence[str]:
]
@classmethod
def determine_version(cls, exe: Path) -> str:
def determine_version(cls, exe: Path):
version_argument = cls.compiler_version_argument
if isinstance(version_argument, str):
version_argument = (version_argument,)
for va in version_argument:
try:
output = compiler_output(exe, version_argument=va)
output = spack.compiler.get_compiler_version_output(exe, va)
match = re.search(cls.compiler_version_regex, output)
if match:
return ".".join(match.groups())
@@ -104,7 +95,6 @@ def determine_version(cls, exe: Path) -> str:
f"[{__file__}] Cannot detect a valid version for the executable "
f"{str(exe)}, for package '{cls.name}': {e}"
)
return ""
@classmethod
def compiler_bindir(cls, prefix: Path) -> Path:
@@ -152,183 +142,3 @@ def determine_compiler_paths(cls, exes: Sequence[Path]) -> Dict[str, Path]:
def determine_variants(cls, exes: Sequence[Path], version_str: str) -> Tuple:
# path determination is separated so it can be reused in subclasses
return "", {"compilers": cls.determine_compiler_paths(exes=exes)}
#: Returns the argument needed to set the RPATH, or None if it does not exist
rpath_arg: Optional[str] = "-Wl,-rpath,"
#: Flag that needs to be used to pass an argument to the linker
linker_arg: str = "-Wl,"
#: Flag used to produce Position Independent Code
pic_flag: str = "-fPIC"
#: Flag used to get verbose output
verbose_flags: str = "-v"
#: Flag to activate OpenMP support
openmp_flag: str = "-fopenmp"
def standard_flag(self, *, language: str, standard: str) -> str:
"""Returns the flag used to enforce a given standard for a language"""
if language not in self.supported_languages:
# FIXME (compiler as nodes): Use UnsupportedCompilerFlag ?
raise RuntimeError(f"{self.spec} does not provide the '{language}' language")
try:
return self._standard_flag(language=language, standard=standard)
except (KeyError, RuntimeError) as e:
raise RuntimeError(
f"{self.spec} does not provide the '{language}' standard {standard}"
) from e
def _standard_flag(self, *, language: str, standard: str) -> str:
raise NotImplementedError("Must be implemented by derived classes")
@property
def disable_new_dtags(self) -> str:
if platform.system() == "Darwin":
return ""
return "--disable-new-dtags"
@property
def enable_new_dtags(self) -> str:
if platform.system() == "Darwin":
return ""
return "--enable-new-dtags"
def setup_dependent_build_environment(self, env, dependent_spec):
# FIXME (compiler as nodes): check if this is good enough or should be made more general
# The package is not used as a compiler, so skip this setup
if not any(
lang in dependent_spec and dependent_spec[lang].name == self.spec.name
for lang in ("c", "cxx", "fortran")
):
return
# Populate an object with the list of environment modifications and return it
link_dir = pathlib.Path(spack.paths.build_env_path)
for language, attr_name, wrapper_var_name, spack_var_name in [
("c", "cc", "CC", "SPACK_CC"),
("cxx", "cxx", "CXX", "SPACK_CXX"),
("fortran", "fortran", "F77", "SPACK_F77"),
("fortran", "fortran", "FC", "SPACK_FC"),
]:
if language not in dependent_spec or dependent_spec[language].name != self.spec.name:
continue
if not hasattr(self, attr_name):
continue
compiler = getattr(self, attr_name)
env.set(spack_var_name, compiler)
if language not in self.link_paths:
continue
wrapper_path = link_dir / self.link_paths.get(language)
env.set(wrapper_var_name, str(wrapper_path))
env.set(f"SPACK_{wrapper_var_name}_RPATH_ARG", self.rpath_arg)
uarch = dependent_spec.architecture.target
version_number, _ = archspec.cpu.version_components(
self.spec.version.dotted_numeric_string
)
try:
isa_arg = uarch.optimization_flags(self.spec.name, version_number)
except (ValueError, archspec.cpu.UnsupportedMicroarchitecture):
isa_arg = ""
if isa_arg:
env.set(f"SPACK_TARGET_ARGS_{attr_name.upper()}", isa_arg)
# FIXME (compiler as nodes): make these paths language specific
env.set("SPACK_LINKER_ARG", self.linker_arg)
paths = _implicit_rpaths(pkg=self)
if paths:
env.set("SPACK_COMPILER_IMPLICIT_RPATHS", ":".join(paths))
# Check whether we want to force RPATH or RUNPATH
if spack.config.CONFIG.get("config:shared_linking:type") == "rpath":
env.set("SPACK_DTAGS_TO_STRIP", self.enable_new_dtags)
env.set("SPACK_DTAGS_TO_ADD", self.disable_new_dtags)
else:
env.set("SPACK_DTAGS_TO_STRIP", self.disable_new_dtags)
env.set("SPACK_DTAGS_TO_ADD", self.enable_new_dtags)
spec = self.spec
env.set("SPACK_COMPILER_SPEC", spec.format("{name}{@version}{variants}{/hash:7}"))
if spec.extra_attributes:
extra_rpaths = spec.extra_attributes.get("extra_rpaths")
if extra_rpaths:
extra_rpaths = ":".join(compiler.extra_rpaths)
env.append_path("SPACK_COMPILER_EXTRA_RPATHS", extra_rpaths)
# Add spack build environment path with compiler wrappers first in
# the path. We add the compiler wrapper path, which includes default
# wrappers (cc, c++, f77, f90), AND a subdirectory containing
# compiler-specific symlinks. The latter ensures that builds that
# are sensitive to the *name* of the compiler see the right name when
# we're building with the wrappers.
#
# Conflicts on case-insensitive systems (like "CC" and "cc") are
# handled by putting one in the <build_env_path>/case-insensitive
# directory. Add that to the path too.
env_paths = []
compiler_specific = os.path.join(
spack.paths.build_env_path, os.path.dirname(self.link_paths["c"])
)
for item in [spack.paths.build_env_path, compiler_specific]:
env_paths.append(item)
ci = os.path.join(item, "case-insensitive")
if os.path.isdir(ci):
env_paths.append(ci)
tty.debug("Adding compiler bin/ paths: " + " ".join(env_paths))
for item in env_paths:
env.prepend_path("PATH", item)
env.set_path("SPACK_ENV_PATH", env_paths)
def _implicit_rpaths(pkg: spack.package_base.PackageBase) -> List[str]:
detector = spack.compilers.libraries.CompilerPropertyDetector(pkg.spec)
paths = detector.implicit_rpaths()
return paths
@memoized
def _compiler_output(
compiler_path: Path, *, version_argument: str, ignore_errors: Tuple[int, ...] = ()
) -> str:
"""Returns the output from the compiler invoked with the given version argument.
Args:
compiler_path: path of the compiler to be invoked
version_argument: the argument used to extract version information
"""
compiler = spack.util.executable.Executable(compiler_path)
compiler_invocation_args = {
"output": str,
"error": str,
"ignore_errors": ignore_errors,
"timeout": 120,
"fail_on_error": True,
}
if version_argument:
output = compiler(version_argument, **compiler_invocation_args)
else:
output = compiler(**compiler_invocation_args)
return output
def compiler_output(
compiler_path: Path, *, version_argument: str, ignore_errors: Tuple[int, ...] = ()
) -> str:
"""Wrapper for _get_compiler_version_output()."""
# This ensures that we memoize compiler output by *absolute path*,
# not just executable name. If we don't do this, and the path changes
# (e.g., during testing), we can get incorrect results.
if not os.path.isabs(compiler_path):
compiler_path = spack.util.executable.which_string(compiler_path, required=True)
return _compiler_output(
compiler_path, version_argument=version_argument, ignore_errors=ignore_errors
)

Some files were not shown because too many files have changed in this diff Show More