* "spack install" now uses cache by default, update examples accordingly
* Replace some example packages with others
* Packing tutorial reference to "spack env" replaced with "spack build-env"
* Command line prompts in examples are shortened
* Example output (including paths) are updated to be more relevant to training environment
Update all examples that need an MPI provider to build with MPICH; reorganize so that fixing MPICH (as part of environment section) comes first in the tutorial (most examples in the tutorial use an MPI provider).
- previously, uninstall would complain if a spec was needed by an
environment.
- Now, we analyze dependents and dependent environments and simply remove
(not uninstall) specs that are needed by environments
- with no arguments, these commands will now edit or dump the
environment's `spack.yaml` file.
- users may not know where named environments live
- this makes it convenient for users to get to the spack.yaml
configuration file for their named environment.
* Update Makefile to use property methods ("build_targets"/"install_targets")
to demonstrate their usage
* Fix highlighting
* Change cbench example to ESMF:
CBench package file was changed and no longer uses the example shown in
the old docs
Scopes added with -C are now referred to as "custom scopes"
rather than "command line scopes". "command line scope" now refers
to specific config options that are set on the command line (like
"--insecure")
- default is still to use the cache, but we've added back the
`--use-cache` argument so that scripts that used it are still correct.
- `--no-cache` is stil present and is mutually exclusive with `--use-cache`
* Update Dyninst package file for TBB dependency
Dyninst master and 10.0 release
versions require the tbb package. This mod
updates the tbb versions adding the one that
dyninst uses and adds the required spack package
changes to dyninst/package.py.
* Put new version of tbb in proper location, per reviewer comment.
* Only depend on tbb if building the development version.
* Changes for the release of Dyninst 10.0.0
* Introduce FFTW2 and FFT3 providers for Intel-MKL and FFTW Spack packages.
* make fftw default package for fftw-api virtual package
* virtual package test assertion now provides location of default virtual packages.
* Change name of virtual package to fftw-api and used versioned interface.
- all commands (except `spack find`, through `ConstraintAction`) now go
through get_env() to get the active environment
- ev.active was hard to read -- and the name wasn't descriptive.
- rename it to _active_environment to be more descriptive and to strongly
indicate that spack.environment manages it
- to aovid changing spec hashes drastically, only add this attribute to
differentiated abstract specs.
- othherwise assume that read-in specs are concrete
- spack.yaml files in the current directory were picked up inconsistently
-- make this a sure thing by moving that logic into find_environment()
and moving find_environment() to main()
- simplify arguments to Spack command:
- remove short args for infrequently used commands (--pdb/-D, -P, -s)
- `spack -D` now forces an env with a directory
- The `Spec` class maintains a special `_patches_in_order_of_appearance`
attribute on patch variants, but it is was preserved when specs are
copied.
- This caused issues for some builds
- Add special logic to `Spec` to preserve this variant on copy
- TODO: in the long term we should get rid of the special variant and
make it the responsibility of one of the variant classes.
- split 'environment' section into 'environments' and 'modules'
- move location to 'query packages' section
- move cd to developer section
- --env-dir no longer has a short optino (was -E)
- -E now means "run without an environment" (no longer same as --env-dir)
- -D now means "run with this directory environment"
- remove short options for may infrequently used top-level commands
- `spack env status` used to show install status; consolidate that into
`spack find`.
- `spack env status` will still print out whether there is an active
environment
- uninstall now:
- restricts its spec search to the current environment
- removes uninstalled specs from the current environment
- reports envs that still need specs you're trying to uninstall
- removed spack env uninstall command
- updated tests
- moved get_env from cmd/env.py to environment.py
- spack install will now install into the active environment when no
arguments are provided. It looks:
1. at the command line
2. for a local spack.yaml file
3. for any currently activated environment
- `spack env create <name>` works as before
- `spack env create <path>` now works as well -- environments can be
created in their own directories outside of Spack.
- `spack install` will look for a `spack.yaml` file in the current
directory, and will install the entire project from the environment
- The Environment class has been refactored so that it does not depend on
the internal Spack environment root; it just takes a path and operates
on an environment in that path (so internal and external envs are
handled the same)
- The named environment interface has been hoisted to the
spack.environment module level.
- env.yaml is now spack.yaml in all places. It was easier to go with one
name for these files than to try to handle logic for both env.yaml and
spack.yaml.
- `spack env activate foo`: sets SPACK_ENV to the current active env name
- `spack env deactivate`: unsets SPACK_ENV, deactivates the environment
- added support to setup_env.sh and setup_env.csh
- other env commands work properly with SPACK_ENV, as with an environment
arguments.
- command-line --env arguments take precedence over the active
environment, if given.
- env.yaml is now meaningful; it contains authoritative user specs
- concretize diffs user specs in env.yaml and env.json to allow user to
add/remove by simply updating env.yaml
- comments are preserved when env.yaml is updated by add/unadd
- env.yaml can contain configuration and include external configuration
either from merged files or from config scopes
- there is only one file format to remember (env.yaml, no separate init
format)
- env.json is now env.lock, and it stores the *last* user specs to be
concretized, along with full provenance.
- internal structure was modified slightly for readability
- env.lock contains a _meta section with metadata, in case needed
- added more tests for environments
- env commands follow Spack conventions; no more `spack env foo install`
- add `SingleFileScope` to configuration, which allows us to pull config
sections from a single file.
- update `env.yaml` and tests to ensure that the env.yaml schema works
when pulling configurtion from the env file.
- Each schema now has a top-level `properties` and `schema` attribute.
- The `properties` is a fragment that can be included in other
jsonschemas, via Python, not via '$ref'
- Th `schema` is a complete `jsonschema` with `title` and `$schema`
properties.
- add a common argument for `-e/--env`
- modify the database to support queries on subsets of hashes
- allow `spack find` to be filtered by hashes in an environment
- logic used in `spack find` was hiding duplicate installations if their
hashes were different
- short hash doesn't work in this scenario, since specs are structurally
identical
- ConstraintAction always works on a DB query, so use the DAG hash to
ensure uniqueness
- `spack.environment` is now the home for most of the infrastructure
around Spack environments
- refactor `cmd/env.py` to use everything from spack.environment
- refactor the cmd/env test to use pytest and fixtures
- `spack.util.environment` is the new home for routines that modify
environment variables.
- This is to make room for `spack.environment` to contain new routines
for dealing with spack environments
- Instead of one method with all parsers, each subcommand gets two
functions: `setup_<cmd>_parser()` and `environment_<cmd>()`
- the `setup_parser()` and `env()` functions now generate the parser
based on these and a list of subcommands.
- it is now easier to associate the arguments with the subcommand.
* modified tutorial packages
* update hint in hdf5 tutorial file (typo for suggested argument)
* add repo.yaml to tutorial repository
* update tutorial docs to refer user to tutorial package repository
* flake edits
* recommend site scope vs. defaults
* you don't specify the repo's name when adding a repo, just the path
* omit symlinks and create file copies when making a binary cache of a package
* unrelated flake edits involving regexes that recent flake is now angry about
* Record stdout for packages without errors
Previously our reporter only stored stdout if something went wrong
while installing a package. This prevented us from properly reporting
on steps where everything went as expected.
* More robustly report all phases to CDash
Previously if a phase generated no output it would not be reported to CDash.
For example, consider the following output:
==> Executing phase: 'configure'
==> Executing phase: 'build'
This would not generate a report for the configure phase. Now it does.
* Add test case for CDash reporting clean builds
* Fix default directory for CDash reports
The default 'cdash_report' directory name was getting overwritten
by 'junit-report'.
* Upload the build phase first to CDash
Older versions of CDash expect Build.xml to be the first file uploaded
for any given build.
* Define cdash_phase before referring to it
- Delete references to ruamel.yaml at Spack start-up, if they are present
- ruamel.yaml generates a .pth file when installed via pip that has the
effect of always preferring the version of this package installed at
site scope (effectively preventing us from vendoring it).
- This mechanism triggers when implicitly importing the 'site' module
when the python interpreter is started. In this PR we explicitly delete
references to 'ruamel.yaml' and 'ruamel' in sys.modules, if any, after
we set 'sys.path' to search from the directory where we store vendored
packages. This ensures that the imports after those statements will be
done from our vendored version.
- See #9206 for further details
fixes#9739
The non-daemonic pool relies heavily on implementation details of the
multiprocessing package. In this commit we provide an implementation
that fits recent python versions.
* henson: new package
* henson: change github path to henson-insitu
* henson: make mpi-wrappers=off by default
* henson: remove unsued variable and spaces to make linter happy
* henson: rename version master to develop
* xeus: new package
cppzmq: add version 4.3.0
zeromq: make libsodium optional, on by default
* xeus: add patch so it builds, add new version
nlohmann-json: add more versions
cryptopp: add more versions
* xeus: flake8
* xeus: fix license
This allows installing software on a namespace basis by including ${NAMESPACE} in `install_path_scheme`. e.g.,
```
cat ~/.spack/config.yaml
config:
install_path_scheme:
"${ARCHITECTURE}/${NAMESPACE}/${COMPILERNAME}-${COMPILERVER}/${PACKAGE}-${VERSION}-${HASH}"
```
The 'static_to_shared_library' function takes a compiler Executable,
which is intended to be invoked with a list of arguments; the
arguments must be separated from their values in the list, given
the way that 'Executable.__call__' invokes the underlying executable.
'static_to_shared_library' was not doing this, which this commit fixes.
* Provide a newer doxygen version.
+ Also provide a patch that allows collaboration diagrams to work properly with
C++ std::shared_ptr<T>. Ref:
6c380ba91a
Clang has support for using different fortran compilers with the Clang executable.
Spack includes logic to select a compiler wrapper symlink which refers to the fortran executable (since some build systems depend on the name of the compiler, e.g. 'gfortran' or 'flang').
This selection was previously based on the architecture, and chose incorrectly in some situations (e.g. for clang/gfortran on Linux). This replaces architecture-based wrapper selection with a selection that is based on the name of the Fortran compiler executable.
* Add the setting of CBTF_MPI_IMPLEMENTATION to the cbtf-krell
and openspeeedshop package files, so that the mpi implementation
can be found by the tool. This allows users to not have to
set this manually if only one mpi implementation is specified
in the build.
* Update version ranges for dependencies of cbtf packages
* GCC: constrain version 6, 7, and 8 to build with ISL version at
most 0.18
* GCC version 9 (not yet released) will be compatible with
ISL version 0.20 so future GCC releases are constrained to build
with ISL version at most 0.20
* ISL: Add v0.15, replace all md5 sums with sha256 sums for ISL
versions
* GCC versions before 6 were constrained to build with ISL 0.14
but have been confirmed to work with ISL 0.15
* GCC: Place an upper bound on GCC's mpfr dependency
* packages/phist: add variant 'fortran' to enable/disable building the Fortran bindings
* xsdk: add phist
* packages/xsdk: add a comment reg. updating phist dependency
* packages/phist: +fortran variant is only for versions 1.7:
* packages/phist: update comments
* xsdk: forgot to specify kernel library for phist (picked tpetra, assuming trilinos+tpetra will be installed)
* packages/phist: add variant 'openmp' to allow disabling OpenMP
* phist: fix compile error due to missing -lm
* flake8: fix warnings
* packages/phist: fix cmake flag: XSDK_ENABLE_Fortran (rather than XSDK_BUILD_Fortran)
* packages/phist: allow disabling building the scamac matrix generator (which causes trouble on some systems)
* packages/xsdk: disable scamac within phist because it caused problems for @balay and is not essential for now
* packages/phist: actually disable scamac via cmake if ~scamac
* phist: disable openmp from xsdk
This commit overhauls the libMesh package script in a number of ways:
* We now inherit from AutotoolsPackage.
* libMesh contains about a dozen third-party dependencies that can
only be provided by the copy bundled with libMesh. This patch
enumerates all such bundled dependencies and disables all of them
by default.
* libMesh supports compilation with external installations of boost,
eigen, hdf5, slepc, and tbb; this patch adds all of them to
package.py.
* libMesh will look for a few system executables (xz, bzip, gdb) by
default. This is now disabled.
* libMesh will now used a copy of perl installed by spack instead of
trying to find the system copy.
* Compilation with non-system MPI installations requires an additional
flag (--with-mpi=); this has been added.
setup-env includes a call to 'ps' to determine what shell is being
used. 'ps' can be instructed to use a different default output format
via the 'PS_FORMAT' env variable. Thus unset this variable before
calling 'ps'.
* Unite Dockerfiles - add build/run/push scripts
* update docker documentation
* update .travis.yml
* switch to using a preprocessor on Dockerfiles
* skip building docker images on pull requests
* update files with copyright info
* tweak when travis builds for docker files are done
The install() function in the miniconda2 & 3 package.py files referenced
self.url, which is hard-coded to 4.3.11. That's not necessary the
version that the user requested, though. Changed the install() function
to reference self.stage.archive_file. Also added a version string for
4.5.11.
Move suite-sparse gcc version constraint from dealii to suite-sparse
package and update it (gcc 4.9 is required as of version 5.2.0 rather
than 5.1.0). The constraint is now expressed as a conflict rather
than a dependency.
W503 and W504 are issued for line breaks before or after binary
operators (respectively). This allows either approach (explicitly
instructing flake8 to ignore both of these cases).
fixes#9624
merge_config_rules was using `strict=False` to check if a spec
satisfies a constraint, which loosely translates to "this spec has
no conflict with the constraint, so I can potentially add it to the
spec". We want instead `strict=True` which means "the spec satisfies
the constraint right now".
* updated visit package
* make hdf5 and silo conflict with ~gui variant
* Changed parallel => mpi and quotes in description
* Corrected +mpi variant in VisIt package
The previous commit introduced a bug (+parallel variant wasn't renamed +mpi), and the +mpi variant wasn't True by default. This is corrected.
- #8773 made the default mode 0o777, which is what's documented but
mkdirp actually takes the OS default or umask by default
- revert to the Python default by default, and only set the mode when
asked explicitly.
arrow: add missing dependency for python build
The Parquet library moved into the Arrow organisation, hence add a
parquet flavor and adapt dependencies.
* flux: ensure git tags are always available
previous `--unshallow` would only run when `--depth 0` was used. If
1--single-branch` was used, then the `.git/shallow` file would not exist
and `--unshallow` would not work.
add v0.6.0 and update dependencies accordingly
update dependencies to match latest flux versions
adds support for python 3
increases precision of czmq version required for older flux versions
adds new lz4 dependency
* Add stat version 4.0.1
* stat 4.0.1: requires py-xdot
* stat versions at or above 4.0.0 dont require python 2 (it still
requires python but can build with 2 or 3)
* stat versions at or above 4.0.0 dont require py-pygtk and py-enum34
* Replace py-xdot version 0.9.1 with 1.0
* py-xdot: use lib directory vs. lib64 for atk dependency
#9100 added a warning message when a path extracted from a module file
did not appear to be a valid filesystem path. This check was applied
to a variable which could be a list of paths, which would erroneously
trigger the warning. This commit updates the check to run at the
actual point where the path has been extracted.
* Add version 18.9.0 (w/CMake)
* Add version dependent install methods to handle transition from
Make-based package (17.1) to CMake-based package, using the NEST
package as an example
* Remove unnecessary build_targets method for older Make-based
version
* Don't retrieve just the C/Fortran interfaces for netlib-lapack -
blas/lapack libs are now handled the same for all implementations
* Remove netlib-lapack detection patch
* add new version 3.1.0
* add conflict for parmetis without mpi
* strumpack >= 3.1.0 uses -DTPL_ENABLE_<package> instead of
-DSTRUMPACK_USE_<package>
* require cmake at least 3.2
* Add a build_language config.yaml option which controls the language
of compiler messages
* build_language defaults to "C", in which case the compiler messages
will be in English. This allows Spack log parsing to detect and
highlight error messages (since the regular expressions to find
error messages are in English)
* The user can use the default language in their environment by setting
the build_language config variable to null or ''
* Added support for the Aluminum library to LBANN and Hydrogen. Also
fixed several bugs in the grouping of dependencies of both packages.
* Updated the conduit package to have the proper dependency on the
python variant.
* Added new versions for NCCL
* Fixed a bug in how Hydrogen set the path for OpenBLAS.
* Added support for conduit in LBANN.
* vtk: add variants for Xdmf IO support, ffmpeg support, and MPI
* vtk: depends on boost when Xdmf support is enabled
* vtk: add backported patch for python3.7
* vtk: New policy only for cmake >= 3.12
* mesa: add py-argparse dependency (for build)
* lz4: use MOREFLAGS instead of LIBS to add -lrt
scalasca actually requires cubew, not the full cube bundle. However,
cubew@4.4 causes the build of previous versions to fail because there is
no cube-config script anymore.
- `spack license list-files`: list all files that should have license headers
- `spack license list-lgpl`: list files still under LGPL-2.1
- `spack license verify`: check that license headers are correct
- Added `spack license verify` to style tests
- remove the old LGPL license headers from all files in Spack
- add SPDX headers to all files
- core and most packages are (Apache-2.0 OR MIT)
- a very small number of remaining packages are LGPL-2.1-only
- This is bundled into one commit to make it easy to revert if we do get
consent from peetsv's company to relicense these contributions.
- Revert "dialign-tx adding patch for gcc@5.4.0 (#6413)"
- This reverts commit 09eeb991a2.
- Remove the following LGPL-2.1 packages because we have not gotten
consent to relicense them:
- fast-gbs
- igv
- metasv
- perl-xml-libxml
- py-cnvkit
- py-crossmap
- py-pybedtools
- py-pyfaidx
- py-pyvcf
- py-vmd-python
- r-cghflasso
- r-dismo
- r-proc
- r-spatial
- r-utf8
* omega-h: fix dependencies
1. Remove Gmodel altogether. Omega_h doesn't
actually call Gmodel for anything, it was
more or less just doing additional unit
testing for Gmodel
2. Make Trilinos off by default. Most codes do
not need the ParameterList interface, and
Kokkos through Trilinos is not well explored
in Spack
3. Use modern CMake style to choose the zlib
dependency
compilers.yaml can track a module that is needed for a compiler, but
Spack does not fill this in automatically. This adds a note to the
documentation informing the user how to do this.
getopts.pl is not listed as a Spack dependency of sspace-standard,
so without this change, this package was only working on systems
where getopts.pl was already installed.
If we do not specify libdir explicitly, Meson chooses something like
lib/x86_64-linux-gnu, which causes problems when trying to find libraries
and pkg-config files.
* Added conflict in Trilinos for STK when building as shared on Darwin.
* Created shared variant for Nalu and Nalu-Wind to automatically build
Trilinos as static library on Darwin.
* Moved ECP tags from Nalu to Nalu-Wind (the latter is the official
ECP application).
* Remove py-setuptools dependency from py-spyder (fixes#9205)
* Add constraint on transitive qt dependency to build with webkit
extension (required for this package)
* Update pycodestyle dependency constraints
* reordered dependencies (py-zmq, py-chardet) to confirm to the
ordering in setup.py
Spack packages installed using spack buildcache were not running
post-install hooks, which create module files and manage licenses
(if necessary).
This was already occurring for Spack packages installed with
spack install --use-cache
Spack can now be configured to assign permissions to the files installed by a package.
In the `packages.yaml` file under `permissions`, the attributes `read`, `write`, and `group` control the package permissions. These attributes can be set per-package, or for all packages under `all`. If permissions are set under `all` and for a specific package, the package-specific settings take precedence. The `read` and `write` attributes take one of `user`, `group`, and `world`.
packages:
all:
permissions:
write: group
group: spack
my_app:
permissions:
read: group
group: my_team
* superlu-dist: Update package.py for superlu-dist v6.0.0 using CMake
* superlu_dist: Update the header of package.py
* Specify lapack_blas and DCMAKE_INSTALL_LIBDIR
disable support for superlu-dist before v5
current stable releases of PETSc and Trilinos have non-overlapping
requirements on Superlu-dist. For now turn it off by default in
Trilinos, which requires older versions.
* Better default CLI arguments for CDash reporting
--log-format=cdash is now implied if you specify the --cdash-upload-url
option to spack install.
We also now default to writing CTest XML files to cdash_report/ when using
the CDash reporter if no --log-file argument was specified.
* Improved documentation on how to use the CDash reporter
* bison: Add missing build dependencies
bison also depends on cmp, which is currently not available in Spack.
* help2man: Add missing build dependency
* m4: Fix build with newer versions of glibc
* openssl: Add missing build dependency
openssl's configure script is actually a Perl script.
* texinfo: Add missing perl dependency
* diffutils: New package
* findutils: Fix build with newer versions of glibc
* mvapich2, mpich: Add missing findutils dependency
* Add option to disable threads for Geant4
* Add option to enable DagMC in MOAB
* Explicitly add OFF flag when threading is disabled
* Fix unintentional enabling of C++14 in Geant4
The following always evaluated to True:
'c++14' or 'c++1y' in spec:
* Push default flag handlers into module scope
* Preserve backwards compatibility of builtin flag handler names
Ensure Spack continues to work for packages using the `Package.env_flags` idiom and equivalent.
* update docs and tests to match
* Update packages to match new syntax
* xerces-c: Add variant for choice of transcoder (gnuiconv, iconv, icu,
macos, windows). It's important to specify a --enable-transcoder
option on the configure line or else xerces may make a different
choice when multiple transcoders are available.
* Pass the compile flags to configure. For cflags and cxxflags, this is
necessary to respect the value from the spack install line.
Otherwise, xerces (and any autotools package) will choose a default
value that overrides the spack compiler wrapper.
* Add xerces version 3.2.2.
* icu4c needs the --enable-rpath configure option on Darwin
* New package: tldd
* https://gitlab.com/miscripts/tldd
* Improvements
* make it prefetchable by specifying a hash based version
* specify the constraints on pstreams versions (from INSTALL)
* Improvements
* Make it a Makefile Package and only overwrite the install_targets
* Remove the URL as there's really no downloadable 'tar ball' version
* Initial commit for new package STRUMPACK, a linear solver library.
* Update to STRUMPACK version 3.0.3.
This adds a check for OpenMP task priority support (since OpenMP 4.5)
* treesub: new package starting at 0.2
* treesub: fix ant build, add exec script
* treesub: use portable octal format
* treesub: add license header
not sure how it even went missing.
The build process of Catalyst requires Python at build time (see line 86) even when the +python variant is not selected.
When the +python variant is selected, Python becomes required at build, link, and run time. There doesn't seem to be any restriction on the version of Python required when +python is not selected.
* Fix: bug in VTK package
Fix for issue 9423 (https://github.com/spack/spack/issues/9423): compilers other than clang don't have the `is_apple` property.
* VTK package: removing trailing whitespace
So flake8 stops complaining.
* Release 2.0 of the ECP Proxy Apps suite
* update laghos version
* make install dir structure consistent across all ecp-proxy-apps
* minor
* update macsio version
* patching p4est to select the proper openmp lib
* For when compiler.openmp_flag isn't supported
* authored by @nrichart, I only tested it
* Adding openmp variant
Signed-off-by: Ricardo Silva <ricardo.silva@epfl.ch>
Fix two bugs with module file parsing:
* Detection of the CRAY_LD_LIBRARY_PATH variable was broken by #9100.
This fixes it and adds a test for it.
* For module names like "foo-bar/1.0", the associated PACKAGE_DIR
environment variable name would be "FOO_BAR_DIR", but Spack was not
parsing the components and not converting "-" to "_"
Building bazel yielded errors since commit cd9691d. The first commit
fixes the build for older bazel versions, the second commit adds new
versions and ports the patches to these versions.
Fixes#9166
This is intended to reduce errors related to lock timeouts by making
the following changes:
* Improves error reporting when acquiring a lock fails (addressing
#9166) - there is no longer an attempt to release the lock if an
acquire fails
* By default locks taken on individual packages no longer have a
timeout. This allows multiple spack instances to install overlapping
dependency DAGs. For debugging purposes, a timeout can be added by
setting 'package_lock_timeout' in config.yaml
* Reduces the polling frequency when trying to acquire a lock, to
reduce impact in the case where NFS is overtaxed. A simple
adaptive strategy is implemented, which starts with a polling
interval of .1 seconds and quickly increases to .5 seconds
(originally it would poll up to 10^5 times per second).
A test is added to check the polling interval generation logic.
* The timeout for Spack's whole-database lock (e.g. for managing
information about installed packages) is increased from 60s to
120s
* Users can configure the whole-database lock timeout using the
'db_lock_timout' setting in config.yaml
Generally, Spack locks (those created using spack.llnl.util.lock.Lock)
now have no timeout by default
This does not address implementations of NFS that do not support file
locking, or detect cases where services that may be required
(nfslock/statd) aren't running.
Users may want to be able to more-aggressively release locks when
they know they are the only one using their Spack instance, and they
encounter lock errors after a crash (e.g. a remote terminal disconnect
mentioned in #8915).
* petsc: add new versions 3.9.4 and 3.10.0
* petsc:update superlu_dist dependency for version 3.10 [might have to change this later for xsdk]
* slepc: add in new releases 3.9.2 and 3.10.0
- Many container builds are timing out frequently during Spack tests in
Travis CI.
- Travis recommends to try `sudo: required` to see whether this is an
infrastructure issue or something else.
- added `sudo: required` to all Linux builds.
- added --verbose to `spack test` invocation so that we can see more
easily what tests it's timing out on.
Signed-off-by: Todd Gamblin <tgamblin@llnl.gov>
* Catch2: 2.3.0+
Add the latest two releases of Catch2.
In 2.3.0+, the include changed to a future-proof
`<catch2/catch.hpp>` path which dependent projects will break upon.
* openPMD-api: Catch2 2.3.0+
openPMD-api just updated to use the new, future-proof Catch2
headers `<catch2/catch.hpp>`.
When a Spack Executable was configured to capture stderr and the
process failed, the error messages of the process were discarded.
This made it difficult to understand why the process failed. The
exception is now updated to include the stderr of the process when
the Executable captures stderr.
Adds 'code' to the list of suffixes that are excluded from version
parsing of URLs, such that if a URL contains the string
'cistem-1.0.0-beta-source-code', a version X will substitute in to
produce a URL with cistem-X-source-code ('source' was already excluded).
The 'cistem' package version is updated to make use of this (and fix
a fetching bug with the cistem package). A unit test is added to check
this parsing case.
Turns out that there was a configury problem in libhio
which caused issues building the package on non-cray
systems, aka when using mpicc and friends.
Signed-off-by: Howard Pritchard <hppritcha@gmail.com>
This commit adds a patch that prevents the package from hanging. The package would hang because CMakeLists would look for a system MPI. This patch changes the behavior.
* add version 1.3.0
* change build parameters, install process
* mxnet will not compile against spack's nnvm and dmlc so we must use
the distributions provided by the source package.
* also removed 'bin' installation as mxnet is just a shared library.
* improve BLAS linking
* remove unnecessary args
* added variant for python bindings
* r-nloptr: Explicit depends_on(nlopt). Previously there was a hidden dependency, masked by the package's configure script.
* r-nloptr: Change spelling of variables to satisfy pyflake8
* r-nloptr: Break a long line to satisfy pyflake8
* r-nloptr: larger indent on continuation, to satisfy pyflake8
* r-nloptr: another attempt at larger indent on continuation, to satisfy pyflake8
* r-nloptr: now a smaller indent on continuation, to satisfy pyflake8
* r-nloptr: Another attempt at right amount of indent.
* r-nloptr: Another attempt at right amount of indent.
* Use the correct attributes for finding the nlopt includes and libs.
* Lines too long, split into pieces
Improve Spack's parsing of module show to eliminate some false
positives (e.g. accepting MODULEPATH when it is in fact looking for
PATH). This makes the following changes:
* Updates the pattern searching for several paths to avoid the case
where they are prefixes of unwanted paths
* Adds a warning message when an extracted path doesn't exist (which
may help catch future module parsing bugs faster)
* Adds a test with the content mentioned in #9083
This patch does not apply cleanly anymore and breaks clang/Mac builds
balay@asterix /home/balay/git-repo/github/trilinos (develop=)
$ patch -Np1 < /home/balay/git-repo/github/spack/var/spack/repos/builtin/packages/trilinos/xlf_tpetra.patch
can't find file to patch at input line 3
Perhaps you used the wrong -p or --strip option?
The text leading up to this was:
--------------------------
|--- a/packages/tpetra/core/src/Tpetra_Details_libGemm.cpp
|+++ b/packages/tpetra/core/src/Tpetra_Details_libGemm.cpp
--------------------------
File to patch:
* Added jdk version, added r-ks and r-multicool packages, made some r packages less conservative in their r version requirement
* flake8 fixes
* fix formatting in jdk version cmd
* Added type=('build', 'run') to some R package dependencies,
added some missing dependencies to r-ks. Thanks Adam.
* fix flake8 error
Due to an internal bug regarding the handling of scalars
(ndim must be 0) in numpy prior to 1.15.0, openPMD-api
will require the latest numpy 1.15+.
Since there is a small regression with datetime pickling in
1.15.0 from the required fix and spack only packages 1.15.1
currently, we depend on 1.15.1 in the recipe.
Spack originally handled environment modifications in the following
order:
1. clear environment variables
(unless Spack was invoked with --dirty)
2. apply spack-specific environment variable updates,
including variables set by Spack core like CC/PKG_CONFIG_PATH
and those set by installed dependencies (e.g. in
setup_dependent_environment)
3. load all external/compiler modules
1 and 2 were done together. This splits 1 into its own function and
imposes the following order for environment modifications:
1. clear environment variables
2. load all external/compiler modules
3. apply spack-specific environment variable updates
As a result, prepend-path actions taken by Spack (or installed Spack
dependencies) take precedence over prepend-path actions from compiler
and external modules. Additionally, when Spack (or a package
dependency) sets/unsets an environment variable, that will override
the actions of external/compiler modules.
* add 5 most recent versions and git master branch
* add trilinos, gmodel, throw, examples, optimize, symbols,
and warnings variants
* disable XSDK so that Trilinos dependencies are found.
* added py-pycogent
* flake8 edits
* added type=('build', 'run') to deps, changed py-matplotlib and mpi4py
* changed python min version to 2.6
* fixed cython to py-cython dep
* fixed mpi4py to py-mpi4py
* fixed py-sqlalchemy
* added 1.5.3 version
* fixed version 1.5.3 install
* changed deps names to reflect convention
* added variants
* changed download URL and added SHA256 hash
* fixed deps
* added github url for 1.5.3
* removed unneeded deps
* doesn't require py-setuptools
* fixed url and added py-setuptools specific to version 1.9
* fixed variant names and changed cython to normal dependency instead of variant
* removed cython dep, failed to build when it was present. Changed variant name 'mpi4py' to 'mpi' and changed sqlalchemy variant name to mpi
* fixed dependencies and added py-pymysql dep
* added setup-environment step
* changed download url for 1.5.3 to pypi
* fixed flake8 issue
* updated
* added changes
* Addition of latest maestrowf releases.
* Addition of filelock to dependencies.
* Addition of when clause to depends_on for filelock and tabulate.
* Correction of the tar url.
* Add 'extra_env' argument to Executable.__call__: this will be added
to the environment but does not affect whether the current
environment is reused. If 'env' is not set, then the current
environment is copied and the variables from 'extra_env' are added
to it.
* MakeExecutable can take a 'jobs_env' parameter that specifies the
name of an environment variable used to set the level of parallelism.
This is added to 'extra_env' (so does not affect whether the current
environment is reused).
* CMake-based Spack packages set 'jobs_env' when executing the 'test'
target for make and ninja (which does not use -j)
Update the version numbers and now depend on
the develop version of dyninst because of a
feature in openspeedshop depends on new code
in the dyninst develop tree. Without dyninst
develop, openspeedshop will not build.
* py-matplotlib: Add patch to fix FreeType library detection
Adds patch to add $CPATH to locations searched for FreeType header file
Patch has been upstreamed:
https://github.com/matplotlib/matplotlib/pull/11457
* py-matplotlib: Add patch to fix FreeType library detection
* Make patch only apply to versin 2.2.2 (as already upstreamed)
* Fix flake8 format issue
* py-matplotlib: add range to freetype patch
* py-matplotlib: whitespace
1. I presently get errors of the sort
==> Error: KeyError: 'No spec with name mpi in silo@4.10.2%gcc@8.1.0+fortran+mpi
[...]
$HOME/spack/var/spack/repos/builtin/packages/silo/package.py:84, in configure_args:
81 'FCFLAGS={0}'.format(self.compiler.pic_flag)]
82
83 if '+mpi' in self.spec:
>> 84 config_args.append('CC=%s' % self.spec['mpi'].mpicc)
85 config_args.append('CXX=%s' % self.spec['mpi'].mpicxx)
86 config_args.append('FC=%s' % self.spec['mpi'].mpifc)
without the extra explicit MPI dependency.
2. Silo uses zlib, so we should explicitly depend on it.
Added SUNDIALS 3.1.2 and 4.0.0-dev.1 versions and relevant dependencies
(3.1.2 requires CMake 2.8.12). Updated all versions to use sha256 hashes.
Addressed issue specific to clang compiler on macOS.
* Addition of py-filelock and first page of versions.
* Addition of setup.py patch for setuptools.
* flake8 failure fixes.
* Reduced url declaration to a single line.
* Correction to py-filelock to use distutils.
Consolidate prefix calculation logic for intel packages into the
IntelPackage class.
Add documentation on installing Intel packages with Spack an
(alternatively) adding them as external packages in Spack.
Add variant 'nls' for native language support with default True.
The default inside binutils was always on, but this gives a way of
turning it off, if desired.
Adjust the dependencies. Flex is never used for a one-time build.
Bison and m4 should not be needed, except that prior to rev 2.30, gold
did not include its generated files, so bison is needed when +gold.
Drop configure option --enable-interwork. This option does not exist.
This PR includes the following changes:
* Added JDK 10
* Changed the JDK version numbers according to the consensus reached
in #2284
* Added spec['java'].home and spec['java'].libs, similar to #3367
(JDK and IcedTea)
* Added a check to prevent people from installing JDK on macOS
* Set CLASSPATH for packages depending on Java (JDK and IcedTea)
* Add TODO for extending virtual packages (not currently possible)
* Add TODO for adding Java dependents to views
* Add TODO for packages which extend multiple packages (e.g. Java
and Python)
The functions returning the default scope to be modified or listed
have been moved from spack.cmd to spack.config.
Lmod now writes the guessed core compiler in the default modify scope
instead of the 'site' scope.
closes#8916
Currently Spack ends with an error if asked to write lmod modules files
and the 'core_compilers' entry is not found in `modules.yaml`. After
this PR an attempt will be made to guess that entry and the site
configuration file will be updated accordingly.
This is similar to what Spack already does to guess compilers on first
run.
The Trilinos/stk configuration is pretty Nalu specific right
now (and is acknowledged as such in a comment in package.py), and
this commit enables a module that Nalu will be needing.
* r-goseq: new package
* r-goseq: git not url
* r-goseq: adding packages needed at runtime
* r-goseq: package type
* r-goseq: dep types
* r-goseq: one more
- OpenFOAM-v1806/wmake/scripts/have_kahip must check
$KAHIP_ARCH_PATH instead of $METIS_ARCH_PATH to detect the KAHIB
library
- use a local patch file until the issue is hopefully fixed upstream
- Support for Python 3.3 isn't really needed, as nothing uses it as the
default system Python, and nearly everyone will have a newer Python 3
version installed.
* libquo: Packaging updates.
This commit accomplishes a few things:
* Changes distribution fetch source
* Removes a deprecated version (1.2.9)
* Changes dependencies based on version type
* Adds a custom autoreconf stage when using the 'develop' version,
omitting this step when building from a distribution
* Reintroduce v1.2.9.
leveldb headers are typically included via <leveldb/c.h>. The migration
to install_tree in 73c978ddd9 caused the
headers to end up in prefix.include, which breaks existing applications.
#8223 replaced regex-based makefile target parsing with an invocation of
"make -q". #8818 discovered that "make -q" can result in an error for some
packages.
Also, the "make -q" strategy relied on interpreting the error code, which only
worked for GNU Make and not BSD Make (which was deemed acceptable at
the time). As an added bonus, this implementation ignores the exit code and
instead parses STDERR for any indications that the target does not exist; this
works for both GNU Make and BSD Make.
#8223 also updated ninja target detection to use "ninja -t targets". This does
not change that behavior but makes it more-explicit with "ninja -t targets all"
This also adds tests for detection of "make" and "ninja" targets.
This replaces the single github issue-reporting template with templates for the
following types of issues:
* Build issues
* Spack command errors
* Feature request
Each template includes suggestions which are generally useful for the type of
issue encountered.
* QMCPACK now supports QE 6.3. Dropped QE 5.3 due to incompatibilities.
* QMCPACK general clean-up of depends_on.
* Add QMCPACK 3.5.0
* More robust support of MKL detection in QMCPACK.
Fixes#9001#8289 added support for install_tree and copy_tree to merge into an existing
directory structure. However, it did not properly handle relative symlinks and
also removed support for the 'ignore' keyword. Additionally, some of the tests
were overly-strict when checking the permissions on the copied files.
This updates the install_tree/copy_tree methods and their tests:
* copy_tree/install_tree now preserve relative link targets (if the symlink in the
source directory structure is relative, the symlink created in the destination
will be relative)
* Added support for 'ignore' argument back to copy_tree/install_tree (removed
in #8289). It is no longer the object output by shutil.ignore_patterns: you pass a
function that accepts a path relative to the source and returns whether that
path should be copied.
* The openfoam packages (currently the only ones making use of the 'ignore'
argument) are updated for the new API
* When a symlink target is absolute, copy_tree and install_tree now rewrite the
source prefix to be the destination prefix
* copy_tree tests no longer check permissions: copy_tree doesn't enforce
anything about permissions so its tests don't check for that
* install_tree tests no longer check for exact permission matching since it can add
file permissions
* libunwind: add version 1.3-rc1 and variant xz
Version 1.3-rc1 is the latest published snapshot from libunwind.
Variant xz adds spack-built support for reading compressed symbol
tables, or else disables this feature. Without the variant, libunwind
will look for liblzma in /usr/lib.
* Add preferred=True to version 1.2.1 as the last official stable
release.
* gperftools@2.7: new version
* gperftools: update homepage & url
The URL in the homepage field of the gperftools package redirects to
its GitHub repo, and the URL for version 2.3 returns a 404 error. This
commit updates both fields to the most recent sensible values.
* gperftools: update commit hashes to sha256
This commit deletes the url field for all listed versions because it
is no longer needed as a result of updating the package url field. In
addition, all commit hashes are updated to SHA256.
- `imp` is deprecated and seems to have started having some weird
issues on certain Linux versions.
- In particular, the file argument to `load_source` is ignored on
arch linux with Python 3.7.
- `imp` is the only way to do imports in 2.6, so we'll keep it around for
now and use it if importlib won't work.
- `importlib` is the new import system, and it allows us to get
lower-level access to the import implementation.
- This consolidates all import logic into `spack.util.imp`, and make it
use `importlib` if it's avialable.
Replace use of `shutil.copytree` with `copy_tree` and `install_tree` functions in `llnl.util.filesystem`.
- `copy_tree` copies without setting permissions. It should be used to copy files around in the build directory.
- `install_tree` copies files and sets permissions. It should be used to copy files into the installation directory.
- `install` and `copy` are analogous single-file functions.
- add more extensive tests for these functions
- update packages to use these functions.
* zoltan: fix building with Intel compilers
- building with the Intel compilers may require explicit linking
with the Intel Fortran library (libifcore)
* simplify code; get rid of format() for static strings
- dependency patching test didn't attempt to apply patches; just to see
whether they were on the spec.
- it applies the patch now and verifies that that patch was applied.
* r-genelendatabase: new package
* r-genelendatabase: git not url
* r-genelendatabase: add runtime deps and fix package type
* r-genelendatabase: dep type
* Add package for fasta
Add a package for the fasta sequence alignment tools.
The build system is novel (fasta is "venerable", so...). The source
tree includes nearly 50 variant Makefiles, the installer is expected
to pick one that's appropriate and then edit the "XDIR" variable to
specify the installation path. I can only test 64 bit darwin and
linux, so that's all I've included. I'm also assuming that every
64bit x86 system we'll see supports the SSE2 extensions. Other
situations can be dealt with when someone has a test case.
* Switch to MakefilePackage and clean up a bit
- Clean up the description
- Switch to MakefilePackage
- generalize the path to the chosen Makefile
- implement separate `edit`, `build`, and `install` phases.
* Fix call to format()
* Fix broken if logic
* packages/phist: new version 1.6.1
* packages/phist: depends on Python 3.x for generating Fortran and C++ bindings
* packages/phist: new version 1.7.2
* phist: depend on python@3 only for versions later than 1.7
* packages/phist: depend on python@3: rather than 3
* phist: python dependency is type='build', thans @adamjstewart
* Branch with the meson build-system
* Fix build_environment for dual loads and add create code
* Add documentation
* Fixed option list
* Update build_system_guess for meson
* Fixed documentation errors
* Added meson to build and configure and updated documentation
* fix typos
- cc cleanup caused a parsing regression in flag handling
- We added proper quoting to array expansions, but flag variables were
never actually converted to arrays. Old code relied on this.
This commit:
- Adds reads to convert flags to arrays.
- Makes the cc test check for improper space handling to prevent future
regressions.
* libiberty: new package
The libiberty.a library from GNU binutils. Libiberty provides
demangling and support functions for the GNU toolchain.
This package uses the binutils tarfile but only builds the libiberty
subdirectory. This is useful for other packages that want the
demangling functions without the rest of binutils.
Add variant 'fpic' to compile with -fPIC.
Addresses some issues raised in PR #8806.
* libiberty: change variant name to 'pic'.
Allow libiberty to install the library in lib64 and don't try to copy
it to lib.
* update of veloc & dep versions (replace old non-working versions)
* veloc doesn't work with gcc <= 4.9.3
* explicitly pass MPI to CMake for veloc build
- flags were prepended in reverse order to args, but this makes it hard
to see what order they'll be in on the final command line.
- add them in the order they'll appear to make cc easier to maintain.
- simplify code for assembling the command line
- fix separator used in SPACK_SYSTEM_DIRS test
- This corrects most of the issues found by shellcheck
- This also uses ':' as the delimiter for SPACK_SYSTEM_DIRS, for
consistency with other variables.
- filtering using sed causes most builds to slow down quite a bit, as the
compiler wrapper has to run sed many times, and *it* runs many times
- do the system directory parsing directly in bash
- Add tests to ensure that RPATHs are not added in cc mode, which can
cause some builds to fail.
- Change cc.py to use pytest style
- Instead of writing out all the flags, break the flags down into
variables so that it's easy to read what each test is supposed to
check. This should make cc.py more maintainable in the future.
- Adding -L and -Wl,-rpath to compile-only command lines ("cc mode" or
"-c") causes clang (if not also other compilers) to emit warnings that
confuse configure systems.
- Clang will print warnings about unused command-line arguments.
- This fix ensures that -L and -Wl,-rpath are not added if the compile
line is just building an object file with -c
- This also cleans up the cc script in several places.
Spack currently prepends include paths, library paths, and rpaths to the
compile line. This causes problems when a header or library in the package
has the same name as one exported by one of its dependencies. The
*dependency's* header will be preferred over the package's, which is not
what most builds expect. This also breaks some of our production codes.
This restores the original cc behavior (from *very* early Spack) of parsing
compiler arguments out by type (`-L`, `-I`, `-Wl,-rpath`) and reconstituting
the full command at the end.
`<includes> <other_args> <library dirs> <rpaths>`
This differs from the original behavior in one significant way, though: it
*appends* the library arguments so that dependency libraries do not shadow
those in the build.
This is safe because semantics aren't affected by *interleaving* `-I`, `-L`,
and `-Wl,-rpath` arguments with others, only with each other (so the order of
two `-L` args affects the search path, but we search for all libraries on the
command line using the same search path).
We preserve the following:
1. Any system directory in the paths will be listed last.
2. The root package's include/library/RPATH flags come before flags of the
same type for any dependency.
3. Order will be preserved within flags passed by the build (except system
paths, which are moved to be last)
4. Flags for dependencies will appear between the root flags and the system
flags, and the flags for any dependency will come before those for *its*
dependencies (this is for completeness -- we already guarantee this in
`build_environment.py`)
* elfutils: update version and add variants
Add latest version 0.173.
Add variants bzip2, xz and zlib to support reading compressed DWARF
sections, default True.
Move maintainer-mode to a variant with default False. This is only
useful for developers who want to modify the source for generated
files.
* Add dependency on zlib for reading compressed DWARF sections.
Add variants to use bzip2 and xz for compressed sections.
Remove maintainer mode and the dependencies on flex and bison.
These are not used for one-time builds.
Be sure to squash both the commits and the commit messages.
* Fix performance issue when compiling.
Spack was doing active wait when compiling, spoiling one core.
My fix consists in not setting any timeout for select, instead of
the previous 0 second.
* Fix comments about select.select timeout
This commit adds 'libunwind' as the default 'unwind' provider in
spack. This preference should be overridden by the
darwin/packages.yaml file on macOS. Continues the changes discussed in
This commit adds default unwind providers to the default packages.yaml
for darwin. Compiler versions are supplied with the apple-unwind
package so that apple-unwind is only used with Apple's clang fork, and
not with LLVM's clang.
This commit continues the changes discussed in #8823 by creating a new
"apple-libunwind" placeholder package that supplies instructions for
how to configure packages.yaml to register Apple's libunwind
implementation with spack. This package also provides the "unwind"
virtual package representing the libunwind base API. The clang-apple
compiler version should be specified in packages.yaml so that
apple-unwind is only used with Apple's clang compiler, not the stock
LLVM compiler.
This commit begins addressing the ideas discussed in #8823. The
libunwind library now provides the "unwind" virtual package, which
represents the "libunwind base API" common to LLVM libunwind, Apple's
LLVM libunwind, and non-GNU libunwind.
* Switch to perf_event backend to allow user installations but with limited features.
* Fix flake8 issues
* Fix flake8 issues remove filter for INSTALL_CHOWN
* Incorporate the comments
* Add unzip to lua-luaposix
I ended up on an [Ubuntu] system that hadn't had unzip employed
and discovered lua-luaposix requires it (while buildig Lmod).
Closes#8533
* Move unzip prereq to lua, type=run
Rather than touching up each of the lua rocks (packages)
that need unzip, make it available as a run dependency in
lua.
Tested by building lmod on a minimal Ubuntu system.
Flang now uses its own version of llvm and clang (called flang-driver). This is
handled by adding flang-specific versions of the LLVM package and updates flang
to depend on those versions.
* libmonitor: update to version 2018.07.18 and add variant to configure
for hpctoolkit.
* Specify the @2013.02.18 version in cbtf-argonavis and cbtf-krell,
since this is no longer the default version. These are the only
other packages that use libmonitor.
* amg2013: fix homepage and url
* kripke: fix homepage and url
* lcals: fix homepage and url
* lulesh: fix homepage and move to git
* urls on single line
- This was a nasty workaround due to the way our compiler wrappers used
to work. We don't want to have to modify our elfutils installation to
install libdwarf.
- Since cd9691de5, we no longer need this because the package will always
come before dependencies in our include order.
Renames the flux package to flux-core and updates its dependencies, versions,
and variants. Adds a flux-sched package.
This also updates two flux-core dependencies: lua and czmq.
The mumps package was unable to build using the llvm clang compiler
suite, as it defaulted to using mpif90 for linking and mpif90 cannot
be used for linking shared library code. This PR modifies the MUMPS
package.py to allow it to use IBM XL Fortran for linking. It also
eliminates the need for the existing MUMPS IBM XL patches by having
package.py specify the compiler-dependent "shared" flag to the
linker, and always using the compiler suite's Fortran compiler
for linking.
Several packages specified pkg-config as a dependency, which is
a specific implementation of the pkgconfig virtual. Most packages
do not request a specific implementation. The current concretizer
may choose a different implentation of pkgconfig for the nonspecific
dependents, which conflicts with the request for pkg-config. This
replaces all requests for a specific implementation of pkgconfig
with the virtual package as a temporary solution to the issue.
* Update and Bugfix for pexsi/package.py
1. pexsi@0.10.2 is not compatible with superlu-dist@5.4.0 due to [Change LargeDiag to LargeDiag_MC64; Add LargeDiag_AWPM](d7dce5a348).
2. In the 'edit' phase, '@MPICXX_LIB' must be substituted before '@MPICXX' is substituted.
* change dict to list of tuples
Use a list of tuples to remember the order of `substitutions`.
* Update package.py
* Update package.py
* netcdf-fortran: enforce autoreconf when building with NAG.
* netcdf-fortran: patch the configure script instead of running autoreconf to get rid of additional dependencies.
* netcdf-fortran: add comment for the patch.
Spack currently prepends include paths, library paths, and rpaths to the compile line. This causes problems when a header or library in the package has the same name as one exported by one of its dependencies. The *dependency's* header will be preferred over the package's, which is not what most builds expect. This also breaks some of our production codes.
This restores the original cc behavior (from *very* early Spack) of parsing compiler arguments out by type (`-L`, `-I`, `-Wl,-rpath`) and reconstituting the full command at the end.
`<includes> <other_args> <library dirs> <rpaths>`
This differs from the original behavior in one significant way, though: it *appends* the library arguments so that dependency libraries do not shadow those in the build.
This is safe because semantics aren't affected by *interleaving* `-I`, `-L`, and `-Wl,-rpath` arguments with others, only with each other (so the order fo two `-L` args affects the search path, but we search for all libraries on the command line using the same search path).
We preserve the following:
1. Any system directory in the paths will be listed last.
2. The root package's include/library/RPATH flags come before flags of the same type for any dependency.
3. Order will be preserved within flags passed by the build (except system paths, which are moved to be last)
4. Flags for dependencies will appear between the root flags and the system flags, and the flags for any dependency will come before those for *its* dependencies (this is for completeness -- we already guarantee this in `build_environment.py`)
If the OpenMPI build finds the infiniband drivers in /usr/lib64, it adds
-Wl,-rpath -Wl,/usr/lib64 to the OpenMPI wrappers. If the wrappers are using
a compiler outside of /usr, and the OpenMPI wrappers are used to build software
outside of Spack, they will rpath /usr/lib64 into the executable which then has
GLIBC, GLIBCXX runtime errors due to it picking up GCC libraries in /usr/lib64.
This adds the directories specified in "extra_rpaths" to the OpenMPI wrappers,
which allows them to use the correct compiler when invoked outside of Spack
builds.
- previously, output could be confusing when deptypes were only shown for
one dependent when a node had *multiple* dependents
- also fix default coverage of `Spec.tree()`: it previously defaulted to
cover only build and link dependencies, but this is a holdover from
when those were the only types.
* Updated llvm to version 6.0.1. The previous 6.0.0 had an incorrectly declared symbol, discussed at https://reviews.llvm.org/D44140, which, amongst other things, broke py-numba. This version works fine with py-numba.
* Flag the conflict between py-numba and llvm@6.0.0
* Removed a single trailing space to satisfy checks
* First draft of a templight recipe
* Let's be explicit for now, we can be clever later on
* The clang6 version does not work, drop it
* Reorder FIXMEs more logically
* Add patch for current templight develop branch
* Create an artificial stable release of templight
* Take some additional inspiration from llvm package
* Added the templight supporting tools
* Remove old notes from the no-url branch
* Avoid unnecessary join_path
* intel-tbb: Add variant `tm` to disable transactional memory.
Some AMD or very old Intel systems don't support transactional memory. This commit adds a variant `tm` that defaults to `True`, but can be switched off to allow running on those systems.
- Previously, Spack didn't check the arguments you put in version()
directives.
- So, you could do something like this, where there are arguments for a
URL fetcher AND for a git fetcher:
version('1.0', md5='abc123', git='https://foo.bar', commit='feda2343')
- Now, we check the arguments before constructing a fetcher, to ensure
that each package has *only* arguments for a single type of fetcher.
- Also added `test_package_version_consistency()` to the `package_sanity`
test, so that all builtin packages are required to have valid
`version()` directives.
- packagers can specify two top-level fetch URLs if one is `url`
- e.g., `url` and `git` or `url` and `svn`
- allow only one VCS fetcher so we can differentiate between URL and VCS.
- also clean up fetcher logic and class structure
- Packages can remove the top-level `url` attribute and still work
- These are now legal:
- Packages with *only* version-specific URLs (even with gaps)
- Packages with a top-level git/hg/svn attribute and `version`
directives for that.
- If a package has both a top-level hg/git/svn attribute AND a top-level
url attribute, the url attribute takes precedence.
Some packages do not have a `url` and are instead downloaded via `git`,
`hg`, or `svn`. Some packages like `spectrum-mpi` cannot be downloaded at
all, and are placeholder packages for system installations. Previously,
`__init__()` in `PackageBase` crashed if a package did not have a `url`
attribute defined.
I hacked this section of code out, but I have no idea what the
repercussions of that are.
- This hard-codes the hash lengths rather than computing them on import.
- Also cleans up the code in `spack.util.crypto` to make it easier to
understand.
Fix this output error:
```
$ spack -m module loads mpileaks
==> Error: `spack module loads -m t -m c -m l ...` has been moved. Try this instead:
$ spack module t loads mpileaks
$ spack module c loads mpileaks
$ spack module l loads mpileaks
```
In case a deprecated form of the module command is used, the program
will exit non-zero and print an informative error message suggesting
which command should be used instead.
As requested in the review all the commands meant to manage module
files have been grouped under the `spack module` command.
Unit tests have been refactored to match the new command structure.
fixes#2215fixes#2570fixes#6676fixes#7281closes#3827
This PR reverts the use of `spack module loads` in favor of
`spack module find` when loading module files via Spack. After this PR
`spack load` will accept a single spec at a time, and will be able
to interpret correctly the `--dependencies` option.
fixes#4400
The feature requested in #4400 was already part of the module file
configuration, but it was neither tested nor documented. This
commit takes care of adding a few lines in the documentation and a
regression test.
This just because the fixture has been moved one level above the one
it was originally defined. In this more general context there's more
than one configuration file that could be patched for tests.
'spack module' has been split into multiple commands, each one tied to a
specific module type. This permits the specialization of the new
commands with features that are module type specific (e.g. set the
default module file in lmod when multiple versions of the same package
are installed at the same time).
- repo membership test was broken by the refactor of spack/__init__.py
- refactor singleton so that 'spec in repo' works again for `spack.repo.path`
- fix spec command and add basic tests for `spack spec` and `spack spec --yaml`
* Test Spack on Python 3.7 as part of Travis CI
* Currently using xenial to pull-in python 3.7
* As xenial is not officially supported yet, Travis tolerates failures on it.
- There was a lot of documentation in `PackageBase` dating back to the
very first versions of Spack.
- It was repetitive and out of date, and the docs at spack.readthedocs.io
are better.
- Remove the outdated specifics, and leave the minimal useful set of
developer docs in `package.py`.
- This changes `get_checksums_for_versions` to generate code that uses an
explicit `sha256` argument instead if the bare `md5` hash we used to
generate.
- also use a generic digest parameter for the `version` directive, rather
than a specific `md5` parameter.
- Frequently, the documentation build will fail mysteriously in some
spack command.
- The cause is some new bug introduced by the PR, but this is not
apparent because the unit tests haven't run and the doc tests aren't
targeted at code bugs.
- Users end up puzzled by doc failures when they're really code failures.
- Move the doc tests parallel with the code tests, so that we can more
easily see bugs like this.
* fix to build libmatheval with guile@2.0:
* fixed formatting issues
* fixed formatting issues2
* fixed check for autoreconf and renaming of configure.in
* fixed force_autoreconf
* now patching configure, which is hopefully more robust
* minimal patches for guile 2.0 and 2.2 (disabeling unit testing for libmatheval)
* minimal patches for guile 2.0 and 2.2 (disabeling unit testing for libmatheval)
* add mofem-cephas package
* add mofem fracture module
* add user modules build and fracture modyle
* add minimal surface module
* add slepc variant
* bump mofem core lib version
* bump mofem core lib version
* bump version
* fix bug
* set upper bound to petsc version and other chanes
* fix indentation
* add minimal med file installation
* chcekc with flake8 and installation with spack packages
* add variants to med package
* upper bound to adol-c and remove obsolete internal package install
* fix basic module install
* module install from external source in extended prefix
* remove obsolte code and reverse to variant doxygen
* fix git adress
* improve packaging for mofem users modules
* fix flake8
* move dependencies after variants
* move root_cmakelists_dir right before cmake_args
* remove unused variants
* use append for single element
* replace root_cmakelists_dir
* use install_tree instead copy tree
* simplify code
* remove phase and mkdirp
* add run tests
* instal ext modules to ext_users_modules directory
* move version below url
* simplify directory name
* use underscore in variant name
* remove unused variable
* fix link to blas libs
* add missing boost dependence
* fix problem with copying module source code
* change variant name form doxygen to docs
* add expanded description
* make installation consistent with spack
* fix flake8
* make extensions installed
* code comments and minor corrections
* make slepc variant false by default
- Add command-line scope option to Spack
- Rework structure of main to allow configuration system to raise
errors more naturally
Co-authored-by: Todd Gamblin <tgamblin@llnl.gov>
- Fixes a bug in `llnl.util.lock`
- Locks in the current directory would fail because the parent directory
was the empty string.
- Fix this and return '.' for the parent of locks in the current
directory.
* New package: Verrou, a floating point error checker
* Clean up usage of patch
* Account for future patches in version ranges
* Manual call to autogen is not needed anymore, but extra dependencies are needed
* Use trick from automake to evade shebang length limit
* Use a glob instead of an explicit file list
* Turns out conflict() does not do what I want...
* Suggestions from @citibeth reviews
- Use 'develop' convention for the master branch
- Prefer tarballs over git repositories
- Increase robustness against default configuration
* Drop a couple of older releases with a different patching procedure
* Playing coding style golf
* Various bug squashing
- Missed master -> develop substitution
- Turns out manually calling autogen.sh is needed after all
- Missed + in spec
* More style golfing to fit in 80 columns
* Remove unnecessary mention of the develop branch
Enforce PEP8 naming conventions for things like variables, methods,
classes, etc.
See the table here:
https://pypi.org/project/pep8-naming/
...for error codes emitted, in case some should be added as
exceptions in the flake8 configuration files.
* py-pytest: Added dependency on py-setuptools-scm, as well as py-setuptools, to prevent the build from silently installing its own copy of setuptools-scm
* py-pytest: Dependency on setuptools-scm only needs to be 'build'
Replace regex-based target detection for Makefiles with a preliminary "make -q"
to check if a target exists. This does not work for NetBSD make; additional work
is required to detect if NetBSD make is present and to use a regex in that case.
The affected makefile target checks are only performed when the "--test" flag is
added to a "spack install" invocation.
* The xlf_seacas.patch fails on the latest trilinos develop branch so confine it up to version 12.12.1.
* Fixing likely mistak in seacas patch version range.
Fixes#8036
Before this PR Package.installed was returning True if the spec prefix
existed, without checking the DB. This is wrong for external packages,
whose prefix exists before being registered into the DB. Now the property
checks for both the prefix and a DB entry.
Spack provides a number of classes based on commonly-used build systems
that users can extend when writing packages; the classes provide functionality
to perform the actions relevant to the build system (e.g. running "configure" for
an Autotools-based package). This adds documentation for classes supporting the
following build systems:
* Makefile
* Autotools
* CMake
* QMake
* SCons
* Waf
This includes build systems for managing extensions of the following packages:
* Perl
* Python
* R
* Octave
This also adds documentation on implementing packages that use a custom build
system (e.g. Perl/CMake).
Spack also provides extendable classes which aggregate functionality for related
sets of packages, e.g. those using CUDA. Documentation is added for
CudaPackage.
- The setup-env.sh script currently makes two calls to spack, but it
should only need to make one.
- Add a fast-path shell setup routine in `main.py` to allow the shell
setup to happen in a single, fast call that doesn't load more than it
needs to.
- This simplifies setup code, as it has to eval what Spack prints
- TODO: consider eventually making the whole setup script the output of a
spack command
* update help of `clean --all` to include `-p`
* remove old orphaned `.pyc` removal
* restrict removal or orphaned pyc files to `lib/spack` and `var/spack`
- Clean up error messages for when a lock can't be created, or when an
exclusive (write) lock can't be taken on a file.
- Add a number of subclasses of LockError to distinguish timeouts from
permission issues.
- Add an explicit check to prevent the user from taking a write lock on a
read-only file.
- We had a check for this for when we try to *upgrade* a lock on an RO
file, but not for an initial write lock attempt.
- Add more tests for different lock permission scenarios.
- write locks previously wrote information about the lock holder (host
and pid), and read locks woudl read this in.
- This is really only for debugging, so only enable it then
- add some tests that target debug info, and improve multiproc lock test
output
* initial version of the package pyTorch (#171)
* adding version 0.4
* flake8 correction
* reordering version, removing run dependecies
* Changing depends_on type to default
When a user specifies a URL for a specific version of a package, Spack originally
would use that URL for all newer versions of the package. This behavior has
proven to be generally more harmful than useful, so this PR removes the feature
such that a version-specific URL override affects only that version.
* Update kaldi
always use the --fst-version otherwise it does the wrong thing with
selecting the fst version.
also enable speex by default
* Give version a more meaningful name
* Update older version with a date string instead
* gtkorvo-atl: Updated to version 2.2
* gtkorvo-dill: Updated to version 2.4
* gtkorvo-enet: Updated to version 1.3.14
* libevpath: Updated to version 4.4.0
* libffs: Updated to version 1.5
* gtkorvo-atl: Bugfix: Fixed cmake args
* gtkorvo-enet: Switched order of versions to have the latest version on the top
* gtkorvo-atl: Cleaned up cmake_args. Added conditional self.run_tests
* gtkorvo-dill: Cleaned up cmake_args. Added conditional self.run_tests
* libffs: Correction in dependency versions. Cleaned up cmake_args. Added conditional self.run_tests
* libevpath: Correction in dependency versions. Cleaned up cmake_args. Added conditional self.run_tests
* gtkorvo-atl: Removed unused option from cmake args
If the user sets "ccache: true" in spack's config.yaml, Spack will use an available
ccache executable when compiling c/c++ code. This feature is disabled by default
(i.e. "ccache: false") and the documentation is updated with how to enable
ccache support
* Added support for uncrustify 0.67.
Since the build system changed for uncrustify at version 0.64,
I had to change the package from an AutotoolsPackage to a plain
Package and use @when annotations to build the packages differently.
* Removed blank line at end of file
* Added support for python2.6 in uncrustify configuration.
- Versions before 2.5.2 suffer from a serious security problem
(https://nvd.nist.gov/vuln/detail/CVE-2018-12021). Deleted all
versions prior to v2.5.2.
- Update homepage (looks like singularity's going Freemium).
- Use release tarball instead of automagically generated archive
tarball (digest values are more stable).
- No longer need autotools for released versions, just @devel.
- Install needs to be single-threaded, otherwise when installed as
root the various 'chmod' operations race against the copies.
preCICE (Precise Code Interaction Coupling Environment) is a
coupling library for partitioned multi-physics simulations.
Partitioned means that preCICE couples existing programs (solvers)
capable of simulating a subpart of the complete physics involved in
a simulation.
of-precice: new package
preCICE adapter for OpenFOAM
* First draft of amrvis package file.
* More additions to amrvis.
* Formatting.
* Forcing compiler environment variables to point to spack mpi compilers when using mpi.
* Disabling intel compiler for amrvis.
* Comments.
* Refining amrvis package file.
* Moving library and include locations variables to be inserted at the first lines of the makefile.
* Globbing amrvis binary instead of constructing its name.
* Making env variable setting consistent and fixing globbing of executable.
* Using iglob instead of glob.
* Turning MPI on by default for Amrvis.
* ENH: Building OpenFOAM sub-packages (issue #8579)
* Some support for packages building with OpenFOAM
- Adjust the wrappers calling the OpenFOAM Allwmake script. Have them
look for a Allwmake-spack file first, which is assumed to contain
special adjustments for compiling with spack.
This file could be delivered as part of a tarball (which is unlikely)
or generated on the fly by the spack sub-package as part of its
patch or configure stage.
CONFIG: change the default paraview variant for openfoam to be False
- the different combinations of paraview backends, off-screen etc
make it difficult to suggest that building with paraview as
a standard dependency makes much sense.
Additionally, building paraview with qt can become quite an issue.
So it makes much more sense to only enable that upon request.
ENH: add a +vtk variant.
- for VTK with off-screen rendering to be used by the runTimePostProcessing
function object, which is a fairly simple framework for generating images of
some OpenFOAM derived objects (eg, sampling planes).
SPACK spec problem:
- reflect the flex restriction impose by the scotch dependency within
the openfoam spec as well, as partial workaround for buggy or annoying
spec resolution.
OTHER:
- updated the backstop foamEtcFile file to include args handling
as per the OpenFOAM-v1806 updates.
* new version: OpenFOAM-v1806
- https://www.openfoam.com/releases/openfoam-v1806/
Functional updates:
- `python` now creates a copy of the `python` binaries when it is added
to a view
- Python extensions (packages which subclass `PythonPackage`) rewrite
their shebang lines to refer to python in the view
- Python packages in the same namespace will not generate conflicts if
both have `...lib/site-packages/namespace-example/__init__.py`
- These `__init__` files will also remain when removing any package in
the namespace until the last package in the namespace is removed
Generally (Updated 2/16):
- Any package can define `add_files_to_view` to customize how it is added
to a view (and at the moment custom definitions are included for
`python` and `PythonPackage`)
- Likewise any package can define `remove_files_from_view` to customize
which files are removed (e.g. you don't always want to remove the
namespace `__init__`)
- Any package can define `view_file_conflicts` to customize what it
considers a merge conflict
- Global activations are handled like views (where the view root is the
spec prefix of the extendee)
- Benefit: filesystem-management aspects of activating extensions are
now placed in views (e.g. now one can hardlink a global activation)
- Benefit: overriding `Package.activate` is more straightforward (see
`Python.activate`)
- Complication: extension packages which have special-purpose logic
*only* when activated outside of the extendee prefix must check for
this in their `add_files_to_view` method (see `PythonPackage`)
- `LinkTree` is refactored to have separate methods for copying a
directory structure and for copying files (since it was found that
generally packages may want to alter how files are copied but still
wanted to copy directories in the same way)
TODOs (updated 2/20):
- [x] additional testing (there is some unit testing added at this point
but more would be useful)
- [x] refactor or reorganize `LinkTree` methods: currently there is a
separate set of methods for replicating just the directory structure
without the files, and a set for replicating everything
- [x] Right now external views (i.e. those not used for global
activations) call `view.add_extension`, but global activations do not
to avoid some extra work that goes into maintaining external views. I'm
not sure if addressing that needs to be done here but I'd like to
clarify it in the comments (UPDATE: for now I have added a TODO and in
my opinion this can be merged now and the refactor handled later)
- [x] Several method descriptions (e.g. for `Package.activate`) are out
of date and reference a distinction between global activations and
views, they need to be updated
- [x] Update aspell package activations
* Add specific version to package verilator
Change-Id: If7645410ec192f92a5eed83ee9b317b569576b4a
* fix dependency types
Change-Id: Ib36c72257c1fa6678c8553225ca21a010d7ae6d1
- Spack was assuming that a group with gid == current uid would always exist.
- This was breaking the travis build for macos.
- also fix issue with the DB tarball test finding coverage filesx
* A new package: perl-compress-raw-zlib.
* A new package: perl-compress-raw-bzip2.
* A new perl package: perl-io-compress.
* flake8.
* Add zlib and bzip2 dependency.
* petsc: remove dependency on boost - its required only as a dependency on trilinos
Also set COPTFLAGS='' etc - this way CFLAGS is the primary way optimization flags are set by spack
* petsc: update pkgs with petsc~boost dependency
* fix flake8
* boost is optional for trilinos - so adjust petsc dependency on boost [via trilinos] accordingly
* do not add boost dependency on petsc [due to trilinos requirement]. Trilinos adds this dependency anyway.
plasma@develop has another flag 'DHAVE_MKL' that needs to be disabled [for non-mkl build]
add dependency on gcc-6.0
add in new dependency on readline
plasma: requires both LIBS in make.inc and command line?
If cmd line is missing -some internal values from plasma makefile gets used - causing link errors [on Mac]
tools/lua-5.3.4/src/Makefile: $(MAKE) $(ALL) SYSCFLAGS="-DLUA_USE_LINUX" SYSLIBS="-Wl,-E -lreadline"
if make.inc is missing - I get:
/usr/bin/ld: cannot find -lmkl_intel_lp64
/usr/bin/ld: cannot find -lmkl_sequential
/usr/bin/ld: cannot find -lmkl_core
collect2: error: ld returned 1 exit status
Also add a new dependency -ldl
m4 exited immediately (causing the nettle build to fail).
```
(alice)[11:09:46]spack>>/Users/hartzell/tmp/spack-mac/spack/opt/spack/darwin-highsierra-x86_64/clang-9.1.0-apple/m4-1.4.18-caxsf7l5yd7qbq2bown6bzi5el3ltfwf/bin/m4
Abort trap: 6
(alice)[11:10:18]spack>>fg
```
Tracked down a fix via the Homebrew recipe. I'll send a note to the
`bug-m4@gnu.org` list to ensure that upstream knows about it.
* Fix gcc@5.5.0 on OS X 10.13.5
gcc@5.5.0 failed to build on my OS X 10.13.5 box.
I found/modified a patch in the MacPorts world that traced back to
Homebrewy, https://trac.macports.org/ticket/56502#no1
The issue has also been reported to gcc,
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=83531
I eventually discovered a nearly identical version of the patch at
Homebrew.
https://github.com/Homebrew/formula-patches/blob/master/gcc%405/10.13_headers.patch
gcc builds with this and I can compile many things, but e.g. curl and
cmake both fail. As @davydden and others observed in #1847, it seems
to be an apple thing.
* Fix mistaken change to patch when clause (added :)
I mistakenly deleted a colon from the
'darwin/gcc-7.1.0-headerpad.patch'.
cd /tmp/balay/spack-stage/spack-stage-joxtpr/omega_h-9.13.4/spack-build/src && /home/balay/spack/opt/spack/linux-ubuntu16.04-x86_64/gcc-5.4.0/openmpi-3.1.0-7ws4tgx4z3o
ru5g5skjrxs5sobpeiyef/bin/mpic++ -Domega_h_EXPORTS -I/home/balay/spack/var/spack/stage/omega-h-9.13.4-th4b67fz5es6cbgcar742jxamjt3tqjb/omega_h-9.13.4/src -I/tmp/bala
y/spack-stage/spack-stage-joxtpr/omega_h-9.13.4/spack-build/src -I/home/balay/spack/var/spack/stage/omega-h-9.13.4-th4b67fz5es6cbgcar742jxamjt3tqjb/omega_h-9.13.4/tpl
-I/home/balay/spack/opt/spack/linux-ubuntu16.04-x86_64/gcc-5.4.0/zlib-1.2.11-5nus6knzumx4ik2yl44jxtgtsl7d54xb/include -O2 -g -DNDEBUG -fPIC -o CMakeFiles/omega_h.di
r/Omega_h_map.cpp.o -c /home/balay/spack/var/spack/stage/omega-h-9.13.4-th4b67fz5es6cbgcar742jxamjt3tqjb/omega_h-9.13.4/src/Omega_h_map.cpp
In file included from /usr/include/c++/5/chrono:35:0,
from /home/balay/spack/var/spack/stage/omega-h-9.13.4-th4b67fz5es6cbgcar742jxamjt3tqjb/omega_h-9.13.4/src/Omega_h_timer.hpp:4,
from /home/balay/spack/var/spack/stage/omega-h-9.13.4-th4b67fz5es6cbgcar742jxamjt3tqjb/omega_h-9.13.4/src/Omega_h_timer.cpp:1:
/usr/include/c++/5/bits/c++0x_warning.h:32:2: error: #error This file requires compiler and library support for the ISO C++ 2011 standard. This support must be enabled with the -std=c++11 or -std=gnu++11 compiler options.
#error This file requires compiler and library support \
^
* Added minuit package
* Removed default method and template
* Different URLs for old versions
* Patches for versions down to 1.5.0
* Reverted ROOT package changes
* Removed print statement
* Added checksums for older versions
* Consistent version names
- pytest was not reporing the correct version from pytest.__version__.
It reported 'unknown'
- this fixes issues on some systems where system-installed pytest plugins
would try to use the version and convert it to an int
* SAMRAI: remove dependency on boost for samrai 3.12.0 or later versions, but keep the older versions ccontinuing working with boost 1.64.0 or earlier. Also, the older versions cannot compile with boost 1.65.0 or above. (#8491)
* SAMRAI: fix typos of comments and remove obsolete code.
* Updating Nalu now that it can build with shared libraries and adding the new Nalu-Wind wind energy focused application.
* Explicitly turn off TPLs for Nalu when not enabled.
* Adding patch for OpenCV to fix a conflict between an OpenCV and a CUDA header file
* Added patch command to package.py file for OpenCV SPACK package.
* Adding version range for patch.
* shortening comment lines, since testsuite failed
* Add changes to the Krell packages with version updates and ranges for the latest release.
* Add reviewer requested changes. Remove preferred clauses and update the qt version for the cbtf-argonavis-gui@develop
* Trigger integration/test restart
* Add patching for xlf when %clang. Changes are effective only when Fortran compiler is 'XL'.
* Use the right version for trilinos when patching for seacas.
setting to use, instead of the spack compiler.
%clang can use multiple Fortran compilers with different options, so
we need to know what compiler is used in order to provide the correct
options.
* Add new version (3.4.0) with new variant (pumi).
* Fix an issue: skip the tests in examples/pumi as they require
files that are not included in the mfem distribution/repository.
* Use patches to fix the testing issues with PUMI and PETSc v3.9.
* Use the new bit.ly short link for v3.4.
* Add the ability to build the dyninst master branch under the develop version name. Fix elfutils and libdwarf to work with the various dyninst versions and vice-versa.
* Add reviewer suggested change to dyninst package file. Remove colon in spec.satisfies clause.
* Add reviewer suggested change to dyninst package file. Remove duplicate lines.
* Add reviewer suggested change to dyninst package file. Encompass all versions under 10.0 with respect to using libdwarf in the build.
Fixes the following problem:
==> Installing trilinos
==> Warning: Suspicious requests to set or unset 'LD_LIBRARY_PATH' found
==> Warning: env.unset('LD_LIBRARY_PATH') at /home/balay/spack.new/lib/spack/spack/build_environment.py:269
==> Warning: ---> env.set('%s' % key, value) at /home/balay/spack.new/lib/spack/spack/build_environment.py:292
==> Trying to clone git repository: https://github.com/trilinos/Trilinos.git at tag develop
Fetching tags only, you probably meant:
git fetch --tags
error: pathspec 'develop' did not match any file(s) known to git.
==> Fetching from [git] https://github.com/trilinos/Trilinos.git failed.
==> Error: FetchError: All fetchers failed for trilinos-develop-tgkpnctp6nn4imoizxlw4ymqyztcie4n
The patch to select ELPA API needs to be applied to
version 6.2.0 too.
The link line of hdf5 is computed incorrectly from
the configuration part of the build-system. To fix
this the relevant file (make.inc) is patched.
* py-pybind11: Additionally install python extensions e.g. run python setup
* py-projectq: Added new package for quantum computing
* py-projectq: corrected typos
* edited flake8 incompatibilities
* additional flake8 faults
* changes according to suggestions
* flake8 errors fixed
* typo and missing environment variable set
* py-pybind11:added setup_environemt
* py-projectq:changed version according to request
* flake8 correction
* py-pybind11:changed to setup_py and removed noop
* py-projectq:removed dependency and changed type of py-pytest dependency
* py-pybind11: flake8 error corrected
* py-projectq: Small error correction
* Git's gitk needs TK's wish to be found in PATH
* Add a _runtime_ dependency on `tk`
* Add an environment rule to add the path to TK's `wish` program to $PATH for
the generated `git` modulefile.
* Make TclTk run environment an optional variant for git.
* Cleanup based on PR recommendations.
The following improvements are made to cxx standard support
(e.g. compiler.cxxNN_flag functions) in compilers:
* Add cxx98_flag property
* Add support for throwing an exception when a flag is not supported (previously
if a flag was not supported the application was terminated with tty.die)
* The name of the flag associated with e.g. c++14 standard support changes for
different compiler versions (e.g. c++1y vs c++14). This makes a few corrections
on what flag to return for which version.
* Added tests to confirm that versions report expected flags for various c++
standards (or raise an exception for versions that don't provide a given cxx
standard)
Note that if a given cxx standard is the default, the associated flag property will
return ""; cxx98 is assumed to be the default standard so this is the behavior for
the associated property in the base compiler class.
Package changes:
* Improvements to the boost spec to take advantage of the improved standard
flag facility.
* Update the clingo spec to catch the new exception rather than look for an
empty flag to indicate non-support (which is not part of the compiler flag API)
The tcl package references the original sources upon install, and the tk build
requires the tcl sources. This updates the tcl package to install its sources following
the example of #4102, and also updates the tclConfig.sh file to properly reference
the installed sources (rather than the staging directory created by Spack).
* Update to 2018.3 and improve version URLs, dependencies and patching.
Improve to version -> URL translation to be algorithmic rather than manual.
Make binutils build dependency explicit where appropriate because we're sensitive to the assembler used.
Build's use of '-mrtm' is only patched away if it needs to be because the assembler is too old.
Incorporate @adamjstewart request to convert GCC version check to a conflict.
* Allow specification of C++ standard.
* Improve TBB patch per PR comments.
* Restrict RTM patch application to older OS only.
* New version 2018_U4.
* Delete support for old versions of Elemental in Hydrogen
* Fix cmake_args versions in lbann
* Remove unused import (again)
* Revert elemental to pre-llnl fork
* Strip tailing whitespace from libgfortran.so
* Fix flake8
* Remove debug print
* * Add back elemental@develop
* Prohibit installation of hydrogen previous to 0.99
* Fix flake8
* Change cmake_args error to a conflicts for old versions of Hydrogen
* Use ~ not -
* fermisciencetools: new package
* fermisciencetools: new package
* added comments about the error in building the source distribution; fixed flake8 errors
* add sqlitebrowser
Change-Id: I20cac709509e82d348cebc11b0d584b53b2ad3d4
* add message and clean up unnecessary code
Change-Id: Ib9d2229ee25d751274680d06824891ee78fa5970
* and make flake8 happy
Change-Id: I814eb89eefe1767af4d8043b08b61991750516df
* make gcc-independence great again
Change-Id: I75f7271757782a6735eb7f03c0551d190722974d
* remove explicit import of dso_suffix
Change-Id: I626c6391b22524895199db5508606c4cf22bf9b5
* add a shared variant for all supported versions
* address comment from @citibeth about MacOS lib installs
* address additional comments from @citibeth about MacOS lib installs
* set +shared variant to be default behavior
* add package py-cartopy
Change-Id: Ifbe541e852921c384a381594a2c1110814556311
* fix comments and flake8
Change-Id: I8d5e994580e8252e00654e68ad79d74e06089939
* remove pip dependency... not sure why added in the first place
Change-Id: Ie2a788887dac30778cc43931a0732980a347514d
* move build, link, run to default dependencies
Change-Id: I1ec0cb7dc4a697c4ca9e68a0cc0d8694cc21c100
Fixes#7885#7193 added the patches_to_apply function to collect patches which are then
applied in Package.do_patch. However this only collects patches that are
associated with the Package object and does not include Spec-related patches
(which are applied by dependents, added in #5476).
Spec.patches already collects patches from the package as well as those applied
by dependents, so the Package.patches_to_apply function isn't necessary. All
uses of Package.patches_to_apply are replaced with Package.spec.patches.
This also updates Package.content_hash to require the associated spec to be
concrete: Spec.patches is only set after concretization. Before this PR, it was
possible for Package.content_hash to be valid before concretizing the associated
Spec if all patches were associated with the Package (vs. being applied by
dependents). This behavior was unreliable though so the change is unlikely to
be disruptive.
Provides a patch for the fix in
https://github.com/ornladios/ADIOS/pull/182
Installed ADIOS MPI-enabled libraries are with that able to also
perform serial I/O without the need to start the whole app via
`mpiexec`.
* trilinos: add more package conflicts
This was done by putting in required lib dependencies from each
package's `cmake/Dependencies.cmake`.
* trilinos: added few more trilinos packages
* trilinos: hiccup fix
* add 1.4.13.4 for SZ
* undo spack link (change llnl/spack to spack/spack) and change .zip to .tar.gz
* revise package.py for SZ 1.4.13.4 based on Adam's comments.
Fixes#8345
Spack environment modifications are applied before modules are loaded; this
includes settings to CC, FC, F77, and CXX, which point to the Spack compiler
wrappers. If the loaded modules set CC, this overrides the Spack compiler
wrappers. This PR adds a context manager to preserve the values of CC etc. that
are set by Spack: any effects on the CC, FC, F77, and CXX variables from modules
are undone and their original values are restored.
* adding partially completed recipe for phast. will need clapack as dependency
* try to install phast
* seems to still need CLAPCK
* using CLAPACK instead
* phast: updated
* fixed flake8 error
* phast cannot build with clapack built with external blas?
* phast: depend on clapack built without external blas
* fixed flake8 errors
* phast: updated build_directory
* pybind11: test support
Add a test functionality to pybind11.
* CMake: test also on "make check"
Some projects use non-CTest manual targets for tests.
* Fix digest for abyss@1.5.2
The digest value for v1.5.2 appears to have changed. This value works
for me, today.
The existing value v2.0.2 works as is.
* Use "download" URL for ABYSS
Following the suggestion of the project team, switch to using the
release tarballs from the "download" URL instead of the automagically
generated ones at the "archive" URL.
Converstation here: https://github.com/bcgsc/abyss/issues/232
* extend Prefix class with join() member to support dynamic directories
* add more tests for Prefix.join()
* more tests for Prefix.join()
* add docstring
* add example to docstring of Prefix class
* cleanup Prefix.join() tests
* use Prefix.join() in Packaging Guide
The URL that we were using for pigz only provides the current
version. When the author released v2.4 the package became
uninstallable (unless you have the tarball cached).
It turns out that their are tarballs of the tagged source tree
available on GitHub and they're sufficient to build things (no
additional bits needed).
This commit switches the URL to the GitHub archive URL.
Sadly, this changes the v2.3.4 digest, but that version was no longer
available/install-able anyway, so "net-net we're still in positive
territory".
Installing the current version of Picard fails with:
```
==> Error: Unable to parse extension from https://github.com/broadinstitute/picard/releases/download/2.18.3/picard.jar.
If this URL is for a tarball but does not include the file extension
in the name, you can explicitly declare it with the following syntax:
version('1.2.3', 'hash', extension='tar.gz')
If this URL is for a download like a .jar or .whl that does not need
to be expanded, or an uncompressed installation script, you can tell
Spack not to expand it with the following syntax:
version('1.2.3', 'hash', expand=False)
```
Fix it, but adding `expand=False`, like the other versions.
Fixes#8217
Trying to relocate a distribution when the new and old paths are
equal leads to failure, because the test that ensures that no
unrelocated bits are left over always fails. As an example, this
occurs if a user installs a package, generates a binary with it
using 'spack buildcache', uninstalls it, and then attempts to
reinstall into the same spack installation using the generated
binary package.
This updates the relocation check to accept the presence of the
old prefix in binaries if the package is being reinstalled into
its original location.
* fix missing -lrt for Centos6, issue: #7552
* fix missing -lrt for Centos6, issue: #7552
* fix using make() as opposed to patching
* remove patch, as the fix is provided withing package.py instead
* make flake8 happy
* allow user to constrain dependencies that are added conditionally
* remove check for not-visited deps from normalize, move it to concretize. The check now runs after the concretization loop completes (so an error is only reported if the user-mentioned spec doesnt appear anywhere in the dag)
* remove separate full_spec_deps variable; rename spec_deps to all_spec_deps to clarify that it merges user-specified dependencies with derived dependencies
* add unit test to confirm new functionality
- `spack config blame` is similar to `spack config get`, but it prints
out the config file and line number that each line of the merged
configuration came from.
- This is a debugging tool for understanding where Spack config settings
come from.
- add tests for config blame
* Update package: add info for emacs@26.1
In addition to the digest, it needs gnutls (perhaps there's a way to
turn it off, but it seems like a Good Thing to encourage, so...).
* Clean up tls variant handling
- Don't to specify the dependency for @26.1:, just need to fix what's
already there.
- Don't assume that configure will DTRT w.r.t. gnutls when it's
discovered in the environmnet, use explicit configure switches.
- Clean up some unnecessary dependencies when +tls. Why would you
need libxpm when building +tls? Will ask the original author about
this.
* Use append when adding configure switches
* Add dependency on pcre
When one builds +X, Spack's pcre is pulled into the build via
the X dependencies. When one builds without it, the build discovers
and links against the system library.
We might as well just depend on it and have a consistent outcome.
The recipe previously did not include -DWITH_JASPER
Other edits:
* BUILD_IPP_IW is BUILD instead of WITH; it was in a different section of the CMakeLists.txt. Move it to the top
* Several options were missing :BOOL, add them
* -DWITH_ZLIB doesn't actually exist as an option, at least in the latest version.
* Added new package Camellia.
* Fixed an issue with a missing blank line.
Added a couple more Trilinos packages (things I expect to depend on eventually).
* Fixed a couple issues raised by @adamjstewart.
* Specified 4.9.2 as the supported MOAB version.
5.0.0 causes build issues on Macs during the link of Camellia's drivers. The issue is that we end up with -framework -lAccelerate in the link line, where it should read -framework Accelerate, I believe. I'm not sure where this comes from, but 4.9.2 does not have this issue.
* Trimmed slightly the Trilinos requirements.
Also deleted a blank line that had spaces in it because the Travis-CI thing didn't like that.
* Reduced length of MOAB description (Travis-CI complained).
* Reduced the MOAB line further; didn't quite get there last time!
This is to satisfy Travis-CI's requirement that the description be less than 80 characters long (E501).
* Put in my GitHub handle, in place of an allusion to my email address.
Also added a new argument for Camellia build, and made the MOAB description longer (thanks to @adamjstewart's note that long lines are allowed so long as they are not split).
* Added MPI dependency.
Added explicit OFF for MOAB enablement when ~moab.
Swapped order of variant and version.
* Add Tophat v2.1.2, fixes automake issues
The Tophat team merged the patch from
https://github.com/spack/spack/pull/8244 and release v2.1.2
This change requires the old automake if you're building the older
release and is relaxed about the automake it requires if you're
building the v2.1.2 release.
Building v2.1.1 and v2.1.2 works on a CentOS 7 system, I don't have
the necessary bit to actually test the resulting binaries.
Here's the commit note from the change that was ultimately merged
upstream.
Fixes#8025
> Tophat was failing to build with automake@1.16.1, it worked with the
> older automake@1.15.1. This commit adds a patch to Tophat's
> src/Makefile.am which cleans up a few things. The result builds
> successfully with both automake@1.15.1 and automake@1.16.1. I have no
> way to check that the resulting builds Do The Right Thing.
>
> It changes two things:
>
> the original Makefile.am had a rule for $(SAMPROG) with no
> actions, and since there was a directory there with a tempting name
> the newer automake was trying to do something with it, but lacked
> the appropriate clues. Since that target is actually made as a side
> effect of making the library (sigh...), it seems to work to just
> give that rule something harmless to do (the Peter Principle
> triumphs again...).
>
> a bunch of the targets need a libtophat.a and libgc.a; the older
> automake was probably able to guess what to do given the list of
> sources but the newer automake apparently won't make the necessary
> assumptions. This patch wires up a simple rule and cleans up the
> appropriate dependencies so that things work.
>
> While it may appear that I'm someone who understands automake, keep in
> mind that I only play such a person on a TV reality show. YMMV.
* Remove extraneous when constraint
Fixes: #8258#8090 altered import behavior so that import spack no longer
provides access to many other Spack modules. This addresses
a case which depended on the prior behavior and was not
updated as part of #8090. This particular import error only
came up when users were setting compiler flags on specs.
See also: #8194
* kokkos: update kokkos spackage to add current host/gpu architectures.
* kokkos: simplify host/gpu architecture selection and make intent of arg population logic more clear.
* kokkos: mollify flake8.
* Make strings and logic more pythonic. Add error for GPU arch without CUDA.
* kokkos: simplify conflicts behaviour. make args population more pythonic.
* kokkos: move gpu values to list for use in variant and conflicts checking.
* kokkos: How fortunate the man with None.
* kokkos: fix gpu_arch conflict loop error. add conflict to guard against +cuda and 2.5.00:develop versions, until kokkos issue #1296 is resolved.
* kokkos: mollify flake8.
* kokkos: add descriptive message to version conflict with +cuda.
* update globalarrays package to current releases
* rename class to Globalarrays to match package folder
* rename bytes to int64, minor cleanup of descriptions
* int64 False by default
* remove reference to external as we aren't supplying paths, and use args.append()
* correcting args.append syntax
* add conflicts
* Fix bug; don't know why this ever worked in the first place.
* Add fix to netcdf-fortran too; flake8
* Simplify configure_args()
* netcdf-fortran: simplify that too
- there were some leftover spack.* names being used after we removed
globals and moved everything in the top-level namespace to spack.pkgkit
- point those references to their new homes
- remove most `import spack` statements, except for files that need
`spack_version`
- import spack is no longer sufficient to use submodules
(e.g. spack.directives).
- these submodules must be imported directly. Update references
accordingly.
- Spack packages were originally expected to call `from spack import *`
themselves, but it has become difficult to manage imports in the
Spack core.
- the top-level namespace polluted by package symbols, and it's not
possible to avoid circular dependencies and unnecessary module loads in
the core, given all the stuff the packages need.
- This makes the top-level `spack` package essentially empty, save for a
version tuple and a version string, and `from spack import *` is now
essentially a no-op.
- The common routines and directives that packages need are now in
`spack.pkgkit`, and the import system forces packages to automatically
include this so that old packages that call `from spack import *`
will continue to work without modification.
- Since `from spack import *` is no longer required, we could consider
removing ``from spack import *`` from packages in the future and
shifting to ``from spack.pkgkit import *``, but we can wait a while to
do this.
* Use libs[0] and headers.directories[0] instead of hardcoded paths in opencv dependencies
* Fix library finding in libjpeg-turbo
* Use build_type=Debug instead of +debug to determine build_type
- spack.util.lock behaves the same as llnl.util.lock, but Lock._lock and
Lock._unlock do nothing.
- can be disabled with a control variable.
- configuration options can enable/disable locking:
- `locks` option in spack configuration controls whether Spack will use filesystem locks or not.
- `-l` and `-L` command-line options can force-disable or force-enable locking.
- Spack will check for group- and world-writability before disabling
locks, and it will not allow a group- or world-writable instance to
have locks disabled.
- update documentation
Spack doesn't have a CSharp package, so the only time ANTLR would
succeed is if a system-installed CSharp was available. This disables
CSharp support, which enables building on systems without
system-installed CSharp.
* PR #4945 did not make this work on SuSE 11. Adding the latest version of libexpat did. Also update for Expat's move to GitHub.
* Update package.py
* Update package.py
Move to url_for_version()
* Changed netlib-lapack to fix an error in their CMake setup. Changed
netlib-lapack spackage file to add Fortran flags correctly for XL
compiler.
* Fixed line length -- forgot abt flake8...
* Fix the package version where the ibm patch applies.
- Spack core has long used llnl.util.filesystem.join_path, but
os.path.join is pretty much the same thing, and is more efficient.
- Use os.path.join in the core Spack code from now on.
- simplify the singleton pattern across the codebase
- reduce lines of code needed for crufty initialization
- reduce functions that need to mess with a global
- Singletons whose semantics changed:
- spack.store.store() -> spack.store
- spack.repo.path() -> spack.repo.path
- spack.config.config() -> spack.config.config
- spack.caches.fetch_cache() -> spack.caches.fetch_cache
- spack.caches.misc_cache() -> spack.caches.misc_cache
- `spack.cmd.all_commands` does a directory listing on
`lib/spack/spack/cmd`, regardless of whether it is needed
- make this lazy so that the directory listing won't happen unless it's
necessary.
- It turns out that jsonschema is one of the more expensive imports.
- move imports of jsonschema into functions to avoid the performance hits
for calls that don't need config.
- spack.store was previously initialized at the spack.store module level,
but this means the store has to be initialized on every spack call.
- this moves the state in spack.store to a singleton so that the store is
only initialized when needed.
- spack.repository module is now spack.repo
- `spack.repo` is now `spack.repo.path()` and loaded lazily
- Added `spack.repo.get()` and `spack.repo.all_package_names()` as
convenience functions to simplify the new lazy interface.
- updated tests and code
- no longer require `spack_version` to be a Version (it isn't used that
way anyway)
- use a simple tuple `spack_version_info` with major, minor, patch
versions
- generate `spack_version` from the tuple
- replace `spack.config.get_configuration()` with `spack.config.config()`
- replace `get_config`/`update_config` with `get`, `set`
- add a path syntax that can be used to refer to specific config options
without firt getting the entire configuration dict
- update usages of `get_config` and `update_config` to use `get` and `set`
- Current configuration code forces the config system to be initialized
at module scope, so configs are parsed on every Spack run, essentially
before anything else.
- We need more control over configuration init order, so move the config
scopes into a class and reduce global state in config.py
Fixes#8152
Flex 2.6.4 doesn't build with gcc@7:. This sets Flex 2.6.3 as the
preferred version. The flex package had a conflict directive for
this but originally the recorded conflict was specific to ubuntu;
this appears to be an issue on multiple operating systems so this
also updates the conflict to be more general and adds a comment
which links to the relevant issues.
Fixes#7946
Creates an openspeedshop-utils package that does not depend on qt3
and can be used to resolve the needs of cbtf-argonavis-gui.
Changes for creating a release versus develop build are also
included. There are package versions which are no longer relevant
(e.g. 1.3.0 for cbtf-argonavis-gui has been replaced with 1.3.0.0)
but these versions need to be kept to allow uninstalling them;
issue #8173 has been created to investigate this.
- After spack install, these test cases are built in a seperate step, outside of spack, but they need to patched after staging so users don't have to do it in a seperate step.
- The test cases are installed under `<petsc install directory>/share/petsc/examples` and can be built and executed by invoking `gmake -f gmakefile.test` in that directory.
* Added Shim to Support hdf5+mpi in Spack Silo
Added compatibility for hdf5+mpi to silo spackage to support spackages
with dependency chains that include silo and hf5+mpi
* Fixed fixable flake8 issue with silo spackage
Fixed missing whitespace around operator
* Setting MPI Variant to Default to True
As per guidance changing variant to default to MPI builds
* Create custom PATH that texlive requires in its module.
* Updating texlive to 2018 distribution.
* Using platform instead of sys for easier portability across linux and osx.
* Add package for "bumpversion"
bumpversion is a tool that automagically increments version numbers
for projects. It can update files and do other magic along the way.
* Remove py- package prefix and use pypi.io URL
Renamed the package 'bumpversion'.
Fetch the package from pypi.io.
* Add version info for bumpverion@0.5.3
Fixes#8088#7012 added a @when condition for a @run_before check to constrain
that check to only run on Darwin. @when is intended to be used to
choose one of several different implementations of a given function
and cannot be used to conditionally deactivate a check altogether.
This replaces the external decorator with a check that executes at
the beginning of the function.
* dealii: disable tpetra when +cuda; check for '' in cuda_arch
* trilinos: fix Kokkos CUDA
* Revert "trilinos: fix Kokkos CUDA"
This reverts commit 6aae894bc7d49c9e736c06f65b81b8c219443920.
* update according to @aprokop comments
* minor
* fix flake8
For Spectrum MPI, if -np is omitted, the default behavior is to
assign one process per process slot, where the default process slot
allocation is one per core. On systems with many cores, the number
of processes can exceed the size of the grid specified when the
testcase is run and the test case fails. This specifies a small
number of processes when running the test case to prevent failure.
* llvm+lldb plaform=darwin: check for lldb_codesign
Building LLVM with LLDB requires that the "lldb_codesign" code
certificate be created (see
https://llvm.org/svn/llvm-project/lldb/trunk/docs/code-signing.txt for
details). This commit checks for this certificate on Darwin if LLDB is
to be built, and returns an informative error message if this
certificate is unavailable.
Add variants to control building of many 3rd-party modules in opencv.
They are enabled by default, which matches the defaults chosen by
opencv when these options are not set explicitly. This also updates
the lbann opencv dependency to disable most of these new variants by
default in order to build faster.
Fixes#2781
This PR introduces a new attribute for packages called
`archive_files`, which designates files that should be saved from
a package build (e.g. the config.log generated during autotools
builds).
The attribute contains a list of glob expressions; Any file that
matches will be archived in the `<prefix>/.spack/archived-files`
directory. Errors that occur when archiving files are collected and
reported in a file named `<prefix>/.spack/archived-files/errors.txt`.
`AutotoolsPackage` and `CMakePackage` provide a sensible default
override for this attribute.
- Adds v1.3.0 and v1.4.0
- List explicit CMake version dependency (from Benchmark CMake files)
- Disable the building of tests. Starting in v1.4.0 this introduced a dependency on GoogleTest
Tested by building all versions on macOS High Sierra
fixes#7941
Modified string representation of Specs to add a space before deps
Unit-tests have been modified accordingly
Added a test for regression on #7941
* New Package: svn2git
https://github.com/nirvdrum/svn2git
svn2git is a tiny utility for migrating projects from Subversion to Git while
keeping the trunk, branches and tags where they should be. It uses git-svn to
clone an svn repository and does some clean-up to make sure branches and tags
are imported in a meaningful way, and that the code checked into master ends up
being what's currently in your svn trunk rather than whichever svn branch your
last commit was in.
* Actually requires +perl variant of subversion.
* fix style issue.
* Added magma package
* Incorporated Serban's change
* Remove deleted files
* argparse.py
* Sync forked Spack repo in github
* Use reported version of IBM XL Fortran compiler for compiler versions
>= 16.0.
Starting with the April 2018 release, the IBM XL C and Fortran
compilers report the same version, 16.0. Consequently, there is
no need to downgrade the Fortran compiler version to match that
of the C compiler.
* Eliminate blank lines.
* Update the Spectrum MPI package.py file to specify the correct compilation
scripts for the PGI compilers.
Fixes#7924
Spack's yaml schema validator was initializing all instances of
unspecified properties with the same object, so that updating the
property for one entry was updating it for others (e.g. updating
versions for one package would update it for other packages).
This updates the schema validator to instantiate defaults with
separate object instances and adds a test to confirm this behavior
(and also confirms #7924 without this change).
* mvapich2: add registration cache and file systems variants, avoid empty --with-pm= if no process manager is specified, add missing leading dash in --disable-silent-rules
* mpifileutils: roll back to dtcmp v1.0.3 due to API change, add option to build master branch
* add comment to explain why certain dtcmp is required, rename version for git branch
* dtcmp v1.0.2 is missing github tarball
When building eospac with %clang, spack's Fortran compiler wrapper is called
'flang', which is rejected by the eospac build system..Allow compilation with
%clang by patching the eospac build system.
My testing only covers %clang where flang is actually gfortran. However, this
patch only registers 'flang' as a known Fortran compiler and does not set any
flags or options that are compiler specific.
.
Add configurable HTTP, Python and readline support.
Add previously missing dependencies.
Minor patch to improve Python support (see https://github.com/xrootd/xrootd/pull/695).
* Use reported version of IBM XL Fortran compiler for compiler versions >= 16.0.
Starting with the April 2018 release, the IBM XL C and Fortran compilers report the same version, 16.0. Consequently, there is no need to downgrade the Fortran compiler version to match that of the C compiler.
* Use GitLab's API endpoint for fetching a git snapshot.
* More GitLab packages use the API.
* find_list_url for GitLab's API URLs.
* Flake8
* Url for 'hacckernels'.
* Check GitLab API regexps before the non-API ones.
* Add Trinity package
New package to install Trinity - another odd installer (assumes installation to /usr/local/bin and doesn't copy dot files hence deleting them)
* Update package.py
added dependencies from https://github.com/trinityrnaseq/trinityrnaseq/wiki/Installing-Trinity
* Update package.py
Copy tree to prefix.bin so that PATH can correctly be set for the Trinity command - can't split it off as it seems everything needs to stay together.
* Update package.py
Fixed over-length lines and space before comma
This commit fixes#7900.
In short: in the absence of constraints to the contrary, spack will
download the preferred version of a package, which is usually the most
recent non-develop version. On a clean spack installation without
hypre installed, `spack install mfem` will download hypre 2.14.0,
which is API-incompatible with mfem 3.3.2 (and earlier versions of
mfem), so spack will throw compiler errors re: the API mismatch.
This commit fixes that error by restricting mfem release versions
3.3.2 and earlier to hypre version 2.10.b to 2.13.999 (only 2.13.0
currently exists, but if it has a non-2.14.x maintenance release,
these versions of hypre should also be permitted).
* automake: added version 1.16.1
* lmod: added version 7.7.29
* git: added version 2.17.0
* tmux: added version 2.7
* scala: added version 2.12.5, fixed typo in version 2.10.6
* Added spark@2.3.0, hadoop@3.1.0, jdk@8u172-b11
* picard: added version 2.18.3
* tar: added version 1.30
Added OpenMP variant for kokkos that can be disabled to build and test
on platforms where OpenMP is not available or desired. Also updated
version list
* Committer: sajid <sajid@xrmlite.phys.northwestern.edu>
modified: var/spack/repos/builtin/packages/py-numpy/package.py
* Corrected typo
* removing whitespace
* remove whitespace
* remove whitespace, if not will do this in vim
* fabtests: new package
* fabtests: fix hashes for github archive behavior & add libfabric dep
* fabtests: both versions confirmed working
* fabtests: use developer provided archive
* Fix build issue with PMIx on aarch64
Work around a bug in PMIx that causes build issues on aarch64. Bug was
fixed in PMIx 2.1.1, so the spack package checks both the architecture
and the PMIx version. Workaround is to just not build the PMIx test and
example code.
* Minor code cleanups
Minor corrections to make 'spack flake8' happy.
* Add support for version 2.1.1
* r-matr: Create new package.
* shocklibs: Create new package.
* r-biom-utils: Create new package.
* r-mgraster: Create new package
* py-poster: Create new package.
* py-requests-toolbelt: Create new package.
* perl-libwww-perl: Create new package.
* py-mg-rast-tools: Create new package.
* mg-rast-tools: Fix version constraints for dependencies.
* Change dependency type of py-setuptools.
* openPMD-api
Add a new package for openPMD-api.
openPMD is a meta-data standard (markup, schema) for hierarchical,
scientific particle and mesh based data.
openPMD-api is a library wrapping the underlying backends (HDF5,
ADIOS, ...) into a single, object-aware API.
* Repository: https://github.com/openPMD/openPMD-api
* Project: https://github.com/openPMDhttp://www.openPMD.org
* mpi variant forwarding: fix concretization
+ I am unable to build suite-sparse@5.2.0 with clang@6.0.0. The build fails
with the error:
```
2397 [ 99%] Linking C executable wildtype_demo
2398 make[4]: Leaving directory
`/tmp/kellyt/spack-stage/spack-stage-w0XYhK/SuiteSparse/GraphBLAS/build'
>> 2399 libgraphblas.so.2.0.1: undefined reference to `__fpclassify'
>> 2400 clang-6.0: error: linker command failed with exit code 1 (use -v to
see invocation)
>> 2401 make[4]: *** [wildtype_demo] Error 1
```
+ This error appears because libgraphblas.so requires `-lm`, but that link
dependency is not explicitly listed in GraphBLAS\CMakeLists.txt.
+ A patch file is provided to add this dependency. The patch is only applied
when the compiler is `%clang`.
* ADIOS 1.13.1
ADIOS 1.13.1 fixes several issues with zero-sized block
reading and writing when compression transports are used.
* SZ 1.4.12.3
Add the lastest SZ release used in ADIOS1.
* geopm: Added initial files for package geopm
and py-natsort dependency
* geopm: Updated variants for hwloc and mpi defaults
* geopm: Cleaned up unwanted hyphenation in description.
* geopm: updates according to change request.
* geopm: Updated minor change requests.
Clarification for ruby-ronn dependency included as a TODO.
This requires changes in the current configure and build process
which will be changed in a future version.
* geopm: changed to AutotoolsPackage and fixed hwloc
The changes for hwloc required to add a specific version bc of a
conflicting version of openmpi's hwloc dependency being below
1.99. This works but updates to this package should check if this
restriction is still required. (Thus TODO added)
* geopm: Added link dependencies (thanks for pointing this out).
Also removed empty line below class definition.
* geopm: Fixed depends_on according to package guide.
Removed wrong type 'run' and removed types for numactl, mpi and hwloc
to use default build & link type.
* geopm: Fixed and updated +doc dependencies
new pacakge ruby-ronn (package specified as in ruby-narray package)
Updated second +doc dependency doxygen.
* geopm: Changed args for Autotools specific feature.
Added enable_or_disable and with_or_without instead of if cascade.
Added overhead variant.
Also added missing dependency (default: build&link) for json-c).
* glvis@3.3: new package
* mesa-glu: now provides 'glu' (GL Utility library)
* openglu: add package, provides glu
* [GLVis] Use MakefilePackage as base. Add 'develop' version.
Small tweaks in the makefile arguments.
* [GLVis] Locate the X11, GL and GLU libraries within their respective
prefixes. Set the C compiler in the makefile arguments.
* [GLVis] Fix a bug: the same variable was used for two different things.
* [libx11] Add the libs property
* [OpenGL] Add the libs property
* [OpenGLU] Add the libs property
* [GLVis] Use the libs properties of libx11, opengl, and openglu
Activating a package that is already activated now sends a `tty.msg`
and returns.
```
-bash-4.2$ ~/spack/bin/spack activate aspell6-en
==> Package aspell6-en/lc4v24f is already activated.
```
* Better error message for spack providers
fixes#1355
`spack providers` now outputs a sensible error message if non-virtual
specs are provided as arguments:
```
$ spack providers mpi zlib petsc
==> Error: non-virtual specs cannot be part of the query [zlib, petsc]
```
Formatting of the output changed slightly.
* Calling 'spack providers' without arguments print the virtual pkg list
Also, the error message in case of a wrong parameter has been improved
to show the list of valid packages.
* Avoid printing headers if stdout is not a tty
* The provider list is formatted with colify if not in a tty
* Added a test to check the list of providers returned from the command
Popping the when spec from kwargs in the extends directive breaks
class inheritance. Inheriting classes do not find their when spec.
We now get the when spec from kwargs instead, leaving it to be found
by any downstream package classes.
* opengl: make error throw on install, not fetch
* opengl: add configuration advice for darwin
* opengl: add more versions of GL/OpenGL
* opengl: improve documentation in comments
* [OpenGL] Fix formatting
fixes#7593
Unit tests on OSX are trying to concretize mpileaks, and they fail due
to a conflict in the package:
"%gcc@7.2.0:" conflicts with "elfutils@0.163"
This solves the issue asking explicitly to concretize against
elfutils@1.170
fixes#7705
Package.provides now checks constraints to ensure that a spec provides
a given virtual package. Note that 'strict=True' is not passed to
satisfies as this function is also used during concretization.
* new package "DSRC" (bioinformatics)
* remove uncommented line
* get rid of unnecessary import
* convert to MakefilePackage class
* get rid of make() in install phase, remove spaces
* make flake8 happy!
Add a proper install variant for Catch(2).
This allows to use Catch2 downstram with `pkg-config` and
CMake's `find_package` commands via standard environment hints.
Fixes#7548
This updates the "spack view" command to use the same parsing logic
as "spack install" on the user-provided specs. For example you can
provide a DAG hash to refer to an exact installed spec instead of
specifying name, compiler, etc.
* fix boost and muparser for Clang 9.1.0 with C++17
* muparser: add C++11 flags
* dealii: temporary disable python by default
* dealii: make CMake find right python
* Current versions of kokkos conflict with hwloc 2.x
Current versions of kokkos cannot build with hwloc v2.x due to changes
in the associated structs. Requiring v1.11.9 for the time being
until it becomes clear what the kokkos model will be
* Explicitly commenting fix for posterity
Explicitly commented version number of hwloc dependency for kokkos for
posterity purposes and to pass code coverage for speedy merging
* Generalized Kokkos's Hwloc Dependency
Generalized dependency to v1.x of hwloc
* Fixed Flake8 issue
Fixed comment to conform to flake8
* Fixed ordering for Version Requirements
Fixed ordering to indicate <= 1.11.9, not >1
* Generalized kokkos dependency
* explicitly specify user-config.jam as it isn't found
via the environment variable on older boost versions
(cf. what debian/rules does)
=> fixes +mpi build
* fix "error: Ambiguous key" error on :1.58.999 boost
(cf. https://github.com/boostorg/boost/blob/boost-1.59.0/bootstrap.sh#L357)
* ceed bundle package including OCCA, MAGMA, Nek5000, mfem, and PETSc
Very preliminary, a lot of ad hoc decisions, fragile, missing functionality
All packages do build on my MacOS laptop with Cuda
Funded-by: ECP
Project: CEED
Time: 4 hours
Reported-by: Tzanio Kolev <kolev1@llnl.gov>
* Some adjustments in a few CEED-related packages.
This is still very preliminary, but with these changes I'm able to build from
scratch on a Mac laptop and a Linux desktop (RHEL7).
Note that there doesn't seem to be a good way to support CUDA in Spack, so
you'll have to install that manually yourself. You will also need a Fortran
compiler, in particular on a Mac where you also have to make sure that it
is being picked up in ~/.spack/darwin/compilers.yaml.
* ceed bundle package including OCCA, MAGMA, Nek5000, mfem, and PETSc
Very preliminary, a lot of ad hoc decisions, fragile, missing functionality
All packages do build on my MacOS laptop with Cuda
Funded-by: ECP
Project: CEED
Time: 4 hours
Reported-by: Tzanio Kolev <kolev1@llnl.gov>
* Reverting Mac-related CUDA checks in the OCCA and MAGMA packages.
A much better solution is to install CUDA manually and add it to
~/.spack/darwin/packages.yaml:
packages:
cuda:
paths:
cuda@8.0.54: /usr/local/cuda
* Fix flake8 warnings
Funded-by: CEED
Project: CEED/Spack
* remove unneeded commented out code
Funded-by: CEED
Project: CEED/Spack
* Don't have PETSc use MUMPS to speedup install time, try parallel installs for Magma
Funded-by: CEED
Project: CEED/Spack
* Added libceed package and made changes to get the ceed metapackage up and running.
* Get the ceed package up and running and added installing the .h file in the libceed package.
* All packages are now pointed at specific versions (either tagged or git hashed.
* Changed some of the specific versions in the CEED metapackage.
* Some version updates.
* Applied Cameron Smith's patch for the latest pumi.
* Merge
* Nek5000: Use tarball instead of git tag
* Major update of the mfem package - needs more testing. Add a testing
shell script for mfem, test_builds.sh, in the mfem directory. It builds
a number of mfem configurations to ensure they all work.
Add 'suite-sparse' variant to petsc.
Add 'develop' version of libceed.
Add 'metis' variant of laghos.
Update the dependencies of ceed.
* Changes in petsc's handling of zlib to make the latest build work on Mac.
* [MFEM] When using '+libunwind', mfem needs '-ldl'.
* [OCCA] Remove debug print statements.
* [PETSc] Replace the check ('zlib' in spec) with ('+zlib' in spec) - the
latter checks if 'zlib' variant of petsc is enabled which is incorrect;
the former checks if petsc is configured to depend on 'zlib'.
* [MFEM] Add conduit variant.
* [libCEED] Enable testing using 'make prove'.
* [MFEM] Tweak test_builds.sh - all builds pass on fedora 27.
* Add a post install test
* Add a post installation check for NekCEM
* pumi: bump git hash, remove unused patch
* Fixed testing of Laghos package
* Adding variants in CEED package
* Empty file for URL didn't work on LLNL/LC. Replacing with bundle-package.txt -> README.md.
* [MFEM] If the spec defines the 'cxxflags' compiler flags, use that value
to set the CXXFLAGS value in the mfem build system.
* Disabling HDF5 by default.
You can still build it with "spack install ceed+hdf5".
* libceed: fix use of prefix versus DESTDIR
* Added cuda variant for libCEED (off by default)
* [libCEED] When building v0.1, fix ceed.pc before installing it.
* [CEED] Add variants for OCCA and PUMI. Replace the hdf5 variant with
a variant called quickbuild that disables variants in packages to
speedup the build - currently hdf5, boost, and superlu-dist in PETSc.
* [libCEED] Remove unused 'import os'.
* [MAGMA] Add version 2.3.0.
* [OCCA] Major update of occa/package.py.
* [libCEED] Several improvements.
* [CEED] Use fixed versions for occa and libceed.
* [OCCA] Fix a bug in CUDA setup.
* NekCEM doesn't need depends_on('python')
* [OCCA,libCEED,MFEM,CEED] Update occa and libceed versions plus a few
tweaks in occa and mfem.
* Switching to gslib-1.0.2 to fix summitdev bug.
* [Nek5000] Use the Fortran 77 compiler instead of the 'fc' compiler.
Propagate any 'fflags' and 'cflags' specifications to the Nek tools.
* [MFEM] Tweak for 'ppc64le'.
* attempt to override fips compatibility error for md5
* Compiler option changes in libCEED to fix summitdev build.
* Disable the nekcem install test until it can be fixed.
* [NekCEM] Major update of the package to fix various issues.
* [NekCEM] Fix formatting
* zoltan: remove non-portable -g0 flag
the pgi compiler does not accept it
* pumi: specify dependency on cmake > 3
* [Nek5000] Add dependency on libx11 when needed. Various hacks to
support IBM XL compilers.
* [NekCEM] Link with the pthreads library - it is required.
* [MAGMA] Add a patch for magma v2.3.0 when building with gcc <= 4.8.
Add dependency on BLAS.
* [CEED] Update to magma 2.3.0. When building with gcc < 4.9 constrain
the used suite-sparse version to <= 5.1.0 - starting with v5.2.0,
suite-sparse requires gcc >= 4.9.
* Small updates
* [libceed, gslib] Fix style
* [Nek5000] Need 'libxt' as a dependency as well.
* [MUMPS] Temporary workaround for ray.
* Updating occa and libceed to laters pre-1.0.0 and pre-0.2 hashes
* petsc: add pkg-config patch for variable quoting
Variable definitions in pkg-config are processed more like make than a
shell, so don't need quoting. Older versions of pkg-config (<= 0.28)
did not remove the quotes when printing the value.
Freedesktop #67904 (https://bugs.freedesktop.org/show_bug.cgi?id=67904)
Reported-by: Tzanio Kolev <kolev1@llnl.gov>
* petsc: add 3.8.4 and myself as co-maintainer
* Updating libceed to latest pre-0.2 hash.
* [PETSc] Add the 'headers' and 'libs' properties.
* [MFEM] Explicitly add rpaths to link flags for external packages.
This is necessary when MFEM's exported options (in config.mk) are
used outside of Spack for linking against MFEM.
* Unset MFEM_DIR before building MFEM
* [PETSc] As observed by @jedbrown, there is no need to define the 'libs'
property explicitly - the default handler works fine too.
* Build MAGMA shared by default, install a few additional headers
* Forgot a self
* Switching to OCCA tag v1.0.0-alpha.4.
Adding urls for OCCA and libCEED ('spack uninstall' complains otherwise).
* Removed FIXMEs for OCCA tag
* [occa] Cleaned up tagged versions
* [libceed] Updated occa to v1.0.0-alpha.5
* [ceed] Updated occa to v1.0.0-alpha.5
* [libCEED] Always define the 'NDEBUG' makefile option based on the
'debug' variant setting - this should always work regardless of the
default setting inside the libceed makefile.
* [MUMPS] Revert a temporary workaround.
* [CEED, libCEED] Minor tweaks.
* libCEED v0.2 release
* [CEED] Use version 0.2 of libCEED.
* [HPGMG] Remove duplicate version.
* [CEED] Update the hpgmg version
* hpgmg: use tarball for 0.3
The Git repository contains somewhat heavy documentation so the tarballs
are much faster.
* hpgmg: +fe was default for the 0.3 release
* hpgmg: explicitly name build directory to avoid use of ambient PETSC_ARCH
* [HPGMG] Fix flake8 formatting
* libceed@0.2: work around occaFree issue
* Add ignore_conflicts to package activation
includes:
* with_dependency flag for do_activate
* ignore_conflicts flag for do_activate
* start of cleanup of activate cmd
* ensure that apr is picked up
* Update package.py
* add apr explicitly as dependency
* remove unnecessary newlines
* Update package.py
fixing dependencies
Fixes#7593
By default MacOS concretizes using the clang compiler. The unit tests
include a call to "spack spec mpileaks", which has elfutils as a
dependency; #7096 added a conflict in elfutils to avoid building
with clang, which lead to the MacOS unit tests to start failing.
This updates the concretization to force using gcc when concretizing
mpileaks.
Modifications:
* Added zlib dependency, starting from version 3.0.0
* Added memchecker support for debugging
* Remove mpirun and similar links if slurm is selected as a scheduler
This fixes a check that decides when to skip buildcache relocation.
Originally the check was flawed in two ways: it would skip if the
source prefix matched the destination prefix, which no longer
matters since the source prefix is replaced with a placeholder
(so it always needs to be updated); it also would skip relocation
if the rpaths were not relative, when in fact it should be the
opposite (binaries without relative rpaths *should* be relocated,
and those without don't need it).
- FastPackageChecker was being called at startup every time Spack runs,
which takes a long time on networked filesystems. Startup was taking
5-7 seconds due to this call.
- The checker was intended to avaoid importing all packages (which is
really expensive) when all it needs is to stat them. So it's only
"fast" for parts of the code that *need* it.
- This commit makes repositories instantiate the checker lazily, so it's
only constructed when needed.
- This was needed when we transitioned to all lowercase packages because
git didn't handle case changes well on case-insensitive filesystems.
- Now it just adds extra stat calls to startup, and we check for
all-lowercase package names in tests, so we'll remove it.
- people using really old versions of Spack can re-clone.
* Create unload_module method
Extract code from load_module into unload_module.
* Unload modules to create a clean env on Cray
removes cray-libsci, cray-mpich and darshan to prevent any silent
linking with those packages.
* Add format to separate target and os for path
spec format can now handle separations of target and os for setting
up the path.
* Added ${PLATFORM} et al to spec.format()
${PLATFORM}, ${OS}, ${TARGET}
* Update tests
Updated tests and got rid of unnecessary code.
* Also update documentation to reflect this new ability.
* Add default path scheme to config.yaml
Added default path scheme to config.yaml. Users can overwrite this
section if they want.
* Speedup the default 'libs' property search - important for external
packages.
* As advised by @alalazo, use tuples instead of lists inside
_libs_default_handler.
* [hypre] Add the 'headers' property plus a small tweak in the 'libs'
property.
* [hypre] Add fallbacks for searching for '.libs' in 'prefix.lib64' and
in all of 'prefix'.
* [hypre] Fix style.
* [hypre] Use find_headers instead of find + HeaderList.
* added a new package file for PHIST (Pipelined, Hybrid-parallel Iterative Solver Toolkit)
* packages/phist: fix some formatting errors (too long lines etc)
and set allowed values for option "outlev"
* packages/phist: fix some more flake8 errors
* add headers property to netlib-lapack and intel-mkl
* ghost: fix finding cblas header and libs (at least for mkl and netlib-lapack, which provide headers())
* added a new package file for PHIST (Pipelined, Hybrid-parallel Iterative Solver Toolkit)
* packages/phist: fix some formatting errors (too long lines etc)
and set allowed values for option "outlev"
* packages/phist: fix some more flake8 errors
* phist: use headers property to find lapacke include dir
* phist: fix flake8 error
* phist: be more specific about dependencies and update version
* phist: state trilinos minimum version when building with Tpetra kernels
* gmp: add oldish version 5.1.3 with checksum
* phist: get rid of the 'petsc+complex' variant, it can be installed by kernel_lib=petsc ^petsc+complex
* phist: new version
* phist: fix version checksum
* Revert "add headers property to netlib-lapack and intel-mkl"
This reverts commit d8cb62905e.
* Revert "ghost: fix finding cblas header and libs (at least for mkl and netlib-lapack, which provide headers())"
This reverts commit d406a77ed7.
* phist: remove unused function
* phist: reorder package file versions/variants/depends_on/rest
* phist: some minor changes requested by @adamjsteward
* Add latest version of MPFR and patches for older versions
* Fetch and checksum patches instead of storing them
* Fix bug, version attribute was being overridden
* Added installation date and time to the database
Information on the date and time of installation of a spec is recorded
into the database. The information is retained on reindexing.
* Expose the possibility to query for installation date
The DB can now be queried for specs that have been installed in a given
time window. This query possibility is exposed to command line via two
new options of the `find` command.
* Extended docstring for Database._add
* Use timestamps since the epoch instead of formatted date in the DB
* Allow 'pretty date' formats from command line
* Substituted kwargs with explicit arguments
* Simplified regex for pretty date strings. Added unit tests.
* meme: added perl-xml-parser dependency, resolved issue. still need to fix ghostscript or image-magick, conversion not working.
* meme: fixed issue with libxslt/libxml2 conflict, and added imagemagick support
* meme: removing libxslt/libxml2 dependency lines
* meme: configure args
* meme: separating configure args
* meme: removing a comment, and hoping for travis to rebuild. site seems back up.
* meme: make image-magick optional
* meme: add comment referencing PR
* meme: bah flake8
This updates architecture concretization to
* Search for the nearest parent in the DAG for architecture information
rather than defaulting to the root of the DAG
* Propagate architecture settings transitively, such that if for
example the target is set at the root of the dag it will set the
same target on indirect dependencies (assuming no intermediate
dependency specifies a separate target). Previously this occurred
in general but under some conditions did not, for example if an
intermediate dependency specified some subset of architecture
properties.
* Promote to Autotools for simplicity; broaden Intel CPU opt targeting from SSE4.2 to AVX2
* make Intel CPU opt targeting same as in ../libxc
* flake8 W291
* use canonical means to pass (ahem) the test phase, h/t @adamjstewart
* revert f25d598 (unrelated merge)
* re-merge
* Create mirror for system with different compilers
Spack concretizes the spec provided by the user in
"spack mirror create" to ensure downloading the right
dependencies. Under normal circumstances concretization
requires that the chosen compiler exists on the system,
but this is not required when creating download mirrors
for other systems, so this requirement is removed in that
case.
* Add test for disabling compiler existence check
* Update compiler existence checking logic
* improve test for disabling compiler existence check
* make py-setuptools a run-time-only dep for py-basemap and patch python package to only apply setuptools flag for build deps
* py-qtconsole does not require setuptools
* This allows Spack to work with MD5 hashes on machines with openssl in FIPS mode.
* We are still using MD5 for validation in many places, and a later PR will replace all uses of MD5 with SHA256.
* This is a quick fix until that happens.
* intel-xed: new package
The Intel X86 Encoder Decoder library for encoding and decoding
x86 machine instructions.
* intel-xed: style fixes for flake8.
* intel-xed: better style fix
* intel-xed: update to version 2018.02.14, add PYTHONPATH for
building with python 3, cleanup the opt flag.
The restrict compilation option with icc causes deadlock when multithreading is used. This issue has already been reported to the Scotch development team, but for current versions it is more reliable not to use the restrict compilation option.
* libtool: actually do symlinking correctly on darwin
Version 4 of the libtool/darwin debacle:
AutotoolsPackage hardcodes libtool & libtoolize into the autoreconf
stage, so the commands `libtool` and `libtoolize` MUST be present, and
shimming in glibtoolize into AutotoolsPackage when sys.platfrom ==
'darwin' does not work.
join_path(spec['libtool'].prefix.bin, 'libtool') still shadows system
BSD libtool (in apple-cctools, see PR #7177), but this shadowing could
be okay, depending on the combination of dependent specs.
* librom: new package
* Adding unzip package.
* Removing bzip2 dependency in unzip and raising an error when on Cray.
* Fixing error.
* Using conflicts statement for Cray detection in unzip package.
* add headers property to netlib-lapack and intel-mkl
* ghost: fix finding cblas header and libs (at least for mkl and netlib-lapack, which provide headers())
* fix flake8 errors
* ghost: remove unnecessary query parameter
* fix flake8 errors
* ghost: make it a CudaPackage (as suggested by @davydden, thanks!)
* ghost: missing whitespace
* Fix a bug when checking for 'xl' or 'xl_r' compiler.
* Add support for parallel build - the 's', 'c', 'd', and 'z' targets
are build separately allowing parallel builds.
* When build '+shared', inject all dependencies into the link lines of
the mumps libraries.
* Run the examples only when installing with the '--test' Spack option.
It seems there has been a change in concretization that resulted in
a request for a llvm latest (5.0.1) while not satisfying the later
more precise request for a version.
Remove the first line and only rely on the last.
A build of python@3.X had the following in the logs:
```
The necessary bits to build these optional modules were not found:
_dbm _gdbm _tkinter
```
As Tkinter is already a variant, we adopt the same strategy for dbm.
This includes:
* Added latest available versions (1.11.9, 2.0.0)
* Added variants to enable graphical output from lstopo
* Enabled build of the bundled netloc package if @2.0.0:
* Added dependency from numactl if @:1.11.9
* Added a constraint to avoid using hwloc@2.0.0 with openmpi
cube's configure only recognizes qt versions with single digit version
components and breaks with qt 5.10. Patch the configure files directly
to avoid having to regenerate them.
Mscgen is a small program that parses Message Sequence Chart descriptions and
produces PNG, SVG, EPS or server side image maps (ismaps) as the output.
See http://www.mcternan.me.uk/mscgen/
- transitive dependencies were not being handled correctly
- restructure code to do recursion and mark visited packages properly
- add `-V` option to *not* expand virtuals in spack dependencies
This re-adds the option to create and install unsigned tarballs, now
with the -u option (--unsigned) rather than the -y option.
This also changes the "keys" command, replacing the -y/--yes-to-all
option with the -t/--trust option (which has the same effect but is
more-clearly named).
Fixes#7130
shutil.move expects a source path like "/x/y/" to be a directory and
fails if "/x/y" is a symlink. This invokes realpath on the source
path to avoid the issue.
Fixes#7356
In some cases OperatingSystem (e.g. LinuxDistro) was getting
instantiated with a version that contains dashes. This breaks because
the concretizer later converts this value to a string and re-parses
it, and the '-' character is used to separate architecture components.
This adds a guard in the initializer to convert '-' to '_'.
This is needed when using system lua on RHEL and clones where lua
is installed in /usr/lib64. Note that the path entry for lua in
packages.yaml should specify the major/minor version of lua (since
this is used to calculate the lua lib directory)
updated hoomd-blue to latest tagged release in repo. This version also supports newer gcc6 compilers, so added constraint for older version to avoid breaking existing installs.
+ Newer versions of zoltan2 are doing test compiles that link to parmetis. If
these checks don't find `metis.h`, the check fails.
+ This small tweak ensures that the zoltan2 configure step can find `metis.h`
that is provided by spack.
Fixes#7237Fixes#6404Fixes#6418Fixes#6369
Identify when binary relocation fails to remove all instances of the
source prefix (and report an error to the user unless they specify
-a to allow the old root to appear). This check occurs at two stages:
during "bincache create" all instances of the root are replaced with
a special placeholder string ("@@@@..."), and a failure occurs if the
root is detected at this point; when the binary package is extracted
there is a second check. This addresses #7237 and #6418.
This is intended to be compatible with previously-created binary
packages.
This also adds:
* Better error messages for "spack install --use-cache" (#6404)
* Faster relocation on Mac OS (using a single call to
install_name_tool for all files rather than a call for each file)
* Clean up when "buildcache create" fails (addresses #6369)
* Explicit error message when the spack instance extracting the binary
package uses a different install layout than the spack instance that
created the binary package (since this is currently not supported)
* Remove the option to create unsigned binary packages with -y
For hdf5 `1.8.10` and some compilers, the hdf5 build system adds the
'-ansi' flag. hdf5 1.8.10 is not actually ansi compliant so this
adds a patch to modify the places where it violates ansi
compatibility (which ends up being a few comments).
Make neuron package more generic
* find the bin directory dynamically for use in run_env and spack_env
* replace filter_compilers after install with filter_compiler_wrappers
* update checksum for _current_ 7.5 version
* make +python conflict with ~shared
* prepend the architecture specific lib directory to the LD_LIBRARY_PATH
Signed-off-by: Ricardo Silva <ricardo.silva@epfl.ch>
Fix installs as "root" user for the last libSplash release.
Missing quoting of string compares resulted in a CMake error
when run as root (e.g. in a Docker build).
This updates Cray.setup_platform_environment to use cray-specific
pkgconfig paths so that all providers of 'pkgconfig' have access
to them (previously the 'pkg-config' provider had this but the
'pkgconf' provider did not).
* [OpenMPI] Add the 'headers' property. This removes some redundant
headers from sub-directories, returned by the default '.headers'
handler.
* [OpenMPI] In the .headers property, add a fallback to search all of
prefix for the mpi.h header. Recommended by @junghans to support some
external configurations, see #7268.
* [SPACK/spec.py] When a query through ForwardQueryToPackage returns
'None', treat that as query failure and raise RuntimeError with
suitable message. This overrides the current behavior to raise an
AttributeError which is now triggered only when no suitable query
property is found and there is no default handler.
* [spack/spec.py] Fix style.
* [SPACK/spec.py] In case of query failure, i.e. property returning
'None', raise AttributeError instead of RuntimeError in order to
pass the unit test. Also, small update in the logic distinguishing
query failure and lack of relevant property/attribute handling.
* add verilator package
* try to add environment variable and include folder
* manually install existing scripts and patch out spack's CXX wrapper
* added some comments
* make flake8 happy
Adds Google's new cpu_features library.
Install was added recently, so don't take the `0.1.0` release but
the current `master` until a new version is published.
* Combined llvm and llvm-lld: removed the separate llvm-lld package
and added llvm-lld as an optional add on to the llvm package (the
way several other llvm tools are maintained e.g. lldb)
* Added more lld hashes to llvm package
* The 'CppBackend' target doesn't exist for version 3.9.0 or later
so exclude it for later versions
* Was incorrectly specifying 'sparc' as a target for the 'sparc'
architecture - needed to specify 'Sparc'
* Fix issue #7248 building llvm: don't make the LLVMDemangle target
for llvm < 4.0.0
This updates the fix_darwin_install_name function to use the Spack
Executable object to run install_name_tool, which ensures that
process output is formatted as a 'str' for python2 and python3.
Originally fix_darwin_install_name was invoking subprocess.Popen
directly.
* No longer need to patch CMake. Using CMAKE_CXX_FLAGS to include BLAS/LAPACK/MKL header files properly.
* Backport two patches to QMCPACK versions prior to 3.3.0: https://github.com/QMCPACK/qmcpack/pull/621, https://github.com/QMCPACK/qmcpack/pull/623
* Insufficient to include lapack_dir in CMAKE_CXX_FLAGS, CMAKE_C_FLAGS is also needed.
* Make patched Espresso a default variant instead of a hard dependency. This is easier on computer architectures where Espresso is hard to build.
* Fix patching with QE variant. Now using correct form of patch command.
* Add QMCPACK v3.4.0
* Backport patch to QMCPACK versions prior to 3.3.0: https://github.com/QMCPACK/qmcpack/pull/643
Yaml-cpp hasn't tagged a release in over 2 years, yet they have fixed
many things and now have the ability to use find_package in CMake.
Since we use this feature in Nalu, we require the latest yaml-cpp
until they tag a new release. I have also deprecated the now
unnecessary ENABLE_INSTALL flag.
* Added a package for the MDAnalysis toolkit.
* Adding a package for the Python virtualenvwrapper pacakge. Included
dependent virtualenv-clone and stevedore packages. Updated py-pbr with
a new version to support this.
* Added extra necessary dependencies and fixed the file download paths.
* Added a package for the MDAnalysis toolkit.
* Added a new package for pygpu, which is the python front end of the
libgpuarray. Additionally changed the dependency for theano so that
it properly picks up the pygpu library.
* Added the libcheck package. Fixed flake8 errors in py-pygpu recipe
and removed spurious dependency on cudnn for libgpuarray.
The edit phase of cbench was used to set environment variables. Though
this works, the variables set this way are not dumped to build.env.
This commit fixes the issue by setting those variables in the
`setup_environment` function.
* [SUNDIALS] Add the 'headers' and 'libs' properties, plus a couple of
small tweaks.
* [SUNDIALS] Revert incorrectly changed conflicts() directive. Fix style.
Fixes#5189
When working with non-normalized paths containing ".." on some
file systems, Spack was found to encounter a permission error when
writing to the path. This normalizes a path written by the
intel-parallel-studio package and also normalizes all paths
written by the license install hook (for all packages) to avoid
this issue for intel-parallel-studio.
Adds the first stable API release of ADIOS2, v.2.1.0.
Also adds a missing `pybind11` CMake flag for proper Python
selection (needed in both the new and previous version).
To accommodate build systems on Darwin which look for glibtool/ize,
#7060 attempted to provide glibtool/ize symlinks for libtool/ize
but accidentally created the same symlink twice. #7073 reverted this
and replaced libtool/ize with glibtool/ize on Darwin. This led to
#7163 since AutotoolsPackages depend on the presence of the 'libtool'
binary, so this returns to the approach of #7160 but makes the
correct symlinks.
XBraid installs `libbraid.a`, but the default `libs` method
implementation will search for `libxbraid`, which does not exist. This
commit fixes the behavior of the `libs` method for the `xbraid`
package by overriding package's `libs` method.
Prior to this commit, running `spack module refresh` with `gdbm`
installed throws the warning:
```console
==> Warning: Could not write module file [/Users/oxberry1/spack/share/spack/lmod/darwin-sierra-x86_64/Core/gdbm/1.14.1.lua]
==> Warning: --> global name 'spack_cc' is not defined <--
```
This warning is repeated for every `gdbm` installation in Spack.
The cause of the issue seems to be the same as in spack/spack (#6089):
using `spack_cc`, `spack_cxx`, or similar inside `configure_args`
causes the warnings to throw when module files are generated.
Moving the use of `spack_cc` into `setup_environment` seems to fix the
problem, even though it's a bit of a kludge.
Running `spack module refresh -m lmod --delete-tree -y` throws the
following warning:
```console
==> Warning: Could not write module file [/Users/oxberry1/spack/share/spack/lmod/darwin-sierra-x86_64/openblas/0.2.20-ibhtl5q/clang/5.0.1/ipopt/3.12.9.lua]
==> Warning: --> global name 'prefix' is not defined <--
```
This warning arises because the variable expansions in
`spack.modules.BaseFileWriter` occur at module/package scope. Even
though `prefix` is an argument to `install` methods in e.g.,
`MakefilePackage`, so using this argument within package methods is
legal, doing so defeats the introspection logic in
`spack.modules.BaseFileWriter`. Replacing `prefix` with `self.prefix`
resolves the problem, because the introspection logic in
`spack.modules.BaseFileWriter` can use introspection to query IPOPT's
prefix.
Following the discussion with Todd and Adam, find has been modified to
accept glob expressions. This should not affect performance as every
glob implementation I inspected has 3 cases (no wildcard, wildcard but
no directories involved, wildcard and directories involved) and uses
fnmatch underneath.
Mixins have been changed to do by default a non-recursive search (but
a recursive search can still be triggered using the recursive keyword).
Following a comment from Todd, the search path for the files listed in
`filter_compiler_wrappers` can now be narrowed. Anyhow, the function
implementation still makes use of `find`, the rationale being that we
have already seen packages that install artifacts in e.g. architecture
dependent folders. The possibility to have a relative search path might
be a good compromise between the previous approach and the one suggested
in the review.
Also: 'ignore_absent' and 'backup' keyword arguments can be optionally
forwarded to `filter_file`.
Following comments from Todd:
- the call to tty.debug has been moved deeper, to log the filtering of each file
- the shadowing on the name "kwargs" is avoided
Implemented a declarative syntax for the additional behavior that can
get attached to classes. Implemented a function to filter compiler
wrappers that uses the mechanism above.
- command reference now includes usage for all Spack commands as output
by `spack help`. Each command usage links to any related section in
the docs.
- added `spack commands` command which can list command names,
subcommands, and generate RST docs for commands.
- added `llnl.util.argparsewriter`, which analyzes an argparse parser and
calls hooks for description, usage, options, and subcommands
- Shorten Spack command usage for short options. Short options are now
shown as [-abc] instead of as [-a] [-b] [-c]
- fix bug that mixed long and short options for top-level `spack help`
- Add proper help for `spack buildcache` subcommands
- Reorganize the help categories of Spack commands so that buildcache is
in packaging and diy and setup are now in build.
- previously commands with this argument showed a long list of choices
that were platform specific.
- use a better metavar: {defaults,system,site,user}[/PLATFORM]
* Remove variants disabling blas and lapack for py-numpy, issues
building these have been resolved
* For CMake greater than 3.10, FindMPI changed, so use
MPIEXE_EXECUTABLE instead of MPIEXE for 3.10 and newer
I think the main issue here is that we ship a custom version of a system
library (`argparse`), and this is prone to fail if `argparse` is
imported before we hack `sys.path` internally.
Probably a better solution would be not to customize argparse, but
instead have a wrapper on top of whatever the system provides.
fixes#7128
Before this PR packages that were indirectly dependent on python might
have failed due to inconsistency between the python found in the
environment and the standard libraries set in PYTHONHOME
This adds a static version of hpgmg (up to now only the latest commit
from the master branch was available). This also adds a debug variant
and sets optimization flags for ~debug when the compiler is gcc,
intel, or clang
Adds the latest release of libSplash (1.7.0) and newly learned
CMake flags for explicit dependency control.
Modifies HDF5 dependency: non-MPI splash does (and did) never run
with a parallel HDF5, due to it's internal cross-dependencies on
MPI.
Fixes#7159
When activating extensions in external views, the --ignore-conflicts
option was being ignored. In this particular issue the conflict was
for the duplicate __init__ file for multiple python packages in the
same namespace, but in general any conflict for extensions would
cause an error whether or not --ignore-conflicts was set.
This also renames the 'force' option of do_activate to
'with_dependencies' and updates views to call do_activate with this
set to False (since it traverses the dependency dag anyway). This
isn't strictly required, it just avoids redundant calls.
This reorganizes most sections and rewords a significant portion of
the content (including all introductions) but keeps all the examples.
* Remove section 'What happens at subscript time' from tutorial:
it is too detailed for a tutorial
* Move the 'Extra query parameters' and 'Attach attributes to other
packages' sections into a separate grouping 'Other packaging topics'
* move the 'Set variables at build time yourself' section after
'Set environment variables in dependents' section since the latter
is more motivating
* start the 'set environment variables at build-time for yourself'
section with qt as an example
* renamed section 'specs build interface' to 'retrieving library
information' and updated section introduction
* renamed section 'a motivating example' to 'accessing library
dependencies'; split out the material which deals with implementing
.libs for netlib-lapack into a separate section called 'providing
libraries to dependents'. consolidated in material from the section
'single package providing multiple virtual specs' since
netlib-lapack is an example of this (this removes the material
about intel-parallel studio)
* Allow dashes in command names and fix command name handling
- Command should allow dashes in their names like the reest of spack,
e.g. `spack log-parse`
- It might be too late for `spack build-cache` (since it is already
called `spack buildcache`), but we should try a bit to avoid
inconsistencies in naming conventions
- The code was inconsistent about where commands should be called by
their python module name (e.g. `log_parse`) and where the actual
command name should be used (e.g. `log-parse`).
- This made it hard to make a command with a dash in the name, and it
made `SpackCommand` fail to recognize commands with dashes.
- The code now uses the user-facing name with dashes for function
parameters, then converts that the module name when needed.
* Improve performance of log parsing
- A number of regular expressions from ctest_log_parser have really poor
performance, most due to untethered expressions with * or + (i.e., they
don't start with ^, so the repetition has to be checked for every
position in the string with Python's backtracking regex implementation)
- I can't verify that CTest's regexes work with an added ^, so I don't
really want to touch them. I tried adding this and found that it
caused some tests to break.
- Instead of using only "efficient" regular expressions, Added a
prefilter() class that allows the parser to quickly check a
precondition before evaluating any of the expensive regexes.
- Preconditions do things like check whether the string contains "error"
or "warning" (linear time things) before evaluating regexes that would
require them. It's sad that Python doesn't use Thompson string
matching (see https://swtch.com/~rsc/regexp/regexp1.html)
- Even with Python's slow implementation, this makes the parser ~200x
faster on the input we tried it on.
* Add `spack log-parse` command and improve the display of parsed logs
- Add better coloring and line wrapping to the log parse output. This
makes nasty build output look better with the line numbers.
- `spack log-parse` allows the log parsing logic used at the end of
builds to be executed on arbitrary files, which is handy even outside
of spack.
- Also provides a profile option -- we can profile arbitrary files and
show which regular expressions in the magic CTest parser take the most
time.
* Parallelize log parsing
- Log parsing now uses multiple threads for long logs
- Lines from logs are divided into chnks and farmed out to <ncpus>
- Add -j option to `spack log-parse`
* libdwarf: fix linker flags
* libdwarf: fix header install names
* libdwarf: add zlib dependency
Prior to this commit, libdwarf was linked to the system zlib. This
commit links it to spack-installed zlib.
* libdwarf: fix dynamic library suffix for Darwin
* Adding support for the Accelerate library on OS X
* Fixed guard on variable and added another depends_on statement to make sure that openblas gets selected even when openmp_blas variant is not turned on.
* Added a package for the MDAnalysis toolkit.
* Added some flags to fix the build RPATHs and OpenMP for LBANN on OS X.
* Fixed Hydrogen branch of Elemental to find OpenMP on OS X.
* Marking database tests as slow
* Marking url command tests as slow
* Marking every test that uses database as slow
* Marking tests that import files as slow
* Marking gpg tests as slow
* Marking all versions and one list tests as slow
* Added more markers to unit tests + cli option to skip slow tests
Following a discussion with Axel, the generic 'slowtest' marker has been
split into 'db', 'network' and 'maybeslow'. A brief description of the
meaning of each marker has been added to pytest.ini.
A command line option to run only fast tests has been added to
'spack test'
* Don't use classes to group tests together
Reverted grouping tests under a class, as required in the review
* Minor style changes
This attempts to address one of the complaints at #5996 (comment):
> Repo currently caches package instances by Spec, and those Package instances have a Spec.
> This is unnecessary and causes confusion. I think I thought that we'd need to cache instances
> after loading package classes, but really just caching the classes is fine.
With this update, Repo's package cache is removed and Specs cache the package reference themselves. One consequence is that Specs which compare as equal will store separate instances of a Package class (not doing this creates issues for #4595 (comment)).
There were several references to Spec.package that could be replaced with Spec.package_class without any additional modifications. There are still a couple remaining references to Spec.package in Spec that would require adding functionality before replacing (e.g. calling Package.provides and Package.installed).
Note this makes it difficult to mock fetchers for tests which invokes code that reconstructs specs. test_packaging was one example of this where the updates caused a failure (in that case the error was avoided by not making an unnecessary call).
Details:
* Replace instances of spec.package with spec.package_class where a class method is being called
* Remove instances of Repo.get where Spec.package_class can be used in its place
* remove Repo.get caching instances of Package class for specs
* remove redundant check (which is also incorrect now that each spec stores its own copy of its package)
* avoid creating mirror with specs because it copies specs and those copies dont refer to the mocked fetcher (and it is also not required for the test)
* remove checks that are no longer necessary since repo doesn't cache specs
* Cleaned up JUnit report generation on install
The generation of a JUnit report was previously part of the install
command. This commit factors the logic into its own module, and uses
a template for the generation of the report.
It also improves report generation, that now can deal with multiple
specs installed at once. Finally, extending the list of supported
formats is much easier than before, as it entails just writing a
new template.
* Polished report generation + added tests for failures and errors
The generation of a JUnit report has been polished, so that the
stacktrace is correctly displayed with Jenkins JUnit plugin. Standard
error is still not used.
Added unit tests to cover for installation failures and installation
errors.
* libtool: fix linking of libtool on darwin
* binutils, libtool: fix conflicts w/ BSD tools, v2
Fix namespace conflicts with BSD tools in a more elegant fashion,
using a program prefix, similar to Homebrew.
* Tell R's configure about tcltk config
Add configure arguments that specify the location of
the tcl and tk config scripts.
Fixes#7072
* Flake8 cleanup
Looking at build logs, it seems that prior to this commit bzip2 was
using 'cc', and relying on the PATH to be set correctly to find Spack
wrappers. This commit improves the robustness of the recipe, by using
the absolute path of the wrapper.
Avoid adding an "outside" (local home's) python user site directory
during python package installs.
Implements #6611
Fixes packages with auto-find side effects such as `py-setuptools`
that cause `py-matplotlib` to fail to build #6558
On Darwin and other BSD systems, the system 'libtool' and 'libtoolize'
are BSD libtool and libtoolize, respectively. Some build systems
require the GNU versions of these commands, so BSD package systems
tend to name the GNU versions 'glibtool' and 'glibtoolize',
respectively, to avoid namespace collisions.
A problem with the current libtool package is that it installs the GNU
libtool commands as 'libtool' and 'libtoolize', respectively, but
build systems on BSD platforms will attempt to run 'glibtool' and
'glibtoolize'. The expedient solution to this problem is to symlink
'libtool' to 'glibtool' and 'libtoolize' to 'glibtoolize', because
attempting to patch the detection logic one build system at a time
seems impractical.
* Added new link line flag
* Refactored some of the common components out of the different
version's build commands. Also corrected the when commands to
properly capture non-numeric names such as local.
* Added support to the Elemental package for the Hydrogen branch.
* Turning off default Qt support. Restricting Qt version to <= 4.9 else geant4 won't build
* Making vecgeom optional
* Differentiating between cxx11 and cxx14 support
* Making vecgeom optional. Part 2. forgot to move usolids option
* Adding explicit CXX11 and CXX14 support
* Adding more clhep version to support newer GEANT4 versions
* Adding GEANT4 versions 10.03.p03 and 10.04
* Checkpoint. Adding opengl and x11 variant. Lead up to more changes to enable X11 support
* Adding dependency on fixesproto for newer versions
* Making the X11 dependent pieces work without requiring X11 on the system
* Adding motif
This provides options for hardware and software rendering with the
gl API and updates the vtk package to make use of those options:
* Create new "gl" virtual package, provided by mesa
* Add external-only "opengl" package, which represents a system
install that provides gl
* For vtk: prefer system gl implementation by default to get speed
up from hardware rendering. When software rendering is specifically
requested (+osmesa), try to use the llvmpipe approach, as it is
much faster.
* Make qt dependency optional for vtk
* Add basic support for VTK on osx
* qt: Depend on virtual gl package, rather than mesa impl
* visit: add version 2.12.3
* visit: add version 2.13.0
* visit: enable building with hdf5+mpi
* visit: add qwt dependency
Building VisIt requires Qwt, which is not part of Qt. Prior to this
commit, this dependency was not included in the VisIt spack package,
resulting in a configure-time bug with an error message indicating a
missing Qwt dependency.
This commit fixes this bug by adding Qwt dependency information.
* Revert "Travis: use --concurrency=multiprocessing only on build tests (#6872)"
This reverts commit 596d463714.
* Removing 'coverage combine' in test script
According to what was discovered in #6887, one of the problems is
calling 'coverage combine' twice without the '-a' flag. This removes
the first call within our test scripts.
Update the qt4 CoreWLAN patch to apply to macOS/OS X 10.12 and 10.13,
which both still use the CoreWLAN framework in their SDKs (although
it's apparently deprecated?!). Without this patch update, `spack
install qt@4.8.6` fails on macOS/OS X 10.12 because the linker
cannot find headers from the CoreWLAN framework.
Apple removed support for Objective-C garbage collection in binaries
in XCode 5.1, but VTK 6.1.0 and earlier set flags for garbage
collection, resulting in a compile-time bug when compiling VTK using
any recent version of XCode (i.e., Apple Clang).
This commit fixes this bug by removing the garbage collection flags
via setting the variable `VTK_REQUIRED_OBJCXX_FLAGS` to the empty
string. This variable has the same value (i.e., the empty string) in
the root-level CMakeLists.txt file for VTK 6.2.0 and later.
VTK 6.1.0 and earlier do not use paths stored in `NETCDF_CXX_ROOT` to
detect the presence of NetCDF C++ headers and libraries. Consequently,
VTK 6.1.0 does not build.
This commit fixes this bug by setting the advanced variables
`NETCDF_CXX_INCLUDE_DIR` and `NETCDF_CXX_LIBRARY` at the command line
to specify the directory containing headers and the full library path
for the NetCDF C++ bindings.
The flag_handlers method was being set as a bound method, but when
reset in the package.py file it was being set as an unbound method
(all python2 issues). This gets the underlying function information,
which is the same in either case.
The bug was uncovered for parmetis in #6858. This is a partial fix.
Included are changes to the parmetis package.py file to make use of
flag_handlers.
This makes use of the new flag_handler logic from 28d8784a to set
compiler flags for ipopt by passing them as arguments to the build
system rather than injecting them into the compiler wrappers. This
avoids conflicts between flags that are chosen by the build system
and flags that are set by the user.
The feature added in #4611 is currently broken. This commit fixes the
behavior of the command and adds unit tests to ensure the basic semantic
is maintained.
It also changes slightly the behavior of Spec.concretized to avoid
copying caches before the concretization (as this may result in a
wrong hash computation for the DAG).
- Generating the HTML from for >2300 packages from RST in Sphinx seems to
take forever.
- Add an option to `spack list` to generate straight HTML instead.
- This reduces the doc build time to about a minute (from 5 minutes on a mac laptop).
* Vendor ordereddict for python2.6 only
This commit substitutes the custom module 'ordereddict_backport' with
the more known 'ordereddict' and vendors it only for python 2.6. Other
supported versions of python will use 'collections.OrderedDict'.
* Use absolute imports also for python 2.6
See PEP-328 for more information on the subject
* Added provenance of vendored ordereddict
* ruby: fix +openssl & +readline variants
Fix "unqualified variable spec['openssl']" error in the ruby package
that arises when trying to install the `+openssl` variant by
referencing the `spec` field in the `Ruby` class. A similar error
arises when trying to install the `+readline` variant; this error is
also fixed by this patch.
* ruby: make +openssl variant default to on
Ruby's gem command will fetch gems from HTTPS resources by default
(e.g., gem install bundler). Without openssl, request to fetch gems
bounce back with the error
```
ERROR: While executing gem ... (Gem::Exception)
Unable to require openssl, install OpenSSL and rebuild ruby (preferred) or use non-HTTPS sources
```
Without the ability to install gems -- required for some spack
packages -- the ruby installation has limited utility for many users.
* ruby: update rubygems ssl cert to fix ssl errors
The SSL certificate bundled with Ruby 2.2.0 is outdated, so e.g., `gem
install erubis` will fail with an SSL certificate error. This commit
installs the updated SSL certificate to the proper directory so that
gems can be downloaded and installed from RubyGems.
The url in graphviz has been fixed and its version has been
updated to the latest available (2.40.1). Added variants for
expat, qt, gtk+ and ghostscript.
* py-fparser: parser for Fortran 77..2003 code
* py-fparser: guard against external pytest.ini
* py-fparser: define 'develop' version as latest master branch
* py-fparser: fetch tagged versions from git repo
* py-fparser: add latest release 0.0.6
* py-fparser: whitespace adjustments for flake8
* py-fparser: inherit python dependency from PythonPackage
* py-fparser: rely on spack to resolve sub-dependencies
* py-fparser: try type='test' for py-pytest
* py-fparser: fetch release tarballs instead of git commits
* py-fparser: test deptype is available, remove comment
* py-fparser: limit versions that use py-six
* py-fparser: version ranges do not need a lower bound
* qt: move licence acceptance from qt3krell.patch to qt3accept.patch
* qt: ensure that ptrdiff_t is defined for qt@3.3.8b
* qt: disable webglplugin for qt@5.10~opengl
* qt: depend on libxext (for qt@3 and @4)
* qt: avoid X dependencies on darwin platforms
* Revert "qt: avoid X dependencies on darwin platforms"
This reverts commit 05ce08de94.
* qt: limit range of versions depending on libxext
See #6794
This fixes cases where test-only dependencies were omitted from
consideration when modifying the environment at build time. This
includes an update to the python package definition to add
testing-related python extensions to its specialized environment
setup.
* coinhsl: add version 2014.01.10 (stable)
* coinhsl: make 2014.01.10 preferred version
Make Harwell Subroutines Library for IPOPT version 2014.01.10 the
preferred version because it is the latest stable version released.
* coinhsl: add option to link w/ BLAS libraries
This updates the conflict-checking logic to require that the conflict
spec matches exactly and that all fields mentioned in the conflict
spec are present in the concretized spec in order to report a
conflict. This will automatically skip all conflict checks for
dependencies of externals (since externals strip dependencies). This
will not affect non-external packages since all fields and
dependencies are fully specified for such packages.
According to Travis docs the exit code of after_success doesn't affect
the build result. Instead, uploading the coverage data as the last step
of the script will cause the job to fail if the command exits with
non-zero.
https://docs.travis-ci.com/user/customizing-the-build/#Breaking-the-Build
* Keep track of source and versions for external libraries
* Note source of more obscure libraries
* We aren't upgrading jsonschema after all
* Add note on modifications made to pytest
On a local workstation, it seems that tracking multiple processes during
coverage may result in malformed coverage reports for unit tests and not
for build tests.
Given that multiple processes make a difference in coverage mainly for
build tests, try to disable the tracking for unit tests to see if we get
more stable coverage results.
Building emacs on darwin throws an error when trying to build an Emacs
app in the nextstep/Emacs.app path of the build tree. For now, disable
building this app.
It's possible to enable building the app also; Homebrew offers options
to this effect, and also adds Mac-specific options for starting the
emacs daemon. However, for the sake of simplicity and getting a
workable up-to-date emacs installation on my machine as quickly as
possible, this commit focuses on a minimal viable modification.
* Add a new package: unblur.
* Add a new package: unblur.
* Add a new package: unblur.
* Add a new package: unblur.
Conflicts:
var/spack/repos/builtin/packages/unblur/package.py
* Add a new package: unblur.
* Fix git on ubuntu, first cut
Spack needs to pass information about where the linker
can find `libintl`. We're currently using `LDFLAGS` to do so.
The `LDFLAGS` info is pasted into the command line upstream
of the a file (`libgit.a`) that includes unresolved symbols that
need that library. This fails on Ubuntu, although it seems to
work on CentOS (see #6841).
This change allows git to build on a Ubuntu 16.04.3 droplet.
TODO: test on other platforms...
* Add a bit of useful commentary
* Upgrade recipe for parmetis to be a CMakePackage
+ Eliminate `install` method (use the one from CMakePackage).
+ Move configure options to new method `cmake_args`
+ Move special install instructions for DarwinOS to a `run_after` method.
* Fix run_after section; Remove variant +debug.
Control the find_package of the PNG variant explicitly.
This avoids picking up an "external" PNGwriter install in
case `~png` is picked by changing the default "AUTO" search
to explicit "ON" (required) of "OFF" (ignore if exists).
* py-psyclone: Code generation for the PSyKAl framework from the GungHo project
* py-psyclone: add installation tests based on py.test
* py-psyclone: define 'develop' version as latest master branch
* py-psyclone: extract tagged version 1.5.1 from git repo
* py-psyclone: whitespace adjustments for flake8
* py-psyclone: use compatible versions of py-fparser
* py-psyclone: use type='test' for selected dependencies
* Updating OpenFAST to use explicit lapack and blas libraries from Spack.
* Simplifying openfast blas lib logic and fixes flake8 errors.
* Changing order of blas and lapack libs in openfast package.
* py-pybedtools: Create new package
* py-pybedtools: add py-six requirement
* py-pybedtools: remove py-cython dependency for release version
* py-pybedtools: include minimum version of py-pysam
* Provide build_type variant for Metis.
+ Ideally, we would make Metis a CMakePackage, but `metis@:5` doesn't use CMake.
+ For now, provide a `build_type=` variant similar what is found in
CMakePackage.
+ There is a potential for duplicate specification of `CMAKE_BUILD_TYPE` if both
variants `+debug` and `build_type=` are specified. I am looking for advice on
how this can be resolved.
* Update metis recipe in response to flake8 and user comments.
+ Wrap comment lines that used more than 80 columns.
+ Change `+debug` variant to avoid potential for multiple `-DCMAKE_BUID_TYPE=`
arguments provided to cmake. Specifying this variant no longer appends this
configure option. However, if `+debug` is specified, require
`build_type=Debug` to maintain expected behavior.
* For metis, remove +debug variant; use build_type= instead.
+ Update recipe for metis@4 to extract `build_type=` values and set `OPTFLAGS`
accordingly.
+ For metis@5:, the behavior from the previously supported variant `+debug` can
be obtained with the options `+gdb build_type=Debug`
* Conflicts added for metis@4 when build_type != Release|Debug.
* Use spack function conflicts instead of 'raise InstallError'.
* Create a symbolic link for MotionCor2.
* Create a symbolic link for MotionCor2.
* Create a symbolic link for MotionCor2.
* Create a symbolic link for MotionCor2.
* Adding NEURON package widely used in computational neuroscience community
* flake8 fixes
* Add rx3d variant
* Addressed review comments from Adam
* Addressed 2nd review comments
* Added check for +mpi on bg-q
* Filter spack compiler wrappers in NEURON scripts
Add gcc march=native flag to make front-end build robust
* Make develop as preferred version due to bug in long filename handlings in <= v7.4
* Avoid use of gcc in cross compiling environment
Use arch specific options on cray and bg-q arch
* Simplify cross compiling build:
- v7.5 added and no longer need develop as preferred version
- legacy cray related options removed
- cross compiling args reduced
- python arguments reduced as neuron finds proper libraries
* Simplify bg-q configurations
* Cleanup and unify arch options
* Make sure to add commonly available packages as dependency
* Arch path added to dependent_environment as nrniv needed at runtime
* Support for Power 8 architecture
* Fix python3 on OS X
* flake8 fix
* remove unnecessary conflcist directives
* Added a package for the MDAnalysis toolkit.
* Added a patch file to get rid of compiler warnings with intel compiler
and inlining. Added checksum values for other versions of protobuf.
* Added dependency on Intel compiler for intel patch
* added variant examples to paraview package.py
* qt: added libxt dependency
* add variant to use external hdf5, this is needed when compiling with gcc@4.8.5
* paraview: added python dependecies to PYTHONPATH
* Start adding Nek5000 package
* Add tools to install: genmap, prenek, postnek
* Address comments by @davydden
* Use install_tree (thanks @junghans) and polish
* Update url
* Fix flake formatting errors
* Change F77 and CC compilers
Apparently spack compilers can't be called outside spack and we
have an underscore test in our Nek5000/core/makenek.inc file which
calls F77 compiler.
* Add support for visit and profiling options
* Fix a typo
* Fix a typo
* Remove the C check
* Add visit variant
* Update the install of int_tp tool
* Let MAXNEL to be set during installation
* Added magma package
* Incorporated Serban's change
* Remove deleted files
* argparse.py
* Sync forked Spack repo in github
* Add spectrum-mpi lib path to LD_LIBRAY_PATH in spack_env.
As a result of supporting alternate installation directories, spectrum-mpi does
not use rpath anymore, so we need to set LD_LIBRARY_PATH in spack_env to allow
MPI programs to run during dependent package configuration.
* Fix flake8 error (blank line at EOF)
* xios: XML-IO-SERVER library for IO management of climate models
* xios: install to prefix directory
* xios: add a test case
* xios: define versions to use cache and mirror directories
* xios: use proper svn URL for version 910
* xios: package url must be defined, otherwise uninstall fails
* xios: add trunk version to allow building latest revision
* xios: enable parallel build
* xios: allow installation with mpif90 compiler wrappers
* xios: use mpi variants of dependencies
* xios: allow user to specify MPI launcher via MPIRUN environment variable
* xios: prefer spack versions of blitz and boost (needed for gcc@7)
* xios: enable builds with clang+gfortran on OSX
* xios: use numeric versions for releases, @develop for trunk
* xios: different patches required for @1.0 and @develop
* xios: patch function for clang works across revisions
* xios: link blitz library
* xios: workaround bug 17782 in llvm
* xios: whitespace adjustments for flake8
* xios: avoid bare except clause
* xios: netcdf-fortran inherits ^netcdf+mpi from xios dependencies
* combine all patches for 1612 into a single file, with additional
comments to explain the rational.
* Add kahip decomposition variant.
* Add support for ARM64 architecture.
* filter instead of patching etc/config.*/settings. This eliminates
patch files for 1706, perhaps for the future as well.
* URLs at or before version 1612 have a different structure, so
provide a url_for_version implementation to handle this
* singularity - added master branch as this has latest fixes - 2.4 does not compile on centos 6 without use of this branch
* renamed to develop
* added master branch properly
* add OctavePackage
1. remove import CudaPackage which is not needed anymore
2. mention CudaPackage and OctavePackage in packaging guide
3. adjust OctavePackageTemplate
4. add clue file for Octave build
5. sanity check on self.prefix
* use setup_environment
* esmf: put .mod files in the spack compiler search path
* esmf: allow building with Cray mpi
* esmf: create run-path dependent libraries on osx
* esmf: link fortran interface libraries when building with mpich
* Fix detection of blas and lapack: provide libraries/includes as
arguments to CMake rather than using CMake's auto-detection.
This includes a patch to QMCPACK's CMake files to refer to
Spack-built blas/lapack implementations. This also includes
special-case logic for the intel-mkl implementation of blas/lapack
* Break up unit tests and short tests. Only hard failure if the unit
tests fail, if short tests fail issue a warning only.
* Add QMCPACK 3.3.0, get rid of 3.0.0
* Only specify a file as needing relocation if it contains the spack
root as a text string (this constraint also applies to binaries)
* Don't fail if there is an error retrieving RPATH information from a
binary (even if it is specified as requiring relocation)
This adds the ability for packages to apply compiler flags in one of
three ways: by injecting them into the compiler wrapper calls (the
default in this PR and previously the only automated choice);
exporting environment variable definitions for variables with
corresponding names (e.g. CPPFLAGS=...); providing them as arguments
to the build system (e.g. configure).
When applying compiler flags using build system arguments, a package
must implement the 'flags_to_build_system_args" function. This is
provided for CMake and autotools packages, so for packages which
subclass those build systems, they need only update their flag
handler method specify which compiler flags should be specified as
arguments to the build system.
Convenience methods are provided to specify that all flags be applied
in one of the 3 available ways, so a custom implementation is only
required if more than one method of applying compiler flags is
needed.
This also removes redundant build system definitions from tutorial
examples
- define MPICH_SKIP_MPICXX, link libmpich instead of libmpi
(as per disussion in PR #4990)
ENH: more flexible editing/patching of openfoam env files
CONFIG: comment out requirement for openmpi+thread_multiple
- was originally slated for the 1712 release and beyond, but there are
reportedly unstable version/network combinations.
Leave as comment for future reference, since it may be arise again.
STYLE: renamed openfoam-site.patch as 1706-site.patch (and reformatted)
- to preserve against impending changes in the layout of
config.{csh,sh}/settings
Fixes#5940
This PR changes the option '--restage' of 'spack install' to
'--dont-restage', inverting its meaning and the default behavior
of the command.
* fix post install method for py3 in mercurial
* Update package.py
* Update package.py
* use print_str instead of manual print detection
* Update package.py
* use only depends_on
* Update package.py
* Update package.py
State py3 compatibility for @4.3: according to setup.py from the 4.3.tar.gz from pypi
Fixes#6200
For compilers that successfully run a version detection script but
don't actually return a version, Spack was keeping track of the
empty version and then failing when attempting to construct a
compiler spec. This skips any attempt to add a compiler entry when
no version is reported (but logs when a compiler fails to report
a version).
* Support pruning of vars with Env from_sourcing_file (issue #6501)
- Blacklist string literals or regular expressions of environment
variables that are to be removed from consideration as being affect
by the sourcing of the file. Conversely, whitelist modifications
that should not ignored. Whitelisted variables have priority over
blacklisting.
Eg,
EnvironmentModifications.from_sourcing_file
(
bashrc
blacklist=['JUNK_ENV', 'OPTIONAL_.*'],
whitelist=['OPTIONAL_REQUIRED.*']
)
This modification can be used to eliminate environment variables that
are not generalized for modules (eg, user-specific variables).
* BUG: module prepend-path in wrong order (fixes#6501)
* STYLE: module variables in sorted order (issue #6501)
- looks nicer and also helps when comparing the contents of different
module files.
* ENH: remove duplicates from env paths when creating modules (issue #6501)
- this makes for a cleaner module environment and helps avoid some
unnecessary changes to the environment that are only provoked by
redundancies in the PATH.
eg,
before PATH=/usr/bin
after PATH=/usr/bin:/usr/bin:/my/application/bin
should only result in /my/application/bin being added to the PATH
and not /usr/bin:/my/application/bin
Activate via the 'clean' flag (default: False):
EnvironmentModifications.from_sourcing_file(bashrc, clean=True,..
Add preferred version 2.0.3. Added beta version. Added conflict with
gcc versions >= 5 when building with cuda@8 (which is currently the
only supported cuda version). Added cuda architecture variant. Added
comment about current lack of cuda@9 support.
Fixes#6522
Patch psxevars.sh and psxevars.csh only when year > 2016.
pxevars.csh does not exist for versions before 2017. psxevars.sh
does not appear to require this patch for versions before 2017.
Fixes#2440
The "Getting started" guide should be short and sweet. This commit
simplifies the "Environment-Modules" section pruning:
- outdated / wrong suggestions as noted in #2440
- uncommon setups that are better treated in a reference guide
Fixes#6521.
Bazel's build system filters environment variables, including those
set by Spack for use by its compiler wrapper. This updates the patch
to include the SPACK_DEBUG_LOG_ID variable added in #5837.
* optional path to use spacks py-site-pkgs, install host-cfg for provenance
* add more comments and address review requests
* only use st-pkgs-dir when +python
* ascent does not need doxygen
* capture mpiexec if it exists
* added missing import
* flake8 fix
* default adios variant to False
* ascent develop now reqs conduit master
* optional path to use spacks py-site-pkgs, install host-cfg for provenance
* add more comments and address review requests
* only use st-pkgs-dir when +python
* ascent does not need doxygen
* capture mpiexec if it exists
* added missing import
* flake8 fix
* default adios variant to False
* ascent develop now reqs conduit master
* conduit: use hdf5 1.8 for wider output compat
* Add: Py-Plotly
The requirements of py-plotlit are rather strict fixed to specific
versions, so I did not add an "upper versions" `:` to the constrains.
See
https://github.com/plotly/plotly.py/blob/master/requirements.txt
* Py-Requests, Py-Six, Py-Pytz: Older Versions
Add two older versions of requests, pytz and six for plotly.
- depends on GTK+ with X enabled
(btw, non-X backends are deprecated/removed in GTK+)
- patch for 3.0.1-3.0.2: missing include caused multiple
`error: call of overloaded ‘abs(XYPOSITION)’ is ambiguous`
(seen in GCC 6.3.0); fixed in 3.0.3+
According to the documentation here:
http://www.sphinx-doc.org/en/stable/ext/autodoc.html
"For module data members and class attributes, documentation can either
be put into a comment with special formatting (using a #: to start the
comment instead of just #), or in a docstring after the definition."
* Update CGNS to be able to be used in Trilinos.
* Updating Trilinos to be able to use CGNS.
* Fixing wrong cmake defines in trilinos for enabling cgns and cleaning up cgns package.
* Changing cgns parallel variant to mpi. Removing explicit dependence on parallel-netcdf in trilinos.
Add additional transform libraries to ADIOS.
Also modifies existing transforms to the versions they were
introduced (otherwise this was resulting in unknown configure flags).
* Updated function which checks if a binary file needs relocation.
Previously this was incorrectly identifying ELF binaries as symbolic
links (so they were being excluded from relocation). Added test to
check that ELF binaries are not considered symlinks.
* relocate_text was not replacing paths in text files. Added test to
check that text files are relocated properly (i.e. paths in the file
are converted to the new prefix).
* Exclude backup files created by filter_file when installing from
binary cache.
* Update write_buildinfo_file method signature to distinguish between
the spec prefix and the working directory for the binary cache
package.
"spack spec" was providing helpful error information about conflicts
that was missing from "spack install", this updates "spack install"
to provide the same information.
The original docstring had confusing wording re: what is going to
symlinked and where when using the `extend` directive, and how exactly
the symlinking is performed (not automatically on install, but using
`spack activate`). See #5559.
Showing "Normalize" on output doesn't give users additional information,
as this step is essentially an implementation detail of concretization.
This PR skips it and shows just the input spec and the concretized one.
Printing partial hashes for input spec has been disabled.
* xsdk: Added MFEM rc package.
* xsdk: add version entry for xsdk-0.3.0-rc1 - and set it as preferred version
* xsdk: remove 'xsdk' from version - recommended by Todd.
Now the usage is:
'spack install xsdk@0.3.0-rc1'
instead of the previous usage:
'spack install xsdk@xsdk-0.2.0'
* xsdk: add petsc@3.8.0 as a dependency for xsdk-0.3.0-rc1
* mfem: fix dependency to use the corrected xsdk version
* Updated the mfem version tags.
* mfem: fix merge
* fix flake8 warning
* Modified xSDK scripts
* fixing the default package.yaml
* fix bad revert
* revert changes to superlu-dist so that the 5.2.1 update can be merged
* update xsdk to use superlu-dist-5.2.1
* update alquimia for xsdk
* fix whitespace
* update hypre for xsdk
* update sundials for xsdk
* update sundials for xsdk
* Fix checking for MKL and extend range of GCC versions
* pflotran: update to release branch
* plasma: fix spec -> self.spec
* Changes to enable mfem hooks into hypre, superlu-dist, and petsc. Sundials hooks may come in the future.
* adding plasma with comment
* Bumped the version of mfem up to handle integration with sundials3.0.0. Also enabled hypre in sundials.
* Turn on the plasma package definition for testing
* Fix conflict declaration to obey ordering rules
* xsdk: udate superlu-dist dependency to v5.2.2
* plasma: change gcc dependency to workarround potential spack bug #5778
* trilinos: update hdf5 dependency to use 'hdf5+hl'
Without this change - I get the following error [triggered by 161dca6d5c9810f808 changes]
$ ./bin/spack install trilinos+hdf5
==> Error: An unsatisfiable variant constraint has been detected for spec:
hdf5@1.10.1%gcc@7~cxx~debug~fortran~hl+mpi+pic+shared~szip~threadsafe arch=linux-fedora27-x86_64
while trying to concretize the partial spec:
netcdf+mpi
^m4
^mpich@3.3a2%gcc@7 device=ch3 +hydra netmod=tcp +pmi+romio~verbs arch=linux-fedora27-x86_64
^zlib@1.2.11%gcc@7+optimize+pic+shared arch=linux-fedora27-x86_64
netcdf requires hdf5 variant +hl, but spec asked for ~hl
* plasma: add in -lm dependency per #6046
* update sundials version to 3.0.0
* Add fixes to support Netlib LAPACK
* mfem does not use cmake for builds [anymore] - so remove code corresponding to this dependency.
This enables mfem build on OSX [with xcode/clang compilers]
* mfem: update to v3.3.2-rc4
* add patch for building sundials on mac
* xsdk: enable plasma for gcc-4.9+ - same as what plasma/package.py does.
* magma: Fix for cuda-9.0 as it does not support sm20
nvcc fatal : Unsupported gpu architecture 'compute_20'
* add in code for 'spack install xsdk+cuda' - which installs magma
* fix flake8 warning
* fix flake8 warnings
* Updated to the latest MFEM RC.
* alquimia: fix to use petsc-3.8.0+ versions
* petsc: update to version 3.8.1
* petsc: add in conflict between petsc-3.8.0 & mkl
* xsdk: update petsc to 3.8.1 to get MKL build fix
* petsc: fix superlu_dist dependency
* petsc: when hdf5 depends on zlib - petsc+hdf5 should also be configured with zlib
* petsc: fix dependency [of zlib to be on hdf5 - and not on petsc]
* petsc: fix pkgconfig error - with incorrect library link order that can break mfem examples with static build
* petsc: use the fact that hdf5/package.py has a hard dependency on zlib [so no need to check for this dependency in code]
recommended-by: @davydden
* petsc: change hdf5 dependency to 'hdf5+hl' as petsc configure is looing for -lhdf5_hl
https://github.com/spack/spack/issues/5846
* plasma: also detect -lgfortran
* update sundials version to 3.1.0
* alquimia: xsdk-0.3.0-rc -> xsdk-0.3.0
* petsc: update to version 3.8.2
Includes pkgconfig fix [for mfem static build failure]
* Revert "petsc: fix pkgconfig error - with incorrect library link order that can break mfem examples with static build"
This reverts commit d1afa8c20d.
* xsdk: update to use petsc-3.8.2
* Fixed a problem with MFEM trying to use mpicxx instead of the spack defined compiler.
* Minor changes to the mfem build and pointing at mfem's release branch that is currently undergoing internal testing.
* Enabling shared build for Plasma
* plasma: do not look for libgfortran on cray - spack finds the wrong one from the frontend
'-L/usr/lib64/gcc/x86_64-suse-linux/4.3 -L/usr/lib64 -lgfortran'
* Some changes to the mfem package that are in the process of being pulled into spack/develop.
* xsdk: switch from 0.3.0-rc1 to 0.3.0
* fix flake8 warnings
* plasma: extract plasma changes from xsdk-0.3.0
* plasma: cleanup per review
- netlib-lapack: add support for spec['lapack:c'].libs, spec['blas:c'].libs
- add getblaslapacklibs() to eliminate duplicate code in edit() and build_targets()
- eliminate string manipulation of ld_flags
- remove gfortran check (will check later)
- remove build() (will check later)
* netlib-lapack: fix tuple per review
* netlib-lapack: use spec[lapack:c,fortran]
* Adding version 2.4.3 to ViennaRNA
* Replaced tab indent with spaces
Flake8 error correction
* no swig for viennarna v2.4.3 only
* correcting version/argument conditional syntax error
libtool is also needed for linking, otherwise downstream
dependencies such as clfft will fail linking against pocl's
libOpenCL.so with errors such as:
```
>> 182 /usr/bin/ld: warning: libltdl.so.7, needed by /usr/local/opt/spack/linux-ubuntu16.04-x86_64/gcc-5.4.0/pocl-0.14-rwo73qzkcxfupndxoz3pcl2s34o5pt6i/lib/libOpenCL.so, not found (try using -rpath or -rpath-link)
>> 183 /usr/local/opt/spack/linux-ubuntu16.04-x86_64/gcc-5.4.0/pocl-0.14-rwo73qzkcxfupndxoz3pcl2s34o5pt6i/lib/libOpenCL.so: undefined reference to `lt_dlinit'
>> 184 /usr/local/opt/spack/linux-ubuntu16.04-x86_64/gcc-5.4.0/pocl-0.14-rwo73qzkcxfupndxoz3pcl2s34o5pt6i/lib/libOpenCL.so: undefined reference to `lt_dlerror'
>> 185 /usr/local/opt/spack/linux-ubuntu16.04-x86_64/gcc-5.4.0/pocl-0.14-rwo73qzkcxfupndxoz3pcl2s34o5pt6i/lib/libOpenCL.so: undefined reference to `lt_dlopen'
>> 186 /usr/local/opt/spack/linux-ubuntu16.04-x86_64/gcc-5.4.0/pocl-0.14-rwo73qzkcxfupndxoz3pcl2s34o5pt6i/lib/libOpenCL.so: undefined reference to `lt_dlsym'
>> 187 collect2: error: ld returned 1 exit status
```
* Reduce the calls to the python interpreter during initialization
This should reduce the delay the users experience when sourcing the
setup file to activate shell support. It works by generating at once
all the commands that needs to evaluated (they are stored in
a string and later `eval`ed by the shell).
* setup_env.sh: changed `read` with an equivalent magic
For some reason `read` breaks when sourced from a running script.
Change the incantation we use to construct the unique python command
that will be evaluated.
* setup_env.sh: python command now constructed with `printf` for portability
This recovers the support for `zsh` that was broken in previous commits.
* py-petsc4py: Update version to 3.8.0 and add dependencies
- Added explicit dependency on python and numpy
* Kept old version - and corrected minimal python version
* Update to 3.8.1
about: Report a bug in the core of Spack (command not working as expected, etc.)
---
*Explain, in a clear and concise way, the command you ran and the result you were trying to achieve.
Example: "I ran Spack find to list all the installed packages and..."*
### Steps to reproduce the issue
```console
$ spack <command1> <spec>
$ spack <command2> <spec>
...
```
### Error Message
If Spack reported an error, provide the error message. If it did not report an error
but the output appears incorrect, provide the incorrect output. If there was no error
message and no output but the result is incorrect, describe how it does not match
what you expect. To provide more information you might re-run the commands with
the additional -sd flags:
```console
$ spack -sd <command1> <spec>
$ spack -sd <command2> <spec>
...
```
that activate the full debug output.
### Information on your system
This includes:
1. which platform you are using
2. any relevant configuration detail (custom `packages.yaml` or `modules.yaml`, etc.)
-----
We encourage you to try, as much as possible, to reduce your problem to the minimal example that still reproduces the issue. That would help us a lot in fixing it quickly and effectively!
If you want to ask a question about the tool (how to use it, what it can currently do, etc.), try the `#general` channel on our Slack first. We have a welcoming community and chances are you'll get your reply faster and without opening an issue.
Other than that, thanks for taking the time to contribute to Spack!
about: Some package in Spack didn't build correctly
---
*Thanks for taking the time to report this build failure. To proceed with the
report please:*
1. Title the issue "Installation issue: <name-of-the-package>".
2. Provide the information required below.
3. Remove the template instructions before posting the issue.
We encourage you to try, as much as possible, to reduce your problem to the minimal example that still reproduces the issue. That would help us a lot in fixing it quickly and effectively!
---
### Steps to reproduce the issue
```console
$ spack install <spec> # Fill in the exact spec you are using
... # and the relevant part of the error message
```
### Platform and user environment
Please report your OS here:
```commandline
$ uname -a
Linux nuvolari 4.15.0-29-generic #31-Ubuntu SMP Tue Jul 17 15:39:52 UTC 2018 x86_64 x86_64 x86_64 GNU/Linux
$ lsb_release -d
Description: Ubuntu 18.04.1 LTS
```
and, if relevant, post or attach:
- `packages.yaml`
- `compilers.yaml`
to the issue
### Additional information
Sometimes the issue benefits from additional details. In these cases there are
a few things we can suggest doing. First of all, you can post the full output of:
```console
$ spack spec --install-status <spec>
...
```
to show people whether Spack installed a faulty software or if it was not able to
build it at all.
If your build didn't make it past the configure stage, Spack as also commands to parse
about: Suggest adding a feature that is not yet in Spack
---
*Please add a concise summary of your suggestion here.*
### Rationale
*Is your feature request related to a problem? Please describe it!*
### Description
*Describe the solution you'd like and the alternatives you have considered.*
### Additional information
*Add any other context about the feature request here.*
-----
If you want to ask a question about the tool (how to use it, what it can currently do, etc.), try the `#general` channel on our Slack first. We have a welcoming community and chances are you'll get your reply faster and without opening an issue.
Other than that, thanks for taking the time to contribute to Spack!
Written by Todd Gamblin (<ahref="mailto:tgamblin@llnl.gov">tgamblin@llnl.gov</a>) and
many <ahref="https://github.com/spack/spack/graphs/contributors">contributors.</a> LLNL-CODE-647188.
{%- if last_updated %}
<br/>
{% trans last_updated=last_updated|e %}Last updated on {{ last_updated }}.{% endtrans %}
{%- endif %}
</p>
</div>
{%- if show_sphinx %}
{% trans %}Built with <ahref="http://sphinx-doc.org/">Sphinx</a> using a <ahref="https://github.com/snide/sphinx_rtd_theme">theme</a> provided by <ahref="https://readthedocs.org">Read the Docs</a>{% endtrans %}.
{% trans %}Please activate JavaScript to enable the search
functionality.{% endtrans %}
</p>
</div>
</noscript>
{% if search_performed %}
<h2>{{ _('Search Results') }}</h2>
{% if not search_results %}
<p>{{ _('Your search did not match any documents. Please make sure that all words are spelled correctly and that you\'ve selected enough categories.') }}</p>
{% endif %}
{% endif %}
<divid="search-results">
{% if search_results %}
<ul>
{% for href, caption, context in search_results %}
# This is a template package file for Spack. We've put "FIXME"
# next to all the things you'll want to change. Once you've handled
Some files were not shown because too many files have changed in this diff
Show More
Reference in New Issue
Block a user
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
