I get the following error message, if I do not use editline from the system.
```
>> 3090 Undefined symbols for architecture x86_64:
3091 "_tgetent", referenced from:
3092 _terminal_set in libedit.a(terminal.c.o)
3093 "_tgetflag", referenced from:
3094 _terminal_set in libedit.a(terminal.c.o)
3095 "_tgetnum", referenced from:
3096 _terminal_set in libedit.a(terminal.c.o)
...
3110 _terminal_insertwrite in libedit.a(terminal.c.o)
3111 _terminal_clear_EOL in libedit.a(terminal.c.o)
3112 _terminal_clear_screen in libedit.a(terminal.c.o)
3113 _terminal_beep in libedit.a(terminal.c.o)
3114 ...
3115 ld: symbol(s) not found for architecture x86_64
```
Bug fix release:
4.0.4 -- June, 2020
-----------------------
- Fix a memory patcher issue intercepting shmat and shmdt. This was
observed on RHEL 8.x ppc64le (see README for more info).
- Fix an illegal access issue caught using gcc's address sanitizer.
Thanks to Georg Geiser for reporting.
- Add checks to avoid conflicts with a libevent library shipped with LSF.
- Switch to linking against libevent_core rather than libevent, if present.
- Add improved support for UCX 1.9 and later.
- Fix an ABI compatibility issue with the Fortran 2008 bindings.
Thanks to Alastair McKinstry for reporting.
- Fix an issue with rpath of /usr/lib64 when building OMPI on
systems with Lustre. Thanks to David Shrader for reporting.
- Fix a memory leak occurring with certain MPI RMA operations.
- Fix an issue with ORTE's mapping of MPI processes to resources.
Thanks to Alex Margolin for reporting and providing a fix.
- Correct a problem with incorrect error codes being returned
by OMPI MPI_T functions.
- Fix an issue with debugger tools not being able to attach
to mpirun more than once. Thanks to Gregory Lee for reporting.
- Fix an issue with the Fortran compiler wrappers when using
NAG compilers. Thanks to Peter Brady for reporting.
- Fix an issue with the ORTE ssh based process launcher at scale.
Thanks to Benjamín Hernández for reporting.
- Address an issue when using shared MPI I/O operations. OMPIO will
now successfully return from the file open statement but will
raise an error if the file system does not supported shared I/O
operations. Thanks to Romain Hild for reporting.
- Fix an issue with MPI_WIN_DETACH. Thanks to Thomas Naughton for reporting.
Signed-off-by: Howard Pritchard <howardp@lanl.gov>
* Fix how the Conduit detects that the MPI compiler is the same as the
CC compiler and is more careful when it sets the MPI compilers to be
the Cray PE system compilers.
* Remove unnecessary push of the MPI compilers to the C compilers for Hydrogen.
* Added version 8.2.0, added dependency for documentation build, added variants for documentations
* Renamed variant '+man' to '+html'
Co-authored-by: Alexander Knieps <a.knieps@fz-juelich.de>
- Parallel HDF5 isn't required -- the comment seems to be about a
transitive dependency with pnetcdf.
- Boost usage should respect the variant, not automatically be reenabled
when choosing DTK.
Versions of py-tensorflow between versions 1.1 and 1.14 need a patch to
avoid an import error on the cloud package even if built without support
for the cloud package.
* bazel: Update for use with Fujitsu compiler
* bazel: Fix for use with Fujitsu compiler
* bazel: Fix flake8 error
* bazel: add conflicts setting for use with Fujitsu compiler
* fix flake8 error
* fix flake8 error
In Python 3.8, the reserved "tp_print" slot was changed from a function
pointer to a number, which broke the Python wrapping code in vtk@8
(causing "cannot convert 'std::nullptr_t' to 'Py_ssize_t'" errors in
various places). This is fixed in vtk@9.0.0.
This patch:
1) adds vtk@9.0.0
2) updates depends_on constraints to only use python@3.8: for vtk@9:
vtk@:8 depends on python@2, and vtk@8.0.1:8.9.9 depends on python@:3.7
3) Adds CMake flag VTK_PYTHON_VERSION=3 when using python@3 with vtk@9
* python: adding a distutils fix to improve build compatibility for C++ extension modules (e.g. py-matplotlib)
* python: added C/C++ distutils patches for python@3.6:3.8
Allow Spack to build with ROOT as an external dependency by setting
LD_LIBRARY_PATH: given that the external package was not built by
Spack, dependents would not be able to locate libraries using RPATHs
when running ROOT binaries.
Specified Python to be v2.7 only, as Python3 support is not currently
implemented in chill.
Update chill dependency versions for the following libraries to the
specific versions:
* rose: v0.9.13.0
* bison: v3.4.2
Both rose and iegenlib are build time dependencies, but are also run
time dependencies. Added 'run' to the build type for both dependencies.
cuda: 10.1 and onward, installers will crash if /tmp/cuda-installer.log
exists
Try to help if user owns the file, otherwise try to provide useful
info. Clean up the file post-install to try to avoid the whole issue.
* save edits
* tidy up
* Update var/spack/repos/builtin/packages/py-lmfit/package.py
Co-authored-by: Adam J. Stewart <ajstewart426@gmail.com>
* add python version constraints
Co-authored-by: Sinan81 <sbulut@3vgeomatics.com>
Co-authored-by: Adam J. Stewart <ajstewart426@gmail.com>
Co-authored-by: Sinan81 <Sinan81@github>
* Add pygelf Python package
* Update ReFrame package version
* Address styling remarks
* Update var/spack/repos/builtin/packages/reframe/package.py
Co-authored-by: Adam J. Stewart <ajstewart426@gmail.com>
* Update var/spack/repos/builtin/packages/reframe/package.py
Co-authored-by: Adam J. Stewart <ajstewart426@gmail.com>
* Update var/spack/repos/builtin/packages/reframe/package.py
Co-authored-by: Adam J. Stewart <ajstewart426@gmail.com>
* Update var/spack/repos/builtin/packages/reframe/package.py
Co-authored-by: Adam J. Stewart <ajstewart426@gmail.com>
* Update var/spack/repos/builtin/packages/reframe/package.py
Co-authored-by: Adam J. Stewart <ajstewart426@gmail.com>
* Address PR remarks
* Remove setuptools runtime dependency
* Address PR remarks
* Address PR remarks
Co-authored-by: Adam J. Stewart <ajstewart426@gmail.com>
bazel uses gcc's -MF option to write dependencies to a
file. Post-compilation, bazel reads this file and makes some
determinations.
"Since gcc is given only relative paths on the command line,
non-system include paths here should never be absolute. If they
are, it's probably due to a non-hermetic #include, & we should stop
the build with an error."
Spack directly injects absolute paths, which appear in this file and
cause bazel to fail the build despite the fact that compilation
succeeded.
This patch disables this failure mode by default, and allows for it
to be turned back on by using the '~nodepfail' variant.
Whenever attempting to use any ncurses functionality within cscope, a
page fault would result within the ncurses library.
Program received signal SIGSEGV, Segmentation fault.
0x00007ffff7fad3cf in termattrs_sp () from .../lib/libncursesw.so.6
(gdb) bt
#0 0x00007ffff7fad3cf in termattrs_sp () from .../lib/libncursesw.so.6
#1 0x00007ffff7faa794 in _nc_setupscreen_sp () from .../lib/libncursesw.so.6
#2 0x00007ffff7fa614c in newterm_sp () from .../lib/libncursesw.so.6
#3 0x00007ffff7fa65b9 in newterm () from .../lib/libncursesw.so.6
#4 0x00007ffff7fa2970 in initscr () from .../lib/libncursesw.so.6
#5 0x0000000000403dc2 in main (argc=<optimized out>, argv=0x7fffffffcea8) at main.c:574
This is due to a conflict between libtinfo.so and libtinfow.so. Both are
linked into cscope:
$ ldd $(which cscope)
/bin/bash: .../lib/libtinfo.so.6: no version information available (required by /bin/bash)
linux-vdso.so.1 (0x00007fff5dbcb000)
libncursesw.so.6 => .../lib/libncursesw.so.6 (0x00007f435cc69000)
libtinfo.so.6 => .../lib/libtinfo.so.6 (0x00007f435cc2c000)
libc.so.6 => /lib/x86_64-linux-gnu/libc.so.6 (0x00007f435ca29000)
libtinfow.so.6 => .../lib/libtinfow.so.6 (0x00007f435c9e8000)
/lib64/ld-linux-x86-64.so.2 (0x00007f435cca7000)
Specifically linking libtinfow.so instead of libtinfo.so resolves the
issue.
All instances of '...' above represent the path to the installed ncurses
for Spack.
* Adding a package for wcs.
* Turning on sbml for wcs.
* The cpp flag needs to be available for wcs.
* Wcs needs SBML to properly define the namespace.
* Flake8 fixes.
* Fixing the help string with the description.
* Changing cpp to use the new variant syntax.
* Fixing flake8 errors.
* Forgot to delete one last fixme comment.
* Spack "develop" needs to link to repo "devel"
Co-authored-by: Adam J. Stewart <ajstewart426@gmail.com>
Co-authored-by: Robert Blake <rob.c.blake.3@gmail.com>
Co-authored-by: Adam J. Stewart <ajstewart426@gmail.com>
* Cray: fix Blue Waters support
* pkg-config env vars needed on Blue Waters
* cray platform: fix support for user-build MPI on cray machines
* reintroduce cray environment cleaning behind cnl version guard
* cray platform: fix support for user-build MPI on cray machines
Co-authored-by: Gregory <becker33@llnl.gov>
* Add new package: py-devlib
* Update var/spack/repos/builtin/packages/py-devlib/package.py
Co-authored-by: Adam J. Stewart <ajstewart426@gmail.com>
* add depends
Co-authored-by: Adam J. Stewart <ajstewart426@gmail.com>
* [hsf-cmaketools] add package
* fix formatting
* Apply suggestions from code review
Co-authored-by: Adam J. Stewart <ajstewart426@gmail.com>
* [hsf-cmaketools] remove cmake_prefix_path which is set already by spack
Co-authored-by: Adam J. Stewart <ajstewart426@gmail.com>
gcc 9.3.0 and glibc 2.31 (found in the base install of Ubuntu 20.04)
cause the gcc package to error during build with the error:
"size of array 'assertion_failed__####' is negative"
Previous to this fix, the error was resolved for v8.1.0 <= gcc <= v9.2.0
via two patches.
This fix backports those patches for v5.3.0 <= gcc <= v7.4.0
Potentially these patches need to be backported to versions of gcc
before v5.3.0, but other compile issues need to be resolved for earlier
versions of gcc first.
Fixes#16968
* intel-tbb: Fix for #16938 add custom libs method
Override the libs method to look for libraries of form libtbb*
(instead of inherited which looks for libintel-tbb*)
* Fixing pre-existing flake8 issues
* py-flake8: add version 3.8.2
* This version depends on different versions of py-pycodestyle
and py-pyflakes
* When built for python@:3.7, this depends on the
py-importlib-metadata backport library
* py-pycodestyle: add version 2.6.0
* py-pyflakes: add version 2.2.0
Stratimikos is an optional dependency for our project. It depends on
Thyra, and thyra has subpackages that should be enabled based on
tpetra/epetra/epetraext.
* gnuplot: Fix for #16928
Dependency for --with-wx flag mistyped (should be wxwidgets)
* Revert "gnuplot: Fix for #16928"
This reverts commit 2b85814e5c.
* gnuplot: Fix for #16928
Dependency spec for --with-wx flag mistyped (should be wxwidgets, not
wx)
* Adding a module for sbml.
* Adding support for all the languages.
* Update var/spack/repos/builtin/packages/sbml/package.py
Co-authored-by: Greg Becker <becker33@llnl.gov>
* Update var/spack/repos/builtin/packages/sbml/package.py
Co-authored-by: Greg Becker <becker33@llnl.gov>
Co-authored-by: Robert Blake <rob.c.blake.3@gmail.com>
Co-authored-by: Greg Becker <becker33@llnl.gov>
Because of the way abstract variants are implemented, the following
spec matrix does not work as intended:
```
matrix:
- [foo]
- [bar=a, bar=b]
exclude:
- bar=a
```
because abstract variants always satisfy any variant of the same
name, regardless of values.
This PR converts abstract variants to whatever their appropriate
type is before running satisfaction checks for the excludes clause
in a matrix.
fixes#16841
The current package does not work. Also several ocaml versions do
not compile such as 4.09.0. Updated to the new ocaml version to
4.10.0, which is now a prerequisite.
* Add new version of py-astroid
* Add new version of py-lazy-object-proxy
* Apply suggestions from code review
Co-authored-by: Adam J. Stewart <ajstewart426@gmail.com>
* Apply suggestions from code review
Co-authored-by: Adam J. Stewart <ajstewart426@gmail.com>
Update both main and data packages. Data versions have not changed
in new version.
Provide new python variant to build Python bindings for 10.6.2 and
newer. Add dependencies on python and Boost+python required by variant.
- add future-proofing for wmake rules locations:
Accept wmake/rules/{ARCH}{COMP} or wmake/rules/{ARCH}/{COMP}
- compiler option is now '-spack' instead of 'RpathOpt'
which now seems to be a bit harsh on the eyes.
Now have compilations such as 'linux64GccDPInt32-spack',
which is moderately easier to read.
- add OpenFOAM 1912, patch 200506
STYLE: adjust for new flakey8 indentation rules
* openmpi: get rid of implicit system dependencies
* Python 2 compatibility.
* Rename pbspro to openpbs and revert packages.yaml.
* Remove virtual package 'sendmail'.
fixes#16725
The dap, jna, pnetcdf, netcdf4, and ncgen4 variants added in #16047
are _not_ supported by the configure script for netcdf-cxx4 package
(These appear to be configure args for netcdf-c package).
Fujitsu-MPI wrapper commands aren't recognized from 'FindMPI'
function of 'cmake'. If we are using the Fujitsu compiler and Fujitsu MPI,
specify the MPI path information explicitly.
Flux requires `build` for python and many of the python packages because
it builds python bindings. Beyond the bindings, the Flux front-end
commands now use python too, hence the `run` type. Finally, Flux's
`pymod` module is linked against the python interpreter, so the package
requires a `link` dependency on python too.
* IWYU: fix 0.14 build
The CMake patch used for 0.13 hadn't been applied to the master when
0.14 was released, and this version of IWYU requires C++14 or higher.
* Flake8
```
'/var/folders/fy/x2xtwh1n7fn0_0q2kk29xkv9vvmbqb/T/s3j/spack-stage/spack-stage-valgrind-3.15.0-mtir7ubjz7mqmjbb7bogze2qm35hl4ze/spack-src/configure' '--prefix=/ornldev/code/spack/opt/spack/clang-11.0.0-apple/valgrind/mtir7ub' '--enable-only64bit' '--build=amd64-darwin'
1 error found in build log:
43 checking host system type... x86_64-pc-darwin
44 checking for a supported CPU... ok (x86_64)
45 checking for a 64-bit only build... yes
46 checking for a 32-bit only build... no
47 checking for a supported OS... ok (darwin)
48 checking for the kernel version... unsupported (18.7.0)
>> 49 configure: error: Valgrind works on Darwin 10.x, 11.x, 12.x, 13.x, 14.x, 15.x, 16.x and 17.x (Mac OS X 10.6/7/8/9/10/11 and macOS 10.12/13)
```
* ffr: add flag to use fixed format in which the length of one line of the source code is 255 when building with Fujitsu compiler.
* ffr: changed to elif.
The dev branch of UnifyFS now depends on the latest release of
GOTCHA, and will future releases.
This updates our spackage to depend on the correct version of GOTCHA
depending on the version of UnifyFS being installed.
* Create VMD recipe
This is a new recipe to install VMD on Spack-managed hosts.
* Fix lint errors.
* Use plain Package
As per peer-review:
- Use Package to build
- Use configure to create a Makefile
- Use install to copy files to prefix directory
* Move VMD package to correct path, duh...
* Restructure description so first short paragraph can be used by module files.
* Add an empty line as suggested by peer-review. That's how you separate paragraphs.
* Remove extra spaces.
* Use setup_build_environment since that's where you're supposed to export OS variblaes. Thanks to peer-review for spotting this.
* Create VMD recipe
This is a new recipe to install VMD on Spack-managed hosts.
* Fix lint errors.
* Use plain Package
As per peer-review:
- Use Package to build
- Use configure to create a Makefile
- Use install to copy files to prefix directory
* Move VMD package to correct path, duh...
* Add Cubist (#16069)
* Add Cubist
* enhance recipe
* Not using OS module anymore
* remove white space
* Fix build shell
* make Flake8 happy
* use bash shell for build
* Convert it To MakefilePackage as per peer-review
* dbcsr: expose all options, check openblas feats (#16034)
* dbcsr: expose all options, check openblas feats
* dbcsr: use Ninja to build, ensure serialized tests
* dbcsr: add myself as maintainer
* MPark.Variant: GCC 7.3.1 Conflict (#16081)
* MPark.Variant: GCC 7.3.1 Conflict
Due to an ICE in this specific patch-release of GCC, compile
errors in downstream packages should be avoided with a clean
conflict.
* Fix superfluous spaces
Co-Authored-By: Adam J. Stewart <ajstewart426@gmail.com>
* Fix typo
Co-authored-by: Adam J. Stewart <ajstewart426@gmail.com>
* Move VMD package to correct path, duh...
* Add an empty line as suggested by peer-review. That's how you separate paragraphs.
* New matlab versions (#16086)
* Add new version 1.1.1 (#16087)
* New package bonniepp added (#16091)
* openbabel: fix compilation errors (#16090)
- Disable maeparser as it is broken with CMake
- Added missing dependencies
* singularity: updated maintainer list (#16093)
* New version xrootd-4.11.3 (#16092)
* I added Gaussian 16. I also execute bsd/install to fix scripts instead of filtering them.
* revert VMD so only Gaussian is in my PR.
* revert VMD so only Gaussian is in my PR.
* revert VMD so only Gaussian is in my PR.
* I added myself as a package maintainer.
Co-authored-by: asmaahassan90 <31959389+asmaahassan90@users.noreply.github.com>
Co-authored-by: Tiziano Müller <tiziano.mueller@chem.uzh.ch>
Co-authored-by: Axel Huebl <axel.huebl@plasma.ninja>
Co-authored-by: Adam J. Stewart <ajstewart426@gmail.com>
Co-authored-by: Harmen Stoppels <harmenstoppels@gmail.com>
Co-authored-by: Amjad Kotobi <amjadkotbi@gmail.com>
Co-authored-by: athanasio <athanasio@users.noreply.github.com>
Co-authored-by: Carlos Arango Gutierrez <arangogutierrez@gmail.com>
The openssl build process can use the wrong perl for
various reasons, including:
* Wrong value in PERL env var
* The build process first looks for `perl5`, which the
spack system does not provide, but some other
distributions provide it. That way, the build process can
end up using the wrong perl.
Stop all of these problems by explicitly setting PERL to
the to be used dependency.
* Updated versions of COSMA.
* Added an empty line for formatting.
* Switched to sha256.
* Renamed gpu variant to cuda. Extending the CudaPackage base class.
This change also adds a code path through the spack ci pipelines
infrastructure which supports PR testing on the Spack repository.
Gitlab pipelines run as a result of a PR (either creation or pushing
to a PR branch) will only verify that the packages in the environment
build without error. When the PR branch is merged to develop,
another pipeline will run which results in the generated binaries
getting pushed to the binary mirror.
* Fixing the build directory for cardioid.
* These imports are no longer needed due to deletions.
Co-authored-by: Robert Blake <rob.c.blake.3@gmail.com>
* [maloc] created template
* [maloc] added some build dependencies
* [maloc] added homepage and description
* [maloc] added doc variant in an attempt to disable documentation. Still builds doc though... at least when there is a system doxygen found. Having doxygen being a dependency here can bring python in and cause issues for dependents, even when listed as only a build dependency. Maybe this is a bug?
* [maloc] flake8 and fixme removal
* [maloc] shorted description
