* CGNS: Allow use of git branch versions
Add ability to use git `master` and `develop` branches of the CGNS library. The `seacas` package sometimes needs to use the `master` or `develop` branch due to some new changes to the parallel I/O implementation in CGNS which has not yet been released.
* CGNS: Reorder to put newest version on top
Based on review comments, place newest version (develop) first in list.
- some commands were missed in the rollout of spack environments
- this makes all commands that need to disambiguate specs restrict the
disambiguation to installed packages in the active environment, as
users would expect
* dyninst: patch to build dyninst with older gcc
Add 'v9.3.2-auto.patch'. This patch changes some 'auto t: type' usage
to the older but equivalent 'type.begin()'. This allows building
dyninst 9.3.2 with gcc 4.4 which doesn't support the newer syntax.
This patch is harmless with newer gcc.
* Fix typo in patch. (My bad for sloppy cut-and-paste.)
* Restrict the patch to gcc 4.7 and earlier. gcc 4.8 supports the newer
usage and thus doesn't need the patch.
* proper include & linking of libarchive
* due to linux headers, mfu can't build on mac
* hard code deps in configure line
rather than relying on pkg-config
* +xattr during config
* must also link the lib
* find mfu.h
* long lines
* Add thread count variant
Atlas automatically configures itself and optimizes to run with a fixed number of threads. This can be overridden during compile by setting a flag and atlas will tune itself to use the specified number, rather than all system CPUs.
Default value is the existing logic, autoconfigure.
* Fix Linting
* Another lint for longline
* Added rocksdb package
* Fixes based on @tristan0x and @pramodk comments
* Cleanup
* A couple minor fixes based on @adamjstewart 's advice
* Fixed flake8 errors
- previous removal of FOAM_EXT_LIBBIN (ThirdParty) from the environment
during the build could lead to the system scotch library being found
instead of the spack installation.
Avoid this largely cosmetic change to the setting in favour of
improved build robustness. These modifications will be largely
superfluous in the 1812 release anyhow.
- Split the patching algorithm into 'regular' and '@:1806' since the
next versions of openfoam will again need fewer patches.
- Add hook into foamCreateManpage to be used where available
* SEACAS: Add preliminary support for seacas package
* Eliminate leftover from package I copied from.
* And remove another stray line
Sorry, thought I had checked this better...
* SEACAS: Shorten long lines; fixup indentation
* SEACAS: Remove unused os import
* Add PLASMA version 18.11.0 and Lua variant (#9689)
* Add PLASMA 18.11.1 with optional Lua and test for headers
* New versions for a few python modules (#9969)
* py-jupyter-console: Add version 5.2.0
* py-jupyter-core: add version 4.4.0
* py-lxml: add version 4.2.5
* py-zmq: add version 17.1.2
* py-terminado: add version 0.8.1
* py-pexpect: add version 4.6.0
* py-pytables: add version 3.4.4
* SEACAS: Default to mpi on
* SEACAS: Modify to reduce variants
Instead of using an `allpkgs` or explicitly specifying each individual subset,
changed to minimize variants:
* By default, the exodus and IOSS libraries are built and a few
executables directly related to the IOSS library.
* variant `common` defaults to True and builds the "common" seacas
applications. This is based on my definition, so there may be
could be some discussion on what belongs here.
* variant `legacy` defaults to True and builds the remainder of the
seacas applications. These are still used, but mainly in more
esoteric areas. Might be better to default to False, but I wanted
default to just build everything which is what other SEACAS builds
always do...
Removed the `kokkos` variants for now as I need to do some more
thinking on how best to support this. Not often used currently,
so shouldn't be an issue.
Cleaned up other issues raised by Adam and found by me.
* SEACAS: Address flake8 issues
* SEACAS: Clean up variant package logic
Fixed package variant logic.
Consolidated some other variant handling logic
Other minor refactors for hopefully better readability
* SEACAS: Fix flake8 error
Missing whitespace around operator fixed.
* GDL: python integration fixes
* renamed python-related variants to follow the convention
* building the Python module requires patches currently targetting 0.9.8 othwerwise asking for the Python module *only* builds the Pyhton module
* building the python module also requires patching the vendored (with the GDL) antlr to be built as a shared library
* Typo
Co-Authored-By: rmsds <ricardo.d.silva@gmail.com>
* Rename embed-python variant to embed_python
* zoltan: only add gussed MPI libs if no MPI wrappers are used
- if MPI-wrappers are used for compilation, we can assume that
linking works without manually specifying MPI libs
(guessing may result in wrong libs, cf. #8979)
- thus, only guess the NPI libs and add them explicitly if no
MPI-wrappers are used
- use llnl.util.filesystem.find_libraries instead of a locally
defined routine to guess the MPI libs if needed
(cf. #8979)
* zoltan: rely on MPI-wrappers to know the required MPI libs
* This fixes a number of bugs:
* Patches were not properly downloaded and added to mirrors.
* Mirror create didn't respect `list_url` in packages
* Update the `spack mirror` command to add all packages in the
concretized DAG (where originally it only added the package specified
by the user). This is required in order to collect patches that are specified
by dependents. Example:
* if X->Y and X requires a patch on Y called Pxy, then Pxy will only
be discovered if you create a mirror with X.
* replace confusing --one-version-per-spec option for `spack mirror create`
with --versions-per-spec; support retrieving multiple versions for
concrete specs
* Implementation details:
* `spack mirror create` now uses regular staging logic to download files
into a mirror, instead of reimplementing it in `add_single_spec`.
* use a separate resource caching object to keep track of new
resources and already-existing resources; also accepts storing
resources retrieved from a cache (unlike the local cache)
* mirror cache object now stores resources that are considered
non-cachable, like (e.g. the tip of a branch);
* the 'create' function of the mirror module no longer traverses
dependencies since this was already handled by the 'mirror' command;
* Change handling of `--no-checksum`:
* now that 'mirror create' uses stages, the mirror tests disable
checksums when creating the mirror
* remove `no_checksum` argument from library functions - this is now
handled at the Spack-command-level (like for 'spack install')
* Store ccache directory explicitly in Travis.
Despite we started using ccache on `develop`, it seems the cache itself
is not stored from one CI build to the next. This might be du to the
fact that our language on Travis is Python and not C nor C++.
Hence here we store the ccache directory explicitly.
