* [mfem] A few updates: add 'strumpack' variant; add 'zlib'
variant (same as 'gzstream'); fix optmization flag
for v4.0.
* [mfem] flake8 fix
* [mfem] Add version 4.1
* [mfem] Add/tweak some 'conflicts' directives.
* [gslib] Add new release versions + 'develop' version.
* [petsc] Restrict hdf5 version to <= 1.10.99 since 1.12.0 fails
* [metis] Use the original metis url for v4.0.3.
* [conduit] Remove restrictions to the used hdf5 variant to allow
building with other packages that use hdf5, e.g. petsc.
* [mfem] Few updates:
* Replace the 'gzstream' variant with 'zlib' variant.
* Do not add system library paths with -L flags.
* Allow '+pumi+shared' variant.
* Update the 'test_builds.sh' script.
* [occa] Add version 1.0.9.
* [mfem] Some OCCA and RAJA updates.
* [gslib] Fix the build for new versions of the library.
* [mfem] Add 'gslib' variant for GSLIB.
* [mfem] Add 'cuda' variant.
* [mfem] Add 'libceed' variant + a few more tweaks.
* [mfem] Add 'umpire' variant.
* [ceed] Add a draft for v3.0. Not tested. Just made sure that
concretization works for 'ceed' and 'ceed+cuda'.
* [nek] Fix Nek5000/NekCEM
* [nek] Add Nek5000-v19 & polishing Nek packages
* [flake8] Fix flake8 failure
* petsc: use of HDF5 does not care about +hl+fortran
* [petsc] Temporarily allow any hypre version with petsc@develop.
[ceed] Remove the requirement for hypre@develop.
* [libceed] Do not explicitly set NVCCFLAGS for v0.5 and later.
* [laghos] Add version 3.0, pointing to dev branch for now.
Do not set CXX at the make command line.
Simplify the dependecy directives a little.
[ceed] Use laghos v3.0 for ceed v3.0.0.
* [laghos] Keep the injection of CXX in the makefile for laghos
versions <= 2.0.
* [nekcem] Recovert hash-versions used by older versions of the
'ceed' package.
* [occa] Disable hip autodetection because it fails on some machines.
* [laghos] Update v3.0 with the actual release source.
* [suite-sparse] Explicitly add the c11 flag to CFLAGS.
* Update package.py (#15749)
* [magma] Add forgotten specification of the 'cuda_arch' variant.
* [ceed] Use magma v2.5.3 for ceed v3.0.
* libceed-0.6
* mfem: depend on libceed 0.6:, not 0.6.0:
* [libceed] Add 'magma' variant -- enable MAGMA backend.
* [ceed] In v3.0, use '+magma' variant of libceed when cuda is enabled.
* Initial package for Remhos (needs to be updated with actual sha256
* Adding Remhos to CEED-3.0, for now @develop
* petsc: add 3.13.0 (using petsc-lite) and 3.12.5
* ceed: update to petsc@3.13.0:3.13.99
* Temporary fix
* [nekcem] Add hash-version for ceed v3.0.
* [nek5000] Simplify source urls.
* [nektools] Use the same sources and versions as in nek5000.
* [ceed] Update Nek-related package versions.
* libceed: add v0.6 portabilty fix
* libceed: better v0.6 portabilty fix
* Adding Remhos 1.0 release in CEED-3.0
* Updating hash for Remhos-1.0
* [petsc] Add cuda variant.
* [libceed] Flake8 fix.
* [petsc] Add cuda variant.
* [ceed] Fix the OCCA version to 1.0.9. Enable petsc+cuda when
compiling ceed@3.0.0+cuda.
* nek5000: fix python 2.7+ syntax
* [laghos] Fix testing.
* [remhos] Fix testing.
* [remhos] For testing use the 'tests' target instead of 'test'.
* Add/update the maintainers for ceed, libceed, mfem, laghos, and remhos.
* [ceed] Remove unnecessary dependencies.
* libceed: activate AVX when supported
Co-authored-by: Thilina Rathnayake <thilinarmtb@gmail.com>
Co-authored-by: Jed Brown <jed@jedbrown.org>
Co-authored-by: Stan Tomov <tomov@eecs.utk.edu>
Co-authored-by: Tzanio <tzanio@llnl.gov>
This commit sets the `FORCE_UNSAFE_CONFIGURE` environment variable to 1 in autotools builds.
We see a lot of builds popping up and complaining about `FORCE_UNSAFE_CONFIGURE`. This behavior is not actually part of `autoconf` per se. It comes from this patch to `mknod.m4`, which is used by a lot of autoconf builds:
* https://lists.gnu.org/archive/html/bug-gnulib/2010-07/msg00282.html
Which originated from this problem that someone had on AIX:
* https://lists.gnu.org/archive/html/bug-gnulib/2010-07/msg00279.html
The gist of the problem seems to be that they want to check whether `mknod` can do something as root, but instead of checking whether they're running as root and using `su` or something to test this, they just made it harder to run `configure` as root.
This seems very ad hoc and this is one of many checks that are run as root in `configure`. Many of them run before this check, so it's not clear that the `FORCE_UNSAFE_CONFIGURE` thing is even preventing bad things from happening.
So:
1. This only happens in `autotools` builds, so we should go ahead and put it into `autotools.py` instead of in the global build environment, and
2. The variable does too little and provides a false sense of security in the first place, so we'll just disable it and avoid the nuisance. If we really feel strongly about this we can put some warnings in Spack about running as root, but at the top level, not in the middle of an already running script like `configure`.
* Throw an error at spack install invocation instead of most of the way through the build process when cuda_arch is unspecified.
* Clean-up of CMake booleans. No actual change.
* Use CMake variables for hwloc and libelf installation directories and avoid injecting extra flags into CMAKE_CXX_FLAGS
* Conflict should only exist for +cuda variant.
* Biopandas python package
* Update var/spack/repos/builtin/packages/py-biopandas/package.py
Co-Authored-By: Adam J. Stewart <ajstewart426@gmail.com>
* Update var/spack/repos/builtin/packages/py-biopandas/package.py
Co-Authored-By: Adam J. Stewart <ajstewart426@gmail.com>
* remove scipy dependency
Co-authored-by: Adam J. Stewart <ajstewart426@gmail.com>
At least the v2.0.2.0 tar ball contains compiled object files etc, which
cause the build to fail on other architectures (ppc64le in particular), so
this patch adds a `make clean` after configuring first.
* SourceForge: Mirror Mixin
Add a mixing class for direct `CNAME`s to sourceforge mirrors.
Since the main gateway servers are often down, this could reduce
timeouts and fetch errors for sourceforge.net hosted software.
* SourceForge: unspectacular mirror replacement
add mirrors to all sourceforge packages with trivial
download logic.
tested fetch of latest version of each of these packages
with various mirrors before committing.
* SourceForge: xz
the author homepage is chronocially overrun and this is the offical
upload with many mirrors.
* macOS: Fix emacs Linking
Fix linking issue of emacs on macOS (clang and gcc).
Applies the same work-around as conda-forge:
b051f6c928/recipe/build.sh
Homebrew avoids this by linking against the system ncurses lib:
https://github.com/Homebrew/homebrew-core/blob/master/Formula/emacs.rb
* ncurses: fix outdated variant comment
this comment was build on the assumption that gnutls
triggers a termlib dependency in emacs. that's not the
case, ncurses itself depends on termlib when build with
this feature.
libpng still has its sourceforge page but is actively been
developed on github.
since the sourceforge urls are too often down (as seen in
my nightly CI/CD tests), just switch the download source to
GitHub instead.
`DYLD_LIBRARY_PATH` can frequently break builtin macOS software when
pointed at Spack libraries. This is because it takes *higher* precedence
than the default library search paths, which are used by system software.
`DYLD_FALLBACK_LIBRARY_PATH`, on the other hand, takes lower precedence.
At first glance, this might seem bad, because the software installed by
Spack in an environment needs to find *its* libraries, and it should not
use the defaults. However, Spack's isntallations are always `RPATH`'d,
so they do not have this problem.
`DYLD_FALLBACK_LIBRARY_PATH` is thus useful for things built in an
environment that need to use Spack's libraries, that don't set *their*
RPATHs correctly for whatever reason. We now prefer it to
`DYLD_LIBRARY_PATH` in modules and in environments because it helps a
little bit, and it is much less intrusive.
This commit removes the DYLD_LIBRARY_PATH variable from the default
modules.yaml for darwin. The rationale behind deleting this
environment variable is that paths in this environment variable take
precedence over the default locations of libraries (usually the
install path of the library), which can lead to linking errors in some
circumstances. For example, executables intended to link with Apple's
system BLAS and LAPACK will instead link to a spack-installed
implementation (e.g., OpenBLAS), causing runtime errors.
These errors are resolved by instead relying on paths set in
DYLD_FALLBACK_LIBRARY_PATH, which is lower in precedence than default
locations of libraries.
provided (#15662).
Prior to this fix, the checked Spec object would not be populated, and
concretization would fail.
Co-authored-by: Marc Allen <mrcall@amazon.com>
Mesa links against libtinfo so needs to depend on ncurses. It also needs
a little help finding the library directory so an LDFLAGS configure
option is added.
* 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>
* 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
