| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
|
|
|
|
* Initial attempt at flann packaging.
Python2 somehow works. Python3 does not. Still debugging their
obscure setup.py configuration.
* Flann good enough. Python3 broken but close.
Flake8 checks in place. Unsure about CMAKE_BUILD_TYPE and default
spack behavior.
* spack uses RelWithDebInfo as default build type.
* builds py2/3, but direct site-packages install
* prefix working, empty python install dir
* flann package +python installs correctly
* str format {0} instead of {}
* potential doctest fix
* consistency of build env with PythonPackage
* fix python again, test deptype todo, build type
* potentially enable matlab, untested
|
|
|
|
|
|
PR #3367 inadvertently changed the semantics of _find_recursive and
_find_non_recursive so that the returned list are not ordered as the
input search list. This commit restores the original semantic, and adds
tests to verify it.
|
|
|
|
Signed-off-by: Howard Pritchard <howardp@lanl.gov>
|
|
|
|
|
|
* Edits to get setup-env.csh working better.
Autosets the sys_type a la setup-env.sh
* More stealing from bash setup script for module roots
* Add error message if SPACK_ROOT isn't set
* Remove _sp_lmod_root per Adam J Stewart
|
|
* cosmomc: Enable Python bindings
* cosmomc: Make python dependency optional
|
|
* hdf5: Correct spelling
* hdf5: avoid trailing spaces
|
|
* highfive: New package; this is a C++ wrapper for the HDF5 library
The C++ wrapper that is shipped with the HDF5 library has several shortcomings. The highfive library seems to avoid these.
* highfive: Make MPI variant default to true
|
|
* Add latest version of PGI compilers
* Prepend, don't overwrite
|
|
* version updates, remote switch
* trim versions to original and most recent
* proper url discovered and tested
|
|
* enable cuda support for suite-sparse
* do not use spec.satisfies
* cuda restructure, give clearer comment of why
* str format compatibility
* flake8 checks
|
|
|
|
* Fix newline character in GCC package
* Remove the newline completely
|
|
* adding 0.5.1 and variant for uint8 bit stream word type
* Revisions as per davydden
* fixing style (flake8) issues
* - switched to derive from MakefilePackage
- fixed typo in variant check
- Confirmed installs correctly with bswtuint8 variant
* added build method; changed zfp_incdir to incdir
* * Adding uber fastmath package to support 'spack install fastmath'
* Adding dims variant boxlib to control compile-time spatial dimension count
* Explicitly disabling many parts of moab to get lib to build. Don't need tools yet.
* Add logic to spack/setup-env to not refer to $SYS_TYPE if its not defined.
* adding mesquite package
* * Added Chombo (still working on Fortran name mangling)
* Made mesquite depend on mpi
* Fixed use of boxlib's dims numerical variant
* adding PUMI
* flake8 compliance
* adding phasta
* fixing flake8 issues
* undue $SYS_TYPE protection change
* fixing install step for fastmath uber package; resolve issues requested by adamjstewart in PR
* revisions as per adamjstewart
* re-enabling trilinos and phasta dependencies
* all changes requested modulu values= and assert statement
* fixing cmake_args usage in pumi; fixing lib installs for chombo
* first pass at numerical variant
* fixing dims variant as per @adamjstewart
* fixing func/var name collision
* fixing means of disabling -Werror
* fix name collision for cmake_args var/func
* fixing chombo homepage/url; make FASTMath use chombo#3.2; remove extraneous comment about boxlib dim
* remove dummy fastmath package; switch to @BarrySmith soln for install of dummy bundle
* fix flake8 import os issue
* remove extraneous comment
* adding url for fastmath scidac site
* final fixes as per @adamjstewart
* adding 1 dims option for boxlib
* switching make(all) to gmake(all)
|
|
|
|
Make openmpi handle bad ibv_devices return statuses without bailing out completely.
See #4162 for background.
|
|
Otherwise, the install can fail if the system-provided `tic` binary is too old.
|
|
|
|
|
|
Also use a Spack-provided GoogleTest.
|
|
|
|
cp2k: fixed compilation issues for intel stack
|
|
* Travis version of pip is too old
* Also need to upgrade setuptools
|
|
Added DFLAGS to the `make.inc` file being written.
These macros are also added to the language specific variables
like CFLAGS, CXXFLAGS and FCFLAGS. Changed `spec.satisfies('foo')`
with `'foo' in spec` in `intel-mkl`, see #4135. Added a basic
build interface to `intel-mpi`.
|
|
* Add GCC 7
* Allow users to build subset of GCC compilers with multi-value variant.
* Add comment explaining what 'all' actually means
* More specific golang support
* Only require Zip when building Java
* Comment out provide directives that don't currently work
* Add Ada support
* Use conflicts directive
* Fix joining of languages
* Need special flag to build jit
* Explicitly declare GNAT download extension
* Import tty, update lib64 to lib
* BRIG and Go are not supported on macOS
* Simplify formatting and imports
* JIT patch required for newer versions as well
|
|
It seems that parse_anonymous_spec may fail if more than one part
(variant, version range, etc.) is given to the function. Added tests to
code against to fix the problem in #4144.
|
|
- Full help is now only generated lazily, when needed.
- Executing specific commands doesn't require loading all of them.
- All commands are only loaded if we need them for help.
- There is now short and long help:
- short help (spack help) shows only basic spack options
- long help (spack help -a) shows all spack options
- Both divide help on commands into high-level sections
- Commands now specify attributes from which help is auto-generated:
- description: used in help to describe the command.
- section: help section
- level: short or long
- Clean up command descriptions
- Add a `spack docs` command to open full documentation
in the browser.
- move `spack doc` command to `spack pydoc` for clarity
- Add a `spack --spec` command to show documentation on
the spec syntax.
|
|
* qt: add 5.8.0 which builds on Sierra with clang 8.1.0
* enable parallel build, add freetype dependency
* minor
* minor
|
|
Remove python dependency for libxcb and xcbproto
|
|
Seen in https://travis-ci.org/LLNL/spack/builds/229484526, very likely
due to a problem in the Travis builder.
|
|
|
|
The conditionals are repeated in the statement before _add_dependencies
|
|
* adding paramiko and missing dependencies, setup to work with #2548
* adding other deps for paramiko
* fix flake8 errors
* removed spurious add
* address suggestion for proper dependencies
* fix cryptography deps
* remove FIXME comments and commented depends lines
|
|
|
|
|
|
|
|
* Add list_url to mpfr
* Switch to download directly from GNU
* http -> https
|
|
* Updating bamtools to include a dependency for zlib.
In a standard compile, bamtools will fail if zlib headers are not installed on the target machine. In order to maintain compatibility with all systems -- and since zlib is included already as a dependency for cmake -- this patch adds zlib as a link dependency for the bamtools package.
* Modified cmake rpath include.
Bamtools has a non-standard library location, so we need to append $prefix/lib/bamtools to the rpath. Not sure there's a better way to do this...
* Fixing syntax error in package.py
Fixed a non-terminated parenthesis on line 46.
* Updated bamtools to be a CMakePackage
Removed extraneous code, and altered the package to extend cmake_args
including the non-standard library location.
* UpRemoving cmake dependency and removing blank line from end of file
* Updates to cmake_args.
Removed the duplicate definition of std_cmake_args in favor of simply overriding the CMAKE_INSTALL_RPATH variable that is provided. This should allow the package to be linked correctly to itself.
|
|
* Fix -include issue with intel and pgi in zlib
MERGE CANDIDATE
* Patch zlib to compiler with the cray compiler suite.
|
|
|
|
What the user wants, the user gets....
|
|
|
|
|
|
|
