* Fix scikit-learn build with Apple clang
* Update compiler unit tests
* Fix unit tests
* Fix OpenMP dep
* recursive=True no longer necessary
* Add myself as a maintainer
* Specify which versions require OpenMP
* Typo fixes
* LLVM-OpenMP is the correct dependency, no OMPT
* Flake8 fix
* Undo Apple Clang OpenMP flag change, split into a separate PR
Later versions of libjpeg-turbo build with CMake; to build with
user-specified cflags, the user must supply these to CMake as
-DCMAKE_C_FLAGS (Spack's typical approach of injecting these flags
into the compiler wrapper invocation is insufficient in this case).
Currently libjpeg-turbo cannot be implemented as a CMakePackage
(and thereby take advantage of the flag_handler implementation it
provides) because not all versions of libjpeg-turbo use CMake, so
this adds a custom implementation of flag_handler and
flags_to_build_system_args to libjpeg-turbo.
* cmake: change c++ spec when building with Fujitsu compiler.
* cmake: define cxxflags using a Fujitsu compiler's method.
* cmake: Add condition of whether to add cxx11 option.
* from_sourcing_file: fixed a bug + added a few ignored variables
closes#7536
Credits for this change goes to mgsternberg (original author of #7536)
The new variables being ignored are specific to Modules v4.
* Use Spack Executable in 'EnvironmentModifications.from_sourcing_file'
Using this class avoids duplicating lower level logic to decode
stdout and handle non-zero return codes
* Extracted a function that returns the environment after sourcing files
The logic in `EnvironmentModifications.from_sourcing_file` has been
simplified by extracting a function that returns a dictionary with the
environment one would have after sourcing the files passed as argument.
* Further refactoring of EnvironmentModifications.from_sourcing_file
Extracted a function that sanitizes a dictionary removing keys that are
blacklisted, but keeping those that are whitelisted. Blacklisting and
whitelisting can be done on literals or regex.
Extracted a new factory that creates an instance of
EnvironmentModifications from a diff of two environments.
* Added unit tests
* PS1 is blacklisted + more readable names for some variables
It seems that GitHub's output changed slightly for this patch (one more
letter for both short hashes on the index line), probably due to the
number of objects in the repository.
All documentation mentions that `build_jobs` is limited by the number of
cores available in the system. This is also enforced when setting it via
`--jobs`. However, when setting it via `config.yaml`, it can exceed the
number of cores available, making builds run out of memory.
This PR adds the ability to specify the auto-dispatch targets that can
be used by the Intel compilers. The `-ax` flag will be written to the
respective compiler configuration files. This ability is very handy when
wanting to build optimized builds for various architectures. This PR
does not set any optimization flags, however.
Fixes#3690Fixes#5637
Uninstalling dependents of a spec was relying on a traversal of the
parents done by inspecting spec._dependents. This is in turn a
DependencyMap that maps a package name to a single DependencySpec object
(an edge in the DAG) and cannot thus model the case where a spec has
multiple configurations of the same parent package installed (for
example if different versions of the same Python library depend on
the same Python installation).
This commit works around this issue by constructing the list of specs to
be uninstalled in an alternative way, and adds tests to verify the
behavior. The core issue with DependencyMap is not resolved here.
The default library search for a package checks the lib/ and lib64/
directories for libraries before the root prefix, in order to save
time when searching for libraries provided by externals (which e.g.
may have '/usr/' as their root).
This moves that logic into the "find_libraries" utility method so
packages implementing their own custom library search logic can
benefit from it.
This also updates packages which appear to be replicating this logic
exactly, replacing it with a single call to "find_libraries".
Fixes#11782
Spack was not properly resolving relative paths to absolute paths
when a relative path was passed to "spack compiler add [PATH]".
Now, if provided a relative path, the absolute path is written to
compilers.yaml rather than the relative path.
* Add template creation test
* Added --skip-editor option to "spack create": normally
"spack create" opens an editor for the user after generating a
package file; when the --skip-editor option is used, "spack create"
only generates the package file and does not open an editor
* Added --skip-editor option to bash completion
- `setup-env.sh` was not properly detecting a bash shell when bash was run
as /bin/sh.
- Detection routine now always reports bash when bash is run as sh, and
no longer parses the path to the executable indicated in `$BASH`.
- Namepsaces were shown without dots after the new format strings were
added.
- Add a test for `spack find` to ensure that find -N shows the right
output.
This PR has several modifications for the r package.
- The tk package is always depended on but this pulls in X11, making the
'X' variant non-functional. This PR sets a dependency of tk on '+X'.
- There is a missing dependency on libxmu when '+X' is set.
- The libraries for R wind up in a non-standard location and are thus
left out of the RPATH settings. This PR adds that directory to RPATH.
- The MKL integer interface for gfortran is not in the BLAS libs. This
PR replaces the intel interface with the gfortran interface if needed.
- Use the `LibraryList` `ld_flags` method for blas as that is more in
line with th R Installation and Administration manual.
Note that this PR depends on PR #11956. This PR closes#8642.
