Spack generally ignores file-file projection clashes in environment
views, but would eventually error when linking the `.spack` directory
for two specs of the same package.
This leads to obscure errors where users have no clue what the issue is
and how to fix it. On top of that, the error comes very late, since it
happens when the .spack dir contents are linked (which happens after
everything else)
This PR improves that by doing a quick check ahead of time if clashes
are going to be anticipated (by simply checking for clashes in the
projection of each spec's .spack metadir). If there are clashes, a
human-readable error is thrown which shows two of the conflicting specs,
and tells users to user unify:true, view:false, or set up custom
projections.
* add pytng
* black
* add setuptools
* fix
* Update var/spack/repos/builtin/packages/py-pytng/package.py
Co-authored-by: Adam J. Stewart <ajstewart426@gmail.com>
* Update var/spack/repos/builtin/packages/py-pytng/package.py
* Update var/spack/repos/builtin/packages/py-pytng/package.py
Co-authored-by: Adam J. Stewart <ajstewart426@gmail.com>
---------
Co-authored-by: Adam J. Stewart <ajstewart426@gmail.com>
Kokkos when compiled by spack without +wrapper could potentially capture the spack compiler wrappers, resulting in cmake configs and kokkos_launch_compiler trying to run the spack compiler wrapper after installation.
The checksum exception was not detailed enough and not reraised when using cache only, resulting in useless error messages.
Now it dumps the file path, expected
hash, computed hash, and the downloaded file summary.
Batch scripts in general will not function without carriage return line
endings on Windows. We rely on these scripts to support cmd, so we
should not allow these scripts to be converted to lf.
Note: Windows 11 supports lf line endings due to the use of Windows
Terminal. Once support for Windows 10 is dropped, this change can be
reverted.
When running many concurrent spack install processes that need to write
to the db, Spack regularly times out. This is because writing to the DB
after another process has written to it requires deserialization of the
db, mutating it in memory, and serializing it again, which takes some
time. On top of that, I believe there's a 1 second retry when a write
lock cannot be obtained, so I think this means only 3 processes can
really write to the DB at the same time before timing out.
* Style: black 23, skip magic trailing commas
* isort should use same line length as black
* Fix unused import
* Update version of black used in CI
* Update new packages
* Update new packages
* Update package.py
Initial new stuff
* Update package.py
* Update package.py
* Update package.py
* fix targets
* non-llvm backends
* ooops
* fix style
* Somehow that was not caught?
Somehow that was not caught?
* style
* Last fix
make capitalization consistent with Halide not LLVM...
