Age | Commit message (Collapse) | Author | Files | Lines |
|
* acts: new version 19.3.0
* acts: new variants +profilecpu, +profilemem
|
|
* py-cartopy: add v0.20.3
* Add py-setuptools-scm 7.0.3
|
|
Co-authored-by: Tamara Dahlgren <35777542+tldahlgren@users.noreply.github.com>
|
|
|
|
|
|
* Initial implementation of Omnitrace package
* Fix flake8 errors
* Fix run environment when +python
* String normalization and fix for build env when +tau
* Tweak to style
|
|
Fix compilation of @3.14.0+cuda (didn't check older versions).
|
|
Release branches and tags run protected pipelines, and we noticed
that those pipelines were generating all jobs in the stack, even
when the mirror contained all the built specs and an up to date
index. The problem was caused because the override mirror was
not present in spacks mirror configuration at the time when the
binary_distribution.update() method was called. This fixes that
by always adding the mirror override, if present, to the list of
configured mirrors.
|
|
* py-altair: new package
* Update var/spack/repos/builtin/packages/py-altair/package.py
Co-authored-by: Adam J. Stewart <ajstewart426@gmail.com>
* Update var/spack/repos/builtin/packages/py-altair/package.py
Co-authored-by: Adam J. Stewart <ajstewart426@gmail.com>
* Update var/spack/repos/builtin/packages/py-altair/package.py
Co-authored-by: Adam J. Stewart <ajstewart426@gmail.com>
* py-altair: update
* Update var/spack/repos/builtin/packages/py-altair/package.py
Co-authored-by: Adam J. Stewart <ajstewart426@gmail.com>
Co-authored-by: Adam J. Stewart <ajstewart426@gmail.com>
|
|
|
|
|
|
|
|
* boost: do not access the context-impl variant for versions below 1.65.0
* boost: check if spec.variants contains context-impl
* boost: improve error message when the context-impl variant causes a conflict
Executing spack solve boost@1.63.0 +context context-impl=fcontext
triggers the following error message:
==> Error: No version for 'boost' satisfies '@1.63.0' and '@1.79.0'
With this change, the error message becomes the following:
==> Error: Cannot set variant 'context-impl' for package 'boost' because the variant condition cannot be satisfied for the given spec
|
|
|
|
We adopted the convention of putting binaries for each stack into
a dedicated mirror named after the directory in which the stack
(spack.yaml file) resides. This fixes the mirror url of the
radiuss-aws-aarch64 stack to follow that convention.
|
|
* py-cmseq: new package
* py-cmseq: fixed copyright year
* py-cmseq: changed download to pi, added deps
|
|
|
|
`make_target` can be used to instruct Spack to build one of the pre-defined make
targets in the MPAS makefile. Spack will guess the targret based on the value of
`spack_fc`, but the user can overwrite the target as variant. E.g.
```
spack install mpas-model make_target=pgi-nersc
```
`precision` is used to set the `PRECISION` flag in the Makefile to {single,
double}. Default is double.
Co-authored-by: Stephen Sachs <stesachs@amazon.com>
|
|
Co-authored-by: Stephen Sachs <stesachs@amazon.com>
|
|
|
|
|
|
Co-authored-by: Stephen Sachs <stesachs@amazon.com>
|
|
Co-authored-by: Stephen Sachs <stesachs@amazon.com>
|
|
|
|
|
|
* remove unhelpful comment
* Filter compiler duplicates while reading manifest
* more-specific version matching edited to use module-specific version (to avoid an issue where a user might add a compiler with the same version to the initial test configuration
|
|
* cutlass: new package
* cutlass: fix ci
* cutlass: remove `join_path`
Co-authored-by: Tamara Dahlgren <35777542+tldahlgren@users.noreply.github.com>
* cutlass: adding a message for cuda_arch
Co-authored-by: Tamara Dahlgren <35777542+tldahlgren@users.noreply.github.com>
* cutlass: simplify cmake_args
Co-authored-by: Tamara Dahlgren <35777542+tldahlgren@users.noreply.github.com>
* cutlass: fix cmake_args
Co-authored-by: Tamara Dahlgren <35777542+tldahlgren@users.noreply.github.com>
|
|
(#28270)
|
|
* Set R_LIBS_USER='' in dependent build environment
Despite R packages being installed with the --vanilla flag, which
ignores Rprofile and Renviron files, R will still set R_LIBS_USER if the
default directory exists. This could lead to pulling in packages from
that directory during the build which could cause the build to fail. To
avoid that, explicitly set R_LIB_USER='' to ensure that packages from
the HOME/R directory are not in the library path.
* Update var/spack/repos/builtin/packages/r/package.py
Co-authored-by: Adam J. Stewart <ajstewart426@gmail.com>
Co-authored-by: Adam J. Stewart <ajstewart426@gmail.com>
|
|
|
|
|
|
* PythonPackage: add default libs/headers attributes
* Style fix
* libs and headers should be properties
* Check both platlib and include
* Fix variable reference
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
fixes #30965
|
|
|
|
|
|
* new package: py-lark to replace py-lark-parser
* update packages that need py-lark-parser and use py-lark instead
|
|
|
|
* add new py-ats version
* fix copyright date
|
|
|
|
|
|
* Add py-torch v1.12, py-torchvision v0.13
* Style fix
|
|
|
|
* Add AWS RADIUSS builds
* Correct variable naming
* Add two more MFEM specs
* Updates to MFEM spec suggested by @v-dobrev
* Simplify MFEM specs
|