refers #20040
Before this PR optimization rules would have selected default
providers at a higher priority than default variants. Here we
swap this priority and we consider variants that are forced by
any means (root spec or spec in depends_on clause) the same as
if they were with a default value.
This prevents the solver from avoiding expected configurations
just because they contain directives like:
depends_on('pkg+foo')
and `+foo` is not the default variant value for pkg.
* OpenBLAS: More Precise GCC Conflicts
Add more precise GCC conflicts so e.g. GCC 6 and GCC 7.5 don't fail.
* Compact syntax
Co-authored-by: Adam J. Stewart <ajstewart426@gmail.com>
Co-authored-by: Adam J. Stewart <ajstewart426@gmail.com>
As part of pull request #19452, a patch method was added to the mfem
package to delete byte order marks from 3 mfem source files. These
files first appeared in a stable release of mfem as of version
4.1. Consequently, attempts to install mfem 3.4 or mfem 4.0 fail
because no files exist at the path arguments of the filter_file
commands used to execute this operation. Decorating the patch method
so it runs only on mfem versions 4.1 and later resolves the errors
that were thrown due to files not found.
This commit adds that decorator.
* Qt: add options to disable docs and gui
- Add `~gui` option for minimal build
- Add `+doc` option to install docs, and attempt to disable the implicit
llvm dependency if not
- Removes the 'freetype' option which hasn't worked reliably in qt5, as
many of the gui components implicitly rely on freetype.
- Add and test version 5.15 (and skip qtlocation if disabling opengl)
- Refactor some of the dependency logic
I've tested this on linux with 5.15.2 and 4.8.7 in a couple of different
configurations.
* Address reviewer feedback and correctly disable llvm
* Fix qt doc generation
* py-rosdep: add new package
* setuptools needed at run-time
Co-authored-by: Andrew W Elble <aweits@rit.edu>
Co-authored-by: Andrew W Elble <aweits@rit.edu>
* py-rospkg: add new package
* setuptools needed at run-time
Co-authored-by: Andrew W Elble <aweits@rit.edu>
Co-authored-by: Andrew W Elble <aweits@rit.edu>
* py-catkin-pkg: add new package
* setuptools is needed at run-time
Co-authored-by: Andrew W Elble <aweits@rit.edu>
Co-authored-by: Andrew W Elble <aweits@rit.edu>
fixes#19981
This commit adds support for target ranges in directives,
for instance:
conflicts('+foo', when='target=x86_64:,aarch64:')
If any target in a spec body is not a known target the
following clause will be emitted:
node_target_satisfies(Package, TargetConstraint)
when traversing the spec and a definition of
the clause will then be printed at the end similarly
to what is done for package and compiler versions.
* spack recipe for gromacs with aocc compiler support
Change-Id: I364aab4a0aa2dcd44bc47eb50c81b2d94c99cfbd
* Removed arch and other associated compilers flags
Added cycle_subcounters variant
Co-authored-by: vkallesh <Vijay-teekinavar.Kallesh@amd.com>
fixes#20019
Before this modification having a newer version of a node came
at higher priority in the optimization than having matching
compilers. This could result in unexpected configurations for
packages with conflict directives on compilers of the type:
conflicts('%gcc@X.Y:', when='@:A.B')
where changing the compiler for just that node is preferred to
lower the node version to less than 'A.B'. Now the priority has
been switched so the solver will try to lower the version of the
nodes in question before changing their compiler.
* llvm-amdgpu: fix the build for version 3.9.0
Adapt the fix-system-zlib-ncurses.patch for version 3.9.0. Without
the patch, llvm-amdgpu builds, but then rocm-device-libs fails with
"cannot find -ltinfo."
Tighten the version requirements for cmake according to the
llvm/CMakeLists.txt file.
* Add a conflict for cmake 3.19.0.
