solver: requirements on virtual can influence provider weights#51282
Merged
haampie merged 5 commits intospack:developfrom Sep 10, 2025
Merged
solver: requirements on virtual can influence provider weights#51282haampie merged 5 commits intospack:developfrom
haampie merged 5 commits intospack:developfrom
Conversation
Signed-off-by: Massimiliano Culpo <[email protected]>
Before this change, provider weights were determined just by package preferences. This could cause issues in cases where the preferred, or required, provider is not the default package preference. Signed-off-by: Massimiliano Culpo <[email protected]>
Member
Author
Signed-off-by: Massimiliano Culpo <[email protected]>
haampie
reviewed
Sep 9, 2025
haampie
reviewed
Sep 9, 2025
Member
|
Needs a bit more work to support It should drop the conflicting providers from the list of preferred ones to start counting at 0 for these penalties. |
Signed-off-by: Massimiliano Culpo <[email protected]>
Signed-off-by: Massimiliano Culpo <[email protected]>
Contributor
|
@alalazo Things are looking better with this PR. As a reminder, these are the four versions I tried. Before this PR, v001 and v003 were bad, v002 and v004 were good. Results: v001 is still bad (ok), v002/v003/v004 are good. There are slight differences between the concretization of v002 and v003/v004, but v003 and v004 are now identical |
Member
Author
|
I'll check the other issue in another PR. Currently we hardcoded a translation of |
Contributor
Personally, I'd be ok with exiting with an error when running into the legacy syntax given that the new syntax works with your changes, but of course that's up to you and the other spack developers. Thanks for fixing this! |
Member
|
You could compare the minor differences between v002 and v003/v004. The latter should now be superior, because the preference for a dependency on oneapi makes the concretizer toggle variants to minimize the number of "data only" packages (that do not depend on any compiler). |
Contributor
Yes, I am happy with v003/v004, I don't think v002 is any better. In v002, it chooses to use a different libbsd that doesn't depend on libmd, but that doesn't affect us. |
Member
|
This PR does not apply cleanly to 1.0 and will be dropped from the 1.0.3 release |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
fixes #51267
Before this change, provider weights were determined just by package preferences. This could cause issues in cases where the preferred, or required, provider is not the "default" provider.
This PR changes the ASP setup to take into account the most common cases of requirements and strong preferences, so that the provider weights are set accordingly.