powerpc/ppc64: unmask media-libs/libva #37198
Conversation
Pull Request assignmentSubmitter: @a17r @gentoo/github Linked bugsBugs linked: 934215 New packagesThis Pull Request appears to be introducing new packages only. Due to limited manpower, adding new packages is considered low priority. This does not mean that your Pull Request will not receive any attention, however, it might take quite some time for it to be reviewed. In the meantime, your new ebuild might find a home in the GURU project repository: the ebuild repository maintained collaboratively by Gentoo users. GURU offers your ebuild a place to be reviewed and improved by other Gentoo users, while making it easy for Gentoo users to install it and enjoy the software it adds. In order to force reassignment and/or bug reference scan, please append Docs: Code of Conduct ● Copyright policy (expl.) ● Devmanual ● GitHub PRs ● Proxy-maint guide |
gentoo-bot
added
need assignment
Pull request CI reportReport generated at: 2024-06-17 17:41 UTC There are existing issues already. Please look into the report to make sure none of them affect the packages in question: |
| # Georgy Yakovlev <[email protected]> (2019-02-04) | ||
| # Temurin ppc64 build is little-endian only | ||
| # so we can use bins. | ||
| -dev-java/openjdk-bin |
There was a problem hiding this comment.
this was reported as UnmatchedProfilePackageUnmask, so I assumed it was free to cleanup.
There was a problem hiding this comment.
rechecked all the stuff, and all those unmasks are "anti mask" for ppc64be, but for those we cleaned up the arches well, so no need for those unmasks.
Thank you.
There was a problem hiding this comment.
please for now move those masks you drop here to profiles/arch/powerpc/ppc64/64ul/package.mask (meaning for big endian ppc64) until further verification of libva working on big endian.
ppc64le is much stabler and usable than ppc64be.
There was a problem hiding this comment.
it seems they are already masked on 64ul.
There was a problem hiding this comment.
similary, please move those masks to the big endian profile: profiles/arch/powerpc/ppc64/64ul/package.use.mask until further investigation.
|
Pinged powerpc users (in #gentoo-powerpc IRC) if they know how well libva works on big endian ppc64. |
Mask taken from powerpc/ppc64/64le/package.mask, inverted. Signed-off-by: Andreas Sturmlechner <[email protected]>
Mask taken from powerpc/ppc64/64le/package.use.mask, inverted. Signed-off-by: Andreas Sturmlechner <[email protected]>
Bug: https://bugs.gentoo.org/934215 Signed-off-by: Andreas Sturmlechner <[email protected]>
Signed-off-by: Andreas Sturmlechner <[email protected]>
Pull request CI reportReport generated at: 2024-06-17 19:31 UTC There are existing issues already. Please look into the report to make sure none of them affect the packages in question: |
|
@a17r after some checks with powerpc users, libva somewhat works and somewhat not on powerpc64 big-endian. I think this is good enough to simplify the profiles and just unmask it across ppc64 profiles. |
Prior to this PR, a package was masked in ppc64ul (both endians) and then unmasked in ppc64le. This had the intended effect of masking it only on ppc64 (be). In this PR, that work was undone. At least one unmask was removed from ppc64le, while the mask remained in ppc64 (be) - https://bugs.gentoo.org/925111#c8 Was the pkgcheck tool reporting UnmatchedProfilePackageUnmask incorrectly? If so, please follow up with a bug. Also you said that even though a package is not working on ppc64 (be), it can be unmasked. The reasoning is hard to follow. |
I already did at pkgcore/pkgcheck#685 for https://bugs.gentoo.org/918647#c5.
I covered that in the commit message (86238ae). Anyway, please file a bug for the broken unmasks. |
|
I'll just revert it now as matoro asked (now done in 5429444). |
No description provided.