media/gfx-cura-4.6.1 : Fix Startup Crash at Runtime #17679
Conversation
gentoo-bot
added
assigned
digitall
force-pushed
the
20200926-cura-fixup
branch
from
a7e4493
to
077942e
Compare
|
I will look at adding a futher Pull Request for cura v4.7.0+ shortly... |
|
I also see that the gentoo-bot doesn't update the "no signoff" label on a force-push to fix the commits :/ |
|
Right have unmerged cura and done a depclean to remove all dependencies and tried a full re-emerge of "emerge -av cura" with successful build and startup. Two minor issues:
|
|
It was easier to add the cura v4.7.0 ebuild to this PR and this also provides another way of solving the overall issue i.e. currently the portage tree has no working version of cura... until / unless this is merged. Also, the @gentoo-repo-qa-bot broken status is likely spurious as it relates to missing dependencies when building on ARM64 which was likely the case with the builds prior to my changes... Anyway, will now await human comment :) |
|
Related... https://bugs.gentoo.org/731274 |
|
The label gets updated if you add '[please reassign]'. The issue on arm64 is that libarcus-1.6.1 isn't keyworded on it. So drop keywords for arm64 on cura, then file a bug to re-add it once this is merged. Please also squash the QA fix(es) with git rebase? |
|
@thesamesam : OK. Will squash the v4.7.0 QA fix into the previous commit with rebase... Will check the other changes. [please reassign] |
digitall
force-pushed
the
20200926-cura-fixup
branch
from
7caa4f0
to
7b26713
Compare
|
(Sorry, you have to put it in the PR title -- I wasn't very clear!) |
|
@thesamesam : Ah OK! |
digitall
changed the title
gentoo-bot
changed the title
Pull Request assignmentSubmitter: @digitall dev-python/uranium: @gentoo/3dprint, @gentoo/python, @alexxy, @gentoo/proxy-maint, @MathyV, @Perlovka Linked bugsNo bugs to link found. If your pull request references any of the Gentoo bug reports, please add appropriate GLEP 66 tags to the commit message and request reassignment. If you do not receive any reply to this pull request, please open or link a bug to attract the attention of maintainers. 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
assigned
Pull request CI reportReport generated at: 2020-09-28 19:21 UTC There are existing issues already. Please look into the report to make sure none of them affect the packages in question: |
|
@thesamesam : All good now, prior to review? |
This reverts commit 9e3189e. Removing these and the associated ebuild for v4.6.1 broke Cura v4.6.1 which was the most recent in the portage tree, but only at runtime. Signed-off-by: David G Turner <digitall@scummvm.org>
This reverts commit b9d94fe. Removing v4.6.1 of this broke cura v4.6.1 which was the latest version in tree. This was not detected as the cura RDEPEND is not correctly limiting to v4.6.1 and thus v4.7.0 would match, but fail at runtime. Signed-off-by: David G Turner <digitall@scummvm.org>
This needs to resolve the correct dependencies i.e. v4.6.1, not later such as v4.7.0 as this will likely cause issues. Signed-off-by: David G Turner <digitall@scummvm.org>
This should really be a direct dependency ensuring the python use flag is set as the sub-dependency does not enforce this is compatible with the top level cura version. Signed-off-by: David G Turner <digitall@scummvm.org>
The other v4.7.0 dependencies i.e. curaengine-4.7.0, libcharon etc. are already present in tree, so along with fdm-materials, this completes the changes required to build cura v4.7.0 Signed-off-by: David G Turner <digitall@scummvm.org>
libarcus and other dependencies are not keyworded as ~arm64 so this would not build until these are resolved. Signed-off-by: David G Turner <digitall@scummvm.org>
digitall
force-pushed
the
20200926-cura-fixup
branch
from
a513d7f
to
e278e76
Compare
|
@expeditioneer : I see you added media-gfx/fdm-materials-4.7.0 a few days ago as d756f13 so my change in this PR to add this was not needed. It was identical anyway. I have now rebased this to remove that ebuild and rebase onto the latest master to keep the tree consistent. I hope the fixes to the cura ebuilds can get merged shortly so users have a cura build that actually works in tree :| |
Pull request CI reportReport generated at: 2020-10-01 15:56 UTC There are existing issues already. Please look into the report to make sure none of them affect the packages in question: |
The removal of dev-python/uranium-4.6.1 resulted in cura-4.6.1 using the wrong library dependency. This built fine, but crashes at runtime with various incompatiblities in the Python library API.
This is fixed by restoring the v4.6.1 uranium and associated patches and fixing the cura ebuild to correctly match the dependencies to the version of cura i.e. use "~....${PV}" to match to v4.6.1(-r), but not match to higher or lower versions...