as in Summary libx86 can be built on ARM (armv7a currently). Reproducible: Always
Created attachment 430200 [details, diff] patch from Debian
Created attachment 430202 [details] ebuild
Created attachment 430204 [details] ebuild-2
Created attachment 430206 [details, diff] ebuild-diff
Created attachment 430208 [details] ebuild
Created attachment 430210 [details, diff] patch
ANY progress on this. All required to fix this bug is provided. Waiting billion years for assignee is beyond any good user experience. Any gentoo devs who can test this on arm able to fix.
There is no need to be rude, if you think that it should be fixed ASAP - feel free to ping us on #gentoo-embedded Anyway, CCed arm team
it is not the arm team's job to maintain the ebuild. if the maintainer wants to accept these patches then do it and then open a keyword request. if the maintainer needs help testing then ask for help testing. don't call it a keyword request when there is no ebuild in the tree to test. feel free to re-add arm when this is a keyword request or when you say what you need from us.
commit c96dc7cf3ab2fb58fa22c7fd9c1557dda14d5290 Author: Sergey Popov <pinkbyte@gentoo.org> Date: Fri Apr 29 10:56:18 2016 +0300 dev-libs/libx86: revision bump Port to EAPI 6, bring wider arch support Reported-by: Oleg <oleg@funtoo.org> Gentoo-Bug: 579682 Package-Manager: portage-2.2.28 @arm team: please proceed with testing
applied changes ignoring ARM: +KEYWORDS="amd64 x86 arm" +use arm && ARGS="BACKEND=x86emu" so, bumped ebuild actually does nothing for ARM.
(In reply to Oleg from comment #11) > applied changes ignoring ARM: > > +KEYWORDS="amd64 x86 arm" > +use arm && ARGS="BACKEND=x86emu" > > so, bumped ebuild actually does nothing for ARM. Yes, that's why we call this keyword request. ARM team will test if the package actually works on their architecture and then - add ~arm keyword for ebuild. As for 'use arm ...' sentence - look at 'use x86 || ARGS="BACKEND=x86emu"' - basically it does the same, but allows adding other arches if it would be needed.
i don't think it's any good policy described here. Ebuild has nothing to test, means ARM team have nothing to test. This looks like half-baked ebuild for ARM team, so they are need to cook it more.
(In reply to Oleg from comment #13) > i don't think it's any good policy described here. Ebuild has nothing to > test, means ARM team have nothing to test. They need to test that it works on arm, no? Do you really understand meaning of keywords? From https://wiki.gentoo.org/wiki/KEYWORDS: In an ebuild the KEYWORDS variable informs in which architectures the ebuild is stable or still in testing phase. From https://devmanual.gentoo.org/keywording/index.html: No keyword If a package has no keyword for a given arch, it means it is not known whether the package will work, or that insufficient testing has occurred for ~arch. I do not have access to arm hardware right now, so i can not test it myself. So, it's up to arch team to install this package(doing this without a keyword in ebuild is trivial task) and verify that it works.
i perfectly understand what is keyword request. But what is actually seriously bother me is absolutely unacceptable policy about how things fixed. SO, to be clear. I reported this bug with complete (or at least easily understandable plan) on how to add ARM support for dev-lisb/libx86. It is assigned to someone. Now someone, added ebuild directly into portage tree, with commit message, which mention my report but now it has original meaning of report screwed: dev-libs/libx86: revision bump Port to EAPI 6, bring wider arch support Reported-by: Oleg <oleg@funtoo.org> Gentoo-Bug: 579682 Package-Manager: portage-2.2.28 Does it really bring wider support? No, patch, which allows building on ARM, added. But rest of ebuild does not really add any wider support. Now, apparently, someone, has no ARM device to test and this half-baked ebuild added to ARM team, so someone else need to waste another hours to add ARM finally. No spare device to test, then un-assign yourself and let other bug wranglers know to not bother you about assigns.
(In reply to Oleg from comment #15) > Does it really bring wider support? No, patch, which allows building on ARM, > added. But rest of ebuild does not really add any wider support. Nope, you are wrong. I told you which parts of ebuild are responsible for wider arch support. Missing keywords is nothing to do with support it self, it only says where package was tested. For this purpose there is -arch keyword, which means "tested and proved not working" > Now, > apparently, someone, has no ARM device to test and this half-baked ebuild > added to ARM team, so someone else need to waste another hours to add ARM > finally. No spare device to test, then un-assign yourself and let other bug > wranglers know to not bother you about assigns. This bug is not marked as resolved, i did my part of work - added new ebuild, patch and all apropriate changes. Now it's arm team work, they are CCed to this bug, so it will get their attention. After they will add keyword - bug will be marked as fixed
build tested, keyword added.
