in a configure phase: * QA Notice: epause is not defined in EAPI=3, please file a bug at http://bugs.gentoo.org
hmm, must have missed that and can't check right now, but: The ebuild doesn't call epause at all, maybe some eclass calls it (which then should be a bug for this eclass)
(In reply to comment #0) > in a configure phase: > > > * QA Notice: epause is not defined in EAPI=3, please file a bug at > http://bugs.gentoo.org Maybe I'm missing something (not awfully familiar with the whole QA thing) but I simply don't see this message in my build log. Also as I stated before, Libreoffice simply doesn't call epause at all. So in short: I really need some more explanation to do something about this bug.
I think this is a problem of the check-reqs eclass as it shows up if your HDD or RAM is too small. Then the following is shown: * Messages for package app-office/libreoffice-3.3.2: * * If you experience a build break, please make sure to retry * with MAKEOPTS=-j1 before filing a bug. * * Don't have at least 7144MBytes disk space at /var/tmp/portage/app-office/libreoffice-3.3.2/temp * Bad things may happen! You may abort the build by pressing ctrl+c in * the next 15 seconds. * * QA Notice: epause is not defined in EAPI=3, please file a bug at http://bugs.gentoo.org
(In reply to comment #3) > I think this is a problem of the check-reqs eclass as it shows up if your HDD > or RAM is too small. Then the following is shown: > > * Messages for package app-office/libreoffice-3.3.2: > > * > * If you experience a build break, please make sure to retry > * with MAKEOPTS=-j1 before filing a bug. > * > * Don't have at least 7144MBytes disk space at > /var/tmp/portage/app-office/libreoffice-3.3.2/temp > * Bad things may happen! You may abort the build by pressing ctrl+c in > * the next 15 seconds. > * > * QA Notice: epause is not defined in EAPI=3, please file a bug at > http://bugs.gentoo.org Indeed. Also I think the HDD requirements should be stepped down to 4 GB.
It is in check-reqs eclass which has no maintainer. Reassigning to qa.
*** Bug 377891 has been marked as a duplicate of this bug. ***
*** Bug 374173 has been marked as a duplicate of this bug. ***
*** Bug 378015 has been marked as a duplicate of this bug. ***
Created attachment 282767 [details, diff] check-reqs.eclass-BJA-epause.diff Here is a tricky fix
(In reply to comment #9) > Created attachment 282767 [details, diff] > check-reqs.eclass-BJA-epause.diff > > Here is a tricky fix The original idea behind epause was that it should be possible to disable the sleep. See the old discussion (from 2004): <http://marc.info/?l=gentoo-dev&m=109407032217989>. Replacing epause by sleep again would be a step in the wrong direction. EAPI 4 has introduced the pkg_pretend function. IMO a proper solution would move the eclass's sanity checks to that function.
*** Bug 380455 has been marked as a duplicate of this bug. ***
Created attachment 284751 [details, diff] check-reqs.patch My update for the eclass. Properly uses pkg_pretend and pkg_setup phases. Adds eclassdoc everywhere. Make the variables entering more sane. Adds QA warn for old usage and explain the new approach.
Created attachment 284753 [details, diff] check-reqs.patch Accidentaly attached older version, only changes in comments.
The updated eclass is in main tree so closing.