The PV == 9999 packages are meant to build directly as in git. Therefore they are always masked. Moreover, it does not make any sense to backport a patch from there and try to epatch it again. It will certainly fail as the patch is already applied! Therefore, don't screw build and epatch only when desirable: # diff -u /tmp/libvirt-9999.ebuild /usr/portage/app-emulation/libvirt/libvirt-9999.ebuild --- /tmp/libvirt-9999.ebuild 2015-11-30 11:10:33.230580285 +0100 +++ /usr/portage/app-emulation/libvirt/libvirt-9999.ebuild 2015-11-30 11:14:20.529688224 +0100 @@ -217,14 +217,16 @@ ) >.git-module-status fi - epatch \ - "${FILESDIR}"/${PN}-1.2.9-do_not_use_sysconf.patch \ - "${FILESDIR}"/${PN}-1.2.16-fix_paths_in_libvirt-guests_sh.patch \ - "${FILESDIR}"/${PN}-1.2.17-fix_paths_for_apparmor.patch + if [[ ${PV} != *9999* ]]; then + epatch \ + "${FILESDIR}"/${PN}-1.2.9-do_not_use_sysconf.patch \ + "${FILESDIR}"/${PN}-1.2.16-fix_paths_in_libvirt-guests_sh.patch \ + "${FILESDIR}"/${PN}-1.2.17-fix_paths_for_apparmor.patch - [[ -n ${BACKPORTS} ]] && - EPATCH_FORCE=yes EPATCH_SUFFIX="patch" \ - EPATCH_SOURCE="${WORKDIR}/patches" epatch + [[ -n ${BACKPORTS} ]] && + EPATCH_FORCE=yes EPATCH_SUFFIX="patch" \ + EPATCH_SOURCE="${WORKDIR}/patches" epatch + fi epatch_user
Created attachment 418188 [details, diff] fix.patch Seems like previous diff got mangled.
*** Bug 567134 has been marked as a duplicate of this bug. ***
(In reply to Michal Privoznik from comment #0) > [...] > Moreover, it does not make any sense to backport a patch from there and try > to epatch it again. It will certainly fail as the patch is already applied! > Therefore, don't screw build and epatch only when desirable: Well, except that those three patches: "${FILESDIR}"/${PN}-1.2.22-do_not_use_sysconf.patch \ "${FILESDIR}"/${PN}-1.2.16-fix_paths_in_libvirt-guests_sh.patch \ "${FILESDIR}"/${PN}-1.2.17-fix_paths_for_apparmor.patch are "downstream" patches that ensure compatibility with Gentoo's data layout and init system. They will never be applied upstream. The issue you are encountering is due to the introduction of "virtlogd" and accompanying changes that lead the old patch fail. commit d99c871d82ddda973093333f5dcadc6d3dd19d48 Author: Matthias Maier <tamiko@gentoo.org> Date: Fri Dec 4 00:15:47 2015 -0600 app-emulation/libvirt: update live build to latest changes (bug #567152) - Due to the introduction of virtlogd one of Gentoo's downstream patches did not apply any more. Package-Manager: portage-2.2.26 Things left to do: [ ] write runscript for virtlogd [ ] update systemd units to gentoo's file locations
commit 4d582919852ebc5cb1ee5cedc1a02cdb8d199158 Author: Matthias Maier <tamiko@gentoo.org> Date: Sat Dec 26 23:59:06 2015 +0100 dev-python/libvirt-python: version bump to 1.3.0 Package-Manager: portage-2.2.26 commit fd9a637c9353db923cdea07c0455f6431d177aae Author: Matthias Maier <tamiko@gentoo.org> Date: Sat Dec 26 23:57:40 2015 +0100 app-emulation/libvirt: version bump to 1.3.0, provde service file for virtlogd - Bump to version 1.3.0 (including backport for CVE-2015-5313) - Add runscript for freshly introduce virtlogd Gentoo-Bugs: 569526, 568860, 567152 Package-Manager: portage-2.2.26