Currently only support full systemd and consolekit
There is a patch here: https://gist.githubusercontent.com/jorgicio/dcc3f2f42f15b7426b9497ab68262a0f/raw/901620204d8e82d5d63bb3efaa051cd4cb504831/mate-patchesmate-session-managermate-session-manager-1.20-enable-elogind.patch
Hi! I created a overlay called "mate-elogind" which solves this problem. Just upgraded to the 1.22.x branch. Check it in https://github.com/jorgicio/mate-elogind. Tested here and working. Give it a try. Also, yes, I'm waiting for an official elogind-related support for the Portage tree. The packages with elogind support are: * mate-base/mate-session-manager * mate-extra/mate-power-manager I hope it helps.
Now even gnome-session got support for elogind Mikle, could you add the missing bits to MATE as well? The needed bits are in https://github.com/jorgicio/mate-elogind
(In reply to Joakim Tjernlund from comment #3) > Now even gnome-session got support for elogind > Mikle, could you add the missing bits to MATE as well? > The needed bits are in https://github.com/jorgicio/mate-elogind the link you provided actually shows 404 :)
(In reply to Mikle Kolyada from comment #4) > (In reply to Joakim Tjernlund from comment #3) > > Now even gnome-session got support for elogind > > Mikle, could you add the missing bits to MATE as well? > > The needed bits are in https://github.com/jorgicio/mate-elogind > > the link you provided actually shows 404 :) OOPS, should be https://github.com/jorgicio/mate-elogind-gentoo
ok, I will test your patch against the package itself, the only thing is that I am not running elogind with mate (my relitives do run mate but with systemd, and they are fond of systemd), so I would not be able to give a patch wide testing myself.
(In reply to Mikle Kolyada from comment #6) > ok, I will test your patch against the package itself, the only thing is > that I am not running elogind with mate (my relitives do run mate but with > systemd, and they are fond of systemd), so I would not be able to give a > patch wide testing myself. great, that patch has been tested by several already so it should be good.
Yes, I saw you tried to upstream this one
(In reply to Mikle Kolyada from comment #8) > Yes, I saw you tried to upstream this one right, it is queued. While in mate-session you could fix https://bugs.gentoo.org/633334 ?
(In reply to Mikle Kolyada from comment #6) > ok, I will test your patch against the package itself, the only thing is > that I am not running elogind with mate (my relitives do run mate but with > systemd, and they are fond of systemd), so I would not be able to give a > patch wide testing myself. I had the same issue at first. You may enable the elogind support by adding this line: session optional pam_elogind.so in the /etc/pam.d/system-auth. I hope it helps.
(In reply to jorgicio from comment #10) > session optional pam_elogind.so > > in the /etc/pam.d/system-auth. > > I hope it helps. This gets done by pambase[elogind]. Potentially need to take care of your etc-update/dispatch-conf/whatnot, though.
Where is the patch queued upstream? I'd have some constructive criticism. In the lack of another quickly known place to me, some quick comments here. I don't see why the patch is so invasive in adding the support, as elogind is a drop-in replacement for logind stuff. If it would set the same HAVE_SYSTEMD AC_DEFINE (with whatever autotools magic is necessary to make it work; probably just calling it once if either of them is enabled), one wouldn't need to add more complications to the source files. Though for clarification a separate commit could change HAVE_SYSTEMD to HAVE_LOGIND (to mean any logind interface provider) in all the places instead.
(In reply to Mart Raudsepp from comment #12) > Where is the patch queued upstream? I'd have some constructive criticism. > In the lack of another quickly known place to me, some quick comments here. > I don't see why the patch is so invasive in adding the support, as elogind > is a drop-in replacement for logind stuff. If it would set the same > HAVE_SYSTEMD AC_DEFINE (with whatever autotools magic is necessary to make > it work; probably just calling it once if either of them is enabled), one > wouldn't need to add more complications to the source files. Though for > clarification a separate commit could change HAVE_SYSTEMD to HAVE_LOGIND (to > mean any logind interface provider) in all the places instead. It starts here: https://github.com/mate-desktop/mate-session-manager/issues/146#issuecomment-471130324 a bit messy to follow, but the final patch is a bit different. It does not have the version testing of elogind. Ends here: https://github.com/mate-desktop/mate-session-manager/pull/199
(In reply to Mart Raudsepp from comment #12) > Where is the patch queued upstream? I'd have some constructive criticism. > In the lack of another quickly known place to me, some quick comments here. > I don't see why the patch is so invasive in adding the support, as elogind > is a drop-in replacement for logind stuff. If it would set the same > HAVE_SYSTEMD AC_DEFINE (with whatever autotools magic is necessary to make > it work; probably just calling it once if either of them is enabled), one > wouldn't need to add more complications to the source files. Though for > clarification a separate commit could change HAVE_SYSTEMD to HAVE_LOGIND (to > mean any logind interface provider) in all the places instead. oh, I think you are looking an an early version. My patch just touches the build system(configure.ac/Makefile.am)
Yeah, I was looking at comment #1 link. Btw, to support elogind in ebuild without an imposed eautoreconf, one could also probably do something like this until the patches hit a release: local myconf=( --enable-other-opts ) if use elogind; then myconf+=( SYSTEMD_CFLAGS=`pkg-config --cflags "libelogind" 2>/dev/null` SYSTEMD_LIBS=`pkg-config --libs "libelogind" 2>/dev/null` ) fi mate_src_configure "${myconf[@]}"
(In reply to Mart Raudsepp from comment #15) > Yeah, I was looking at comment #1 link. > > Btw, to support elogind in ebuild without an imposed eautoreconf, one could > also probably do something like this until the patches hit a release: > > local myconf=( > --enable-other-opts > ) > > if use elogind; then > myconf+=( > SYSTEMD_CFLAGS=`pkg-config --cflags "libelogind" 2>/dev/null` > SYSTEMD_LIBS=`pkg-config --libs "libelogind" 2>/dev/null` > ) > fi > > mate_src_configure "${myconf[@]}" Yes, that would work too.
elogind was just added upstream, will be in next 1.22.x release
(In reply to Joakim Tjernlund from comment #17) > elogind was just added upstream, will be in next 1.22.x release ok, probably the best course here is to pick up the upstream patch then
I just noted mate-screensaver has configure flags for consolekit/systemd. Should there not be one for elogind too? jorgicio ?
Created attachment 570898 [details, diff] mate-screensaver patch for elogind
Upstream patch here: https://github.com/mate-desktop/mate-screensaver/issues/187
(In reply to Joakim Tjernlund from comment #19) > I just noted mate-screensaver has configure flags for consolekit/systemd. > Should there not be one for elogind too? Different bug: bug 681334
(In reply to Joakim Tjernlund from comment #19) > I just noted mate-screensaver has configure flags for consolekit/systemd. > Should there not be one for elogind too? > > jorgicio ? I just checked this, so this is a work in progress.
https://github.com/jorgicio/mate-elogind-gentoo For a while, the mate-extra/mate-screensaver package was added too. Also, I had to dropped 1.20.x support because "calculate" overlay did it first.
(In reply to jorgicio from comment #24) > https://github.com/jorgicio/mate-elogind-gentoo > > For a while, the mate-extra/mate-screensaver package was added too. Also, I > had to dropped 1.20.x support because "calculate" overlay did it first. Don't you need to run eautoreconf now that elogind patch touches configure.ac ?
(In reply to Joakim Tjernlund from comment #25) > (In reply to jorgicio from comment #24) > > https://github.com/jorgicio/mate-elogind-gentoo > > > > For a while, the mate-extra/mate-screensaver package was added too. Also, I > > had to dropped 1.20.x support because "calculate" overlay did it first. > > Don't you need to run eautoreconf now that elogind patch touches > configure.ac ? That's not necessary because the ebuild already does that. Just specify the PATCHES variable and you're ready.
(In reply to jorgicio from comment #26) > (In reply to Joakim Tjernlund from comment #25) > > (In reply to jorgicio from comment #24) > > > https://github.com/jorgicio/mate-elogind-gentoo > > > > > > For a while, the mate-extra/mate-screensaver package was added too. Also, I > > > had to dropped 1.20.x support because "calculate" overlay did it first. > > > > Don't you need to run eautoreconf now that elogind patch touches > > configure.ac ? > > That's not necessary because the ebuild already does that. Just specify the > PATCHES variable and you're ready. OK, good. Mikle, can we have MATE and elogind in Gentoo repo soon?
The bug has been closed via the following commit(s): https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=e833a3cb5ddd06db4819af66b8c2a513438aa2b7 commit e833a3cb5ddd06db4819af66b8c2a513438aa2b7 Author: Andreas Sturmlechner <asturm@gentoo.org> AuthorDate: 2019-03-30 22:50:17 +0000 Commit: Andreas Sturmlechner <asturm@gentoo.org> CommitDate: 2019-03-31 10:47:30 +0000 mate-extra/mate-power-manager: Add USE elogind logind is detected at runtime, so move the session manager trifecta to RDEPEND. Closes: https://bugs.gentoo.org/633336 Package-Manager: Portage-2.3.62, Repoman-2.3.12 Signed-off-by: Andreas Sturmlechner <asturm@gentoo.org> .../mate-power-manager/mate-power-manager-1.22.0.ebuild | 15 ++++++++++----- 1 file changed, 10 insertions(+), 5 deletions(-) Additionally, it has been referenced in the following commit(s): https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=5bd270f4a595255b62e34e17375c50d5ef82f15f commit 5bd270f4a595255b62e34e17375c50d5ef82f15f Author: Andreas Sturmlechner <asturm@gentoo.org> AuthorDate: 2019-03-30 22:32:54 +0000 Commit: Andreas Sturmlechner <asturm@gentoo.org> CommitDate: 2019-03-31 10:39:36 +0000 mate-base/mate-session-manager: Add USE elogind Bug: https://bugs.gentoo.org/633336 Package-Manager: Portage-2.3.62, Repoman-2.3.12 Signed-off-by: Andreas Sturmlechner <asturm@gentoo.org> .../mate-session-manager-1.22.0-elogind.patch | 84 ++++++++++++++++++++++ .../mate-session-manager-1.22.0.ebuild | 16 +++-- 2 files changed, 96 insertions(+), 4 deletions(-)
*** Bug 633338 has been marked as a duplicate of this bug. ***