[00:00] <ochosi> robert_ancell: not sure you know this, but despite my warnings, lubuntu is currently (13.10) using lightdm-only to "lock" their session...
[00:01] <robert_ancell> ochosi, oh, what were the warnings?
[00:01] <ochosi> robert_ancell: "you can easily access a 'locked' session with ctrl+alt+f7 if you don't use something to actually *lock* your session"
[00:01] <ochosi> the typical issue
[00:01] <robert_ancell> right
[00:02] <ochosi> they ignored it and thought lightdm is enough
[00:02] <ochosi> hope they pick up some solution for the lts
[00:02] <robert_ancell> ochosi, they're running nothing in-session to do the locking?
[00:02] <ochosi> nope
[00:02] <robert_ancell> hmm
[00:02] <ochosi> i was wondering whether a "warning label" on lightdm would help against these cases
[00:02] <robert_ancell> ochosi, was there a reason why they didn't use light-locker? Perhaps because it wasn't in the archive
[00:03] <robert_ancell> ochosi, where would you put the warning?
[00:03] <ochosi> robert_ancell: well, yes, it wasn't in the archive by the time. but it would've been ready in time
[00:03] <robert_ancell> in the non-existent docs? /me feels bad
[00:03] <ochosi> robert_ancell: phu, not sure where to put the warning. maybe in the docs/help that mention the locking? :D
[00:03] <robert_ancell> I must write a manual
[00:03] <ochosi> yeah, that should be fine
[00:04] <ochosi> there's also some api docs iirc, you could put it there as a note
[00:04] <ochosi> (not really the right place, but better than nothing)
[00:04] <ochosi> btw, kudos on how many fixes you backported to older lightdm versions lately!
[00:05] <robert_ancell> Been trying to keep them fresh. Thanks for noticing!
[00:09] <ochosi> robert_ancell: actually, does that plan you mentioned earlier (the session/greeter sending a signal to lightdm that it's ready to be displayed/VT-switched to) not a bit obsolete with Mir?
[00:09] <ochosi> oops, language-fail :)
[00:09] <ochosi> s/does/is/
[00:10] <robert_ancell> ochosi, I'm not sure - the compositor notifies lightdm when each session sends the first frame which is a lot closer. That may not necessarily mean the session is ready though
[00:10] <robert_ancell> ochosi, the main problem is there's not a good comms channel for a session to notify the daemon and no sessions in general have a good hook to tell when they think they're ready
[00:11] <robert_ancell> we tried this ages ago with xsplash in Ubuntu and it sort of worked by hooking into gnome-session
[00:12] <ochosi> robert_ancell: right, true that. we're also hoping to make the transition from the greeter to the desktop a bit smoother in 14.04 (right now things just pop up on the screen one after the other...)
[00:12] <ochosi> (in xubuntu, that is)
[00:13] <robert_ancell> ochosi, right - to do that you need to run the session in a new X server and VT switch to it. We're not doing that currently due to the VT switch problems
[00:13] <robert_ancell> then the greeter could show a spinner and switch once the session is ready which would be nice
[00:14] <ochosi> robert_ancell: yeah, these sort of things would be nice if VT switching were entirely unproblematic...
[06:00] <pitti> Good morning
[08:16] <mlankhorst> morning
[09:02] <Laney> greetings
[09:06] <larsu> Laney: good morning
[09:08] <seb128> good morning desktopers!
[09:08] <sil2100> Morning! :)
[09:08] <didrocks> hey guys :)
[09:11]  * pitti grouphugs Laney, seb128, sil2100, and didrocks
[09:12] <pitti> #u-desktop comes to life \o/
[09:12]  * didrocks hugs pitti
[09:12]  * seb128 grouphugs pitti&co back
[09:12] <Laney> \o/
[09:12] <didrocks> pitti: well, here for some hours, just quiet catching up :)
[09:13] <ogra_> very holbachy in here today :)
[09:14] <pitti> yeah, just in time for dholbach to get back from India :)
[09:14]  * didrocks adds the adjective to his vocabulary
[09:14] <seb128> ogra_, speaking of which, I see that Daniel is back online ;-)
[09:14] <ogra_> yeah
[09:14] <pitti> it just takes a while to ride on an elephant from India to Berlin :)
[09:14] <ogra_> probably very badly jetlagged
[09:15] <ogra_> pitti, ah, did he fly elephant class ?
[09:17] <Laney> seb128: can you bounce me that email about folks?
[09:17] <Laney> I think I deleted it by mistake ¬_¬
[09:17] <Laney> WAIT, I found it
[09:18] <seb128> Laney, I'm going to reply, I had eds 3.10 on my week TODO
[09:18] <Laney> https://launchpad.net/~laney/+archive/gnome-transition/
[09:18] <seb128> shrug
[09:18] <Laney> that's a start
[09:18] <seb128> Laney, good that you almost lost that email and pointed that ppa in return
[09:18] <Laney> didn't do e-d-s stuff though really, mainly goa
[09:18]  * seb128 does like dupping work
[09:18] <Laney> I was going to reply, thought I already said about it last week
[09:19] <seb128> I didn't record the info it seems ;-)
[09:19] <seb128> do you still do webkit? how did the armhf build go?
[09:19] <Laney> yeah it all spidered out of that, don't remember the exact path
[09:19] <Laney> armhf worked
[09:19] <darkxst> webkit is a mess on the ppa builders atleast
[09:20] <Laney> it's a three way webkitgtk + gnome-online-accounts + e-d-s thing
[09:20] <darkxst> we havent been able to get a amd64 build to succeed in weeks
[09:20] <Laney> darkxst: sad, luckily I have access to a PPA which builds on the real builders
[09:20] <Laney> feel free to copy the webkitgtk out of the one I just linked if you need it
[09:21] <seb128> darkxst, is webkit incompatible in some way?
[09:21] <seb128> ups
[09:21] <seb128> Laney, ^
[09:21] <Laney> don't think so, it just adds webkit2 packages
[09:22] <Laney> they renamed the source to follow upstream
[09:22] <seb128> ok
[09:22] <seb128> so it's not linked to goa/eds updates?
[09:22] <Laney> let me remember
[09:22] <seb128> it would be nice to get it out of the way first
[09:22] <seb128> the new evo needs it
[09:22] <Laney> Perhaps it was the other way
[09:23] <seb128> so it's probably a pre-required
[09:23] <Laney> like one of those made me upgrade webkit
[09:23] <seb128> right
[09:23] <seb128> that makes sense
[09:23] <seb128> which is one of the reasons I want it "out of the way"
[09:23] <Laney> yeah
[09:24] <Laney> I'll look at uploading that next
[09:24] <seb128> thanks
[09:24]  * seb128 is looking at GTK 3.10 (thanks larsu for dealing with the update) this week
[09:24] <Laney> if you want to do some e-d-s stuff then you can copy things from that PPA or make one which deps on it
[09:24] <seb128> ok
[09:25] <seb128> I didn't start on it, but I guess from the comment/ppa that there are some soname changes :/
[09:25] <Laney> or add it to a chroot or whatever
[09:25] <Laney> of course
[09:25] <larsu> seb128: I hope to have it ready for you today
[09:25] <Laney> g-o-a too but the only source change I found was in g-c-c which I already patched in there
[09:25] <larsu> depending on how much work the theming stuff will be
[09:25] <seb128> larsu, once the menu and scrollbar issues are fixed we can push to the ppa for more testing I think
[09:26] <larsu> cool
[09:26] <seb128> Laney, I'm pondering dropping goa support from a bunch of stuff, but that wouldn't be popular with the GNOME remix I think
[09:26] <Laney> certainly not
[09:28] <seb128> Laney, well, we can argue that uoa is what provide online accounts in Ubuntu, the same way that upstart is what provides you an init...
[09:30] <seb128> Laney, reality is that we need to make eds stop depending on webkitgtk because that brings a GNOME stack on Ubuntu Touch, and I failed to due that because of goa, not sure what to do next
[09:30] <seb128> Laney, we could dual build 2 version of e-d-s and make them conflict, but that sucks as well...
[09:33] <Laney> actually it might be possible to split goa
[09:34] <Laney> all of the eds stuff depends on libgoa-1.0.so.0 which is not the one which depends on webkitgtk
[09:36] <seb128> oh, I didn't look at the goa side, thanks for the suggestion
[09:37]  * Laney will look into that
[09:38] <seb128> thanks
[09:47] <Laney> js> print("Hello darkxst");
[09:47] <Laney> Hello darkxst
[09:49] <seb128> Laney, python> python -c 'print("Hello Laney")'?
[09:49]  * seb128 tries to guess the game
[09:49] <Laney> js breakage on ppc
[09:49] <seb128> oh
[09:51] <seb128> larsu, I found another bug with the new GTK...
[09:51] <larsu> shoot!
[09:51] <seb128> larsu, update-manager, the list of packages doesn't resize to match the windows
[09:52] <seb128> larsu, http://ubuntuone.com/6mUJidrQNJEuBjWP80bc8L
[09:52] <seb128> larsu, if I downgrade to 3.8 it works as it should, e.g no shrinked list
[09:53] <larsu> "oops something broke while generating thi page"
[09:53] <larsu> I'll try to reproduce here
[09:53] <seb128> larsu, retry, it was still uploading
[09:53] <seb128> it's working now
[09:53] <larsu> same problem
[09:53] <seb128> k
[09:53] <larsu> ah, now! took a few times
[09:54] <larsu> I can reproduce. Thanks for letting me know!
[09:54] <seb128> yw, thanks for working on those issues!
[09:56] <Laney> darkxst: I lost your non-fatal-ppc patch - do you have it?
[10:32] <darkxst> seb128, seems to just be a ppa builder issue
[10:33] <darkxst> it builds fine on debian atleast
[10:34] <darkxst> Laney, https://code.launchpad.net/~darkxst/ubuntu/trusty/gjs/1.38.1-2
[10:35] <Laney> ta
[10:42] <darkxst> Laney, the "imports" stuff is gjs specific
[10:43] <Laney> yeah, saw
[10:43] <seb128> shrug, ports.u.c is sloooow
[10:43]  * seb128 tries to install build-depends on his nexus for an hour
[10:43] <seb128> it keeps hanging
[10:45] <darkxst> Laney, the gnome guys only have one ppc machine left and that is used for building packages
[10:45] <darkxst> ^well only one that still works
[11:12] <aslan> HI, I have problem with grub the issue is I cant see the grub menu at the time of booting even after giving the grub timeout
[11:13] <aslan> in grub.cfg
[11:15] <seb128> aslan, hey, try #ubuntu for user questions
[12:00] <aslan> seb128, ya I will
[12:00] <aslan>  thankyou
[12:04] <dednick> didrocks: any status update for landing ubuntu-settings-components?
[12:04] <didrocks> dednick: we are relying on having the infra back
[12:04] <dednick> didrocks: ok. thought it might be.
[12:05] <didrocks> dednick: if you look at the ubuntu phone image, the CI guys are still working on it
[12:05] <didrocks> ML*
[12:05]  * didrocks autotypes image now :)
[12:05] <didrocks> dednick: as soon it's in, I'll get that landed
[12:05] <dednick> didrocks: cool. thanks
[12:05] <ogra_> as long as it doesnt turn *all* words into image thats fine :)
[12:05] <didrocks> yw ;)
[12:05] <Laney> but an image is worth 1000 words
[12:05] <didrocks> ogra_: did you image the image for the image thing?
[12:06] <ogra_> image image image, image image ?
[12:06] <didrocks> image is the new smurfs
[12:06] <didrocks> ;)
[12:06] <ogra_> :)
[12:07] <seb128> didrocks, green is the new blue?
[12:07] <didrocks> in a jenkins sense, yeah! :)
[12:07] <ogra_> image !
[12:07] <didrocks> ahah
[13:24] <Laney> hmm
[13:24] <Laney> what team should we subscribe to package bugs?
[13:24] <seb128> Laney, desktop-bug
[13:24] <seb128> s
[13:24] <Laney> ah
[13:25] <Laney> seb128: you look like the owner there - can you subscribe it to libwebp?
[13:25] <Laney> bug #1186553
[13:25] <ubot2> Launchpad bug 1186553 in libwebp (Ubuntu) "[MIR] libwebp" [Undecided,New] https://launchpad.net/bugs/1186553
[13:28] <seb128> Laney, done
[13:28] <seb128> https://bugs.launchpad.net/ubuntu/+source/libwebp/+subscriptions
[13:30] <Laney> merci
[13:30] <seb128> de rien ;-)
[13:32] <Laney> lunch time
[13:32] <Laney> we can fiddle the archive for webkit → webkitgtk after that
[13:32] <seb128> great! enjoy lunch ;-)
[13:47] <didrocks> tjaalton: yeah, I will look at them probably tomorrow
[13:51] <tjaalton> didrocks: thanks
[13:51] <didrocks> yw :)
[13:51] <tjaalton> :)
[15:06] <seb128> bah, writing that email took we a solid 1h
[15:06] <seb128> we->me
[15:06] <seb128> (keyboard layouts handling for the LTS just sent to ubuntu-devel@)
[15:06] <seb128> let's see if anyone is wanting to read that much text :/
[15:08] <pitti> seb128: very nice mail!
[15:08] <Laney> speed reader
[15:09] <seb128> pitti, thanks ;-)
[15:10] <davmor2> seb128: I've seen how fast you type, that's not an email it's a novel ;)
[15:10] <seb128> pitti, do you have any opinion on the topic ?(out of "just make it work" ;-)
[15:10] <seb128> lol
[15:11] <happyaron> reading
[15:12] <pitti> seb128: I haven't read it in that detail yet (just back from giving lighting talk, and debugging systemd-shim now); but I'll read it tomorrow in the train when I have some quiet time to think about it
[15:15] <seb128> pitti, thanks ;-)
[15:18] <kenvandine> seb128, that should have been a blog post :)
[15:18] <kenvandine> seb128, i guess ubuntu-devel is your blog
[15:18] <kenvandine> :-p
[15:19]  * happyaron would like to vote for reverting the behavior.
[15:20] <seb128> kenvandine, lol, I don't know enough about the topic to blog ;-) and I'm interested in technical feedback/a decision, so the dev mailing list makes sense
[15:20] <seb128> happyaron, because you liked the old UI/design better? or because you think we are going to still have bugs at the LTS time?
[15:20] <Laney> should try to keep it on list
[15:21] <larsu> I bet attente has an opinion :D
[15:21] <Laney> ooh
[15:21]  * Laney writes summary
[15:21] <Laney> weekly summary, nothing to do with keyboards :-)
[15:21] <seb128> larsu, I'm sure his opinion his "I wish I could go back one cycle and *not* take on this topic"
[15:21] <happyaron> seb128: I think UI is less important than functions at this point, since old UI isn't way too bad :)
[15:22] <happyaron> seb128: reverting to the old way would at least don't cause regression for those upgrading from previous LTS.
[15:22] <seb128> happyaron, what is missing in the UI infrastructure (assuming we resolve we flackyness in the key handling and the grabbing issues)?
[15:22] <larsu> seb128: I don't know. I think it's an awesome thing to work on. Just a bit ... complex
[15:22] <seb128> larsu, yeah...
[15:22] <happyaron> seb128: user cannot access settings of every IM now
[15:23] <happyaron> lemme find the bug for you...
[15:23] <seb128> happyaron, how did they access those before? (e.g ibus is still there and should be working no?)
[15:23] <happyaron> seb128: https://bugs.launchpad.net/ubuntu/+source/ibus/+bug/1188509
[15:23] <ubot2> Launchpad bug 1188509 in indicator-keyboard (Ubuntu) "Embedded in menu doesn't work in IBus indicator" [High,Triaged]
[15:23] <seb128> happyaron, or is that because users used to access the ibus settings through the ibus keybinding, but the new system made them go through g-s-d which doesn't handle those options?
[15:24] <seb128> happyaron, oh, it's an indicator lack-of-feature it seems
[15:24] <happyaron> seb128: there are two kind of settings, one is for ibus (the framework), the other is for IM (like ibus-pinyin). For the second one, it's usually accessed from the embedded in menu
[15:25] <happyaron> yeah
[15:25] <seb128> I guess gnome-shell do expose pinyin settings from their indicator?
[15:25] <happyaron> yes
[15:27] <seb128> happyaron, how are things working on Kylin today? do they have indicator-keyboard and a fcitx indicator? what about keybindings?
[15:27] <happyaron> seb128: fcitx and most of it's UIs support indicator directly, so no indicator-keyboard needed.
[15:28] <desrt> pitti: good morning
[15:28] <happyaron> seb128: keybindings are handled directly by fcitx
[15:28] <seb128> happyaron, I see, the fcitx stuff is another argument for having layouts and IMs decoupled
[15:29] <seb128> though it seems, from what you say, that Kylin doesn't need a layout indicator
[15:29] <happyaron> yes
[15:30] <seb128> qengho, Sweetshark, mlankhorst, Laney, tkamppeter_, desrt, attente, larsu: hey, it's meeting time
[15:30] <desrt> pitti: you seem to have contradictory reports in your logs
[15:30] <desrt> the indicator sends the request to systemd-shim but systemd-shim never receives such a request?
[15:31] <seb128> qengho, Sweetshark, mlankhorst, Laney, tkamppeter_, desrt, attente, larsu: hey hey hey
[15:31] <Laney> hello
[15:31] <Sweetshark> seb128:o/
[15:31] <seb128> better ;-)
[15:32] <Laney> happyaron is here too, let him join in ;-)
[15:32] <happyaron> :)
[15:32] <seb128> let's get started (/me wonders if that's DST confusing people)
[15:33] <seb128> happyaron, yeah, feel free to join the meetings if you want (we usually don't list people who are in tz where it's not a decent time)
[15:33] <seb128> qengho, hey
[15:33] <qengho> - fixed a few bugs in chromium. handed source packages to #security.
[15:33] <happyaron> seb128: thanks, :)
[15:33] <qengho> - getting an early start on Cr 32.
[15:33] <qengho> EOF
[15:34] <Laney> did the chromium autopkgtest get fixed in this one?
[15:34] <seb128> qengho, thanks (good to see that chromium issues dropped from the most frequent issues, out a fglrx one)
[15:34] <qengho> seb128, work calendar has this an hour later. maybe adjust it h to know if timezone.
[15:34] <qengho> of
[15:34] <qengho> phone keyboard.
[15:34]  * seb128 looks at calendar
[15:35] <seb128> qengho, what about autopkgtests? (thanks Laney for asking ;-)
[15:35] <qengho> seb, new packages fix those.
[15:35] <seb128> great
[15:35] <seb128> qengho, thanks
[15:35] <seb128> Sweetshark, hey
[15:35] <qengho> thanks Laney
[15:36] <Sweetshark> LibreOffice Hackfest Freiburg, finished with a bad case of ubuflu
[15:36] <Sweetshark> did some LibreOffice dev mentoring with Marcos and Jose, two brazilian guests in Hamburg (sponsored by TDF)
[15:36] <Sweetshark> LibreOffice 4.1.3 ppa with some more vendor patching (empty label on quicklist)
[15:36] <Sweetshark> LibreOffice 4.2.0alpha1 tagged upstream and building (with some caveats) on trusty right now: https://launchpad.net/~bjoern-michaelsen/+archive/libreoffice-staging
[15:36] <Sweetshark> got some requests wrt LibreOffice without Java, dunno exactly what to make of it
[15:36] <Sweetshark> some mindless lp bugwrangling
[15:36] <Sweetshark> stated candidacy for TDF board again: http://nabble.documentfoundation.org/Candidacy-for-a-BoD-seat-Bjoern-Michaelsen-tp4082608.html
[15:36] <Sweetshark> EOF
[15:36] <larsu> Sweetshark: ubuflu at a libreoffice hackfest? Shouldn't it be libreflu?
[15:36]  * desrt votes björn!
[15:37] <Laney> free as in flu
[15:37] <larsu> ha
[15:37] <larsu> Sweetshark: in any case: get well soon!
[15:37] <seb128> Sweetshark, no toolchain issue on trusty? how boring!
[15:38] <Sweetshark> larsu: Im just fighting down remaining pockets of resistence ...
[15:38] <Sweetshark> seb128: there will be toolchain fun. there always is. maybe those changes are not in yet ;)
[15:38] <Laney> gute besserung
[15:38] <Sweetshark> seb128: also Im cutting some corners with the alpha release.
[15:39] <seb128> ok
[15:39] <seb128> that one is targeted at the ppa only right?
[15:39] <Sweetshark> seb128: yes
[15:39] <seb128> ok, good
[15:39] <seb128> let me know when you need sponsoring for the SRU
[15:40] <seb128> oh, and get some rest until you feels better!
[15:40] <seb128> Sweetshark, thanks
[15:40] <seb128> mlankhorst, hey
[15:40] <Sweetshark> seb128: last beta/4.2.0rc2 would be first for the archive IMHO ...
[15:40] <seb128> right, we discussed that during the rally, agreed
[15:41] <mlankhorst> hey
[15:41] <mlankhorst> oh I missed the earlier ping? :P
[15:41] <mlankhorst> getting some graphics patches upstream for this merge window
[15:41] <mlankhorst> glamor-egl sru's, libdrm sru
[15:42] <mlankhorst> working on android syncpoints, converting fence api to be compatible with it
[15:42] <mlankhorst> eod :P
[15:42] <mlankhorst> (last one is more kernel work, getting my own stuff upstreamed)
[15:42] <seb128> mlankhorst, thanks
[15:43] <seb128> Laney, hey
[15:43] <mlankhorst> android had their own sync mechanisms, and by using my sync mechanism as a backend for android the dupes are removed
[15:43] <Laney> hallo
[15:43] <Laney> • Patch pilot; some nagging about sponsor queue. Down to 52 now, well done all!
[15:43] <Laney> • Fix gdk-pixbuf FTBFS caused by missing some environment when running make check.
[15:43] <Laney> • Look at webkitgtk / e-d-s / g-o-a updates/transitions.
[15:43] <Laney> ∘ g-c-c requires a patch for new g-o-a; backported from upstream and modified to work on our g-c-c.
[15:43] <Laney> ∘ g-o-a transition is otherwise fine
[15:43] <Laney> ∘ e-d-s transition needs staging still
[15:43] <Laney> ∘ webkitgtk good to upload, doing that imminently
[15:43] <Laney> • Misc fixes for upgrade failures: screen-resolution-extra redshift
[15:43] <Laney> • Work on AS patch for vendor extension overrides.
[15:43] <Laney> • Some attempts at debugging gjs on powerpc, not entirely successful but provided some information to upstream.
[15:43] <Laney> EOF
[15:44] <seb128> Laney, thanks
[15:44] <seb128> tkamppeter, hey
[15:46] <seb128> no tkamppeter?
[15:46] <seb128> ok, let's move to the next one
[15:46] <seb128> desrt, hey ;-)
[15:46] <desrt> that's me!
[15:46] <desrt> so mostly i listened to a lot of lisa loeb
[15:46] <desrt> but... i also got some work done
[15:46] <desrt> the desktop file efficency changes (for the unindexed case) landed, along with the search API
[15:47] <desrt> and i've started reviews on the gcleanup branch that i started a while ago and stef picked up recently... that looks like it will land in the coming weeks -- basically it's a way of giving glib (and glib-based libraries) an easy way to clean up after themselves on exit
[15:47] <larsu> \o/
[15:47] <desrt> that's nice for valgrind-friendly reasons, but it will also let us use a counting allocation checking cheme (+1 for malloc, -1 for free) and assert on exit if they're not equal
[15:48] <desrt> we'd do this from testcases, for example, so we can type 'make check' and get failures on any leak anywhere
[15:48] <desrt> there are a lot of tricky issues there, though
[15:48] <desrt> aside from that, i did a bunch of releases yesterday (systemd-shim, glib unstable, glib stable, dconf)
[15:48] <desrt> that's about it
[15:49] <desrt> uh... eof, i mean
[15:49] <seb128> ;-)
[15:49] <tkamppeter> - cups-filters: Merged CUPS-only foomatic-rip from GSoC project into cups-filters, making cups-filters the new upstream home for it.
[15:49] <tkamppeter> - cups-filters: Cleaned up foomatic-rip code eliminating warnings, sharing colord support with gstoraster, fixing small bugs.
[15:49] <tkamppeter> - cups-filters: Several bug fixes and clean-ups.
[15:49] <tkamppeter> - CUPS SRU for Saucy to get lpadmin -m to work.
[15:49] <tkamppeter> - Bugs.
[15:49] <seb128> new glib/dconf, we should probably get going on those at some point, we maybe let's clear some of the "start of the cycle" backlog first
[15:49] <seb128> Laney, ^ agreed?
[15:50] <seb128> like end of the month seems good to me (once we get webkit/eds/gtk in)
[15:50] <desrt> seb128: new glib is a party... dconf is very very boring... the only change of consequence there was to support a glib change
[15:50] <seb128> tkamppeter, thanks
[15:50] <Laney> mmm, yeah
[15:50] <desrt> (the new "is this key set?" API that was requested by robert_ancell and didrocks for some time now)
[15:50] <Laney> need to fix the build failures in debian before we can go to a new series really
[15:50] <Laney> so would be uploading to ubuntu only
[15:50] <desrt> Laney: i was planning to take a look there today
[15:50] <seb128> tkamppeter, thanks
[15:50] <Laney> neat
[15:51] <desrt> Laney: i'll poke you after the meeting
[15:51] <seb128> desrt, thanks
[15:51] <seb128> attente, hey
[15:51] <Laney> there's also https://buildd.debian.org/status/fetch.php?pkg=glib2.0&arch=armel&ver=2.38.1-2&stamp=1383832557 a different failure on arm that happens some of the time only
[15:51] <attente> hi seb128
[15:52] <attente> just working on the input switcher
[15:52] <seb128> attente, how is the compiz grabber issue going? still stucked on this one?
[15:52] <attente> i uploaded a ppa, but there's one problem that prevents it from being usable
[15:53] <seb128> what's the issue?
[15:53] <attente> seb128, i wouldn't hesitate to advertise the ppa if only for this one issue being fixed
[15:53] <desrt> Laney: damn testcases always failing on weird architectures... i almost think we should just get rid of the test suite...
[15:54] <Laney> make -k check || true
[15:54] <attente> when you log into unity, it always starts with the shortcuts set to super+space and shift+super+space, and doesn't read it from gsettings until it's changed
[15:54] <seb128> attente, oh, so you found the other issues where you were getting the wrong event? what was it?
[15:54] <attente> seb128, yeah, the other problem is a non-issue any more
[15:55] <seb128> great
[15:55] <attente> it was that compiz only cycled through CompOptions for checking key events
[15:55] <attente> so even if you create an action and add it so that compiz can do the grab
[15:55] <attente> if you're doing that manually, compiz still won't forward the events unless it's attached to a plugin option
[15:56] <attente> seb128, feel free to try to ppa out
[15:56] <attente> it's just that one issue i'm working on now
[15:56] <attente> https://launchpad.net/~attente/+archive/modifier-only-input-switch
[15:56] <seb128> attente, well done on fixing that one, I'm going to try the ppa, seems like that the dynamic update shouldn't be a stopper to play with it
[15:56] <seb128> attente, thanks
[15:57] <attente> seb128, it would probably frustrate users to always have to set it every time they log back in
[15:57] <seb128> attente, oh, are you subscribed to ubuntu-devel@ list? I just send a (long) email there to summarize the keyboard changes, the issues, and what to do for the LTS
[15:57] <desrt> 'stop shipping compiz' ;)
[15:57] <attente> seb128, i saw that, but didn't read it in detail yet
[15:57] <seb128> attente, right, it's probably not user friendly, but for debug testing it seems it should do the job
[15:57] <seb128> desrt, gnome-shell has some of the same issues...
[15:57] <attente> seb128, thanks for starting that discussion though :)
[15:57] <larsu> desrt: be nice.
[15:57] <seb128> attente, yw ;-)
[15:58] <desrt> seb128: window managers are trouble, i'm telling you...
[15:58] <seb128> desrt, xkb is the issue there
[15:58] <seb128> desrt, https://bugs.freedesktop.org/show_bug.cgi?id=55585#c0 is the issue I'm most concerned about
[15:58] <ubot2> Freedesktop bug 55585 in UI "Should check all XKB group indexes when matching key events for accelerators and mnemonics" [Normal,Resolved: duplicate]
[15:58] <cyphermox> seb128: yo, I tried all I could with the ejabberd bug, but I can't reproduce the fix
[15:59] <cyphermox> seb128: I'm going to try a few more minutes, but so far no luck :(
[15:59] <seb128> cyphermox, hey, ok, thanks ... can you comment on the bug saying that?
[15:59] <happyaron> except Kylin-related stuff, I went over almost all bugs of ibus on LP and did some triaging work, and borrow what desrt said, "input methods are touble"...
[15:59] <cyphermox> yeah
[15:59] <seb128> cyphermox, thanks for trying
[15:59] <seb128> attente, thanks
[15:59] <seb128> larsu, hey ;-)
[15:59] <larsu> hi!
[15:59] <larsu> - prepared and tested gtk 3.10 to be ready for prime time (updating patches and fixing the new test suite to work in tree); it's coming along nicely with only a few minor bugs left to fix
[15:59] <larsu> - fixed a couple of bugs in ido, indicator-{network,messages,appmenu}, and notify-osd
[15:59] <larsu> eof
[15:59] <larsu> short one :)
[16:00] <seb128> desrt, (not specifically only libreoffice, but eclipse and who-know-what-other-apps have the same issue and I'm not even sure they are all opensource/stuff we can fix)
[16:00] <desrt> larsu: didn't see that gtk action_namespace bug yet...
[16:00] <larsu> desrt: pings, meetings
[16:00] <desrt> fair enough :)
[16:00] <seb128> larsu, thanks, I'm waiting for the next GTK update round (hopping that gives me my menus back :p)
[16:00] <larsu> desrt: I'll try to get it out before leaving
[16:01] <larsu> seb128: I already have a patch for i-appmenu to fix that. It's been a ping-happy two hours
[16:01] <larsu> and I have a meeting right now :)
[16:01] <seb128> happyaron, did you want to do a status update with us? ;-)
[16:01] <desrt> larsu: if it's really going to take too long, just commit it.... but make sure you have a good message, and stable-release backport it
[16:01] <desrt> the code is obviously-incorrect as stands...
[16:02] <happyaron> seb128: already said a bit, :) "except Kylin-related stuff, I went over almost all bugs of ibus on LP and did some triaging work, and borrow what desrt said, "input methods are touble"..."
[16:02] <larsu> desrt: I'll file a bug
[16:02] <seb128> happyaron, oh, right, thanks! ;-)
[16:02] <desrt> larsu: thanks
[16:02] <seb128> my turn
[16:02] <seb128> * continued on desktop merges/syncs for trusty (libzip, opencc, curl, gnome-keyring)
[16:02] <seb128> * some desktop updates (gdk-pixbuf, fontconfig, freetype, dbus, file-roller, evince)
[16:02] <seb128> * some bug fixes and SRUs
[16:02] <seb128> * looked a bit to autopilot tests for ubuntu-system-settings, they are having issues due to the new autopilot version, trying to get that merged in this week
[16:02] <seb128> * spent some time thinking about keyboard layout/IMs and what to do next, wrote detailed email to ubuntu-devel@
[16:02] <seb128> * reviewed GTK 3.10 update from larsu (some packaging nitpicks but it was mostly good), thanks Lars for the work!
[16:02] <seb128> * tested a bit GTK, reported some issues that need to be addressed before ppa upload
[16:02] <seb128> * continued looking at errors.ubuntu.com top issues and nagging maintainers about their bugs

[16:04] <seb128> ok
[16:04] <desrt> seb128: thanks for the nags :)
[16:04] <seb128> so, any question/comments/...?
[16:04] <desrt> (none this week, but in general...)
[16:04] <seb128> desrt, yw ;-)
[16:04]  * happyaron is not quite familiar with this kind of status update on IRC, cuz we use phone conference call previously...
[16:04] <seb128> desrt, speaking of nag, did you see my gnome-keyring one to you/pitti yesterday?
[16:05] <desrt> happyaron: irc meeting is best meeting
[16:05] <desrt> seb128: ya.  i hoped pitti would look at it :)
[16:05] <desrt> what was the link again?
[16:05] <seb128> desrt, hehe ;-)
[16:05]  * desrt wanted to sort these ppc failures today
[16:05] <happyaron> desrt: /me is trying to get familiar with doing updates on IRC from now, :)
[16:06] <desrt> happyaron: are you new on the team, or...?
[16:06] <seb128> desrt, I didn't give a link, basically gnome-keyring build never finish because the gnome-keyring-daemon spawned by the testsuit doesn't exit when the bus goes away
[16:07] <desrt> seb128: some weird test-dbus interaction?
[16:07] <seb128> desrt, stefw said it's a known issue and he would welcome a patch for it (to make the daemon exit when the bug goes away)
[16:07] <desrt> dbustestrunner, whatever
[16:07] <seb128> bug->bus
[16:07] <happyaron> desrt: I joint 3months, but most of my work is for Kylin project, and we have status update meeting at that side already. But also good to do another one here, though.
[16:07] <seb128> desrt, bz 708765
[16:07] <desrt> happyaron: i guess you're attente's natural enemy :)
[16:07] <desrt> seb128: thanks
[16:08] <seb128> desrt, thank you for maybe looking at it ;-)
[16:08] <happyaron> desrt: why? ;-)
[16:08] <happyaron> for the keyboard related stuff?
[16:08] <desrt> opposing input method frameworks
[16:08] <happyaron> haha
[16:09] <desrt> seb128: maybe you're welcome :)
[16:09]  * desrt conducts a very important experiment
[16:09] <desrt> FOILED
[16:09] <desrt> (my excessively long headphone cord is not long enough to allow me to reach the coffee pot in the next room)
[16:12] <happyaron> desrt: call with your phone app, and you can move around with wifi, :)
[16:12] <desrt> seb128: i hate you
[16:13] <seb128> desrt, did you found that gnome-keyring's bug is a glib issue or what? ;-)
[16:13] <desrt> no
[16:13] <desrt> i found that it's using eggdbus
[16:13] <seb128> eggdbus
[16:13] <seb128> lol
[16:13] <larsu> what's that?
[16:13] <seb128> I didn't even know that was a think
[16:13] <larsu> do I want o know?
[16:13] <seb128> thing
[16:13] <desrt> it's not really
[16:14] <desrt> it's just common copy/paste code to handle the mainloop integration between libdbus-1 and glib
[16:14] <desrt> for the most part it's using libdbus-1 API directly
[16:14] <Laney> ooh, that sounds like a hairy yak
[16:14]  * desrt will _not_ rewrite against gdbus
[16:14]  * desrt will _not_ rewrite against gdbus
[16:14]  * desrt will _not_ rewrite against gdbus
[16:15] <seb128> lol
[16:15] <mlankhorst> desrt->override();
[16:15] <Laney> mterry: please could you re-check over https://bugs.launchpad.net/ubuntu/+source/libwebp/+bug/1186553 ?
[16:15] <ubot2> Launchpad bug 1186553 in libwebp (Ubuntu) "[MIR] libwebp" [Undecided,New]
[16:15] <mlankhorst> works in movies! Now rewrite against gdbus. :P
[16:16] <desrt> so uh.  i think i'm done
[16:16] <mterry> Laney, ah an easy one!  :)
[16:17] <Laney> \o/
[16:17] <mterry> Laney, I like it when past-me does all the work
[16:17] <Laney> unless you mean → Opinion
[16:17] <Laney> :P
[16:17] <mterry> heh
[16:18] <desrt> seb128: so uh... how would i go about testing such a patch?
[16:19] <desrt> 'cause it's a one-liner
[16:19] <cyphermox> seb128: of course, because I suddenly stop being an idiot it finally works
[16:20] <seb128> desrt, if you give it to me I can throw it to the builder
[16:20] <desrt> https://bugzilla.gnome.org/show_bug.cgi?id=708765
[16:20] <seb128> cyphermox, ;-)
[16:20] <ubot2> Gnome bug 708765 in general "D-Bus-activated keyring-daemon remains even when its bus terminates" [Normal,Unconfirmed]
[16:20] <desrt> https://bug708765.bugzilla-attachments.gnome.org/attachment.cgi?id=259673
[16:21] <seb128> desrt, the bug description suggests that Stef only wants that behaviour in the case where it's dbus activated though, your patch would always make it exit?
[16:21] <seb128> or is that a different codepath in the source?
[16:22] <desrt> why would we want it to persist when explicitly started?
[16:22] <seb128> well, I guess he has a reason, maybe better to ask him
[16:22] <desrt> oh.  i see.
[16:22] <desrt> ya. i'll get on that
[16:22] <seb128> thanks
[16:22] <desrt> meanwhile, if you could test the patch anyway...
[16:23] <desrt> ie: to find out if this mechanism even works
[16:23] <larsu> desrt: greyback will forward you some question from tvoss (from a mailing list) regarding the desktop file cache
[16:23] <larsu> desrt: just to let you know ;)
[16:23] <desrt> k
[16:24] <Laney> pitti: looks like you forgot to push rhythmbox
[16:30] <Laney> seb128: webkitgtk is up
[16:30] <seb128> Laney, \o/
[16:30] <Laney> for all your NEWing pleasure
[16:33] <pitti> Laney: how do you mean? there's no Vcs-Bzr
[16:33] <seb128> Laney, NEWed, let's the buildds have some fun ;-)
[16:35] <seb128> jdstrand, hey, I want to update telepathy-mission-control-5 to the current stable serie (5.16) ... is that something you started on/are interested doing (I guess not)? if not, how do you usually test that the apparmor profile needs tweak? just running empathy&co and checking that everything work/the logs have no DENY?
[16:38] <Laney> pitti: oh, maybe I dropped it by error?
[16:38] <pitti> Laney: I didn't really look, as often packages get synced, and then UDDed, etc.
[16:38] <Laney> mmm
[16:38] <Laney> nm, will fix
[16:38] <pitti> Laney: I'll import my upload
[16:40] <pitti> Laney: ^ done
[16:41] <pitti> Laney: committed Vcs-Bzr: fix
[16:41] <Laney> danke!
[16:41] <Laney> oh, you beat me
[16:41] <Laney> sorry for dropping it
[16:41] <pitti> np
[16:52] <seb128> desrt, hum, patch doesn't seem to make builders happier ... I'm going to need to try to find a machine/setup where I can reproduce I guess
[17:01] <larsu> seb128: this indicator-appmenu branch makes menus work with 3.10: https://code.launchpad.net/~larsu/indicator-appmenu/insert-action-groups-on-menus/+merge/194901
[17:01] <seb128> larsu, thanks
[17:02] <larsu> seb128: I'll post a gtk patch to make the indicator-messages problem go away
[17:02] <seb128> larsu, great
[17:02] <larsu> didn't end up having time for the theming/scrollbar stuff today, sorry
[17:05] <seb128> larsu, no worry, I count on that update to be work for the week
[17:10] <larsu> desrt: https://bugzilla.gnome.org/show_bug.cgi?id=712164
[17:10] <ubot2> Gnome bug 712164 in GtkApplication "GtkMenuTracker: set action namespace correctly" [Normal,New]
[17:13] <desrt> thx.  will review today.
[17:13] <seb128> larsu, indicator-appmenu.c:548:2: error: 'GtkStock' is deprecated [-Werror=deprecated-declarations]
[17:13]  * seb128 disables Werror
[17:13] <desrt> ...
[17:14] <desrt> didn't we take a '-Wno-error=deprecated-declarations' shotgun to everything already?
[17:14] <larsu> seb128: see the prerequisite branch
[17:14]  * larsu ran into the same issue
[17:15] <seb128> shrug
[17:15] <seb128> larsu, ignore me, good that I hit that, I autofinger-ed "cd indicator-appmenu; bzr bd"
[17:15] <seb128> larsu, where I mean to cd insert-...
[17:16] <larsu> seb128: heh, no problem
[17:16] <seb128> yeah, working menus!
[17:16] <larsu> seb128: gotta run now. See you in the morning!
[17:16] <seb128> larsu, have fun, see you tomorrow ;-)
[17:17] <larsu> desrt: take your time, I won't need it until tomorrow morning anyway
[17:17] <desrt> larsu: have a beer for me :)
[17:17] <larsu> seb128: thanks :)
[17:17] <larsu> desrt: will do.
[17:17] <seb128> larsu, don't troll too much :p
[17:17]  * larsu will be drunk
[17:17] <seb128> haha
[17:17] <desrt> oh
[17:17] <desrt> in that case, don't smoke crack cocaine and get caught on video
[17:17] <larsu> okay..... I'll try
[17:17] <desrt> i hear that's sometimes a problem when people get drunk
[17:18] <larsu> 1 beer for me + 1 beer for desrt = 2 beers --> larsu == drunk
[17:18] <larsu> desrt: only in Toronto
[17:18] <desrt> :)
[17:18] <larsu> that's why I got out of there!
[17:19] <seb128> desrt, do you see any issue with https://git.gnome.org/browse/gnome-menus/commit/libmenu/desktop-entries.c?id=7d90221e27fcc72de497f45db4f3ef3c2d623fd6?
[17:19] <seb128> desrt, just having a random guess at what could create https://bugzilla.gnome.org/show_bug.cgi?id=712161 (assuming it's an issue in gnome-menus, I had a look in the commit in 3.6->3.8)
[17:19] <ubot2> Gnome bug 712161 in libgnome-menu "software-center hits segfault in cached_dir_ref()" [Normal,Unconfirmed]
[17:20] <desrt> seb128: pretty hard to imagine that causing a problem
[17:20] <seb128> k, thanks
[17:21] <seb128> oh, the full commit is https://git.gnome.org/browse/gnome-menus/commit/?h=gnome-3-8&id=7d90221e27fcc72de497f45db4f3ef3c2d623fd6
[17:21] <seb128> I started from the file where the issue is happening and had a filter on that
[17:21] <desrt> oh.  fun.
[17:21] <desrt> the notify part looks suspicious
[17:22] <seb128> that's where the segfault is happening :p
[17:22] <desrt> this patch could possibly cause issues, indeed
[17:23] <desrt> i feel like smacking jasper for writing this
[17:24] <seb128> desrt, oh, speaking of Jasper, he just suggested in the bug to try to revert that commit ... ;-)
[17:24] <desrt> ya
[17:24] <desrt> that's my suggestion too :)
[17:24] <desrt> there's all kinds of suspicious in this patch
[17:24]  * seb128 tries that
[17:24] <desrt> nothing overtly wrong, but a very difficult-to-read change
[17:25] <desrt> his singleton management strategy is utterly unthreadsafe, as an example
[17:25] <desrt> in a patch that appears to be attempting to introduce threadsafety
[17:51] <pitti> desrt: I'm beginning to get a handle on bug 1211514
[17:51] <ubot2> Launchpad bug 1211514 in systemd-shim (Ubuntu Trusty) "Shutdowns fail to finish if laptop lid is closed before completely shutdown" [High,Triaged] https://launchpad.net/bugs/1211514
[17:51] <pitti> desrt: so apparently the impossible is possible after all
[17:52] <desrt> pitti: your log said that we weren't getting the call...
[17:53] <pitti> desrt: both logs together say that we get both calls
[17:53] <pitti> desrt: but it seems the first shim (for poweroff) immediately gets terminated again
[17:53] <pitti> and then the second shim gets started for suspend
[17:54] <pitti> desrt: I suspect it races with D-BUS shutdown; does the shim listen to d-bus going down and kill itself? or is that our "kill everything on powerdown" script?
[17:54] <desrt> pitti: if the system dbus went down then we'd not be able to get a second message through
[17:54] <pitti> well, I'll test both hypotheses
[17:54] <desrt> but maybe it's the SIGTERM that shutdown sends...
[17:55] <desrt> that would shut us down indeed
[17:55] <pitti> desrt: not sure if d-bus would still spawn new services while it's being shut down and it is killing existing services
[17:55] <pitti> so the "kill -TERM $world" during shutdown is more plausible
[17:55] <pitti> I'll try with adding that to sendsigs.omit
[17:55] <desrt> pitti: ya.  i'm pretty sure it's the SIGTERM
[17:58] <pitti> desrt: my first theory was that logind objects that calling suspend.target doesn't actually fail
[17:58] <pitti> desrt: but I monkey-patched it to actually send a D-BUS error instead of a (silent) success, that didn't help
[17:58]  * pitti shelves that, in case it's still relevant after not killing the shim any more
[18:01] <pitti> oh the joy of (not) being pid 1..
[18:02] <desrt> :)
[18:02] <desrt> i heard that having everything in pid 1 is a bad design decision
[18:02] <desrt> seb128: hey... you know our weird gettext desktop file business?
[18:03] <pitti> desrt: yeah, Lennart really ought to take the mail server out of it
[18:03] <didrocks> seb128: do you know if we dropped the patch that desrt has done in gtk for setting alt as a shortcut in g-c-c?
[18:04] <didrocks> if not, no worry, I'll check… one day :)
[18:04] <desrt> didrocks: we kept it
[18:04] <didrocks> interesting… something else regressed it seems
[18:04] <desrt> didrocks: larsu wanted to drop it yesterday (or get it upstream or fix the situation) and i talked him out of it
[18:05] <didrocks> I'm sure it's larsu's fault then!
[18:06] <didrocks> desrt: where is seb128, do you think he's drinking again? :)
[18:06] <desrt> :)
[18:13] <pitti> desrt: et voilà !
[18:14] <desrt> pitti: systemd-shim now on the no-kill list and it works?
[18:14] <pitti> http://paste.ubuntu.com/6406628/
[18:14] <pitti> desrt: yes, tested two times
[18:14] <desrt> ugh.  we have to write our pid there?
[18:14] <pitti> desrt: that's my current hack, not factored out into a function and using glibc-y calls
[18:14] <desrt> pitti: also problematic from the standpoint of not erasing the file when we exit
[18:14] <pitti> I suppose you'd rather want some glib-y/Gio stuff for that?
[18:15] <desrt> i guess we should only write it just before we call shutdown
[18:15] <pitti> desrt: we do
[18:15] <desrt> oh.  unlink.  hi.
[18:15] <pitti> desrt: second hunk
[18:15] <desrt> sorry.  missed it :)
[18:15] <desrt> pitti: i think i'll adjust it to only write just before calling shutdown
[18:15] <pitti> desrt: anyway, need to run now; we can discuss the finer details tomorrow or so
[18:15] <desrt> we avoid the write/unlink in the other cases this way
[18:15] <desrt> pitti: thanks!
[18:15] <pitti> desrt: yes, indeed
[18:16] <pitti> as I said, just a first-cut "does this work?" it's not really pretty yet
[18:16]  * pitti waves good night
[18:53] <xclaesse> seb128, I've heard that you uploaded the empathy update to fix facebook
[18:53] <seb128> xclaesse, yeah, the day you pinged me about it
[18:53] <xclaesse> seb128, but I don't see update coming here... is it stuck somewhere?
[18:53] <seb128> xclaesse, it's out for some weeks? what serie do you use?
[18:54] <xclaesse> seb128, argh, forget me, it actually works, had another issue because I was running stuff from master
[18:54] <xclaesse> nevermind
[18:54] <seb128> xclaesse, https://launchpad.net/ubuntu/+source/empathy/3.8.4-1ubuntu2
[18:54] <seb128> xclaesse, ok ;-)
[18:55] <xclaesse> after kill the stuff running in jhbuild, and restart the distro version, all is fine :)
[18:58] <xclaesse> now it is MSN that does not connect anymore
[18:58] <xclaesse> SRV connect failed: g-io-error-quark:39 Could not connect to messenger.live.com: Connection refused
[18:58] <xclaesse> looks like they finally removed their xmpp server for good
[18:59] <seb128> :-(
[18:59] <seb128> we should probably drop the UI option for it then...
[19:05] <seb128> xclaesse, ^ do you what needs to happen for that?
[19:06] <xclaesse> seb128, on ubuntu that should be enough I think: rm /usr/share/accounts/providers/windows-live.provider
[19:07] <xclaesse> you probably want to inform the user though
[19:07] <seb128> kenvandine, ^ should we do/SRU that?
[19:07] <seb128> xclaesse, right ... do you plan to add something upstream for informing users?
[19:07] <xclaesse> seb128, maybe we should wait to see if there is an official announce from MS, it could be just temporary failure
[19:07] <seb128> yeah, let's see
[19:07] <xclaesse> I don't remember the date they announced months ago
[19:07] <seb128> though I'm sure they are going to close it at some point
[19:08] <Laney> If you have a Mac operating system earlier than OSX or a Windows operating system earlier than Windows XP, you won't be able to update to Skype. Messenger will continue to be available on these operating systems until it is discontinued later in 2013.
[19:11]  * desrt realises the symbolic beauty of the battle between systemd-shim and upstart
[19:12] <desrt> systemd-shim fires the first shot: i shutdown you!
[19:12] <desrt> upstart fires back: SIGTERM
[19:12] <xclaesse> http://social.msdn.microsoft.com/Forums/live/en-US/44ad3195-6e7e-4192-95e5-bf5f282eab01/support-for-xmpp-ends-in-october-2013?forum=messengerconnect
[19:12] <desrt> systemd-shim dies
[19:12] <desrt> dbus: no!  my friend! *revives*
[19:12] <desrt> systemd-shim: i'm back!  i will suspend you!
[19:12] <desrt> *system suspends*
[19:13] <seb128> desrt, why does it suspend it when being respawned?
[19:14] <seb128> oh, I guess because the lid get closed and it doesn't have the flag telling him shutdown is happening anymore I guess?
[19:14] <desrt> yup
[19:14] <desrt> the SIGTERM from upstart wipes that out
[19:14] <desrt> and then because it no longer exists, dbus reactivates it
[19:14] <seb128> is any client trying to use it?
[19:15] <desrt> ya... when the lid closes, logind sends a message
[19:15] <desrt> to systemd
[19:15] <desrt> which restarts the shim, with the flag gone
[19:15] <desrt> it's really hilarious how complicated such a simple thing is
[19:15] <desrt> due to the various processes involved
[19:16] <desrt> and how 'abnormal' things are during shutdown
[19:18] <seb128> yeah
[19:18] <seb128> what do you think is the right fix there then?
[19:18] <desrt> i'm going to do something close to pitti's fix
[19:18] <desrt> instruct upstart not to kill us
[19:18] <seb128> can you do that?
[19:18] <desrt> yup
[19:18] <desrt> that's this sendsigs.omit.d stuff
[19:19] <desrt> fwiw, i don't really fully understand why this old unix style of "kill all the things" is still used
[19:19] <desrt> i guess SIGTERMing world+dog does do two useful things...
[19:19] <desrt> 1) reduces the use of filesystems allowing unmounts/readonly-remounts
[19:20] <desrt> 2) gives everyone a chance to save state
[19:20] <desrt> #1 is a bit silly in the modern day though... we could just explicitly target those processes
[19:20] <desrt> since we have /proc to tell us...
[19:21] <sarnold> /proc is racy..
[19:21] <desrt> sarnold: so is sigterm(*)
[19:22] <desrt> since soemthing could fork between you getting the list of things to kill and sending the signals
[19:22] <desrt> so in any case you have to do multiple passes
[19:22] <sarnold> desrt: which is why kill() and killpg() let you do some selectors themselves.. 0, -1, etc..