=== chihchun_afk is now known as chihchun [03:35] RAOF: Got a fix for devel cooking? https://bugs.launchpad.net/mir/+bug/1351133 [03:35] Launchpad bug 1351133 in Mir "Incomplete removal of libmirprotobuf causes reverse dependencies to FTBFS" [High,Triaged] [03:36] duflu: Have I not proposed that? [03:36] * duflu double checks [03:36] Can't see it [03:37] Bah. [03:37] I did it first, but didn't push it. [03:40] duflu: Enjoy [03:41] * duflu prefers other forms of enjoyment [03:41] https://code.launchpad.net/~raof/mir/finish-off-mirprotobuf-removal/+merge/229374 [03:47] RAOF: Glad to have a proposal to approve. ;) [05:37] RAOF: Err we appear to have broken mirclient dependencies used by gtk-3.0 in the coming Mir 0.6.0 :( [05:38] So things outside-the-silo (gtk 3) won't work any more [05:38] duflu: Really? What's the error? [05:38] https://bugs.launchpad.net/mir/+bug/1352149 [05:38] Launchpad bug 1352149 in Mir 0.6 "libmirclient.so.8 ABI broken but not incremented in Mir 0.6.0" [High,New] [05:39] No, that's ok. [05:39] Or, at least, it should be. [05:39] RAOF: What? The old one can linger? [05:39] Absolutely. [05:40] It'll be NBS (not built from source), but it's not an ABI break. [05:40] * duflu tries it [06:06] RAOF, nbs means the new source is going to be blocker in proposed by britney (if I understand correctly what you are saying) [06:07] blocked even [06:07] seb128: And so gtk3 gets a no-change rebuild, and we go back to base for debriefing and cocktails? [06:08] yes [06:08] just need that no change rebuild ;-) [06:08] well, if it's really no change [06:08] It really is. [06:08] wait [06:09] you are dropping a .so from a binary without renaming it? [06:09] Yes. [06:09] shrug [06:09] Dropping libmirprotobuf entirely. [06:09] that's going to make gtk apps fail to start [06:09] nothing is going to block mir to migrate in that context [06:10] but gtk, until rebuild, is going to try to load that library and fail to load [06:10] you can't do that [06:10] gtk will have a dependency on the libmirprotobuf package, right? [06:10] is that a package [06:10] Yup. [06:10] or is a .so in a binary that is not going away? [06:10] It's a .so in a package that *is* going away. [06:10] oh, ok, it was properly split [06:11] k, in this case all good [06:11] thanks for replying, sorry for the noise [06:11] No problem :) [06:12] Good to have you in to UTC+lots :) [06:20] seb128: I'm just checking that issue ;) https://bugs.launchpad.net/mir/+bug/1352149 [06:20] Launchpad bug 1352149 in Mir 0.6 "libmirclient.so.8 ABI broken but not incremented in Mir 0.6.0" [High,Incomplete] [06:41] RAOF: Do you remember the ldd equivalent command that doesn't list dependencies of dependencies? [06:48] objdump -p bin? [06:52] anpok_: Ah yes, thanks [06:52] Also: objdump -x /usr/lib/x86_64-linux-gnu/libgdk-3.so.0 | grep NEEDED [06:55] Oh, -p is shorter :) [07:18] camako: Anything blocking 0.6.0 that needs my help today? [07:18] duflu, yeah there was yet another blocker [07:19] duflu, alan_g is working on it though [07:19] camako: Awesome. Why haven't I seen it? [07:19] OK [07:19] (or will work on it) [07:19] duflu, it was in the downstream libs [07:20] camako: I'm accumulating branches that need a lot of fixing before reproposing them. Just checking on the state of things before I dive back into those large tasks [07:20] alan_g's branch is up but not finalized [07:20] duflu, understood. I think I might pick up the last couple of commits made on devel, too. [07:20] * duflu looks [07:21] duflu, after I make sure they don't break downstream. [07:21] specifically, would be nice to pickup r1811 [07:22] * camako is not sure if downstream components were tested before r1811 was approved. [07:23] camako: Best to keep 0.6 unmodified. I've started lining up items for 0.6.1 later (which is easier since there's no ABI break then) [07:24] duflu, yeah probably [07:25] OK, I've now reached my code review argument limit for the day. [07:25] * duflu goes to working on actual code [07:30] RAOF: Can you attach any known branches to this? https://bugs.launchpad.net/mir/+bug/1293944 [07:30] Launchpad bug 1293944 in Mir "Mir deb packages with versioned names cannot be installed simultaneously any more" [Medium,In progress] [07:30] duflu: WIP [07:30] RAOF: OK, so long as no proposed ones should be there already [07:31] duflu: That was https://code.launchpad.net/~raof/mir/privatise-all-the-things/+merge/228796 , but it's WIP while I get proper platform probing going. [07:53] RAOF: Hi! It turns out that we send a lifecycle connection lost event to ourselves when we release a client connection. Is this on purpose? [07:54] alf_: Yes [07:55] alf_: Because there didn't seem to be any reason not to, and this allows cleanup code to be in one place: the lifecycle event handler. [08:00] RAOF: the only complication is that we send a SIGTERM in the default lifecycle handler (which wasn't working well before, but is fixed by https://code.launchpad.net/~afrantzis/mir/client-lifecycle-terminate-properly/+merge/229234) [08:01] Yeah, I noticed that. Why are we doing that, and why aren't we sending SIGQUIT? [08:04] I guess having it overridable is better than libX11's “I KILL YOU WITH A SWORD” SIGPIPE, but... [08:07] alf_: Need me to do anything now? Otherwise I'll EOD. [08:09] RAOF: No, thanks, just wanted to ensure we send the event on purpose. [08:09] RAOF: Enjoy your day! === tvoss is now known as tvoss|test === tvoss|test is now known as tvoss === Ranomier2 is now known as Ranomier === pete-woods is now known as pete-woods|lunch === tedg` is now known as tedg === chihchun is now known as chihchun_afk === pete-woods|lunch is now known as pete-woods === pete-woods is now known as pete-woods|at-do === dandrader_ is now known as dandrader|lunch === pete-woods|at-do is now known as pete-woods === dandrader|lunch is now known as dandrader === alan_g is now known as alan_g|EOD [22:23] racarr, hey, do you know that cursors: mir_omnidirectional_resize_cursor_name and mir_closed_hand_cursor_name are the same? [22:23] also have you had the time to get a crosshair cursor landed :)? [22:34] bschaefer: I think omnidirectional resize cursor [22:34] and closed hand are only the same [22:34] in the ubuntu theme [22:35] but omnidirectional resize maps to "fleur" [22:35] and closed hand maps to "grabhand" or something [22:35] grabbing [22:35] ill look in to crosshair I cant remember why I skipped it [22:35] maybe ubuntu default them didnt have it... [22:37] interesting, yeah x11 uses fleur in place of grab hand [22:37] * bschaefer switches to omnidirection [22:37] racarr, theres also the case of [22:37] case SDL_SYSTEM_CURSOR_NO: shape = XC_pirate; break; [22:37] missing [22:37] the CURSOR NO seems to be like an X marks the spot on a treasure map [22:37] racarr, but cool, and thanks! [22:42] cursor no lol... [22:42] bschaefer: Maybe it's "disabled" [22:42] not as in disabled cursor [22:42] but [22:42] disabled action [22:42] like clicking here is not a valid [22:42] action [22:43] whioch I guess we dont have a cursor for either :) [22:43] lol XC_pirate [22:43] ...plus caret = "xterm", etc... [22:43] its like someone was intentionally trying to be funny with the cursor names -.- [23:45] Bah. [23:45] Why is mir_discover_gtest_tests suddenly segfaulting during global construction? [23:48] * RAOF tries a dist-upgrade. [23:52] haha oh no...I have been all happy that my qtmir acceptance test is supposedly passing but just realized the client should actually get denied connection.... [23:54] Hah [23:54] Ba baw! [23:55] oh no I guess it shouldn't [23:55] Hm. libstdc++6 upgrade, you say? [23:55] that logic is in the other part [23:55] that may be related :p [23:55] mir_discover_gtest_tests is not segfaulting for me