[02:14] robert_ancell: did you want to land that silo 59 xmir ? or is that just for testing.... [02:14] i guess question is, are you gonna land that using the citrain ? [02:14] kgunn, I'm waiting bregma and co to confirm it's really better before landing [02:15] robert_ancell: we were just discussing [02:15] robert_ancell: definitely better aiui [02:15] kgunn, oh, I was expecting to use the CI train - but I've never got that far with a silo before [02:15] Good [02:16] The downside is we can't land it into wily, so it has to go through the CI train to get into the overlay PPA right? [02:18] I think the strategy need to be to branch so we can land in vivid+overlay while wily is frozen, then backport to "upstream" for X (ie. 16.04 X, not x.org X) [02:18] otherwise we're blocked [02:19] and there's a good chance we'll never ship a device with wily (jumping from vivid+overlay to 16.04 LTS + overlay) [02:20] bregma, the package in the silo is a branched version - we'll update the XMir patch in x-series [02:20] anyway, robert_ancell, ChrisTownsend says silo 59 is much, much better (does not crash) but still has damaage problems, and I believe duflu has more bugfixes waiting to go in on top [02:21] landing silo 59 would be a good thing at this point [02:21] bregma, ok, cool. So we should land that into the overlay PPA [02:21] And do another silo when it's improved [02:21] Not more fixes in mind, but I have created three accurate-ish bug lists for Xmir [02:21] Lots of bugs remain [02:26] Does anyone know the next step for this silo. Do I just click publish on https://requests.ci-train.ubuntu.com/#/ticket/405 or do I need to ask someone from the CI team. [02:26] ? [02:31] robert_ancell, you need to change "No QA Needed" to "Publish without QA" [02:32] bregma, ok, done. Do I then publish myself? [02:33] robert_ancell, I think so, robru sent an email last week about that, but it doesn;t hurt to ask on #ubuntu-ci-train for clarification [02:34] I seem to have deleted the email [02:34] bregma, there's no-one in #ubuntu-ci-train it seems... What server? [02:35] ah, #ubuntu-ci-eng [03:07] bregma, kgunn, OK, it appears to be in the overlay PPA now [05:13] RAOF: Going for a diff record? :) [05:13] I see it's gone now [05:13] Turns out when you have a full-file conflict with every file in the repository the diff is pretty big! === chihchun_afk is now known as chihchun [05:28] RAOF: Happened again [05:30] Umm, the mir 0.16.0 tarball is 50MB [05:30] wtf? [05:31] Ooh. Cemil included the .bzr repo :) [05:34] Hehd. [06:10] We should probably work out why linking with ld.gold fails. [06:10] gold is much, *much* faster to link. [06:11] (Like roughly an order of magnitude) [06:11] RAOF: What is gold? [06:11] The gold linker. It's the replacement...ish for the standard bfd linkerl. [06:11] Oh https://en.wikipedia.org/wiki/Gold_(linker) [06:14] I think we might be technically underlinking. [06:16] Also have multiple wildcards in the symbols.map, which gold dislikes (and *might* be UB) [06:19] Well, avoiding the wildcards would at least make us think about ABIs more :) [06:19] Hopefully it would be semi-automated tho [06:20] No, I mean we have MIR_CLIENT_9 { global: stuff... local: * } MIR_CLIENT_9.1 { global: stuff... local: *}; [06:20] Oh [06:20] The two wildcards overlap, so it's not defined which block matches. [06:21] * duflu thinks he's not the only one who never bothered to learn about local:* [06:21] It's a poor-man's -fvisibility=hidden :) [06:22] Hm. We should totally use that for mirclient. [07:40] i guess only the last one should have local:*? [07:44] or only the first? === chihchun is now known as chihchun_afk [09:27] alan_g, alf_: https://code.launchpad.net/~andreas-pokorny/mir/libinput-platform-pointer-settings/+merge/272501/comments/687606 thoughts? [09:28] anpok: on my list [09:28] the actual answer spans over to comments.. one that was started by a sleepy-me.. [09:28] *two === marcusto_ is now known as marcustomlinson === marcusto_ is now known as marcustomlinson === chihchun_afk is now known as chihchun === alan_g is now known as alan_g|lunch === alan_g|lunch is now known as alan_g [13:35] guys, did i do this right? [13:35] https://code.launchpad.net/~aacid/mir/fix_forward_declarations/+merge/272751 [13:36] proper target branch and stuff? [14:07] Saviq: Do MPs for unity8 need an explicit changelog change or is the change automatically picked up from the commit messages? [14:08] alf_, no changelog, just the commit message on the merge proposal [14:08] Saviq: thanks [14:09] hi, should unity8 desktop session work in current wily, or did my mesa11 break it?-) [14:10] greyback_: CI now happy with https://code.launchpad.net/~alan-griffiths/qtmir/small-refactoring-of-MirWindowManager/+merge/266698 [14:21] tjaalton: I'm not sure what the state of u8 desktop is, but AFAIK the mesa 11 drop just broke Mir compilation, not runtime. [14:25] tjaalton, https://bugs.launchpad.net/ubuntu/+source/mir/+bug/1499425 [14:25] Ubuntu bug 1499425 in xorg-server (Ubuntu) "xmir black window (mesa 11.0.0)" [High,New] [14:32] tjaalton: hm a lot has changed inXmir .. I believe the wily version requires the -damage to workaround lots of redarw issues [14:32] *parameter [14:33] oh ok duflu already answered that.. [14:34] s/tjaalton/guest42315 [14:35] que dice? oh yeah :D -damage doesn't work [14:38] * guest42315 ubuntuonair in 20 min http://ubuntuonair.com/ === dandrader is now known as dandrader|afk [14:42] ah just compilation, no biggie then :) [14:42] so if the blank screen on login is due to the xmir bug then I guess it's known [14:46] and I did notice that one already, but thought it would be more native than using xmir [14:47] tjaalton: there is a high probability that those are not related [14:48] ok [14:50] could be I'm hitting a skylake display bug === dandrader|afk is now known as dandrader === chihchun is now known as chihchun_afk === alan_g is now known as alan_g|EOD === dandrader is now known as dandrader|afk === dandrader|afk is now known as dandrader === dandrader is now known as dandrader|afk === dandrader|afk is now known as dandrader === Cimi_ is now known as cimi