=== duflu_ is now known as duflu === chihchun_afk is now known as chihchun === chihchun is now known as chihchun_afk === chihchun_afk is now known as chihchun [11:20] alan_g: perhaps we should go with different executables in the existing (acceptance/integration/unit) directories, e.g., acceptance-tests(-main), acceptance-tests-c89, acceptance-tests-symbol-resolution ... [11:20] alan_g: unit-tests, unit-tests-umockdev etc [11:21] alf_: maybe (although needing umockdev seems more integration/acceptance-test to me) [11:23] alan_g: yeah, link-time seams blur the test category lines [11:25] alf_: maybe that's the key idea: "MIR_ENABLE_SPECIAL_LINKAGE_TESTS"? [11:27] alan_g: that would leave out the c89 tests, MIR_ENABLE_SPECIAL_BUILD_TESTS? [11:28] alf_: do we need to include the c89 test in this category? [11:28] * alan_g doesn't think we even need to link c89 test - it should be compile time === chihchun is now known as chihchun_afk [11:31] alan_g: No, we don't need to include it really, but we need to enable it by default somehow. I guess we can start with SPECIAL_LINKAGE + separate c89 and reconsider if new "special" tests arrive. [11:32] alan_g: hmm, or is c89 already built? [11:32] alf_: I thought it was on by default (but haven't looked recently) [11:33] alan_g: ok, it is still there, so yes, SPECIAL_LINKAGE is good enough for now if we want to go the separate category path [11:33] And using link seams does interfere with other tests, so they do make a coherent category of sorts. === chihchun_afk is now known as chihchun === chihchun is now known as chihchun_afk === alan_g is now known as alan_g|lunch === dandrader is now known as dandrader|afk === dandrader|afk is now known as dandrader [14:29] Anyone here privy to the details on why Software Center is going away after 14.04? https://docs.google.com/spreadsheet/ccc?key=0AiT4gOXSkmapdGdFejk0MjFydUlNMDVoMXNRdGdkbFE#gid=1 [14:29] Not sure where to ask - but trying to track down the story on this one - I presume click packages will be a replacement or something? === alan_g|lunch is now known as alan_g [14:34] FunnyLookinHat: the applications lens is replacing it. On desktop that will mean support for debian and click backends aiui, best place is possibly #ubuntuone though as it controls development of software-center now [14:34] davmor2, Cool thanks! === chihchun_afk is now known as chihchun [16:04] kgunn, Saviq http://paste.ubuntu.com/6474449/ [16:05] I followed the instructions in the unity8 wiki and am stuck at this stage [16:05] I feel like I am missing like 1 package but can't see it straight [16:06] kgunn, Saviq stderr is misisng, working on it [16:06] olli: watch out for ancient ppa's.....i had some really old qt edger ppas that hosed me [16:06] kgunn, this is a virign install on my new laptop [16:06] :) [16:07] olli: oh...nvmd === dandrader is now known as dandrader|lunch [16:10] olli, it's possible, we're in transition between in-built shell-facing scopes api and a package outside [16:11] olli, not everything got released yet [16:15] kgunn, ↑↑ [16:30] kdub: are you happy with the shm-buffers-part1 MP? If you also prefer base_ptr() instead of map() it's an easy change... [16:30] alf_, yeah, was just reviewing [16:31] kdub: (I will probably need to resync with trunk anyway) [16:31] kdub: s/trunk/development-branch/ [16:31] * kdub always calls development-branch trunk too [16:36] is anyone having any issues with the display settings after installing mir? [16:43] morning === dandrader|lunch is now known as dandrader [16:56] hey [17:42] alan_g, with more-cleanup-of-ownership-of-client-buffers, the changes to buffer ownership will probably trickle into mc::SwitchingBundle, right? [17:43] kdub: yes. [17:43] i'm finding some of the composition changes for overlay are tied to that class [17:43] trying to anticipate how we'll change [17:43] kdub: I don't have anything in flight right now - so you can sneak in ahead of me [17:44] i don't really want to wade into it if i can help it, but it might be necessary [17:45] * kdub keeps looking [17:56] kdub: I don't anticipate changing the logic - just the way buffers are handed out for clients. (Still playing with raw pointer vs unique handle). So I don't anticipate any conflicts with overlays [17:59] alan_g, okay. at the moment, i'm considering burying the 'frameno' logic back inside of mc::SwitchingBundle [18:02] kdub: is that feasible? Surely there needs to be some logic associated with the (possibly concurrent) compositors? [18:03] * alan_g decides it is too late in the day to worry === alan_g is now known as alan_g|EOD [18:08] the logic is needed, it just seems to me that the buffer tracker is the best place to keep it [18:09] instead of splitting responsibility out to the compositor and the buffer tracker [18:09] i'll see if its cleaner or not :) [18:28] hey folks, since the recent mir release, unity8 is crashing on shutdown [18:29] could you please have a look https://bugs.launchpad.net/mir/+bug/1253685 [18:29] Ubuntu bug 1253685 in unity8 (Ubuntu) "unity8 crashed with SIGSEGV in QMetaObject::activate()" [Medium,New] [18:29] from what I can see surface is deleted (surface=0x0) in inputarea.cpp [18:30] greyback, fyi ↑ [18:31] racarr, could you have a look at this over our night ↑? [18:31] Saviq: hmm, ok will have a look in a bit [18:31] greyback, well, it's EOD for you [18:31] kinda [18:31] did start late tho [18:31] greyback, your call [18:31] racarr, greyback it's very easy to reproduce - just Ctrl+C after running unity8 on the console [18:32] so I've another 30 mins [18:32] (stop unity8 before that) [18:32] * Saviq puts steps to repro [18:42] greyback, please try and find someone to continue past your EOD, though [18:43] unless you fix it, obviously [18:43] Saviq: ok [18:44] also, if you see it's fixed in latest Mir release, maybe we can just go ahead [18:44] (with latest mir + unity-mir + platform-api) [18:44] and we release that [18:45] * didrocks would appreciate an email about it in the morning, even if there is no progress [18:45] Saviq: Ok [19:14] racarr: if you're looking, I suspect this is the line causing the crash: line 90 of inputarea.cpp in unity-mir [19:15] racarr: probably that the InputArea has been destroyed before the MirSurface is [19:26] racarr: I've to run unfortunately, will be back in 4 hours [19:53] how extensively is DepthID used anymore? [19:53] also, recursive mutex, sad [20:17] greyback: Oi thanks [20:17] kdub: Still used to keep theshell surface on top [20:33] racarr, ah, okay... i can look past it for now :) [20:34] was just bemoaning our locking situation in surfacestack === racarr is now known as racarr|lunch [21:10] the new mc::DefaultDisplayBufferCompositor::composite is beautiful === racarr|lunch is now known as racarr