=== maclin1 is now known as maclin [08:32] tsdgeos, ah, it was xenial that had the aborts [08:32] and still does [08:33] somethings up with that xenial CI probably [08:36] indeed [09:01] https://requests.ci-train.ubuntu.com/static/error.gif [09:08] can't branch qtmir from my chroot :S [09:08] http://paste.ubuntu.com/14025322/ [09:08] any idea ? [09:08] actually adding -Ossl.cert_reqs=none after branch worked :D [09:13] Saviq: Is there a bug for the black screen on login? Because I'm finding the same problem with pure Mir demos [09:14] duflu, don't think I know the problem at all [09:14] Saviq: Tried Unity8 on xenial desktop lately? [09:14] duflu, ah that [09:15] duflu, bug #1525285 [09:15] bug 1525285 in clutter-gst-3.0 (Ubuntu) "inspecting clutter plugin hangs outside X11" [High,Confirmed] https://launchpad.net/bugs/1525285 [09:15] duflu, run gst-inspect-1.0 under X11 [09:15] tsdgeos, date issues? [09:16] ah right :D [09:16] that's what you get for changing your clock back one year [09:16] stupid me [09:16] yup, certificate not valid yet [09:16] wait, I didn't mean to say "yup" to your "stupid me" :P [09:18] * duflu hopes bug 1526209 is related to that somehow [09:18] bug 1526209 in mir (Ubuntu) "[regression] Clients of nested Mir servers silently crash/exit instantly" [Critical,Incomplete] https://launchpad.net/bugs/1526209 [09:19] Saviq: Oh I think there are two different black screen bugs. Do you see any part of the shell? [09:24] duflu, i see the launcher and the notifications bar [09:24] corn_field: On xenial desktop? [09:24] duflu, yes [09:24] corn_field: Fully updated? [09:25] duflu, well... updated yesterday i think upgrade & dist-upgrade [09:26] Hurrah. Multiple bugs. [09:27] So the rest of today will be bisection [09:27] fun fun fun! [09:31] duflu, that clutter one is just deadlock on startup, nothing on screen [09:32] except for u-s-c's cursor [09:32] Saviq: Yeah multiple bugs. I see the shell bits at the top but no dash. Looks like that's part of bug 1526209 [09:32] bug 1526209 in mir (Ubuntu) "[regression] Clients of nested Mir servers silently crash/exit instantly" [Critical,Incomplete] https://launchpad.net/bugs/1526209 [09:32] Although different Mir version [09:32] *shruf* [09:32] shrug [09:34] duflu, here, after I got rid of the gst deadlock, everything's fine'n'dandy [10:16] dandrader: morning, have a second to help me? [10:16] tsdgeos, yes [10:16] dandrader: so Gerry yesterday said "events come from mir, are converted from MirEvent to QTouchEvent in qtmir, then converted back to MirEvent to be sent to client." [10:17] i'm trying to find out the place where " converted back to MirEvent to be sent to client" happens [10:17] and i thought it'd be MirSurface::mousePressEvent & friends [10:17] but i've added debugs and breakpoints in there and i can't see them triggering [10:17] so maybe it's not there? [10:18] tsdgeos, yes, MirSurface is doing that. let me check [10:19] dandrader: if i understand correctly those debugs should be in unity8's log, rigiht? [10:19] tsdgeos, you're looking for mouse or touch event generation? [10:19] dandrader: i guess touch maybe? [10:19] tsdgeos, no, they're disabled by default as they're too verbose [10:20] dandrader: yes, but they should be in unity8's log not in the client, right? [10:21] tsdgeos, there's the makeMirEvent family of functions in src/modules/Unity/Application/mirsurface.cpp [10:21] yes [10:21] tsdgeos, right [10:21] but client side also takes the mir events coming from the server and translate into qt ones, naturally [10:21] yes [10:22] ok, understood then, tx [10:23] tsdgeos, ah, actually there's qtmir logging for mirserver -> unity8's qml scene, but not for unity8's qml scene -> mir client [10:23] yes, that's what i am adding [10:23] tsdgeos, guess there as never a need for it [10:24] tsdgeos, iirc, I usually added them in qml code when needed === vrruiz_ is now known as rvr [10:40] greyback, when I said you need to bump qtubuntu-gles, I meant http://bazaar.launchpad.net/~mir-team/qtubuntu/gles-sync/revision/45/debian/changelog [10:40] now why this whole thing went apeshit in the PPA is another question, partly answered by https://requests.ci-train.ubuntu.com/static/error.gif probably [11:21] dednick: greyback: https://code.launchpad.net/~aacid/qtmir/timestampsInPast/+merge/280567 [11:22] and https://code.launchpad.net/~aacid/qtmir/unlikelyResetStartTime/+merge/280569 if you're in the mood of micro-optimizations :D [11:22] tsdgeos: nice work. dednick can you review plz, since you know it better? [11:23] tsdgeos: Q_UNLIKELY exists [11:23] ah right, forgot that code already had Qt includes [11:26] @unity looks like the touch emulation from mouse events is broken in trunk. eg. can't drag down the indicators panel or swipe away the lockscreen in "make tryShell". Can you guys reproduce it? [11:26] hey pstolowski, did it build? [11:27] @unity wonder if it's because I installed latest ubuntu-ui-toolkit which has it's own UbuntuGestures.... [11:28] dandrader, can confirm [11:28] cimi, no, unity8 failed, you probably need to bump debian/control dependencies as it now needs 'unity-shell-application=12 [11:28] or tsdgeos ^ [11:28] pstolowski, that's in trunk [11:30] tsdgeos: , greyback: sure [11:30] pstolowski: ok [11:30] dandrader: but are we loading it? or it loads automagically? [11:31] pstolowski: filters, right? [11:31] Saviq, tsdgeos ah i see what happened, you bumped unity-api to 7.104 in trunk, so now we need to bump to 7.105 in the silo [11:31] tsdgeos, yes; we need 7.105 now, i'm bumping shell plugin and unity-api, pls bump in unity8 [11:32] (in the filters branch) [11:34] tsdgeos, since uitk qml plugin [11:34] tsdgeos, depends on it, I think that lib will be loaded [11:34] Saviq, by "can confirm" you mean you can reproduce the issue or you mean you will check? [11:35] pstolowski: ok, pushed [11:36] thx [11:38] dandrader, I reproduced, yes, sorry [11:39] dandrader, you need to bump unity-api dependency in lp:~dandrader/qtmir/sizeHints [11:44] Saviq, done [11:44] dandrader, tx === dandrader is now known as dandrader|afk [11:51] dandrader|afk, merge trunk, too === dandrader|afk is now known as dandrader [12:34] Saviq, done [12:34] tx [12:46] hello folks, i'm getting these errors when running any X application with Xmir: http://pastie.org/private/evaqzjyuvhwtudi4li5yhw i know there is no support for this, but maybe you might have an idea what could be the problem, or where to start looking at ? thanks in advance [12:49] maikeul: I suggest you ask in #ubuntu-mir [12:50] ha thanks, my bad === alan_g is now known as alan_g|lunch [13:38] Saviq, do you know when we changed from qt 5.4 to 5.5? [13:38] dandrader, it migrated last week [13:38] dandrader, on 10.12 [13:38] and UITK on 11.12 [13:42] ltinkl: it's the same diff as before it was approved, but ok [13:42] tsdgeos, yeah but we found new issues [13:42] Saviq, the problem is that MouseTouchAdaptor is no longer working because the xcb event types have changed with qt 5.5 it seems [13:43] Saviq, thus the "make tryFoo" stuff no longer getting touch emulation [13:43] ltinkl: i see [13:43] dandrader, how can the xcb types change, that's not related to Qt [13:44] ltinkl, qt's 5.5 xcb platform plugin changed the handling of xcb events, so MouseTouchAdaptor should mimic that change [13:45] Hey, I'm trying to add a shortcut to the help overlay with Unity 7, but it isn't working, and I'm not 100% sure why. My best guess is that I have to go into compiz-profile-unity.convert in both compiz and unity and add the relevant shortcut, but I still have yet to try that. Anyone have any insights into this? [13:45] ltinkl, maybe qt's registers itself differently now on the X11 side, causing xserver to send it a different range of events [13:45] ltinkl, seems to be related to Xinput2 events [13:45] dandrader, aha === alan_g|lunch is now known as alan_g === dandrader is now known as dandrader|lunch [14:45] hmmm [14:46] make xvfbtestDragHandle fails in xenial [14:46] can anyone reproduce? dandrader|lunch: mzanetti_: ? [14:48] tsdgeos, there's a bunch of tests failing in X [14:49] checking [14:49] yeah "Received a fatal error" [14:49] tsdgeos, ↑ [14:52] that didn't use to fail locally (i think) [14:54] tsdgeos, could very well be Qt 5.5 [14:55] i guess [14:55] although QMetaProperty::read: Unable to handle unregistered datatype 'PaletteValues_QMLTYPE_21*' for property 'Palette_QMLTYPE_22::selected' looks bad [14:56] i don't have that [14:56] http://paste.ubuntu.com/14027904/ [14:56] is that i have with xvfbtestDragHandle [14:58] oh, the one without xvfb behaves better [14:58] and has the metaproperty things you mention [15:09] cimi, silo 54 rebuilt [15:26] Saviq: something's bad with that metaproperty error, really suspecting a Qt regression or something since if you run the tests one by one them all work, but all together gets us that + crash :/ [15:26] investigating [15:27] :( [16:04] cimi, ping === dandrader|lunch is now known as dandrader [16:09] tsdgeos, "make tryDragHandle" doesn't work as well [16:10] did that use to work? [16:10] tsdgeos, right now I'm working on making the touch emulation work again [16:10] tsdgeos, sure [16:10] dandrader: module "Unity.Test" is not installed [16:10] is that what you get? [16:10] tsdgeos, yes [16:11] tsdgeos, I was specially interested in DragHandle as your nodda changed it quite a bit [16:11] i think that's just the import paths being somehow broken [16:11] don't think it's the same as that weird "unregistered datatype 'PaletteValues_QMLTYPE_21*' for property 'Palette_QMLTYPE_22::selected' " error [16:15] tsdgeos, yes [16:15] dandrader: it's a shame because that nodda branch worked fine on friday, i guess it's all exploded by now :D [16:15] :) === dandrader is now known as dandrader|afk [17:09] ltinkl, yeah I hadn't considered the whole customization aspect of the wizard yet -- we can probably keep compatibility there? Or did we change how pages work that much that it just won't work at all? [17:09] ltinkl, we probably want to keep that first page named the same on disk, as low-hanging compatibility fruit [17:10] mterry, we changed... esp. with the language (and SIM) detection that's needed subsequently for the new timezone page [17:12] ltinkl, it's a big pain to break that for them -- especially imagine if we had more different kind of phones shipping. If we're going to break it, let's also throw in future-proofing so we don't break them again [17:12] ltinkl, I am still confused on why it was necessary, but I trust ya [17:12] ltinkl, how can we change the interface so that we are less likely to break them upon future redesigns? [17:17] mterry, I'll think about it, gtg now, later today === dandrader|afk is now known as dandrader [18:07] Saviq, around? === alan_g is now known as alan_g|EOD [18:39] dandrader, wassup? [19:07] dandrader, just noticed dialer app with a white bar at the bottom, worried it might be us after all [19:07] not telling the client the right surface dimensions [19:23] Saviq, hmm [19:24] Saviq, before or after that fullscreen property fix? [19:26] Saviq, dialer is not fullscreen... [19:27] Saviq, looks normal here in arale. post-fix though [19:32] Saviq, hmm, when over the greeter (emergency call). got t [19:36] Saviq, so unrelated to the Session.fullscreen fix, but still a bug somewhere, maybe this one is indeed in qtubuntu [19:42] dandrader, before fix [19:48] Saviq, and after fix as well, just checked === ljp is now known as lpotter