=== 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 === duflu_ is now known as duflu [10:27] anpok_: hey, any ETA on silo4 landing? [10:29] hmm seems to be building with lots of failures, I'll assume it'll be a few days at least [11:37] greyback: yes [11:37] greyback: waiting on qtubuntu papi .. still need to port the older gtk+ used in vivid.. [12:03] anpok_: You said something about making the input stack treat the device as touchscreen instead of touchpad? [12:03] yes [12:04] the flags array that indicates the available keys (EV_KEY) needs the BTN_TOUCH bit to be set and the evdev properties array needs the DIRECT_TOUCH (or TOUCH_DIRECT?) to be set. === mibofra is now known as Guest58370 [12:07] anpok_: INPUT_PROP_DIRECT ? [12:07] ah yes [12:09] * alf_ wonders where to use this with the evdev python bindings [12:12] AlbertA, thanks for the vulkan video [12:13] there is a capabilities property [12:13] but that seems to be read only [12:14] alf_: but there is a chance that only setting BTN_TOUCH is enought (provided that the ABS_MT_POSITION_{X,Y} touch axes are configured) === chihchun is now known as chihchun_afk [13:25] alf_: hey, when you launch mir with --vt=8, it correctly starts on vt8. But should it also switch the current vt to vt8? [13:25] seb128: ^^ [13:25] thanks [13:25] note that on that snappy image, for some reason the lightdm log has warnings like [13:25] "Error using VT_ACTIVATE 7 on /dev/console: Inappropriate ioctl for [13:25] device" [13:25] hmm, sounds relevant [13:25] so maybe that's the issue [13:26] unsure why it's doing that though [13:26] seb128: what hardware you using? [13:26] snappy is using the standard distro/kernel [13:26] inspiron 3000, all intel [13:26] should be ok [13:26] ubuntu standard doesn't have that issue [13:26] unsure why [13:26] or unsure why snappy has it [13:27] it's the same software stack, just the image is built differently [13:28] seb128: greyback: it should normally switch vts [13:28] is there any log that could indicate why it fails it that's not working? [13:29] seb128: is usc definitely being started after the lightdm x session? In case X is preventing the vt switch somehow (guess) [13:30] I think so [13:30] well, usc is not started before I try to log in as far as I can tell [13:30] and logging in unity8 session doesn't start X, does it? [13:31] nope [13:32] well it shouldn't, unless something tries to spawn it === dandrader_ is now known as dandrader|afk === dandrader|afk is now known as dandrader [15:48] hi. when i run the mir_demo_server using the packaged version in W, mir-demos=0.13.3+15.10.20150617-0ubuntu1, the server just freezes up, no mouse input, no way to switch VTs [16:01] attente: ummm... I just tried here in a haswell laptop, using W and same pkg version: "sudo mir_demo_server --vt 1" it worked ok for me [16:02] attente: do you have a thread backtrace? [16:04] AlbertA: it doesn't crash, and there's very little on output on stdout/stderr [16:04] it's almost as if it doesn't respond to any input [16:04] including the vt switch shortcut [16:05] attente: umm that can happen if you start the server without sudo [16:06] AlbertA: doh. thanks :) [16:48] I'm trying without success to print mir input log for unity8 [16:48] should MIR_SERVER_INPUT_REPORT=log and MIR_SERVER_LEGACY_INPUT_REPORT=log still work? [16:53] dandrader: they should [16:53] oh, wait! I think it did work! [16:53] AlbertA, sorry for the noise [16:54] dandrader: no prob [17:38] Howdy :) [17:51] racarr, so mir still uses android-input's EventHub and InputReader, right? [17:51] dandrader: For now ;) [17:52] trying to get that detailed logging android-input has... [17:52] dandrader: You have to enable it with the ifdefs in the files [17:52] then ANDROID_LOG_TAGS=*:v [17:52] MIR_SERVER_LEGACY_INPUT_REPORT=log [17:52] ahhhh!!!! missed the ANDROID_LOG_TAGS! [17:53] It's so bad :( sorry [17:53] it's been ages since I last used it [17:53] racarr, thanks [17:59] racarr, all those android-input notifications of input devices coming and going etc they will be logged only by usc right? unity8 will get "high level" mir input events from usc like an app, right? [18:00] dandrader: True, yeah...inputreader and eventhub shouldn't be used in U8 [18:02] AlbertA, where does usc logs go? [18:02] not in .cache/upstart, right [18:03] dandrader: they go into /var/log/lightdm/unity-system-compositor.log [18:05] AlbertA, wow, would never find it. thanks! [18:16] AlbertA, do you know how to flush usc log messages? they don't appear in the log file as soon as they happen [18:16] AlbertA, so I have to touch the screen to get some more log happening so that the log file does get updated [18:22] dandrader: I don't know. I usually just do tail -f [18:22] messages show up fine.... [18:23] AlbertA, that's what I'm doing. But the log file doens't get updated/written after every single message [18:23] AlbertA, only after a number of them have been written. and sometimes it even stops in the middle of a log line [18:28] dandrader: ummm strange.. these are the android input stack message or everything? [18:28] AlbertA, android-input. [18:28] AlbertA, didn't enable other logging to tell [18:29] let me see how they print [18:30] racarr, "[InputDispatcher]Dropping event because there is no focused window or focused application." <- this shouldn't ever happen, right? [18:30] racarr, in usc while unity8 is running === mibofra is now known as Guest82426 [18:34] dandrader: Hmm. I don't think it should no [18:36] dandrader: so it goes through the mir logger which doesn't flush after every line so probably explains the behavior [18:36] AlbertA, thanks for looking into it. should I report a bug? [18:37] racarr, this happens when running autopilot tests. as they create a /dev/input/eventX file for the fake keyboard [18:37] racarr, the first time it works. but from the second time onwards it does not [18:37] racarr, https://bugs.launchpad.net/ubuntu/+source/unity8/+bug/1468029 [18:37] Launchpad bug 1468029 in unity8 (Ubuntu) "unity8 crash breaking autopilot tests entering text" [Critical,In progress] [18:39] dandrader: yeah [18:45] AlbertA, https://bugs.launchpad.net/mir/+bug/1470204 [18:45] Launchpad bug 1470204 in Mir "android-input logs are not flushed as soon as they happen" [Undecided,New] [18:46] Trying to figure out what could be going on... [18:48] dandrader: I dunno...unfortunately too a lot of relevant code to key focus selection was just replaced with the new input dispatcher [18:48] unfortunately and or perhaps fortunately [18:51] racarr, at least it's easy to reproduce === dandrader is now known as dandrader|lunch [19:04] racarr: hi [19:04] racarr: hm device introspection is kind of required for raw events because clients would observe/receive them per device rather then per surface? [19:27] racarr: could also be someone not focusing the surface [19:28] i mean .. there were more focus related changes in the past [19:28] in usc.. === dandrader|lunch is now known as dandrader [19:56] racarr: so can i assign you to that bug ? [19:56] https://bugs.launchpad.net/canonical-devices-system-image/+bug/1468029 [19:56] Launchpad bug 1468029 in Mir "unity8 crash breaking autopilot tests entering text" [Critical,New] [19:56] it needs someone to love and care for it [20:08] I'm getting a "std::exception::what: No appropriate client platform module found" [20:08] How do I tell mir client what backend to use? [20:28] Okay, fixed that. [20:29] RAOF, So I seem to be starting Xmir, and it seems to be happy, but Unity8 isn't rendering anything. I don't think it's getting that first buffer swap. [20:29] RAOF, What am I doing wrong? ;-) [20:29] * tedg should be a talk show host. [20:32] racarr: i went ahead and gave that one to you === Guest82426 is now known as mibofra [22:07] kgunn: Oh you're back, I just sent you an email but, roger re:text crash bug [22:07] thanks racarr [23:50] tedg: Have you connected a client to XMir? It's not going to render anything unless a client renders.