/srv/irclogs.ubuntu.com/2014/07/02/#ubuntu-mir.txt

AlbertAdavmor2: FYI, the fix for https://bugs.launchpad.net/ubuntu/+source/unity-system-compositor/+bug/1336411 has already landed05:00
ubot5Ubuntu bug 1336411 in unity-system-compositor (Ubuntu) "Manta on recent images doesn't wake correctly from suspend" [High,Fix released]05:00
=== doko_ is now known as doko
duflualf_: So mir_demo_server_shell doesn't respond to SIGINT any more?08:41
alf_duflu: it should, are you seeing something different?08:42
duflualf_: Yep, as of this week Ctrl+C does nothing08:42
dufluWhich is fine if it's intentional08:42
alf_duflu: hmm, now that I think about it, I actually think that's old? I've always used ctrl-alt-backspace to quit08:43
duflualf_: Yes, well I haven't tried it since May of course :)08:44
alf_duflu: so, I think this was affected by a change in the vt input mode08:47
dufluSounds plausible. Although I'm running in ssh08:48
alf_duflu: then that should work, since you are just sending SIGINT to the process08:49
duflualf_: Boo. That means I have to retest and log a bug08:50
=== chihchun_afk is now known as chihchun
duflualf_: Oh I already reported it. Nothing new... https://bugs.launchpad.net/mir/+bug/129396509:57
ubot5Ubuntu bug 1293965 in Mir "Using the server option "--vt" prevents Ctrl+C (SIGINT) handling" [Medium,Triaged]09:57
dandraderalan_g, about that "surface orientation and surface attributes" discussion: what's a surface attribute? maybe there are (or should be) read-only and read-write surface attributes (from the client point of view) and that's enough?10:23
alan_gdandrader: maybe. I've not had need to think deeply nor read the design docs that motivated them.10:28
dandraderalan_g, from my naive point of view, if you have some information regarding a surface, such that a client could go and ask "hey, what's the value of X for my surface Y?" or "hey, tell me when the value of X for my surface Y changes (i.e. and event)" then that "X" is a surface attribute. If the client has the right to change it or not it's another story.10:40
alan_gcamako: I know you've other things on your stack but... I'm hoping that this will finally make it easy to write a test for your nested lifecycle logic. Let me know if there anything isn't clear. https://code.launchpad.net/~alan-griffiths/mir/more-nested-server-testing/+merge/22517410:40
dandraderalan_g, I find it odd when you are able to know when the value something changes but not its current value10:41
dandraderalan_g, or vice-versa10:41
camakoalan_g, sorry the release took all my time... (too much time if you ask me :-) )10:42
camakoalan_g, still need to test xserver10:42
alan_gdandrader: from my naive point of view attributes is complicated by having an ABI stable implementation (index value pairs) wrapped in an unstable API whose main benefit is that changes are routed through a "configurator". It would make more sense to me with mir_surface_get_attribute(surface, attribute)/mir_surface_set_attribute(surface, attribute, value)10:44
alan_gBut things must be this way for a reason. I just don't know it.10:45
alan_gcamako: np. (Agree about the time it sucks up.)10:45
alan_gdandrader: the guy to talk to about attributes is duflu - he implemented it based on some (now ancient) design docs10:47
dandraderalan_g, ah, ok. so it's about what entails when something is exposed via that attribute system (which has some limitations and expectations)10:49
alan_gdandrader: yes. It may well be that we could now come up with both better requirements and a better implementation. But I've not been involved (yet) and am not sure who the stakeholders are.10:58
davmor2AlbertA: manta fix landed and looks good :)11:04
davmor2kgunn: ^11:04
alan_gdandrader: @"I find it odd when you are able to know when the value something changes but not its current value" are you talking about orientation? doesn't mir_surface_get_orientation() do that for you?11:09
dandraderalan_g, no, that's visibility11:11
dandraderalan_g, that's an example. but I was more like saying in general. not pointing to any specific case11:12
alan_gdandrader: sure. I just wasn't aware of an example and you started by talking about orientation.11:13
=== alan_g is now known as alan_g|lunch
=== dandrader is now known as dandrader|afk
=== renato_ is now known as Guest82807
=== dandrader|afk is now known as dandrader
=== alan_g|lunch is now known as alan_g
=== Guest82807 is now known as renato__
=== alan_g is now known as alan_g|tea
=== dandrader is now known as dandrader|lunch
=== no_mu is now known as Nothing_Much
bregmahey folks I started getting a segfault from Mir server code in Unity 8 (desktop) yesterday on all my machines after updating to the latest packages, reported as #133685416:41
bregmait's not clear to me if this is a Mir bug, a Mesa bug, or what, so I'm pinging people here for some better-educated advice16:41
=== alan_g|tea is now known as alan_g
=== alan_g is now known as alan_g|EOD
=== dandrader|lunch is now known as dandrader
AlbertAummm18:28
AlbertAmy 5sec timer is firing in 6s18:28
AlbertAbut std::chrono thinks it is 5s18:29
AlbertAin the phone devices18:29
racarrstd::chrono thinks it is 5s?18:38
racarrI mean...technically thats allowed right :p18:39
AlbertAracarr: I mean if I measure with an external clock18:40
AlbertAit's 6s...18:40
AlbertAbut std::chrono says its 518:40
AlbertAi'm trying to see what is up18:41
racarra non real time kernel? :p18:41
racarrI guess if it does it every time though18:42
racarrsomething is bad18:42
AlbertAracarr: yeah it's consistent18:42
AlbertAI would understand a few millis here and there18:42
AlbertAnot a full second18:42
AlbertAI wonder if there's some time drift issue in the libstdc++ for armhf18:43
AlbertAnah...It was my inaccurate subjective external clock18:59
racarrhaha wait did you really mean19:00
racarran external clock19:00
AlbertAannotate-output is useful :)19:00
racarrIthought you meant like C time functions or something19:00
AlbertAracarr: nah...19:00
AlbertAracarr: :)19:00
racarrlol19:00
racarryay time still works19:01
AlbertAracarr: when I doubt I usually go to the oscilloscope but have none there :)19:01
racarrlol19:02
AlbertAracarr: maybe it was gravitational time dilation :)19:03
AlbertAgiven that the external clock was closer to my body and it was slower19:03
AlbertA:)19:03
racarrHaha19:03
racarrprobably either that or the Ubuntu NSA backdoor for cryptographic timing attacks19:04
AlbertAracarr: heh19:08
racarrDo we have a C style guide?20:05
racarrimnot 100% sure what alan is talkingabout here https://code.launchpad.net/~mir-team/mir/cursor-spike-phase-7-name-some-cursors/+merge/22524220:06
racarram I being stupid and...missing something or is he just saying20:06
racarrchar const *const instead of char const* const20:06
racarrwhich would then apparently be part of our C style guide?20:06
racarrINSTANTIATE_TEST_CASE_P makes me feel like some sort of test wizard22:33
racarrshooting multipronged branching test lightning bolts.22:33
racarralso in investigating surface attributes found a nice mix of22:34
racarrrace conditions, potential deadlocks, and simple logic emissions22:35
racarr-.-22:35
racarrcest la vie22:35
=== no_mu is now known as Nothing_Much

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!