=== chihchun_afk is now known as chihchun | ||
=== chihchun is now known as chihchun_afk | ||
=== chihchun_afk is now known as chihchun | ||
=== chihchun is now known as chihchun_afk | ||
=== chihchun_afk is now known as chihchun | ||
=== alan_g_ is now known as alan_g | ||
=== ogra_ is now known as 1JTAAEF0J | ||
=== ogra_ is now known as 1JTAAEF3J | ||
=== chihchun is now known as chihchun_afk | ||
=== dandrader is now known as dandrader|afk | ||
=== alan_g is now known as alan_g|lunch | ||
=== dandrader|afk is now known as dandrader | ||
=== alan_g|lunch is now known as alan_g | ||
=== dandrader is now known as dandrader|afk | ||
=== dandrader|afk is now known as dandrader | ||
=== alan_g is now known as alan_g|EOD | ||
=== dandrader is now known as dandrader|afk | ||
=== dandrader|afk is now known as dandrader | ||
=== jono is now known as Guest90711 | ||
mterry | Hey folks! I'm trying to figure out lifecycle event management for mirserver instances. If USC gives the mirserver a life cycle event, which piece of code might end up calling SIGSTOP on the shell? | 19:14 |
---|---|---|
mterry | kdub_, racarr__ ^ ? | 19:17 |
=== dandrader is now known as dandrader|afk | ||
racarr__ | mterry: "Calling SIGSTOP on the shell"A | 20:16 |
racarr__ | ? | 20:16 |
mterry | racarr__, well | 20:19 |
mterry | racarr__, what seems to happen is that the unity8-greeter *sometimes* will be SIGSTOPPED when inactive | 20:19 |
mterry | racarr__, seemingly due to Mir lifecycle management | 20:19 |
mterry | racarr__, the unity8 shell doesn't seem to have this problem | 20:19 |
racarr__ | interesting, and they are both usc clients? | 20:20 |
mterry | racarr__, and the greeter doesn't always. I'm trying to trace down the origin of the SIGSTOP | 20:20 |
mterry | racarr__, yeah | 20:20 |
mterry | racarr__, platform-api has some code to SIGSTOP, but I didn't think USC ran the platform-api code. Does Mir do any default interpretation of those events, or does it simply pass them on to clients? | 20:21 |
racarr__ | prettysure it just passes it on...maybe platform API does some interpretation | 20:21 |
racarr__ | I dont think things are wired to sigstop themselves though?! | 20:22 |
racarr__ | so lets see whats going on | 20:25 |
racarr__ | mterry: So the normal process lifecycle management sigstop stuff | 20:25 |
racarr__ | is processcontroller.cpp in unity-mir... | 20:25 |
racarr__ | there is also this weird | 20:26 |
racarr__ | serverstatuslistener.cpp | 20:26 |
racarr__ | which raises SIGSTOP as part of some dance with upstart... | 20:26 |
racarr__ | maybe that is your offender | 20:26 |
mterry | racarr__, right | 20:26 |
mterry | racarr__, I don't *think* so | 20:26 |
mterry | racarr__, not the upstart bit though. greeter doesn't set that variable and there is no corresponding SIGCONT for that one | 20:26 |
racarr__ | (the second part of that could be the problem though right?) | 20:27 |
racarr__ | second argument to signal handler, siginfo_t | 20:27 |
racarr__ | has pid of sender | 20:27 |
racarr__ | so maybe best to start there | 20:27 |
mterry | racarr__, well the greeter CONTs fine | 20:28 |
mterry | racarr__, and will STOP/CONT as it becomes active/inactive | 20:28 |
mterry | racarr__, and processcontroller looks like it's used for shell apps, not shells themselves. :-/ | 20:29 |
mterry | racarr__, good point about pid of sender | 20:29 |
* mterry goes afk for a bit | 20:30 | |
racarr__ | mterry: Yes...I don't think processcontroller could be involved... | 20:30 |
racarr__ | I was just trying to locate all the sigstops lol | 20:30 |
mterry | fair! | 20:30 |
racarr__ | the one in platform API is for the old surface flinger compatibility | 20:30 |
mterry | racarr__, oh really? You don't think that's still used? | 20:30 |
racarr__ | errr let me look more carefully lol but I dont think so | 20:30 |
mterry | racarr__, the stuff in default_application_manager.cpp is what I was looking at. Not sure when it's used though | 20:31 |
racarr__ | yeah 98% sure thats its the old | 20:31 |
racarr__ | surface flinger out of process window manager | 20:32 |
racarr__ | thing | 20:32 |
=== dandrader|afk is now known as dandrader | ||
mterry | racarr__, I would love it if we just deleted all non-Mir ways of doing things | 20:32 |
racarr__ | I kind of thought I heard we had done that lol | 20:33 |
racarr__ | but apparently now | 20:33 |
racarr__ | not | 20:33 |
* mterry goes afk for realz | 20:33 | |
racarr__ | cya | 20:33 |
kdub_ | mterry, i've seen bad drivers use signals (maybe even sigstop) | 20:38 |
kdub_ | AlbertA, with usc and alpha, is it that the nested display doesn't fill the alpha channel? | 23:46 |
kdub_ | i'm finding I have to disable the alpha blending to get pixels with the overlay work | 23:46 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!