mlankhorst | morning | 06:29 |
---|---|---|
mlankhorst | ricotz: I made geode compile but it's probably broken since I lack the hw to test, still should we upload it? Might be a good way to figure out if anyone actually uses it or not :) | 06:34 |
mlankhorst | I'll try to get my hands on a geode in the meantime | 06:36 |
mlankhorst | but apart from that, xorg 1.13 has the drivers for xserver-xorg-input/video-all ready | 06:37 |
RAOF | mlankhorst: Awesome. | 06:43 |
RAOF | mlankhorst: If you wanted to test that our xserver coredumping patch was working, what would you do? | 06:52 |
jcristau | pkill -SEGV Xorg? | 07:03 |
* RAOF adds that to the set of things to test. | 07:11 | |
RAOF | jcristau: The trick, for of course there is a trick, is that the coredumping patch *sometimes* works :) | 07:13 |
jcristau | ick. | 07:14 |
RAOF | Yeah. | 07:14 |
jcristau | could you get rid of the patch and just always pretend you've been passed -core? | 07:14 |
jcristau | (maybe that's what the patch does. i haven't looked.) | 07:15 |
RAOF | That's actually something I've been thinking aobut. | 07:15 |
RAOF | Although -core would result in everything being killed with SIGABRT, rather than the true cause. | 07:16 |
mlankhorst | RAOF: actually that wouldn't work since the X server checks if it comes from kernel or not | 07:16 |
mlankhorst | Hm maybe why it fails. | 07:17 |
jcristau | oh. bad X server. | 07:17 |
RAOF | mlankhorst: Really? Where? | 07:17 |
mlankhorst | hm let me check | 07:17 |
mlankhorst | OsSigHandler | 07:19 |
jcristau | it does FatalErorr anyway | 07:20 |
jcristau | Error, even | 07:20 |
jcristau | just changes the ErrorF msg afaict | 07:21 |
RAOF | Oh, yeah. That bit. | 07:22 |
mlankhorst | ah k then | 07:22 |
mlankhorst | my preferred way to kill annoying dialog boxes that don't have a close button is with kill -SEGV :) | 07:23 |
RAOF | I wonder how hard it is to use X's apport hook to teach apport how to extract the right signal from the OsAbort backtrace. | 07:24 |
mlankhorst | RAOF: what's stopping you from simply calling apport hook from there directly? Xorg is toast at that point anyhow | 07:25 |
RAOF | mlankhorst: Nothing. Well, except that you don't *know* what the right signal is at that point. | 07:26 |
jcristau | stuff it in a global :p | 07:26 |
mlankhorst | isn't it already with the rethrow patches? | 07:27 |
mlankhorst | hm guess not, but fatalsignal gets it as argument | 07:27 |
RAOF | It is already with the rethrow patches; I'm hoping to *drop* those patches (given they don't work reliably) in favour of post-processing it in apport. | 07:28 |
mlankhorst | oh ok | 07:28 |
mlankhorst | I guess the default hook will fail due to lack of X at that point? | 07:51 |
mlankhorst | else just stop handling sigabrt/segv w/e | 07:52 |
mlankhorst | yay, freedesktop.org account works :) | 08:05 |
jcristau | congrats | 08:05 |
=== shadeslayer is now known as shadeslayer_ | ||
=== shadeslayer_ is now known as shadeslayer | ||
=== yofel_ is now known as yofel | ||
=== shadeslayer is now known as shadeslayer_ | ||
=== shadeslayer_ is now known as shadeslayer | ||
bryceh | yay, got a mesa provisional MRE | 22:16 |
mlankhorst | oh thanks for reminding me i need to test a srgb fix in mesa. it should fix multisampling in nouveau :) | 22:41 |
mlankhorst | gnight :) | 22:41 |
RAOF | bryceh: Wooo! | 23:05 |
bryceh | RAOF, the regression testing plan was their main concern; all that time spent running piglits paid off :-) | 23:13 |
bryceh | RAOF, so at this point can you just wave the mesa package through, or is anything else needed? | 23:13 |
RAOF | At this point it'll be waved through into -proposed, then build, then we'll want to run some tests on the build from -proposed (because we're paranoid), then it'll get waved through to -updates after the 7 day "does anyone notice if it blows up" phase. | 23:15 |
bryceh | ok, I'm going to be on vacation for a week starting wednesday, so won't be able to do any testing really | 23:16 |
RAOF | Ok. | 23:17 |
bryceh | but I'm set up to do the runs so if no one gets to it before I get back, I can certainly fire the chips up after that. | 23:17 |
RAOF | I'll talk to gema and see if I can get it set up in the QA lab, too. | 23:18 |
bryceh | you might check to see if there's already a WI open to do that. I had talked to her about piglit running at sessions last UDS. | 23:20 |
bryceh | however, I'm skeptical it's going to be feasible, given that the tests have been hanging for me and so need manual monitoring | 23:21 |
bryceh | but hey, where there's a will there's a way. | 23:22 |
RAOF | Yeah, we also talked about it at UDS: P. | 23:23 |
RAOF | They surely have the tools to handle a hanging testcase. :) | 23:23 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!