RAOFbryceh, tjaalton: What do you think about switching to r600g by default and announcing a call-for-testing?03:19
RAOF#dri-devel is being very “why are you trying to fix r600c” and “yeah, we just made sure r600g works with gnome-shell”.03:19
brycehRAOF, I'm in favor of switching.  There are trade-offs however03:22
brycehon the one hand, what we have has been tested in the distro and seems reasonably stable aside from a few bugs (which we know about and which look reasonably fixable I guess)03:22
brycehon the other hand, upstream is focusing only on the gallium side, so if we upstream bug reports and patches, they'll be more interested03:23
brycehRAOF, obviously it's insufficient to switch just for the sake of making upstream developer's lives better (c.f. lessons learned with -intel taking this approach)03:24
brycehhowever it's been my experience the ATI guys are much more grounded and give better advice03:25
RAOFWe'd also be following debian and fedora here, so there is at least a bigger testing pool.03:26
brycehRAOF, ultimately I think if we switch in order to get better responsiveness from upstream on bugs, then we really ought to take good advantage of it by making sure we forward -ati bugs aggressively03:26
brycehRAOF, what were the main reasons we had for sticking with r600c?  just stability concerns?03:27
brycehso then yeah, I think it would be ok, so long as we are attentive to forwarding bugs up03:28
RAOFWe weren't sure whether it would be ready for default, and we hadn't tested it like we have r600c.03:28
brycehthe -ati guys are easy to work with and have been dependable for sorting out issues when we've pulled in new ati bits late in the release in the past, so I'd not be uncomfortable doing this.03:29
RAOFOk.  I'll update the new 7.10.1 snapshot with that in mind.03:30
brycehone other thought03:30
brycehactually, nevermind03:31
brycehwhen do you think you'll have 7.10.1 ready for upload?03:31
brycehif it's going to be more than a week, I'd like to roll out cherrypicks that we know fix issues; the less bugs we have to keep track of, the easier it makes triage ;-)03:32
brycehok, wife is chomping at the bit to go out to dinner (belated valentines day or she just doesn't want to cook tonight... hard to say)03:32
RAOFApproximately now :)03:32
RAOFHave dining fun!03:33
brycehok, if you knock something together, email me and I'll upload tomorrow morning03:33
tjaaltonRAOF: yeah, sounds good07:55
RAOFYou wouldn't additionally have any response to the driver selection RFC I sent out to ubuntu-x@ a couple of days ago? :)07:56
tjaaltonah, let me check07:56
tjaaltonRAOF: the galleon switchover stuff?07:58
RAOFtjaalton: Yeah.  Exactly.07:58
* RAOF would *also* like to be involved in a galleon switchover :)07:59
tjaaltonso that patch would be for people to fall back on r600c if they find r600g unstable for them?08:00
RAOFIt would be to fallback to r600c if they have disabled KMS.08:01
tjaaltonoh right08:01
tjaaltonmaybe the second option would work then08:01
RAOFCurrently they can twiddle it with the ForceGallium xorg.conf option.08:01
RAOFOr, rather, the second option would be so that it falls back to r600c if they've disabled KMS.08:02
tjaaltonand ship r600c in the "other" directory, and drop when UMS support is dropped08:02
RAOFThe first option *additionally* allows users to twiddle gallium/classic.08:02
RAOFIndependent of KMS.08:02
RAOFI guess I'm undecided on the utility of providing that knob.08:03
tjaaltonhmm, so for r300 the fallback with UMS already happens?08:05
tjaaltonwithout extra patching08:06
RAOFWith the existing ForceGallium patch, yeah.08:06
RAOFThis is replacing the existing patch, which works but in a way which potentially breaks people who are testing mesa, with a different (upstream approved) solution.08:06
tjaaltonoh right, they all are shipped in the same dir08:06
tjaaltonguess I need to think about it for awhile then :)08:09
* tjaalton grabs a cup of tea08:09
tjaaltonwhy doesn't gallium work with UMS?08:13
tjaaltonmaybe i'm missing something obvious08:13
RAOFProbably because it assumes the whole gem/ttm kit and caboodle.08:14
* RAOF wonders idly what a caboodle is.08:14
RAOFAlso possible: it's only DRI2, and DRI2 is only available on KMS for radeon.08:14
tjaaltonah, that's it then08:15
RAOF(Although that's pretty close to a restatement of requiring gem/ttm)08:15
tjaaltonRAOF: does classic work with KMS?08:25
RAOFYes, it does.08:25
RAOFOtherwise option 2 would be the trivial solution, yes :)08:25
tjaaltonthat makes more sense, yep08:25
tjaaltonthis patch would be useful as long as UMS is also available08:26
tjaaltonno wait. duh08:27
tjaaltonsince kms works with both, the line above is _not_ true :)08:28
tjaaltonthough how long is the classic ones supported08:28
tjaaltonRAOF: I guess the second option is better. when the option becomes obsolete there'll be a number of people with it in their xorg.conf, that probably have forgot about it by then and wonder why the config broke08:32
RAOFHeh, yes.08:33
tjaaltonand the second one can be silently dropped at some point08:33
tjaaltonthough if it makes upstream there's no need08:34
RAOFIt can just sit there, waiting for some other driver to do something similar.08:37
tjaaltonright, like... *chrome? :P08:38
RAOFOh, wow.  Yeah, I guess they might do something like that!08:39
jcristausoon each via owner can use a different driver08:46
tjaaltonmore drivers than users :)08:46
vishfreaky spam's subject!! "Hello alberto.milone! A lot of software is for windows and macOS! Logic Studio 8 Full Pack with Content" 09:18
vishbut i'm not tseliot! ;p09:18
=== alf___ is now known as alf__
tjaaltonI closed bug 710762 as wontfix10:22
ubot4Launchpad bug 710762 in xserver-xorg-input-evdev (Ubuntu) (and 1 other project) "Middle mouse button no longer works (affects: 3) (heat: 18)" [Undecided,Won't fix] https://launchpad.net/bugs/71076210:22
tjaaltonRAOF: you got a new mesa in the works?10:27
RAOFtjaalton: 10:28
RAOFYeah.  I should push it to git.10:28
brycehtjaalton, good call.  Might be worth a brief mention on https://wiki.ubuntu.com/NattyNarwhal/TechnicalOverview10:29
brycehwell, it's late, time for bed.  cya10:30
tjaaltonbryceh: ok, will add10:31
tjaaltonRAOF: ok, so unity works with the nouveau commit by calim?10:32
vishtjaalton: unity depends on middle click10:32
tjaaltonvish: you have three buttons right?10:32
vishiirc it is for "open new instance"10:32
vishtjaalton: yea, but there is what they have done.. removed the item from the context menu and made it depend on middle click10:33
RAOFtjaalton: Yeah, for me at least.10:33
tjaaltonRAOF: cool, so I'll be able to test too10:33
tjaaltonvish: well, there are tradeoffs to re-enable it, as described on the commit (also on the bug)10:34
vishsure, but i think it should be mentioned to the unity folks..10:35
tjaaltonthey should've noticed by now10:35
tjaaltonbut anyway, where can I find them?10:35
vishtjaalton: its Bug #709707 btw10:36
ubot4Launchpad bug 709707 in unity (Ubuntu) (and 2 other projects) "Middle click on application icon should open a new window (affects: 2) (heat: 12)" [Medium,Fix released] https://launchpad.net/bugs/70970710:36
vishtjaalton: on #ayatana10:36
tjaaltonvish: thanks10:36
tjaaltonre-enabling the emulation is possible via input properties, so it should be added to "The GUI" if needed10:38
loolMy desktop doesn't get redrawn properly if I switch to console and back to Xorg, but it does if I switch workspace; is this tracked in a bug already?11:49
tjaaltonbryceh is probably best on top of intel bugs atm, I haven't seen that behaviour on mine11:50
=== schmidtm_ is now known as schmidtm
bjsniderkklimonda, i convinced pitti to change nvidia-common so it's a hard dependency of jockey-common. so starting with natty, to use the nvidia-installer, you'll have to remove all of jockey15:42
kklimondabjsnider: that's great news imo :)15:44
* dupondje thanks bryceh & jcristau 17:13
dupondjeissue seems solved :)17:13
brycehdupondje, good to hear, what was the fix ultimately?17:26
dupondjeinstalling your patched version ? .17:32
brycehdupondje, ahh good17:32
dupondjebtw, you maby have upload rights to fix https://bugs.launchpad.net/ubuntu/+source/gtk-vnc/+bug/711442 ? ;)17:32
ubot4Launchpad bug 711442 in vinagre (Ubuntu) (and 1 other project) "vinagre crashed with SIGSEGV in g_socket_send() (affects: 3) (dups: 1) (heat: 269)" [Medium,New]17:32
brycehdupondje, I do, but there's also a Patch Pilot on call right now (sconklin) who is on deck for sponsoring these sorts of things17:33
brycehdupondje, point him to the bug on #ubuntu-devel and ask for sponsorship17:34
brycehdupondje, that'll be the "right way" to handle this.  :-)  If he doesn't respond or if you have any trouble let me know and I'll lend a hand.17:34
ricotzSarvatt, i am going to update mesa in edgers, there were quite some nvc0 commits ;)19:44
Sarvattricotz: appreciated! :)19:45
brycehweird, after messing around with the keyboard layout prefs dialog debugging some bugs, now my windows no longer take focus when clicked; only when I click the title bars23:21
brycehif I set to focus-follows-mouse, that works, but then unset that and back to broken behavior23:21
RAOFThat's highly odd.23:23
RAOFDoes switching to metacity work?23:23
brycehthis is with metacity running actually23:23
RAOFWell, that bet I was about to make about it being a compiz problem looks pretty silly now. :)23:23

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