[00:00] still terrible scrolling :) [00:03] jbarnes, need anything else from me? [00:03] can you reproduce it on your system? [00:04] still updating [00:04] what site do you use to test the scrolling? [00:04] slashdot [00:04] but its obvious on just about any site [00:04] ok [00:05] slashdot comment threads after the floating element is forced to reposition is really nasty [00:05] hm ok [00:06] to be fair about that one though, thats been nasty forever [00:06] the regression is that even the slashdot homepage doesn't scroll nicely anymore [00:06] it's not super fast on 965 but not horribly slow either (exa though) [00:06] exa/uxa makes no difference in this case [00:07] though oddly Docky (uses lots of cairo) is very very smooth, as smooth is intrepid [00:07] and it has a window as large as a browser [00:08] it's interesting how very specific types of rendering can uncover bugs like that [00:09] though i have a flash video playing everything slows down a lot more than it used to [00:09] ah I think I see it [00:10] it's slower through a list of collapsed comments [00:10] with the floater on the side [00:10] worse with a composting manager running [00:11] yeah, but again to be fair thats been there since day 1 i have had this computer, its also present on windows so that bit is probably gecko's fault [00:11] or maybe not [00:11] still not crazy slow though [00:11] who knows [00:11] on a g45 with jaunty? [00:11] no testing 965 while my gm45 updates [00:12] when you say not crazy slow [00:12] how many lines of text at a time would you say it skips? [00:14] takes about 2s to do one page up [00:14] page up key that is [00:14] thats not crazy slow in your world? =P [00:15] i would try playing music and then scrolling up and down with the mouse wheel [00:15] well not 3-4 that would be twice as slow :) [00:15] see if you can get the music to skip [00:15] but no it's bad just not as bad as I thought [00:15] mine takes about 3s [00:15] just timed it =/ [00:15] yeah it sucks up a bunch of cpu [00:16] hm now does sysprof work on this machine? [00:19] mine or yours? [00:20] either :) [00:21] installing [00:24] looks like it ought to take a little less than a second/scroll [00:24] ok its working [00:24] 945 has this problem too [00:24] oh cool you can collect a profile then [00:24] with debug symbols [00:24] just scroll up and down? [00:24] debug symbols on what? [00:24] on X, xf86-video-intel, kernel & libdrm [00:25] ah yeah probably not right now, I'd have to install a LOT of stuff... [00:25] I'm trying now on my 945 which is even slower to scroll [00:30] alright, presentation time [00:31] I'll come back and bother you later jbarnes, thanks for the help [00:31] sure [00:31] thank you [00:40] man, some day I'm going to write a cron job to check X bugs for the word "randomly" and mark them invalid. [00:41] lol [01:38] * bryce waves to cwillu [01:38] poke poke [01:38] if DBO comes back... I updated fdo bug 18572 with the profile I captured [01:39] Launchpad bug 18572 in gnome-applets "modem_applet: waitpid()'s return value causes wedging." [Medium,Triaged] https://launchpad.net/bugs/18572 [01:39] I must be getting popular or something: bryce waves, bugabundu blew me a kiss, and ubottu just told me that I'm also a bot (written in php) [01:39] :-) [01:40] * cwillu resolves to invent random tags and apply them to random bugs more often :p [01:45] 945 under uxa, black screen on resume from hibernate: ssh'ing in, killing compiz.real and rerunning compiz restores the session (oddly, nothing less than that) [01:46] mm [01:47] * cwillu deletes out the gibberish he typed in his editor while the screen was blanked :p [01:47] I've been ignoring UXA bugs so far. At some point those are going to need deeper attention [01:47] I've been working on exa freezes all day today [01:48] my 945 is due for some punishment if you want me to try anything [01:49] presently, I'm just trying to build a table of all the different freezes that have been reported [01:49] https://wiki.ubuntu.com/X/Bugs/IntelDriver [01:50] just to get some sort of framework in place to discern any patterns [01:54] bugs I have listed as hangs that aren't on your list: 340652, 347527, 316423, 259385, 304871 (fixed I believe), 343690 (also fixed), 340964, 328918, 322613, 320821, 337243, 312776 (intrepid) [01:54] * cwillu starts opening lots of launchpad tabs to check for dups :p [01:55] mind if I sort that table on the bug number? [01:55] lemme save [01:55] sure, I'm about to head off to dinner in a few minutes anyway, so feel free to add to the table [01:55] saved. [01:56] k [01:56] i've also been doing s/hang/freeze/ and s/lockup/freeze/ to make them consistent [01:57] and removing that damn "randomly" word whenever I spot it ;-) [01:57] heh [01:57] I'd suggest "indeterminately", but I'm guessing you wouldn't appreciate it [01:58] the best are, "Freezes when compiz enabled... and with compiz disabled" [02:08] you have 327844 listed twice, is that desired? [02:08] || 327844 || G45 || 2009-02-10 || Frequent || Turn compiz on || Disable compiz || || [02:08] || 327844 || G45 || 2009-02-10 || Rare || Video playback || || || [02:08] * cwillu pings bryce [02:08] yep [02:09] I did that because in the bug they mention two different scenarios where freezes occur [02:09] (ideally we'd handle those as separate bug reports) [02:09] heh, that bug is exactly the one I was thinking about with my "best" comment ;-) [02:10] heh [02:10] btw, save frequently [02:10] yep [02:10] if only we had some collaborative system for managing bugs p [02:10] I made the mistake of not doing so, and firefox crashed when I was half-way through, and had to start all over again :-( [02:10] agreed [02:10] session restore includes form fields though [02:11] didn't work for me [02:11] :/ [02:11] I really want to have some launchpad-like list report that allows filtering by tag [02:11] oh, check browser.sessionstore.privacy_level [02:12] it's set to '1' for me [02:12] which means that it doesn't hold data for https sites [02:12] http://kb.mozillazine.org/Browser.sessionstore.privacy_level [02:12] * bryce sets to 0 [02:12] thanks! [02:13] mm, I think on my laptop I'll leave that to 1 ;-) [02:13] http://kb.mozillazine.org/Browser.sessionstore.postdata might also need a tweaking if you change it from 1 [02:14] it's set to 0 [02:14] I'm thinking that may be what I want. guess I can play with it [02:15] and also, openoffice is a crashy flacky unreliable piece of crap [02:15] * cwillu feels an urge to murder something at least once a day due to openoffice calc [02:16] heh [02:16] the waitlpring table is currently crashing calc [02:16] like, seriosuly [02:16] you trying to edit that IntelDriver page in calc? [02:17] no, just messing with the tables to cross-check against my list [02:17] normally copy/paste of an html table works fine [02:19] I do my heaving lifting in gedit :p [02:22] wow, I don't even trust gedit [02:22] I don't really trust any editor without autosave [02:23] btw 327844 now says was solved with 2.6.3-0ubuntu9 [02:23] I have a ctrl-s reflex burned into my left hand ever since I was 12 years old and watched 3 days of work on a game I was writing go up in a puff of crashing-the-ide [02:23] * cwillu reminisces of gfa-basic [02:25] that's the 'disable dri on everything' patch? [02:27] did you want intrepid bugs in that table? [02:28] (re: gedit, I also wrote my own autosave/restore session plugin for it, among other things) [02:30] nice [02:31] uh, intrepid bugs are okay if they are believed to still affect jaunty [02:31] if they're not occurring in jaunty, shouldn't be in the list [02:31] k, I'm just going to drop them for now then [02:32] 0ubuntu9 was a crash in drm_intel_bo_unpin that occurred after watching videos, relating to DPMS [02:32] okay [02:32] alright, wife says we're late for dinner, gotta go. back in a few hours [02:37] ttyl [03:14] im back jbarnes [03:14] did anything fun get discovered? [03:15] I was discovered hiding behind a rock [03:16] whether I'm fun is still to be determined though [03:16] i hope by that you mean you know fun things about g45 and X [03:17] that would be "still to be determined) [03:17] ah [03:17] can you tell me whats been going on (if its not much trouble) [03:19] exa or uxa? [03:19] either or [03:19] I just want to know whats going on with the scrolling performance issue [03:20] * cwillu doesn't have a performance issue on his buglist for that [03:20] bug number? [03:21] I know of 339233, 339781, 338681, as corruption issues, and 327844 as hang/crashers on that chipset [03:23] mmm I think it was mentioned at some point... [03:23] if DBO comes back... I updated fdo bug 18572 with the profile I captured [03:23] Launchpad bug 18572 in gnome-applets "modem_applet: waitpid()'s return value causes wedging." [Medium,Triaged] https://launchpad.net/bugs/18572 [03:23] just saw that in the scrollback [03:24] nothing else since I rejoined an hour ago [03:24] okay [03:24] thank you cwillu [03:24] i was in the middle of preaching the church of GNOME Do to the freshmen choir at WMU [03:25] bryce, resorted on bug number (which probably isn't helpful to anyone else but me right now :p), and added dates and triggers where I can find them [03:25] I can't bring myself to use gnome-do [03:26] I have it installed, I fire it up once a month or so, but it just feels like it slows me down [03:26] cwillu, its not everyones cup of tea [03:26] cwillu, i dont care if anyone at all uses it to be honest [03:26] I hack it for me and me only [03:26] some evangelical you are [03:26] i was being graded [03:31] * cwillu is away again for a couple hours [07:02] poke [07:50] back [07:51] ended up having to upgrade mom's computer [08:12] heya tseliot [08:12] hi bryce [08:13] tseliot: hey, if no Virtual is set in the xorg.conf, does it take the largest dimension of the outputs attached at boot, or does it use the maximum allowed by the video card (2048x2048 or whatever)? [08:14] bryce: the latter [08:14] if you're referring to the Gnome capplet [08:15] no, just in general by default, before the capplet comes into play [08:16] > Version 2.1.1-0ubuntu2 seems to set the default Virtual size (maximum [08:16] > screen size) to 1920 x 1920, if there is no entry in xorg.conf. I take [08:16] > it the maximum screen size for the i915 chipset family is 2048 x 2048, [08:16] > so why not have it at that? This would make dualscreen setups a bit easier. [08:16] The default settings is found by taking the largest resolution in either [08:16] x or y dimension and making a square from that. This allows for easy [08:16] rotation should you want to do so. [08:16] [08:16] I'm wondering if that is accurate anymore [08:17] it says 2048x2048 on my laptop, even when no other output is plugged in [08:17] (other than LVDS) [08:18] $ DISPLAY=:0 xrandr [08:18] Screen 0: minimum 320 x 200, current 1280 x 800, maximum 1280 x 1280 [08:18] aah, it's version 2.1.1 [08:18] $ apt-cache policy xserver-xorg-video-intel [08:18] xserver-xorg-video-intel: [08:18] Installed: 2:2.6.3-0ubuntu2 [08:18] does it use UXA? [08:19] no [08:19] it has dontzap set, and everything else is default [08:19] this is an i965 [08:19] let me check what my eeepc says [08:20] is that framebuffer size available when you boot with only 1 output (i.e. with no external screens)? [08:21] correct [08:21] my eeepc (running intrepid) says 1024x1024 [08:21] hmm [08:23] on jaunty, with my laptop and UXA I get 2048x2048, let me try with EXA [08:24] it says 1280x1280 with EXA and 2048x2048 with UXA [08:25] but it's not a matter of framebuffer reallocation since that's the framebuffer which is allocated when X starts [08:26] bryce: maybe we should have a look at the UXA code? [08:27] UXA has better memory management behind it, so perhaps that allows it to use larger default virtual [08:27] with EXA, the memory usage is probably more of an issue, thus they don't go to the hardware maximum. [08:27] it can be [08:28] bryce: BTW it looks like X-Kit and dontzap ended up in Fedora: https://bugzilla.redhat.com/show_bug.cgi?id=494517 https://bugzilla.redhat.com/show_bug.cgi?id=494518 [08:28] bugzilla.redhat.com bug 494517 in Package Review "Review Request: python-xkit - A simple, transparent and easy to extend xorg parser" [Medium,Assigned] [08:28] sweet, congrats! [08:28] :-) [08:28] now that's funny [08:28] ? [08:28] we should remember this for the next time they say "Ubuntu never contributes upstream!!1!" [08:29] hehe [08:31] http://who-t.blogspot.com/2009/04/zapping-server.html [08:34] yes but we already have a UI in Kubuntu ;) [08:35] of course that patch is another thing [08:37] I find the whole situation rather amusing [08:44] it definitely is [10:32] tjaalton, jcristau: any ideas as to where libdrm-intel1 gets the dependency on "libdrm2 (>= 2.4.1)" from? [10:35] tseliot: dpkg-shlibdeps [10:38] tseliot: why? [10:38] jcristau: I would like to make it depend on >= 2.3.0 so that psb can replace libdrm [10:38] as we'll have 2 versions of libdrm [10:38] uh [10:38] not going to happen [10:39] I know, it's ugly [10:39] but it's not something that we'll do in Ubuntu [10:40] just for our customised installations [10:53] jcristau: I'm using dh_makeshlibs -plibdrm2 -V'libdrm2 (>= 2.3.0) | libdrm-poulsbo1 (>= 2.3.0)' -- -c4 [10:53] in debian/rules [10:53] that's utterly wrong... [10:54] jcristau: care to explain? [10:54] apart from the ugliness of having libdrm-poulsbo1... [10:55] also if you have a libdrm2.symbols files that won't work anyway [10:56] yes, I have that file [10:56] tseliot: it's wrong because if something needs libdrm 2.4.1, making it depend on 2.3.0 is just asking for breakage [10:57] jcristau: this will not happen in Ubuntu but only in OEM installations [10:57] where -psb is the only thing you can use [10:58] *shrug* [10:59] yes, I know... [11:02] jcristau: thanks for your help, that worked [11:29] * bryce hugs tseliot [11:29] sooo glad someone else is looking at -psb stuff :-) [11:30] bryce: yes, it's my job [11:30] ;) [11:30] \o/ [11:30] but we can share the pain if you miss -psb :-P [11:30] heck no! [11:30] hehe [11:30] btw, volunteered you for something ;-) [11:31] for what? [11:31] jane silber asked for a laptop/projector testing thingee [11:31] oh [11:31] I get it [11:31] hehe [11:31] such as? [11:32] I suggested maybe a better idea would be if I and you gave a short presentation on "projector tips and tricks with ubuntu" [11:32] at AllHands? [11:32] or using the wiki? [11:33] yep. I figure it would be just to demo Screen Resolution, explain about ~/.config/monitors.xml, and how to work around problems [11:33] yep to which one? [11:34] the former? [11:34] i.e. AllHands/UDS [11:34] at AllHands [11:34] ok [11:34] sure [11:34] it should be pretty easy to do [11:35] yeah, that's what I figure [11:35] also, I told her to limit presenters-with-computers to Intel 915 and newer, or ATI (-ati) [11:36] well, I just cc'd you the email, you can see what I said :-) [11:36] yes, right [11:36] I've just received it [11:38] I don't own an Nvidia card (in a laptop) but if someone has an Nvidia card I could give a brief presentation on that too [11:38] with nvidia-settings though [11:38] nice, yes that'd be good [11:39] good, let's gather some ideas for the presentation then [11:39] maybe we should cover only clone mode [11:39] which is what you use with projectors [11:40] yep [11:41] I also like the idea of pre-setting your Virtual to 2048x2048 [11:41] (which we discussed earlier) [11:41] in the driver or in xorg.conf? [11:41] xorg.conf [11:42] shall we set up a wiki page about it? [11:42] so that we can add ideas as they come to our minds [11:42] yes, that's a very good idea [11:43] whoa... can you start it? I just noticed it's 3:45am, I need to get to bed. [11:44] sure, as soon as I find a decent name for it [11:44] yes, it's pretty late there [11:44] try to get some sleep [11:44] :-) [11:44] cya tomorrow [11:44] night bryce [18:28] bryce: any thoughts about bug 282081? [18:28] Launchpad bug 282081 in xserver-xorg-video-intel "[i965] X freezes after screen saving or logout/login on Dell Hybrid Studio" [Unknown,Confirmed] https://launchpad.net/bugs/282081 [18:29] it's been idle since january and is against a pretty old driver... [18:29] * jbarnes would like to close it [18:36] bug #275809 [18:36] Launchpad bug 275809 in xserver-xorg-video-intel "[GM965] X freezes when logging in from screensaver - fails in intelSwapBuffers()/ intelCopyBuffers()" [High,Won't fix] https://launchpad.net/bugs/275809 [18:36] bug #316414 [18:36] Launchpad bug 316414 in xserver-xorg-video-intel "[i965] X freezes at least once a day after leaving screensavers or fullscreen video" [Undecided,Confirmed] https://launchpad.net/bugs/316414 [18:36] (ignore that one) [18:36] those sound like vblank related hangs... should be fixed in recent (2.6.28+ iirc?) kernels and libdrm [18:37] jbarnes, may find dupes on https://wiki.ubuntu.com/X/Bugs/IntelDriver [18:39] but #357908 [18:39] * jbarnes wonders where ubottu got off to [18:39] bug #357908 [18:39] Launchpad bug 357908 in xserver-xorg-video-intel "[i965] X freezes every ~24hr while scrolling in firefox (EXA enabled)" [High,Triaged] https://launchpad.net/bugs/357908 [18:39] heh [18:40] jbarnes, launchpad's fault usually, ubottu doesn't have a special link that works any better than ours :p [18:40] !botsnack [18:40] Yum! Err, I mean, APT! [18:40] helps when I spell 'bug' right too :) [18:40] bug #331596 [18:40] Launchpad bug 331596 in xserver-xorg-video-intel "[i965] X freezes when TripleBuffer On after resume" [Undecided,Incomplete] https://launchpad.net/bugs/331596 [18:41] fyi the old triplebuffer & page flipping implementations are horribly buggy [18:41] and will likely cause corruption and/or hangs after awhile [18:41] so that's a won't fix then? [18:41] yeah [18:41] they've been removed in later versions [18:41] (I'm working on adding them back in though) [18:42] bah, I can't mark won't fix yet :p [18:42] fdo bug 20664 is the rfe for adding it back in [18:42] Launchpad bug 20664 in mozilla-thunderbird "Doesn't work right with Gnome "preferred applications"" [Medium,Fix released] https://launchpad.net/bugs/20664 [18:42] freedesktop 20664? [18:42] Freedesktop bug 20664 in Driver/intel "implement vblank sync'd GL buffer swap" [Enhancement,New] http://bugzilla.freedesktop.org/show_bug.cgi?id=20664 [18:42] hugs [18:43] yeah ubottu should know fdo is shorthand for freedesktop.org :) [18:44] bug #316240 [18:44] Launchpad bug 316240 in xserver-xorg-video-intel "[iG33] X freezes about a second after running "strace glxgears"" [Undecided,Confirmed] https://launchpad.net/bugs/316240 [18:46] bug #327844 [18:46] Launchpad bug 327844 in xserver-xorg-video-intel "[G45] X freezes about 1-5 min after switching compiz on" [High,Triaged] https://launchpad.net/bugs/327844 [18:52] jbarnes: going to be an exceptionally Good Friday?-) (bugwise) [18:52] tjaalton: heh I hope so [18:52] and isn't it a holiday for you too? [18:52] nope [18:52] ah, ok then [18:53] I'm ok with closing the bug, but I guess bryce will be here before too long [18:54] I've got a thinkpad X61 with 965, so if there are any bugs that need confirming being fixed or still there, just shout [18:55] great [18:56] although I'm behind 3G, so pulling stuff will take awhile [18:58] bryce, wonder if that's something we could use, a good directory of willing and responsive testers with any given chipset? [18:58] * cwillu volunteers for 945gm and 815 :p [19:02] yay, 5,5kb/s [19:41] heya [19:44] 282081 --> expired (thanks jbarnes) [19:47] 331596 --> wishlist [19:50] cwillu: there is a hardware database, so if we were really desperate to find testers with given hardware, we could use that. however I've never done it before [19:50] there is also https://wiki.ubuntu.com/X/SwatTeam with a list of hw available for testing [19:51] kinda out of date though