[00:08] cnd, hey [00:08] eruditehermit, howdy [00:09] so a couple of things [00:09] there are 2 sets of acceleration and sensitivity settings in gnome [00:09] mouse and touchpad [00:09] what am I supposed to change as a user [00:10] if I change the mouse stuff, does it only apply to non touchpad things? [00:10] i.e. usb mice [00:10] eruditehermit, correct [00:11] well, I'm not a gnome developer [00:11] but I believe that to be correct [00:11] hrm [00:12] so I think the touchpad accel and sensitivity work [00:12] however, they are really high by default [00:12] the sensitivity [00:12] especially [00:12] as in the lowest sensitivity is really high [00:12] what does accel vs sensitivity mean to you? [00:12] I don't really know the difference [00:13] accel means that if I move my finger fast, I should be able to move my mouse further than if I do it slowly [00:13] a dynamic sensitivity? [00:14] ok [00:14] does that make sense? [00:14] yeah [00:14] so I don't know exactly who's default is bad [00:14] X synaptics input module, or gnome [00:15] probably X synaptics, because the gnome default feels fine here [00:15] so X synaptics should have picked up on some property of your device and set a different scaling factor [00:15] I liked the defaults that the psmouse proto=exps provided [00:15] as far as accel and sensitivity [00:16] is there a way to query that information [00:16] and set it to be the same [00:17] also is there a way to obtain more than the gnome gestures [00:17] how do I find out if the touchpad accepts 3 finger touch [00:19] eruditehermit, no real way to query the exps defaults [00:19] exps used the X evdev driver [00:19] which has a different way of handling acceleration and sensitivity [00:20] eruditehermit, 3 finger gestures are disabled by default in ubuntu because they are reserved for, and may be used by, unity [00:20] unfortunately, the only 3 touch gesture in unity is broken right now [00:20] but it's being fixed :) [00:20] however, you can manually override this if you want [00:20] synclient ClickFinger3=2 [00:20] synclient TapButton3=2 [00:21] will enable middle click emulation through tapping and clicking (if you have a clickpad) [00:21] tapping and clicking? [00:22] on the touchpad [00:22] so if I click with 3 fingers on the touchpad [00:22] I will get middle click? [00:23] wow [00:23] lol [00:23] it worked [00:23] are there ways to make 3 finger gestures? [00:24] eruditehermit, what do you mean? [00:24] how does one make this persist? [00:25] swipe left with 3 fingers closes a program [00:25] lets say [00:25] or something like that [00:25] if you're using unity, you have to live with the unity gestures [00:25] it won't allow you to configure them [00:25] if you're not in unity [00:25] what are the unity gestures? [00:25] you can use the utouch-geis API [00:26] the only three touch unity gesture is dragging the window and spread/pinch to maximize/unmaximize [00:26] but that gesture is broken right now [00:26] ah [00:27] so zooming will be gesture enabled [00:28] eruditehermit: Option "ClickFinger3" "integer" and Option "TapButton3" "integer" [00:29] where do I put that? [00:30] http://paste.ubuntu.com/876835/ [00:30] save that as /etc/X11/xorg.conf [00:30] ah [00:31] xorg really needs a persistent settings mechanism [00:31] I thought the point was to remove xorg.conf [00:34] also [00:34] /etc/init/vgaswitcheroo.switch [00:34] not being executed anymore [00:35] has to have a .conf extension doesn't it? [00:36] err [00:36] thats what I mean [00:36] .conf [00:37] seems very temperamental [00:39] brb [00:39] let me test [00:41] darn idr isn't around, was going to ask if there were any plans to cherry-pick all these intel commits to 8.0 branch anytime soon [00:49] hmm [00:49] that script in /etc/init/vgaswitcheroo.conf isn't working [00:49] is there a way to see if it is being called? [00:49] and if it is being called, if its working [00:49] I swear it was working a few days ago [04:11] hey, does anyone have experience with hybrid graphics? [04:12] plenty of experience hating it but only on nvidia/intel lenovo where its possible to pick the gpu in the bios, none with ati/intel here [04:12] Sarvatt, so remember that init script we worked on a few days ago [04:12] that was RAOF but yeah [04:12] Sarvatt, it stopped working [04:13] yeah collectively [04:13] also [04:13] whenever I wake up from suspend my power consumption goes up a lot [04:13] as if the discrete gpu is switched on again [04:13] sudo service vgaswitcheroo start fix it? [04:13] however vgaswitcheroo reports DIS as off [04:14] ah yeah i have no clue, sounds like it doesnt work over a S3 :( [04:16] Sarvatt, sudo service vgaswitcheroo start works [04:16] but its not working on bootup [04:16] the machines canonical certifies that i have to care about are pretty much just business class ones and dell/lenovo only use ati/intel on consumer class ones [04:17] * Sarvatt has a strange perspective i guess :) [04:17] hrm [04:17] do you know about getting fglrx to work with this? [04:17] supposedly it works [04:17] tseliot is the best person to ask about that, he wrote the scripts fglrx uses [04:18] but he wont be around till monday CET [04:19] btw you said ppa-purge is broken? [04:19] yeah, it doesn't know about multiarch [04:19] it just tries to purge the native arch packages and leaves the :i386 stuff which breaks things bad [04:20] so how do I go about doing this properly? [04:20] http://paste.ubuntu.com/877015/ [04:21] thats what i use to purge edgers, but you it'll install some crap you dont have installed now [04:21] like -dev packages [04:21] * Sarvatt really needs to fix ppa-purge one of these days... [04:23] cairo libdrm and mesa :i386 packages really screw it up [04:24] since anyone with wine installed has that [04:24] hrm [04:24] it doesn't quite work [04:24] which is like everyone that cares to use edgers [04:24] it has some more dependencies that break it [04:24] had even more packages than that installed? [04:24] libpixman-1-dev/precise libpixman-1-0/precise libpciaccess-dev/precise libpciaccess0:i386/precise [04:24] and more [04:24] everytime I add one [04:24] it complains about more [04:24] lol [04:24] oh yeah pixman is new [04:25] since i purged it last [04:26] add libpixman-1-dev/precise libpixman-1-0/precise libpixman-1-0:i386/precise libpciaccess-dev/precise libpciaccess0/precise libpciaccess0:i386/precise to it [04:27] yep that did it [04:27] I was missing one of them [04:27] lol [04:27] cool beans, sorry about the trouble, its such a pain in the ass [04:27] forgot ricotz updated those two libs today [04:27] oh thank you so much for putting up with my questions [04:27] sorry about bothering you guys [04:28] the things not working on my laptop are slowly being fixed [04:28] i uploaded new synaptics and evdev that should work but they still haven't built [04:28] touchpad finally works a year after I bought it [04:28] yeah sforshee fixed alps to not suck [04:28] that was freaking awesome :) [04:28] yeah [04:28] that fixed so many laptops [04:28] finally multitouch on linux [04:28] like all dells in the past year [04:29] all dells in the past 3 years [04:29] so the broken parts on my machine are hybrid GPU [04:29] well alps wasnt used across the whole line till sandybridge, was rare before that [04:29] light sensor [04:29] and power consumption [04:30] that was like entirely rc6 being enabled by default [04:30] well [04:30] I still have issues [04:30] with my hybrid GPU turning on randomly [04:30] and even with all the power savings [04:30] it still only gets 3hrs [04:30] vs 5-6 on windows [04:31] which is a lot better than the 1hr I was getting [04:31] =p [04:31] also [04:31] how to make video players use intel vaapi [04:31] and flash plugin to use vaapi [04:32] sudo apt-get install i965 [04:32] oh flash i dont think that uses vaapi at all [04:33] but the va 965 drivers are a separate package now [04:33] do you mean sudo apt-get install i965* [04:33] also I have that driver installed [04:33] but my CPU usage spikes when playing videos [04:33] i965-va-driver is what you need to make va work [04:33] the package name [04:34] yeah I have that [04:34] that should totally be a recommends or suggests.. [04:35] it should be a depends [04:35] lol [04:35] not sure why someone wouldn't want it [04:35] agreed :) [04:35] because they use fglrx and want to save 50kb disk space? [04:35] or do suggests get pulled automatically [04:35] people are weird [04:35] recommends do i think [04:35] then recommends is fine [04:35] so if someone wants to remove it they can [04:35] but by default it should be installed [04:37] Uncompressed Size: 28.7 k [04:37] even less than i thought [04:38] 2.7kb compressed size on the livecd [04:38] so how do I test if it is being used when I play a video? [04:38] think i'll file that bug :) [04:38] eruditehermit: what playback app? [04:38] lets say totem [04:38] or vlc [04:38] depends on which you're using, its enabled different ways [04:39] hmm [04:39] totem is weird, i think that uses gstreamer-vaapi [04:39] vlc is weird too [04:39] you go to INPUT codecs [04:40] and check the use gpu accelerated decoding [04:40] vainfo will tell you what your gpu supports [04:42] and hope your videos arent like mine encoded in h264 hi10p profile that cant be accelerated by any gpu [04:42] (anime) [04:43] i'm not sure how you make sure the upstart script is used every boot though [04:44] i thought it ran every script in /etc/init/ at boot, maybe the radeon hasn't loaded when it tries to run? [04:44] (sometimes, making it racy) [04:44] there might be a conditional you can throw in there to make it wait until its ready, lessee [04:45] no, i remember what RAOF pasted you waited until radeon loaded [04:46] maybe add the vgaswitcheroo stuff to /etc/init/plymouth.conf? [04:47] hmm no radeon is usually modprobed by xserver which would be well after plymouth starts sometimes, especially in a dual gpu scenario [04:47] err lightdm i meant there instead of plymouth [04:48] and (drm-device-added card0 PRIMARY_DEVICE_FOR_DISPLAY=1 [04:48] would be satisfied by the intel being ready [04:49] hrm [04:49] why wouldn't the upstart script work as is? [04:49] it should wait for both intel and radeon [04:49] radeon being modprobed isnt enough, it takes a long time to finish loading, we hit bugs with that many times in the past [04:50] ah I see [04:50] (just thinking out loud, might be wrong) [04:50] it doesn't wait till its finished? [04:50] it fires off scripts after they just start? [04:51] nope it just checks if its loaded before running and probably runs the same time that its modprobed, can ya paste it? [04:51] it's not critical that the radeon card be turned off *immediately*, right? [04:51] the script? [04:51] so just add a sleep 3 or something [04:51] broder: nope [04:51] oh yeah, friggin easy solution! [04:51] :) [04:51] eruditehermit: do what he says, add a sleep 3 :) [04:52] friday night, beer oclock, brain not working obviously [04:52] http://paste.ubuntu.com/877036/ [04:52] like that? [04:52] eruditehermit: perfect [04:53] so is there a way to do stuff before suspend and after resume [04:53] I want to turn all GPUs on and then off on resume [04:54] have ya googled vgaswitcheroo suspend? [04:55] eruditehermit: put http://paste.ubuntu.com/877039/ in /etc/pm/sleep.d/00_vgaswitcheroo [04:55] possibly chmod +x it [04:55] you did say that echoing OFF back into the switcheroo node fixes it, right? [04:56] I think so [04:56] well I'll do an ON first [04:56] then off [05:01] eruditehermit: its gstreamer0.10-vaapi [05:01] for totem [05:01] i'm not sure if its automatic that it'll use it though, probably is [05:02] hmm I already have that too [05:02] tjaalton: wth is that package name? :P [05:03] but weird [05:03] power usage goes from 13W to 26W when playing videos [05:04] eruditehermit: not surprised, i dont recommend using vaapi at all :) [05:04] how come? [05:04] vaapi is a joke, it uses as much cpu as software rendering [05:05] plus you're running the gpu at higher voltages/clock speeds because its being used [05:05] hmm [05:05] my system is running at 13-15W [05:05] need to get it to 8-10W to match windows [05:06] its nothing like vdpau on nvidia where the cpu isnt stressed [05:06] Linux kernel is power hungry [05:07] ok [05:07] let me test my new startup script [05:07] brb [05:07] thanks Sarvatt, broder [05:07] increase the sleep if it doesnt work, 3 seconds really should be plenty though [05:10] lol [05:10] so it works! [05:10] but another annoyance I forgot about [05:10] brightness settings do not persist [05:10] nor do bluetooth being on/off [05:11] its more like X starts at 8 seconds into the boot on a fast ssd, radeon hasnt finished loading by then, a second later its not ready and there is corruption on the screen from the plymouth->x transition before it was ready that was a problem [05:11] oh cool [05:11] so many things don't persist [05:11] xinput settings too [05:12] eruditehermit: thats what you get for buying a linux unfriendly sony :) [05:12] does it persist on other machines? [05:12] actually this machine has been pretty good [05:12] Dell's were a lot worse [05:12] their mice drove me crazy [05:12] this mouse was usable even with psmouse [05:12] err [05:12] proto=exps [05:13] they're like the only oem besides apple that doesnt have any vested interest in making linux work, fujistsu would be number 2 close behind them in bios bugs and lack of platform drivers taking care of those things :) [05:14] surprising that everything mostly works then [05:14] just the ati thing [05:14] and then general linux kernel power hungry ness [05:15] yea that ati thing is a problem for everyone, hybrid graphics are annoying :) [05:15] I guess the light sensor isn't hooked up [05:15] but the rest of it works [05:15] no driver for it im sure [05:16] honestly dell had shitty laptops last year [05:16] actually everyone did [05:16] when I got this [05:16] I wanted something small and light [05:16] everyone was making big and heavy [05:16] this year people have caught on [05:16] ultrabooks [05:16] choice was apple [05:16] and sony [05:17] yeah agreed [05:17] i went apple [05:17] and sony was lighter/more powerful [05:17] but the only other option was sony [05:17] even apple at the time [05:17] only had the air [05:17] air was 3lbs [05:17] pro was 4.5 [05:17] mine is 3.5 with the power of the pro [05:17] oh ok i waited till june for sandybridge airs, those are awesome :) [05:18] I got it february [05:18] but yeah GPU blows [05:18] i have a 17" laptop with a gtx 460m for that [05:18] lol its good in windows [05:18] i have an i7! [05:18] i7 and discrete GPU [05:18] that you cant use? [05:18] at 3.5lbs [05:18] well [05:19] I can use it in windows [05:19] :( [05:19] yeah [05:19] getting it to work with fglrx is the trick [05:19] would be cool if X could hot swap GPUs [05:19] the thing is, ati only cares about people who ship linux and have the oems escallate problems to them [05:19] read something on phoronix yesterday about it being a possibility [05:20] which pretty much wont ever be sony [05:20] HP and dell have nice ultrabooks now [05:20] more choice [05:21] i like ux31 atm personally [05:21] but really [05:21] brightness not being remembered [05:21] that isn't a sony problem [05:21] its a gnome problem right? [05:21] or x [05:21] no thats actually a bios problem afaik [05:21] i could be wrong [05:22] the brightness level isnt stored anywhere on your system [05:22] its up to the bios to remember what you were at before, is it right on the next POST? [05:22] I just have to lower it on every boot once gnome is up [05:23] if its bright again on the bios screens the next boot linux wouldn't be what you blame [05:24] do you have any experience with fglrx at all? [05:24] I am going to try something crazy [05:24] hardly any [05:24] lol [05:24] to USE it [05:24] i used it maybe 1 hour in my life :P [05:25] just to see if new driver releases worked [05:25] with nvidia [05:25] can you switch GPU without restarting X? [05:26] no [05:27] well [05:27] you cant do that on anything due to how X works [05:27] lets see how fglrx does [05:27] brb [05:39] hrm [05:39] failure [05:41] the fglrx package in precise seems old [05:41] and not to work with hybrid [05:41] it is old, because the one that works properly was released 2 days ago, the goal is to have the one released next week in final precise [05:42] next week? [05:42] they are releasing another one [05:42] or the week after, sometime soon [05:42] they got delayed for the feb release [05:42] I see [05:42] jan release was still as broken as the one in there now [05:43] the link is down [05:43] on their website [05:43] was no feb release, feb catalyst was released 2 days ago [05:43] for 12-2 [05:44] safe to say if its broken in the precise one it'll still not work though :( [05:45] if it is broken in 11-11 in precise? [05:45] hopefully they fixed it all to work [05:45] seems like it does a lot of moving of other libs around [05:45] regular libgl etc [05:45] its just a minor increment to the 12-2 one coming out real soon now, they skipped the february release and released an early version of marches for 12-2 [05:45] yeah [05:46] you can build your own version from the ati.com one easily though no need to use the distro packages [05:46] any idea where to get it from? [05:46] like --build Ubuntu/precise [05:46] their link is down [05:47] http://www2.ati.com/drivers/linux/amd-driver-installer-12-2-x86.x86_64.run works for me [05:47] Duplicate headers received from server [05:47] wow it works with wget [05:48] chrome barfs for some reason [05:48] one sec uploading it [05:48] oh ok [05:48] the website is all kinds of messed up in chrome here [05:48] had to pick the driver with the keyboard [05:48] yeah [05:50] ./amd-driver-installer-12-2-x86.x86_64.run --buildpkg Ubuntu/precise it is [05:50] it doesnt work on i386 [05:50] another reason it probably hasnt been uploaded by now [05:50] --buildandinstallpkg is what I use [05:50] lol [05:53] http://ubuntuone.com/1UXKtX6u65jOGDrNpyWHyY if you didnt manage to get it [05:53] gotta use my 50gb ubuntuone storage somehow [05:53] wget was able to get it for me [05:53] strange that chrome didn't like it [05:54] I thought it was down [05:54] Version 19.0.1061.1 dev [05:54] working fine there [05:54] 19??? [05:54] lol [05:54] I have 17 [05:55] oh i must be 2 days ahead of you then [05:55] dev channel :) [05:55] stable is 17 still [05:55] lol [05:55] but yeah [05:55] chrome was at version 2 a few days ago [05:56] ok [05:56] time to try new fglrx [05:56] brb [06:33] looks like that went well :) [06:49] Sarvatt: the -vaapi part?blame ustream :) [06:49] well i tried to install gstreamer-vaapi like the source package name :) [06:50] was more complaining about the 0.10 part [06:51] but yeah thats how gstreamer crap is namespaced just my own screwup [06:53] yep [06:53] besides, it triggers bug 946742 [06:53] Launchpad bug 946742 in intel-vaapi-driver (Ubuntu) "Shotwell crashes on start with gstreamer0.10-vaapi" [Undecided,Confirmed] https://launchpad.net/bugs/946742 [06:54] oh FUN [06:54] fixed in upstream master, but would rather know which commit [06:56] http://cgit.freedesktop.org/vaapi/intel-driver/log/ is taking forever to load [06:57] at least its hard to hit, have to purposefully install 2 things to hit it [06:58] nah, just browse with nautilus to a folder where you have videos :) [06:58] i965-va-driver and the gstreamer0.10-vaapi thing [06:58] oh [06:58] yeah [07:00] i like how cedarview pvr va crap doesnt even work with it that you went out of your way to add it for :P [07:04] tjaalton: hmm nautilus is fine here [07:04] on snb [07:04] video folder, thumbnails arent killing anything [07:06] was gonna bisect but need to reproduce first :) [07:08] shotwell is just giving generic video thumbnails, bah [07:09] ok, fails here [07:10] http://ubuntuone.com/6BjApmzC9kzm1HyQSVm8fB (the one to the right of to the cursor) [07:11] what kinda videos do you have? avi? [07:12] * Sarvatt browses to the NAS where all his videos are [07:12] guess I tried it with bigbuckbunny [07:12] meh, resume fail [07:14] argh i think i did something so it didnt thumbnail videos and cant remember what it was [07:15] ok so is it unity fail when I get a blank screen, but the pwd dialog is obviously working [07:15] oh [07:15] shotwell-video-thumbnailer: gen6_mfd.c:844: gen6_mfd_avc_ref_idx_state: Assertion `frame_idx < (sizeof(gen6_mfd_context->reference_surface) / sizeof((gen6_mfd_context->reference_surface)[0]))' failed. [07:15] and the mouse cursor changes when it's moved around [07:15] i got that to stderr launching it in a terminal, thats from intel-driver [07:15] yeah, unity replace fixed it.. [07:17] or not [07:17] Sarvatt: yeah that's the one [07:18] http://cgit.freedesktop.org/vaapi/intel-driver/commit/src/gen6_mfd.c?id=368731d104da84605fcf6683d6ce014916fe76b0 [07:22] looks plausible [07:29] only commit that touched that function between the two versions but yeah [07:30] doesn't apply easy to test :) [07:31] yeah I tried something and came to the same conclusion.. not easy to just pull something.. [07:32] so I asked the author if there's going to be a release soon but he didn't reply [07:35] git cherry-pick 99ded53e66af1903f1d58ffbc24404d435a6de84; git cherry-pick 99ded53e66af1903f1d58ffbc24404d435a6de84 gets it to apply easily [07:35] got it building now, wish i could make it crash [07:35] but still i should be able to tell if that assertion is gone [07:35] right [07:36] no assertion after cherry-picking those two [07:36] hehe, thanks [07:36] you can rest now :) [07:39] http://ubuntuone.com/6R5dmBssdYVJtouXzcciv6 if you want to check it [07:40] i965-va-driver and libva are in collab-maint and a friggin nightmare to update [07:40] as bad as libxcb was a year ago [07:41] nightmare how? [07:41] * Sarvatt is spoiled by X packages [07:41] changes have to git in git, autoreconf via a patch? [07:41] err have to go in git [07:41] or did that change too.. [07:41] oh [07:41] * Sarvatt absolutely hates autoreconf via a patch [07:41] like wacom in debian [07:42] yeah that's stupid [07:42] another thing is importing tarballs when upstream uses git.. [07:42] like, what? [07:43] the git-buildpackage workflow? [07:43] aka nouveau? [07:43] xterm [07:43] oh xterm doesnt use git nevermind :) [07:43] well it should work fine with upstream branches from upstream git [07:58] ricotz: tell me if i broke input on edgers :) [07:58] * Sarvatt isnt using it and apparently evdev was broken before [08:04] tjaalton: hmm so the actual driver is in libva-intel-vaapi-driver not i965-va-driver [08:04] * Sarvatt only installed i965-va-driver testing shotwell [08:04] so disregard my saying i didnt hit the assert :) [08:04] no clue what triggered it then [08:05] Sarvatt, havent updated to your uploads yet ;), and the scrolling-brokenness was on the gtk3 side which was fixed in git master [08:06] Sarvatt, maybe you have an opinion to add https://bugs.launchpad.net/ubuntu/+source/libpciaccess/+bug/950985 [08:06] Launchpad bug 950985 in libpciaccess (Ubuntu) "FFe: Sync libpciaccess 0.13-1 (main) from Debian unstable (main)" [Undecided,New] [08:07] http://ubuntuone.com/6kx60S7s3RixNNpIuw6aea is libva-intel-vaapi-driver [11:38] Sarvatt, still about? [19:03] darn..... compiz/unity are fighting over Alt+Tab control :/ [20:33] hello! [20:38] Sarvatt, you about?