=== scientes is now known as paulprotoss === paulprotoss is now known as scientes === BV1AL_ is now known as BV1AL === LetoTheII is now known as LetoThe2nd [07:36] ogra_: you changed Bootloader-sets-root: to yes even if the meaning is currently wrong! :-) [07:36] that's a different behavior from debian [07:38] but will at least now work after the installer finishes? [08:00] rsalveti, i havent tested it on actual images yet, but beyond that it should work [08:01] (as documented in the changelog) === scottb` is now known as scottb [14:40] * ogra_ hugs rsalveti [14:40] yay [14:41] ogra_: there's only one new package that needs to land now, the libdri2 [14:41] which is what I'm testing at this moment [14:41] then we'll be ready to go with pvr-omap4 [14:42] kernel and xorg related packages all in already [14:42] yay [14:43] * ogra_ wonders if omap should be in the xserver-xorg-video-all deps on arm [14:45] I belive so, at least we need to make sure we're installing it by default [15:06] rsalveti: libdri2-omap in the archives for 12.10? [15:06] ndec: that's what I'm doing now [15:07] nice1 [15:07] it's needed by the sgx driver [15:07] nice! [15:07] are you taking the Xserver patches for dri2video as well? sounds more risky ;-) [15:08] ndec: not at this point yet [15:09] want to first make sure it works without any new fancy features :-) [15:09] how about the clutter/totem 3.4? is there any ARM SoC which has support for that? [15:16] ndec: don't think so [15:17] rsalveti: also i wanted to ask you about LEB. so ubuntu is no longer doing pre-installed. how does it impat linaro LEB? [15:17] in fact i don't recall exactly how LEB are generated. [15:18] ndec: doesn't change much, we're still doing pre-installed images [15:19] as well with hwpack + rootfs [15:19] even for quantal [15:20] rsalveti: ok. is there a wiki that explains your image generation? [15:20] we need to change our image generation for TI releases... since we were using pre-installed images ;-) [15:20] ndec: sure, 1 sec [15:21] ogra_: can you grab powertop 2.1 from http://people.linaro.org/~hrw/debian/ and sponsor upload? :D [15:21] ndec: http://git.linaro.org/gitweb?p=ubuntu/ubuntu-build-service.git;a=summary [15:21] http://git.linaro.org/gitweb?p=ubuntu/ubuntu-build-service.git;a=tree [15:21] if you see, this is how we're maintaining our build jobs [15:22] the only thing done by jenkins is a cd to the right folder and ./configure && make [15:22] rsalveti: and it works outside of LP, e.g. I should be able to run this script? [15:22] that runs live-build and generate the tarball in the end [15:22] ndec: yup [15:22] hmm... why did I not look into that before... [15:24] bye [15:24] ndec: the logs from the job handling the quantal images: https://ci.linaro.org/jenkins/view/Ubuntu%20Build%20Service/job/quantal-armhf-ubuntu-desktop/65/consoleFull [15:25] if you want to create a pure ubuntu image (without the linaro sauce), it's quite easy as well [15:25] just need some minor changes, like by not installing linaro-ubuntu-desktop and installing the task ubuntu-desktop [15:25] thanks a bunch! [15:29] hrw, hmm, the .dsc expects a tar.gz tarball [15:30] i cant extract it with dpkg-source [15:45] rsalveti, any chance that we see a pcr for omap3 update (not FF critical indeed) ? [15:46] *pvr [15:46] ogra_: that depends on TI providing us the armhf binaries =\ [15:46] i thought there were published ones [15:47] not for armhf, at least not by the official channel [15:47] hmm, i thought i read something about hf binary drivers [15:47] but i cant findit in my history... [15:52] ogra_: rsalveti: i don't think i have seen a DDK/armhf for OMAP3. [15:52] k [15:52] i think, though, that some TI people are working on adding armhf support in OE, so it is a sign that such binary might show up... [15:53] well, probably someone talked about an inofficial binary or some such [15:53] i know i catched it up somewhere [15:59] * ogra_ adds unity-2d back to the ac100 images [16:01] ogra_: there's a new mesa at the archive today [16:01] might be nice to check if unity3d would work [16:01] or at least how it'd explode [16:01] rsalveti, well, slangasek apparently tested yesterday, with rather disatrous results [16:01] hehe :-) [16:01] it doesnt explode, but itsnt usable either [16:03] hmm, i see a compiz upload but nothing in the changelog indicates its the one that has GLES === prpplague^2 is now known as prpplague [16:28] ogra_: How does adding unity-2d to ac100 work when it's been removed from the archive? Or did I miss someone adopting it and re-uploading it? [16:28] it hasnt been removed [16:28] it was removed from the seeds [16:28] and i dont think anyone plans to remove it from the archive actually [16:29] ogra_: Uhm. It was removed from the archive last week. [16:29] ogra_: Because no one stepped up to fix the dconf->gsettings mess, or something. [16:29] why does apt-cache madison on a current panda still show source and binary then ? [16:30] Because it lies? Or has precise in its sources? Or something? [16:30] The source was removed. [16:30] SH*T [16:30] Some binaries may still linger due to them having rdeps. [16:31] well, rickspencer3's suggestion was actualyl to have it back on the arm images worst case [16:31] Yes, well. The desktop team screwed you there. [16:31] and the results for llvmpipe are far from usable [16:31] damned [16:31] Probably worth an internal discussion about porting the bits that need porting and reintroducing it. [16:31] as I understand it, llvmpipe is completely out of the question for ARM [16:32] we need to wait for the PVR driver to run unity [16:32] rickspencer3, after slangasek's test results definitely [16:32] problem is that we'll be able to solve the unity issue just for panda [16:32] I don't think anyone ever envisioned that it would work, really [16:32] which we have the river [16:32] people can dream hehe :-) [16:32] rickspencer3, haha [16:33] is there another DE that we can use at least to test until the PVR is available? [16:33] just to bridge us until Unity is runnable? [16:33] Gnome 2? [16:33] rsalveti, right, and for ac100 i was just planningg to pull unity-2d in ... i wasnt aware they actually removed it [16:33] rickspencer3: the pvr driver will be available later today [16:33] omap4 will be fine soon [16:33] oh, so Unity 3d should work(ish) later today? [16:33] sorry [16:33] ish, right [16:33] you guys were talking about other boards [16:34] * rickspencer3 backs away quietly [16:34] what wont be fine are all other arm desktop images [16:34] * rickspencer3 backs back in [16:34] I wonder if Lubuntu might be a better flavor for other boards? [16:34] my second choice after re-introducing unity-2d is lubuntu-desktop [16:34] it runs my eeePC 701, which is pretty impressive [16:34] but that requires more work for me [16:35] ogra_: Given that ac100 is a community image, why not just switch it to Xubuntu? :) [16:35] ogra_, you can't just apt-get install lubuntu-desktop? [16:35] infinity, why using that much overhead ? [16:35] ogra_: Or lubuntu, sure. [16:35] :) [16:35] lubuntu works as well [16:35] yeah, and will have a tiny footrprint [16:35] the lubuntu community targets exactly this scenario [16:35] * ppisati +1 for reintroducing unity2d [16:35] ogra_: I really see no reason to ship the full whiz-bang Unity experience for anything except our darling "tech demo" platform (OMAP4). [16:35] it's the only reason that we bought new PowerPC builders, so that lubuntu could make a PowerPC version [16:35] ppisati, its gone from the archive ... i dont see how we woudl get it back without lots fo extra work [16:36] ppisati: And as much as I want unity-2d back, someone needs to commit to the work, it's not self-maintaining. [16:36] infinity, i agree, but someone has to maintain the differences [16:36] ogra_: didn't you just said it was just eliminated from seed? [16:36] *say [16:36] ppisati, i was just proven wron above [16:36] ah [16:36] *wrong [16:36] ogra_: What difference? s/ubnutu-desktop/lubuntu-desktop/ in the ac100 target in livecd-rootfs, profit? [16:37] infinity: i agree, but have you ever tried the "3d experience" on a pandaboard? [16:37] infinity, well, i have to get used to it for testing etc etc [16:37] infinity: if we need to switch for $foobarUbuntu, i prefer U2d [16:37] it's working way better with latest compiz [16:37] ppisati: With the PVR driver, it's not bad. [16:37] you'd be surprised [16:37] and will start to fix arm issues on lubuntu [16:37] since people tend to copmpain to *me* [16:37] *complain [16:37] Fixing ARM issues in the whole archive isn't a bad thing. :P [16:37] But really, it should all Just Work. [16:37] If it doesn't, we suck at this. [16:37] A lot. [16:38] i think its a lot more than changing the target btw [16:38] Either way. I don't want us shipping desktop images for a ton of platforms regardless. ac100 gets a free pass because it's a pain to install it any other way. [16:38] i'm not even sure there is any oem-confi support in the lubuntu ubiquity [16:38] Uhm, it's the same ubiquity... [16:39] * ppisati goes testing another kernel... [16:39] they have overlay packages for everythin in lubuntu, no ? [16:39] Not for the installer itself. [16:39] * ogra_ needs to get used to this new kbd [16:39] ok [16:39] well, then lets just do it :) [16:40] I suspect just swapping xubuntu or lubuntu in will Just Work, as I said, but either way, it's probably less effort than committing to support unity-2d, unless you want to be the new u2d upstream in your spare time. [16:40] And if you do, awesome, lots of us love and miss u2d. :P [16:40] haha [16:40] i'm not that crazy, no [16:40] i really would like to see some community people picking it up though [16:42] infinity, erm, you didnt mean livecd-rootfs but the nusakan crontab (and default-arches), right ? [16:47] * rickspencer3 looks forward to dist-upgrading panda board tomorrow [16:47] ogra_: Oh, fair point, it's not hardcoded anywhere. [16:48] good, you had me confused :) [16:48] ogra_: But yeah, just add armhf+ac100 to lubuntu's default-arches, and give it a try, I guess. Or do a test-spin by hand. [16:48] ogra_: It seems like the sane(ish) way forward. [16:49] yep [16:49] what do we do with mx5 ? [16:49] rsalveti, will HD video work tomorrow too? [16:49] * rickspencer3 hopes [16:49] LOL [16:49] hahahaha [16:49] rickspencer3, we have no codecs [16:49] so, "not yet", right? [16:49] and its unlikely they will ever enter the archive [16:49] ogra_, well, I thought I could install them from somewhere else [16:50] you can, but for precise [16:50] with luck TI will provide gstreamer packages and codecs from the PPA after release [16:50] ogra_: We already dropped mx5 desktop images, didn't we? [16:50] ogra_: If we didn't, I'll do that. [16:50] unfortunately for quantal a lot of stuff changed [16:50] infinity, not sure, i thought that was only temporary [16:50] they are maintaining like hundreds of patches at the gst packages [16:50] wrt livebuilder shortage around a milestone [16:51] ogra_: Nope, the only desktop images we should be producing are omap4 (the blessed tech demo image), and ac100 (cause it sucks installing any other way). [16:51] i think we do build omap still atm [16:51] ogra_: But mx5 reminds me that we still need to do d-i images from universe kernels. [16:51] and i woudl like to keep the opportunity open for an omap3 HF driver [16:51] so i would like to go on buildin them [16:51] no need for testing or milestone releases though [16:52] I see no issues with asking non-omap4 users to use netboot and install desktops by hand. It's not like ubiquity on a Beagle is a pleasant experience anyway. [16:52] infinity, i guess FF day is a little late for that [16:53] according to colin there is some heavy lifting involved to make d-i use universe [16:54] * ogra_ would really like to see some images to actually test flash-kernel-installer [16:58] ogra_: fetch again [16:58] will do [17:02] ogra_: new compiz doesn't include yet the gles branch [17:02] it's one rev before the one we wanted [17:04] D'oh. [17:06] sigh [17:06] oh that endless game [17:08] hm, someone really need to spend some time optimizing the updating of the software catalog [17:08] that always takes a huge time === scottb is now known as scottb_lunch [17:08] improve IO [17:08] its unzipping huge files [17:09] yeah, imagined that [17:13] bah, sigh [17:13] if i hadnt messed up livecd-rootfs i could now do an ac100 tesbuild [17:13] life always strikes back [17:21] hrw, done [17:27] ogra_: cool, thanks [17:50] ogra_: bug 1040611 [17:50] Launchpad bug 1040611 in ubuntu "[needs-packaging] libdri2" [High,Confirmed] https://launchpad.net/bugs/1040611 [17:50] ogra_: last piece before the pvr-omap4 driver [17:53] rsalveti, uh, whats the license [17:53] http://paste.ubuntu.com/1163077/ [17:53] expat [17:54] rsalveti: Is this a generic dri extension thing that's being pushed to fdo/xorg upstream and (eventually) used all over, or some ARM(or TI)-specific madness? [17:54] oh, i misunderstood "as it's binary only, and built against the shared library " [17:54] :P [17:54] i thought that libdri2 had binary bits [17:54] liked against the free one [17:54] infinity: no, idea is to be the reference and generic dri extension [17:54] to avoid duplicates all around [17:54] rsalveti: If it's not actually generic and intended to be used cross-platform, "libdri" is pretty awful namespace pollution. [17:55] robclark: ^ [17:55] he's the father [17:55] robclark: Bring your sideburns over here and tell me about libdri. [18:01] * ogra_ runs an ac199 lubuntu-desktop testbuild [18:01] *ac100 [18:01] If the ac199 comes with an en_US non-android keyboard, I'll take two. [18:02] infinity, rsalveti, libdri2 is basically some code extracted out of mesa so that other things can use it [18:02] infinity, you and your keyboard fixation :P [18:02] robclark: Sure, I grasp *what* it is, but is it being pushed upstream as a generic solution that others intend to use for this same purpose, etc? [18:03] robclark: Cause if it ends up being just an ARM thing, just a TI thing, the namespace pollution of "libdri" is a bit icky. [18:03] robclark: And, if it's partly cargo-culted from mesa, why not build it from mesa, so it can be more readily kept in sync? [18:04] infinity, well, I did send a while back a patch for mesa to use it.. [18:04] robclark: (And so others can learn of its existence and use it) [18:04] anyways, the point was that mesa, vdpau, vaapi, etc where all just copy/pasting the same code [18:04] robclark: Dude, are you going German on me? [18:04] :-P [18:04] "well, I did send a while back a patch for mesa to use it"... I thought Oli typed that, until I read the nick. :P [18:04] Just sayin'. [18:04] I need to kick mesa folks again... [18:05] LOL [18:05] texas is full of germans [18:05] lol [18:05] heheh [18:05] robclark: Anyhow, if you've had any positive response from upstream about it at all, and if it seems likely it'll end up being "a thing" upstream, then I'm less grumpy about including it as-is as a stop-gap in Ubuntu right now. [18:05] might have some bad influence on the language :) [18:06] well.. my reasoning is that currently dri2 is a bit of an exception to the way most x11 extensions work.. ie most you have a proto tree plus a client side lib tree [18:06] dri2 is the only one (that I know of) that doesn't have a client side lib tree.. hence libdri2 [18:07] anyways, why is it a namespace issue? [18:07] robclark: *nod*... Like I said, I'm down with the concept, and your reasoning. Just more curious if you've been able to sell upstream. [18:07] robclark: But if you're pushing upstream to DTRT here, that's cool. [18:07] robclark: It's a namespace issue only because I could totally see someone else introducing a libdri that's an entirely different thing, so it would be nice if yours was on the radar and won. :P [18:07] I do need to kick them again.. if I can make it to xdc then I'll bring it up in person there.. [18:08] I think main issue is that most everyone on xorg/mesa/gfx side of things has too much to do and too little time to care about those sort of little cleanups :-P [18:09] bah, lubuntu build failed [18:09] robclark: You'd think that would make them all the happier about someone else doing it. [18:10] The following packages have unmet dependencies: [18:10] indicator-application-gtk2 : Depends: indicator-application (= 0.5.0-0ubuntu1) but it is not going to be installed [18:10] GRMPF [18:11] i guess lubuntu wasnt a good choice if they have to port the world to gtk3 first [18:11] I assume that's just buildd skew. [18:12] i would assume that indicator-application stopped providing its gtk2 bits [18:12] infinity: so what do we say, trust robclark and push it to the archive? :-) [18:12] Oh, indeed. Very recently. [18:12] rsalveti: Yeah. If robclark fails us, we can wax his face. [18:13] ogra_: Yell at Seb? [18:13] infinity, he will just point me to the unity team [18:13] ogra_: Dropping GTK2 in indicators all over is a nightmare for Xubuntu and Lubuntu, and right before FF, no less. :/ [18:13] and they then point me to sabdfl ... and i wont yell at him i fear :P [18:14] You should. [18:14] In cases like this, he's just another contributor. [18:14] no, yelling is the wrong approach for sure [18:14] Until he formally sabdfls something, which is a different matter. [18:15] there was an email about moving those bits to universe? or was it a complete drop? [18:15] (e.g. separate source package) [18:16] well, seb just referred to that mail [18:16] in -desktop [18:17] so it seems xubuntu wants to take over maintenance of gtk2 parts in universe [18:17] ogra_, so xubuntu is not gtk3? [18:17] but seemingly need to push these in from scratch [18:17] *xfce [18:17] no idea [18:17] they use a lot of gtk2 stuff, that i know [18:18] lubuntu intrestingly as well [18:18] i guess for NM [18:18] infinity: ogra_: and finally, the pvr driver: http://people.linaro.org/~rsalveti/pvr/, but it needs libdri2 [18:18] yay [18:19] http://git.linaro.org/gitweb?p=ubuntu/pvr-omap4.git;a=summary [18:20] scientes: XFCE is GTK2 for now. There was a push for GTK3 porting, and it got derailed upstream. [18:21] Turns out it's a lot more work than people realised when they first got gung-ho about it. :P [18:21] Like, effectively a rewrite of half the world. [18:22] looks like lxde is being ported too [18:22] and i *was* subscribed to the linux gtk3 migration [18:22] but that patch and email flow was just too great === jibel__ is now known as jibel === scottb_lunch is now known as scottb [20:54] ogra_: hey, any news about the driver? [20:55] infinity: for you as well, in case you're willing to sponsor it :-) === rsalveti` is now known as rsalveti === zyga_ is now known as zyga === Ursinha` is now known as Ursinha