[03:21] lilstevie: my btrfs netbook decided to become corrupt with warning this morning, so I think I will be fast-tracking my TF101 now :-P [03:29] lol [03:33] At least the notes were copied to another box === zyga__ is now known as zyga-afk [03:54] Grrr. Looks like hunspell-en-us has clobbered all arm desktop images. [04:47] lilstevie: oh @ evdev, rollup. [04:55] lilstevie: what's the URL for your latest tarball of stuff? [04:55] IIRC you changed its name and suchlike [05:01] Ah, "OLiFE" [09:46] hmm, where does that second omap4 armhf image build come from [09:48] infinity, was that you ? [09:48] * ogra_ definitely didnt trigger a second build [11:01] ogra_: Eh? What second build? [11:02] ogra_: Oh, cdimage? I've been asleep. So, not me. [11:02] k [11:02] i'll add the SUDO_USER stuff today then :P [11:03] but first i would love to know why ubiquity is broken :( [11:03] The build, you mean? [11:03] * ogra_ sees no valid reason for it not finding flash-kernel-installer [11:03] I was going to look at that today. [11:03] After I get this mklibs hackery committed for netboot. [11:03] it complains about f-k-i missing in its source tree [11:03] Yeah, I saw that. Made little sense. [11:04] Ooo, my bootstrap build of ghc finished. [11:04] but f-k-i definitely built for hf and the last ubiquity upload calims it pulled it in in the changelog [11:04] Time to get the archive building it. [11:05] hmm, though it claims it can not create the target in the build log [11:05] not that the source file isnt there [11:05] i wonder if there is just a mkdir missing [11:06] (or -d for the install command) [11:06] But it works on armel. [11:06] Should be the same code. [11:06] yes [11:06] Looking, though. [11:07] well, colin claims in the changelog that he adapted links for hf or some such [11:07] * ogra_ checks for the exacrt wording [11:07] hmm, no, that was d-i [11:07] ubiquity only has * Add armhf support. [11:08] * ogra_ doesnt really get it [11:09] lrwxrwxrwx 1 adconrad adconrad 15 2011-12-08 04:06 armhf -> d-i/lists/armel [11:09] That could be the problem. [11:09] hmm [11:09] That should be a link to the same directory. [11:10] geez, unpacking the xz takes my ac100 close to a halt [11:10] still running [11:11] fun [11:12] infinity, to the same file you mean [11:14] Well, to the file in the same directory, yes. [11:14] yup [11:14] Same problem in debian/ [11:15] you mean the .install file ? yeah [11:15] And that's probably your missing target. [11:15] and .dirs as well it seems [11:15] * infinity is also curious why ./d-i/source/base-installer/kernel/armhf.sh is different from ./d-i/source/base-installer/kernel/armel.sh [11:16] But I'm also not entirely convinced that's used. [11:17] i doubt it, but it will surely caquse d-i issues since it originates there i think [11:20] Well, let's see if fixing those symlinks makes ubiquity happy. [11:20] To test build or not to test build... [11:20] pfft [11:20] just upload [11:21] will just waste 1h of your day to do a testbuilt, if it fails we can upload again [11:21] *testbuild [11:22] testbuilds are for freezes :P [11:24] * ogra_ wonders why gcc-4.5 is in the core seed [11:25] arent we using 4.6 since a while ? [11:25] 4.4 is still in main too. [11:26] main, sure [11:26] but core ? [11:26] Eh? [11:26] "core" is just an upload packageset. [11:26] I don't ever want any version of GCC living outside core, do you? [11:26] i want the default one in core ... [11:26] Packagesets != Seeds. [11:26] why would i want the fallback there [11:27] You want MOTU updating every compiler other than the primary one? [11:27] Possibly accidentally dropping vital patches and breaking them? :) [11:27] well, -snapshot is in universe as well [11:27] Yeah, because we don't build packages with snapshot. [11:27] afaik at least [11:28] Sometimes we fall back to building packages with older versions of gcc. [11:28] (MySQL is built with 4.5, for instance) [11:28] ah, k [11:28] i thought we fobid that a while ago [11:28] *forbid [11:28] We can't, really. [11:28] If a new compiler is broken for some chunk of code, what do we do? [11:28] fix it :) [11:28] Just say "well, we can't ship MySQL unless the compiler is fixed"? [11:29] scream and shout to linaro ... [11:29] That sort of thing doesn't work well with time-based releases. :P [11:29] heh [13:07] infinity, hmm, what is wrong with https://buildd.debian.org/status/package.php?p=gcc-4.4&suite=sid ? [13:07] armel [13:09] doko: You forget to backport your -mfloat-abi=softfp change from 4.6 to earlier versions? [13:09] doko: (The one where you explicitly set that for the Debian/armel build) [13:09] s/forget/forgot/ [13:11] else [13:11] # Debian armel [13:11] CONFARGS += --with-arch=armv4t --with-float=$(float_abi) [13:11] endif [13:11] doko: ^-- That? [13:14] infinity, it's the default anyway, and 4.4 doesn't know about hard-float yet. it's just 4.4 linaro. so I assume I just back out this change for debian [13:15] doko: Eh? softfp wasn't the default, that's why you had to add it to 4.6... [13:16] (Or, rather, the switch wasn't SET by default) [13:18] doko: --with-float=$(float_abi) only happens in the ifeq Ubuntu block, so no float bits set for debian/armel. [13:18] infinity, Debian armel is soft, not softfp [13:19] Well, whatever it was. :P [13:19] I was just copying your else..endif from 4.6 up there. [13:19] Anyhow. You could just revert and not build on armhf at all. [13:19] Or just make rules2's arm* block match 4.6 [13:21] 4.5 is in the same boat, I imagine. [13:45] infinity, ahh, no -mfloat-abi= parameter passed at all, which breaks the patch :-/ [13:46] I know I dislike this default stuffing of config parameters [13:47] I'm sure we can work up a more elegant patch later. But this "works". [13:49] doko: ghc is building on the buildds now, BTW. [13:50] Local build took forever. Silly haskell. [14:12] hmm [14:12] what i wonder is why did ubiquity 2.9.5 build flawless on armhf ... it shuld have had the same issues as .9.6 [14:12] * ogra_ just noticed there is a binary [14:43] I was using Ubuntu-ARM in my Eee pad transformer, now i get this: http://i.imgur.com/6lY5H.png . How can i restore to Android using OLiFE.sh ? 1. Backup 2. Flash 3. Inject [14:45] * ogra_ would suggest to ask that in a transformer channel ... i doubt anyone here knows OLiFE.sh (apart from lilstevie probably) [14:52] Fixed, with OLiFE.sh > Menu update > 1. Android kernel [14:52] But i want to go back to Ubuntu again and see what caused it? [14:54] IamTrying: sec [14:55] wow [14:55] wtf did you do [14:55] 0.o [14:55] Yea i was also thinking what i did. I actually Ubuntu-ARM but it does not auto boot [14:56] So i needed to press always Vol+ to boot [14:56] yeah, but that framebuffer connection is... just not normal [14:56] Then i decided to re-install Ubuntu-ARM again so that auto it boots [14:56] and yeah, that is normal if android is set as default boot [14:56] ok, still doesn't really answer the question [14:56] if you try running it as the recovery boot does it do the same [14:58] probably the antenna is just mis-adjusted :P [14:58] I used this OLiFE.sh script to restore to Android ( https://gist.github.com/1447187 ) and it caused this > http://i.imgur.com/6lY5H.png [14:58] that isn't what I asked [14:59] ogra_: lol [14:59] ogra_: that is like total fb corruption; I don't understand it in the slightest [14:59] bootloader probably [15:00] ogra_: shouldn't be, bootloader goes with the OLiFE scripts [15:00] purposely done so I don't have to deal with bootloader issues [15:00] isnt the frambuffer initialized by the bootloader on the transformer ? [15:00] like on the ac100 [15:00] yeah, but corruption like that would be kernel level [15:01] or above [15:01] Let me freshly install the Ubunt-ARM using OLiFE.sh. But it does not boot auto. [15:01] maybe something done to xorg [15:01] on ac100 we dont re-initialize the fb ... so such a thing cant happen here [15:01] I did always need to press vol+ to boot. and finally today it was failing to boot. [15:01] do you re-init it in the transformer kernel = [15:01] ? [15:01] no, which is why I don't understand [15:01] yeah, weird [15:02] also IamTrying yes, that means you set ubuntu to be secondary boot [15:02] I did those steps and it fixed it now and i am back to Android PRime [15:02] https://gist.github.com/1447187 [15:02] AKA; it is meant to do that [15:02] as I said last time we had this little conversation [15:04] Yes, when i try to setup for option 2, its very odd, it shows a loop of ASUS Screen and Linux unless i press manually down and then up button. 1. Turn device on or 2. Hold vol-down then press vol-up === Ursinha is now known as Ursinha-lunch [15:12] IamTrying: all I can really suggest is post in the QandA thread on XDA developers, maybe someone has come across your condition before and can help [15:13] lilstevie, Yes sure, OLiFE.sh is excellent its genius, its working actually, saved my Android device. I just want to install the Ubuntu-Arm as power-on default boot, but optionally if requires i want to go back like now i did. [15:14] default boot can be a little flakey tbh [15:14] the mem carveout is not nice [15:15] Ok, still fine, but There was a strange crash today. I pressed vol-down, i got the choice then i press vol-up to boot in my Ubuntu-Arm, but it shows lots of error and then nothing just happened. So then i decided to go back to Android. [15:41] I just re-installed Ubuntu-Arm using OLiFE.sh just like Day 1. And i have the same crash everytime when i boot e.g: http://i.imgur.com/AGSzH.png [15:41] Any idea? [15:53] doko_: I'll fix qtmobility right after my next meeting. === Quintasan_ is now known as Quintasan === Ursinha-lunch is now known as Ursinha