| === persia` is now known as persia | ||
| eggonlea | hi mobile team, as we knew, Karmic gave up clutter and netbook-launcher on ARM because of lacking powerful OpenGL (ES) support on ARM platforms. While this has improved much now, will we have any fancy GUI on Lucid ARM port? | 08:36 |
|---|---|---|
| eggonlea | 1) clutter based netbook-launcher | 08:37 |
| eggonlea | 2) OpenGL ES based compiz | 08:37 |
| eggonlea | 3) EFI based 2D launcher | 08:37 |
| eggonlea | which one would be chosen in Lucid? Or just keep gnome based ubuntu-desktop? | 08:37 |
| lool | eggonlea: I can't speak for the mobile team, but I understand the focus is to produce an EFL Ubuntu Netbook image | 09:26 |
| lool | so 2D acceleration only | 09:26 |
| eggonlea | lool: thanks. but till now we didn't see this EFI 2D GUI in livecd. | 09:32 |
| eggonlea | maybe I miss something, or it should be there already since alpha3 is coming. | 09:33 |
| lool | eggonlea: I didn't try the images at http://cdimage.ubuntu.com/ubuntu-netbook/ports/daily-live/current/ | 09:33 |
| lool | eggonlea: But the manifest lists the packages in there | 09:33 |
| lool | netbook-launcher-efl 0.2.2-0ubuntu2 | 09:33 |
| eggonlea | I mean, the default GUI is still gnome desktop after installation. | 09:34 |
| lool | Is it? odd | 09:35 |
| lool | asac: ^ Is it a known bug perhaps? Or perhaps there's an easy work around? | 09:35 |
| eggonlea | y. let me double check the latest livecd. | 09:36 |
| * eggonlea is reinstalling the latest livecd on the board. | 10:24 | |
| asac | latest live images should have the netbook launcher | 10:25 |
| asac | problem is that shortly after we switched to ue images mono breakage killed our images | 10:25 |
| asac | so for a week or so there were no images at all afaict | 10:25 |
| asac | eggonlea: ^^ | 10:26 |
| eggonlea | aha, I see | 10:27 |
| eggonlea | I was wondering why the daily zsync did nothing. | 10:27 |
| asac | yeah. but seems we have images today again \o/ | 10:29 |
| asac | seems we recovered from mono probs ;) | 10:29 |
| lool | asac: Well done on mono | 10:30 |
| lool | I wonder whether it actually works | 10:30 |
| eggonlea | good news indeed. congratulations~. but I'm unlucky to try it for some days because of our LONG Chinese new year holiday. | 10:30 |
| asac | lool: i would think its similar flaky | 10:31 |
| asac | as in karmic | 10:31 |
| asac | in karmic 40 tests failed | 10:31 |
| eggonlea | Hopefully I can catch it tonight. Do you know how long I should wait for it? | 10:31 |
| asac | now 36 | 10:31 |
| eggonlea | I mean, how many hours does it take to generate the coming livecd? | 10:32 |
| asac | you mean the next run? | 10:33 |
| asac | hmm good question | 10:33 |
| asac | ogra: persia: when is the image run? | 10:33 |
| asac | so looking at the timestamp it will be in 18hours or so | 10:34 |
| asac | assuming its done once in 24h | 10:34 |
| asac | eggonlea: todays image isnt good enough? | 10:35 |
| asac | imo that would be ok | 10:35 |
| asac | if not let me know whats broken | 10:36 |
| lool | ARCHES='armel+dove' buildlive ubuntu-netbook && ARCHES='armel+dove' for-project ubuntu-netbook cron.ports_daily-live | 10:38 |
| lool | That's at 1:55 am london time, but after the imx51 build | 10:38 |
| lool | Whoever wrote this line for it wrong though | 10:38 |
| eggonlea | asac: installing... some minutes left. | 10:39 |
| lool | StevenK, ogra: You might want to run a single cron.ports_daily-live; at least that's what I was told should be done back when I added the first desktop armel images | 10:39 |
| lool | StevenK, ogra: see e.g. kubuntu-netbook | 10:41 |
| ogra | lool, i thought StevenK discussed it with slangasek at the sprint when he changed ti | 10:44 |
| ogra | *it | 10:44 |
| ogra | effectively it doesnt do any harm atm though, we dropped a lot of flavours | 10:45 |
| ogra | (and still hope for an additional livefs builder) | 10:45 |
| lool | I think it will cause two revs of the ubuntu-netbook/ports dir | 10:46 |
| ogra | oh, right, i see what you mean | 10:47 |
| ogra | for-project ubuntu-netbook cron.ports_daily-live shouldnt be there twice, i'll talk to StevenK | 10:47 |
| lool | There should be a single for-project ubuntu-netbook cron.ports_daily-live in the whole crontab basically | 10:47 |
| ogra | yeah, that would make sense | 10:48 |
| ogra | but getting the timing right might be problematic | 10:48 |
| ogra | especially for ports | 10:48 |
| ogra | will require some measuring | 10:48 |
| lool | It shouldn't be | 10:48 |
| ogra | well, i have no idea how long a livefs build for sparc, ppc or ia64 takes | 10:49 |
| lool | There currently is only one line whether this occurs | 10:49 |
| ogra | i know hard numbers for armel only | 10:49 |
| lool | It doesn't matter | 10:49 |
| lool | Just change ARCHES='armel+imx51' buildlive ubuntu-netbook && ARCHES='armel+imx51' for-project ubuntu-netbook cron.ports_daily-live ; ARCHES='armel+dove' buildlive ubuntu-netbook && ARCHES='armel+dove' for-project ubuntu-netbook cron.ports_daily-live to ARCHES='armel+imx51' buildlive ubuntu-netbook; ARCHES='armel+dove' buildlive ubuntu-netbook && ARCHES='armel+imx51 armel+dove' for-project ubuntu-netbook cron.ports_daily-live | 10:50 |
| lool | Well actually s/&&/;/ in the new version | 10:50 |
| ogra | no | 10:50 |
| ogra | then dove wont build if imx fails | 10:50 |
| ogra | the && was explicitly added | 10:50 |
| ogra | we had some runs where there were kernel package issues which then teared down all builds | 10:51 |
| lool | You're confused, there's no && between imx51 and dove | 10:51 |
| ogra | there should be :) | 10:51 |
| lool | I disagree | 10:51 |
| ogra | between the two buildlive runs | 10:51 |
| lool | No | 10:51 |
| ogra | but then the second one wont be attempted if the first one fails | 10:51 |
| lool | It's not what we ever did and it's not what buildlive does | 10:51 |
| ogra | it was changed a while ago | 10:52 |
| lool | Precisely, the point is that you do want to try dove even if buildlive imx51 failed | 10:52 |
| ogra | after i had a long discussion with slangasek how to make dove still build if imx51 fails | 10:52 |
| lool | This is currently the case | 10:52 |
| ogra | we want to | 10:52 |
| lool | and would still be after my changes | 10:52 |
| ogra | the ; between the commands makes cron stop the process if the first one exits non zero | 10:53 |
| lool | No | 10:53 |
| ogra | well, thats what happened several times | 10:53 |
| ogra | ; is fine between the buildlive runs and for-project ... since we dont want the latter if the first failed but we want bot buildlive runs to happen inependently so one subarch doesnt block the other | 10:55 |
| StevenK | lool: I've thought about all this | 11:21 |
| StevenK | lool: I can't come up with a shell snippet to do what I want | 11:21 |
| StevenK | lool: I don't like your suggestion either | 11:23 |
| ogra | StevenK, ARCHES='armel+imx51' buildlive ubuntu-netbook && ARCHES='armel+dove' buildlive ubuntu-netbook ; ARCHES='armel+imx51 armel+dove' for-project ubuntu-netbook cron.ports_daily-live | 11:23 |
| ogra | that should be a good start | 11:23 |
| StevenK | ogra: And then imx51 fails and dove doesn't build | 11:23 |
| ogra | he is right in saying we dont want cron.ports_daily-live twice | 11:23 |
| ogra | no | 11:23 |
| ogra | the && should avoid that | 11:23 |
| StevenK | ogra: If imx51 fails buildlive, dove's buildlive won't get attempted | 11:24 |
| ogra | if you use ; it wont, if you use && it will | 11:24 |
| ogra | at least thats what slangasek worked out | 11:24 |
| StevenK | Yes, and read your shell snippet. imx51 buildlive && dove buildlive ; imx51 dove publish | 11:25 |
| ogra | oh, you mean we need another && | 11:26 |
| ogra | right, that might be | 11:26 |
| persia | Um, No. We *don't* want && because that means "only do this if the previous succeeded" | 11:28 |
| StevenK | Yes, we don't want "imx51 buildlive && dove buildlive && imx51 dove publish" | 11:29 |
| persia | Right. | 11:30 |
| persia | If we trusted things to work, we could do that, but we don't. | 11:30 |
| ogra | well we dont want "imx51 buildlive; dove buildlive; imx51 dove publish" either | 11:31 |
| ogra | that has proven to not work often enough | 11:31 |
| StevenK | "I wonder if (imx51 buildlive ; dove buildlive) && imx51 dove publish" would work | 11:31 |
| StevenK | Moving the " to the right place, of course | 11:31 |
| ogra | hmm | 11:32 |
| persia | No. | 11:33 |
| persia | Or at least `(true; false) && echo works` doesn't do what I'd like. | 11:33 |
| persia | Seems () returns the last return code from the series of commands in the subshell. | 11:34 |
| StevenK | Bah | 11:34 |
| persia | What we need is something like || that isn't optimised | 11:38 |
| ogra | StevenK, how about doing all three of them in separate entries at different times ? | 11:57 |
| ogra | (with the necessary gaps indeed) | 11:57 |
| persia | Why is that better than using ;; to do them sequentially? | 12:01 |
| persia | Putting in gaps creates a race condition | 12:01 |
| ogra | not if the gaps are big enough ... putting in ; obviously generates probelms | 12:02 |
| persia | StevenK: `export DRESS="frumpy"; do-this && export DRESS="success"; do-that && export DRESS="success"; [ "$DRESS" == "success" ] && do-it-all` | 12:22 |
| persia | (and yes, this is incredibly ugly | 12:22 |
| StevenK | Hah | 12:26 |
| persia | But now that I've typed that, I've actually started to process backscroll, and realise I'm kinda slow and that better solutions exist. | 12:27 |
| persia | (like just fixing buildlive to not fail on first failure if it does) | 12:27 |
| lool | StevenK: I see that got resolved on #ubuntu-devel since (shell snippet) | 12:53 |
| persia | I think "resolved" isn't quite the right word, but here is no longer the right forum :) | 13:01 |
| lool | Well cjwatson shared a shell snippet to the effect of what StevenK was asking | 13:02 |
| StevenK | Which is ugly as sin | 13:02 |
| lool | You could put that in a script | 13:02 |
| lool | That might even be in moreutils ;-) | 13:02 |
| * StevenK waves his arms | 13:02 | |
| lool | "either_of_do" | 13:03 |
| StevenK | I don't care, I did slangasek suggested, and I'm going to bed | 13:03 |
| persia | lool: It's a rare enough case not to matter. | 13:03 |
| lool | I'm just replying to uglyness in crontab | 13:04 |
| lool | (I agree crontabs are not a proper place to write this kind of shell snippet) | 13:05 |
| persia | heh. context leak :) | 13:06 |
| lamalex | hi everyone, is there a list of windows that dont fit on netbooks that need fixing/have been fixed? | 23:26 |
| persia | lamalex: I don't know of any list of these, but there's been lots of fixes in that area. | 23:32 |
| persia | Some of the 576 veritical pixel machines seem to still have lots of issues, but those aren't getting much sympathy from most deelopers. | 23:32 |
| persia | But there's still some stuff that doesn't fit in 600 pixels, although I don't think anything left in the default images. | 23:33 |
| lamalex | thanks | 23:44 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!