[01:37] superm1: could you reupload ubiquity 2.4.2 without the config.log dropping in the source package, please? (you don't have to bump the version - there can be multiple copies of a package at the same version in the queue) [03:16] cjwatson, sure, not sure how that got in there [09:15] superm1: thanks for the upload [12:37] ubiquity: cjwatson * r4406 ubiquity/debian/ (changelog control ubiquity.install-any): [12:37] ubiquity: pixmaps/cd_in_tray.png has moved back to ubiquity-frontend-gtk. Remove [12:37] ubiquity: it from ubiquity. [13:56] ubiquity: cjwatson * r4407 ubiquity/debian/changelog: releasing version 2.4.3 === ara_ is now known as ara [17:21] ev: is wubi in good shape for RC? [17:30] it was working last week ... [17:30] that's a good sign [17:52] It's nice when things work. [17:57] lol [19:20] cjwatson: got an annoying issue w.r.t. to ubiquity's icon on dove, as in, its not there [19:21] Its getting diverted by oem-config, which is explicately seeded on ARM for pre-installs, which causes my non-preinstall dove images to loose it's ubiquity icon and instead have the finalization icon in its place in System [19:21] cjwatson: ogra says we can't deseed oem-config-gtk because we need it so the right UI gets pulled in [19:21] my first through was to put something in livecd-rootfs, but that's just a hack and I was hoping there was a saner way to solve tihs [20:09] ev: ^ - maybe you know about how best to solve my oem-config issue [20:23] I know this is probly a dumb question.. but is there any way to get *less* email from launchpad... :) [20:25] shtylman_: I recommend email filters, or failing than, finding were the LP devs live and start begging,or failing that, use the source. You also might be able ot get somewhere selling your soul, but the redemption rates for that can be pretty low :-/ [20:30] :( [20:30] I was hoping to be able to keep my soul [20:40] NCommander, what about deseeding oem-config-gtk and preseeding a late_command to touch the file that enables oem config? [20:45] superm1: I don't think that can work for jasper. ogra_cmpc -^ [20:47] NCommander, i dont think that would work [20:48] err, wait [20:48] * NCommander notes that ogra is multipling ... [20:48] that is what we do, the prob is not on our side here but the diversion in the oem-config-gtk package [20:50] superm1, jasper already does what we need wrt enabling oem-config, that works since months ... the prob is that for NCommander the ubiquity ,desktop file vanishes thought the diversion [20:50] ogra_ac: I think oem-config also ends up in the installed image which is annoying [20:50] on dove ? [20:50] er, nm [20:50] no it doesn't [20:51] clean up step does the right thing [20:51] it's definitely removed on the preinstalled inages [20:51] what about undoing the diversion in jasper then? [20:51] right [20:51] ogra_ac: yeah, my brain is fried [20:51] superm1: casper is only used on pre-installed [20:51] er [20:51] jasper [20:51] We're talking bog standard images where we hit this problem [20:51] superm1, NCommander doesnt use jasper (which is the main prob here ) [20:51] oh [20:51] My solution simply to remove oem-config in the CD image step, but thisis obviously a hack [20:51] his image differs from all the opthers [20:51] Which is why I want a better solution hence the consultation with the experts [20:51] so in NCommander's case do what i was saying above - dont include oem-config in the base image [20:52] superm1: well, that would be great, expect we need oem-config-{gtk|kde} seeded :-) [20:52] just let it install from the standard methods and touch the file to enable it in late_command [20:52] superm1, then it is missing for the rest of us too [20:52] oh [20:52] well that's just messy then :) [20:52] the point is, NCommander wants ubiquity [20:52] superm1: yes well, so is our image building :-) [20:52] we dont [20:52] NCommander, so what about doing a casper plugin only used in your instance that cleans this up? [20:53] currently jasper depends on oem-config in the preinstalled images [20:53] superm1: *groan*, that's fugly [20:53] butv to get the right frontend we use the seeds [20:53] Though slightly less fugly [20:53] (which is the proper way imho) [20:53] * NCommander thinks that if he could do it again, he would insist on building alternates vs. live images, lot less pain. [20:53] so kde images get the kde frontend, gtk images get the gtk one [20:53] */2 cents* [20:53] based on seeds [20:54] NCommander, in case you have to build dove in natty i'll be with you [20:54] superm1: I rather fix this before it hits the image if possible. Having to make casper be smart with oem-config going to be frustatingly annoyed [20:54] ogra_ac: \o/. [20:54] ogra_ac: I only built a live image cause there's annoying bugs with building alternate netboot. Given what we've gone through, I'm willing to go fix that instead [20:54] given ther will be no users for that image anyway it doesnt make much sense to invest so much time [20:54] er [20:54] netbook [20:55] ogra_ac: yes, well, that's completely beside the point ;-) [20:55] well when that diversion was added, there wasn't any anticipation of a need to leave the ubiquity icon in place when oem-config was in the image as well, that's for sure [20:55] superm1: yeah, but I don't want the "Prepare and finalize system" icon [20:55] I want oem-config off my image, simply removing the diversion isn't a fix [20:55] right, a diversion removal casper script based on archdetect should do fine here [20:56] NCommander, why ? [20:56] it will never run [20:56] ogra_ac: ugh, that's hidiously specific. My livecd.sh kludge simply checks if we're preinstall, and then apt-get remove's oem-config if we aren't [20:56] and will get removed with ubiquity [20:56] but why ? [20:56] ogra_ac: yes, but while I'm shipping an unused image, I rather it not have fugly unusable icons [20:56] just leave it there and let ubiquity do the job [20:57] so make your script rm it [20:57] ogra_ac: the correct solution is to remove oem-config. That's what I'm going to do, the question is where that will happen, in casper or at image build time [20:57] you are fiddling with the diversion anyway already [20:57] I rather do it at build time because then the correct behavior is selected based on image time, and not voodo [20:57] *voodoo [20:57] sigh [20:58] just leave it where it is, whats so bad about having oem-config installed until ubiquity removes it [20:58] * ogra_ac doesnt really see the point [20:58] ogra_ac: right, I need to do a hack to remove the diversion. I can remove oem-config for the same amount of effort and it doesn't suck as much [20:58] * NCommander decides to wait for cjwatwon's opinion on how to properly fix this. [20:58] yeah, well, then let casper purge it [20:59] ogra_ac: I'd prefer to do it at image build time, but YMMV. Hence waiting for cjwatson [20:59] its an image with no users at all, dont put so much effort in [21:07] ]I'm AFk for a bit [21:13] at this point, I'd recommend a solution that doesn't involve changing ubiquity, because there are way too many semi-conflicting users and I'm scared that we might break something. beyond that, it's 9pm and I'm not sure I have enough brain to work this out [21:15] cjwatson, yeah, i think a csaper script is the least intrusive way atm [21:15] *casper [22:20] ogra_ac: fair enough, casper script it is. I think if we just sense for the squashfs and if we are on ARM is good enough [22:20] well, you have archdetect available [22:21] so you can even do it for dove only [22:22] ogra_ac: I didn't realize we had archdetect in the initrd [22:23] not in th einitrd [22:23] but your script needs a mounted root anyway :) [22:23] so just run your check chrooted [22:35] ogra_ac: heh [22:35] ubiquity: superm1 * r4408 ubiquity/ (debian/changelog src/panel/panel.c): manually update indicators ABI in panel source. [22:36] ubiquity: superm1 * r4409 ubiquity/debian/changelog: releasing version 2.4.4