CIA-3 | ubiquity: cjwatson * r3885 ubiquity/debian/changelog: releasing version 2.1.32 | 00:18 |
---|---|---|
CIA-3 | console-setup: cjwatson * r132 ubuntu/debian/ (changelog config.proto): | 02:03 |
CIA-3 | console-setup: Set default layout for Guernsey, Jersey, and the Isle of Man to gb | 02:03 |
CIA-3 | console-setup: (LP: #533003). | 02:03 |
Santa_Claws | Hi Guys; how do you set the appropriate flags when using the configure scripts included with source software? | 06:59 |
bananeweizen | I have layout issues on page 5 (user account) when running the daily CD image in test drive with German language. Is this a known bug? | 07:45 |
shtylman | does the gtk installer also fail at the language page? | 21:11 |
superm1 | i just ran today's daily mythbuntu w/o troubles a few minutes ago | 21:11 |
superm1 | what's the failure look like? | 21:11 |
shtylman | superm1: http://paste.ubuntu.com/389855/ | 21:12 |
shtylman | maybe I am not running the latest ubiquity... I thought I was... but lemme double check... | 21:13 |
superm1 | the kde_ui is missing support for not_me it looks like | 21:14 |
superm1 | compare Controller.translate in gtk_ui | 21:14 |
shtylman | k.. lemme check that out | 21:14 |
cjwatson | I got partway into fixing Kubuntu ubiquity at the end of the week, but it took me about three hours to figure out why install-only mode was bust and fix that :) hadn't got into the installer proper yet | 21:15 |
shtylman | superm1: yea... kde side is missing a few things... well time to bring it up to speed :) | 21:15 |
cjwatson | it's reputed to need a lot of love at the moment | 21:15 |
shtylman | cjwatson: indeed... hopefuly I have some to give | 21:16 |
cjwatson | please do, I'll chip in start of next week too | 21:16 |
shtylman | cjwatson: cool | 21:18 |
shtylman | cjwatson: I think something of importance we can discuss at UDS ... at least in spare time is documenting this beast and whatnot | 21:18 |
superm1 | cjwatson, mythbuntu daily appears to work properly w/ maybe-ubiquity appended to kernel command line, where exactly do things need to be changed so that it gets a gfxboot with that stuff enabled? | 21:19 |
shtylman | at least think about what we can do to catch code changes better... since we don't have a "compiler" per se | 21:19 |
cjwatson | if we could fix up the partial pychecker stuff that's in there, that would be good | 21:26 |
cjwatson | some of its current reports are bogus so we can't just run it automatically right now | 21:27 |
cjwatson | superm1: debian-cd - I'm happy to do it, just tell me what the first screen should look like | 21:27 |
superm1 | cjwatson, same like how the ubuntu one is doing it is fine (just showing the logo) , thanks | 21:28 |
cjwatson | ok | 21:28 |
cjwatson | on the smartphone right now, will do it later | 21:29 |
shtylman | cjwatson: oooo fancy | 21:31 |
shtylman | is there a problem with manual partitioning at the moment? | 22:24 |
ev | nope | 22:24 |
shtylman | then its a kde side issues :( | 22:24 |
ev | not in the GTK+ frontend, anyway | 22:24 |
shtylman | sadness | 22:24 |
ev | haha | 22:24 |
cjwatson | ok, now kubuntu install-only fails because ibus-daemon isn't running | 22:26 |
cjwatson | I think | 22:26 |
shtylman | heh | 22:26 |
shtylman | cjwatson: did you fix the language page thing? cause I fixed it on my end as well | 22:26 |
shtylman | right now I hit another issue with trying to manually partition | 22:27 |
cjwatson | shtylman: no | 22:27 |
ev | I'm fixing the not_me thing as we speak | 22:27 |
shtylman | cjwatson: ok... was just curious if I would have to resolve conflicts.. no worries then | 22:27 |
shtylman | ev: I fixed it in my copy.. | 22:28 |
shtylman | just havn't commited to trunk | 22:28 |
ev | oh, feel free to commit then | 22:28 |
ev | sure thing | 22:28 |
shtylman | fixed it a few min ago | 22:28 |
ev | good deal | 22:28 |
shtylman | at least added the extra parameter | 22:28 |
shtylman | didn't do anything with it really beyond the one extra if check | 22:28 |
ev | indeed, that's all there is to it | 22:29 |
shtylman | cool | 22:29 |
ev | shtylman: I don't think qt's copy of webkit needs it just yet, but if the slideshow suddenly stops working, you might need to set the property that we enable in the GTK+ frontend. | 22:32 |
ev | enable-file-access-from-file-uris, that is | 22:32 |
shtylman | ev: noted... why was that needed? | 22:32 |
ev | 22:32 | |
ev | context: http://www.gitorious.org/webkit/webkit/commit/624b9463c33adbffa7f6705210384d0d7cf122d6 | 22:33 |
shtylman | interesting | 22:33 |
ev | errr https://bugs.webkit.org/show_bug.cgi?id=34778 | 22:33 |
ubottu | bugs.webkit.org bug 34778 in Platform "Add support to isolate file:/// URIs in their own domain by default" [Normal,Resolved: fixed] | 22:33 |
shtylman | ev: http://paste.ubuntu.com/389893/ | 22:36 |
shtylman | thats what happens when I try manual partitioning in kubuntu | 22:36 |
shtylman | doesn't make much sense to me as I don't know much about the debconf stuff | 22:36 |
ev | ah, that looks like my fault | 22:44 |
* ev digs | 22:44 | |
cjwatson | if you run ubiquity in debug mode, it generally includes a debconf trace which is a lot easier to use to debug that kind of thing | 22:45 |
* shtylman forgives ev ;) | 22:45 | |
cjwatson | METAGET <question> description is in general the right number of arguments, so I suspect that question is an empty string here or something | 22:45 |
ev | debconf, however, is not so forgiving | 22:45 |
shtylman | how does ubiquity talk to debconf? | 22:45 |
shtylman | c api type deal? | 22:45 |
shtylman | dbus? | 22:45 |
cjwatson | no, debconf uses a stdin/stdout protocol | 22:45 |
cjwatson | man debconf-devel | 22:45 |
ev | (and see ubiquity/debconffilter.py) | 22:46 |
cjwatson | write down a file descriptor, read answer back from another file descriptor | 22:46 |
cjwatson | there's a bit about this in doc/README, actually | 22:46 |
shtylman | I see | 22:46 |
shtylman | and debconf does all the work behind the scenes for us? and just tells us what to expect/do next? | 22:47 |
shtylman | it also tells us the capabilities of the system? | 22:48 |
shtylman | like partition options... | 22:48 |
ev | ubiquity communicates with the d-i components via debconf | 22:48 |
cjwatson | kind of ... the code that uses debconf is basically what you see in d-i | 22:48 |
shtylman | I see | 22:48 |
shtylman | ev: just so I am clear... the only thing to add to the frontend to support this new immediate ubiquity mode is a button (or combo) to let the user choose to try or install ... correct? | 23:05 |
ev | not exactly | 23:06 |
ev | you might want to do a bzr diff -c on my changes to ubi-language.py | 23:06 |
shtylman | k | 23:06 |
shtylman | will do that | 23:07 |
ev | you need to hide some things when not in greeter mode | 23:07 |
shtylman | makes sense | 23:07 |
ev | and substitute the MEDIUM and RELEASE variables | 23:07 |
shtylman | k | 23:08 |
shtylman | is there a test mode for this? | 23:08 |
shtylman | like a flag to ubiquity by any chance? | 23:09 |
shtylman | --greeter :) | 23:09 |
ev | yup :) | 23:12 |
CIA-3 | ubiquity: cjwatson * r3886 ubiquity/ (bin/ubiquity-dm debian/changelog): Wait for kwriteconfig to exit before proceeding. | 23:14 |
CIA-3 | ubiquity: cjwatson * r3887 ubiquity/ (bin/ubiquity-dm debian/changelog debian/control): | 23:16 |
CIA-3 | ubiquity: Start ibus-daemon when starting KDE in install-only mode, which should | 23:16 |
CIA-3 | ubiquity: make a few things a bit happier. | 23:16 |
CIA-3 | ubiquity: cjwatson * r3888 ubiquity/ (debian/changelog ubiquity/components/ubi-partman.py): | 23:17 |
CIA-3 | ubiquity: Derive partman component's PageNoninteractive class from PageBase rather | 23:17 |
CIA-3 | ubiquity: than PluginUI, so that it has all the required methods. | 23:17 |
CIA-3 | ubiquity: evand * r3889 ubiquity/ (debian/changelog ubiquity/components/ubi-partman.py): | 23:19 |
CIA-3 | ubiquity: * Fix call to self.progress_start in ubi-partman. | 23:19 |
CIA-3 | ubiquity: * Also display progress titles on the advanced partitioning page. | 23:19 |
ev | shtylman: ^ give that a try, I think it should fix the bug you're hitting | 23:19 |
cjwatson | I have no idea why I ran into the problem in r3888; I was booting Kubuntu, it shouldn't have gone near noninteractive | 23:19 |
ev | ubiquity seems to have a habit of falling into the noninteractive frontend from, presumably, ubiquity-dm, when it fails | 23:20 |
ev | despite the changes we've made to stop that from happening | 23:21 |
shtylman | ev: will do :) | 23:21 |
cjwatson | the odd thing is that it didn't write anything to the debug log when it failed the first time | 23:22 |
cjwatson | oh, although there are two startup lines | 23:22 |
cjwatson | ah, if ubiquity-dm itself fails, then we'll end up in noninteractive | 23:25 |
cjwatson | you know, we should send tracebacks from ubiquity-dm somewhere | 23:26 |
cjwatson | its failure modes are horrendous to debug | 23:26 |
cjwatson | I wonder why I didn't just make it send all its output to /var/log/installer/dm | 23:26 |
shtylman | cjwatson: ev: http://paste.ubuntu.com/389925/ | 23:28 |
shtylman | thats what happens for manual mode now | 23:28 |
ev | bah, on it | 23:28 |
shtylman | :) | 23:28 |
ev | I suspect this is basically the same thing I just fixed, only with progress_stop | 23:29 |
cjwatson | the *args there looks like a bit of a case of believing pychecker | 23:31 |
cjwatson | I should disable the really stupid warnings across the board maybe ... | 23:31 |
cjwatson | pretty sure there's at least one that lies too often to be worthwhile | 23:31 |
CIA-3 | ubiquity: cjwatson * r3890 ubiquity/bin/ubiquity-dm: try to keep to 80 columns a bit more | 23:32 |
CIA-3 | ubiquity: evand * r3891 ubiquity/ (debian/changelog ubiquity/components/ubi-partman.py): Fix call to self.progress_stop in ubi-partman. | 23:34 |
cjwatson | ev: why not just def progress_stop(self):? | 23:35 |
cjwatson | a dummy arg there seems very strange given that PROGRESS STOP does not take any further arguments in debconf | 23:36 |
CIA-3 | ubiquity: cjwatson * r3892 ubiquity/ (bin/ubiquity-dm debian/control): revert r3887, clearly wrong as ibus is not in the kubuntu-desktop task | 23:36 |
ev | ah, I wasn't sure why we were passing an argument in the first place | 23:38 |
CIA-3 | ubiquity: cjwatson * r3893 ubiquity/ (bin/ubiquity-dm debian/changelog): | 23:43 |
CIA-3 | ubiquity: Stop starting gdm and kdm by hand on exit from /var/log/installer/dm | 23:43 |
CIA-3 | ubiquity: (using /etc/init.d/ calls, no less). The Upstart jobs are written so | 23:43 |
CIA-3 | ubiquity: that this happens automatically if necessary. | 23:43 |
CIA-3 | ubiquity: evand * r3894 ubiquity/ (4 files in 3 dirs): Don't accept unnecessary arguments to progress_stop anymore. | 23:43 |
shtylman | ev: from what I can see the pre-release warning is wrapped up into the language page now? correct? | 23:46 |
cjwatson | ev: my fault, by the looks of things; it goes all the way back to r596 | 23:47 |
CIA-3 | ubiquity: superm1 * r3895 ubiquity/ (debian/changelog ubiquity/frontend/gtk_ui.py): Fix wrong page to jump back to when partitioning fails during install. | 23:47 |
ev | cjwatson: lets call it even, given the /etc/init.d/gdm start ;) | 23:47 |
ev | 'night all | 23:48 |
shtylman | peace | 23:48 |
cjwatson | ev: heh | 23:48 |
ev | superm1: thanks for fixing that. I had it on my todo list after I saw the error in the log you posted the other day | 23:49 |
cjwatson | ev: your code was at least correct at the time ;-) | 23:49 |
ev | haha | 23:49 |
superm1 | ev, np, i just remembered it from the other day too | 23:51 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!