| ev | ugh, I'm getting a UnicodeDecodeError, but somehow only when the code in question is run inside ubiquity | 09:51 |
|---|---|---|
| cjwatson | ev: it's too late for 10.04.3, but could you do an SRU for bug 606134, please? | 09:51 |
| ubot2 | Launchpad bug 606134 in ubiquity-slideshow-ubuntu "ubiquity-slideshow-ubuntu package has outdated translations in Ubuntu 10.04.1" [High,In progress] https://launchpad.net/bugs/606134 | 09:51 |
| cjwatson | you assigned it to yourself last year | 09:51 |
| ev | for 10.04.4? | 09:51 |
| cjwatson | yeah | 09:52 |
| ev | okay, will do | 09:52 |
| cjwatson | thanks | 09:52 |
| ev | sure thing | 09:52 |
| * cjwatson is moving unfixed 10.04.3 bugs to 10.04.4 and noticed it | 09:52 | |
| ev | current fun: http://paste.ubuntu.com/647180/ and http://paste.ubuntu.com/647181/ | 10:40 |
| cjwatson | v.decode('utf-8') ? | 10:50 |
| cjwatson | maybe , 'ignore' | 10:51 |
| ev | sure, v.decode works, but I'm just baffled as to how there is any difference in behavior at all between the two programs | 10:52 |
| ev | that is, what on earth is making pygtk work there :) | 10:52 |
| cjwatson | yeah, I can't see explicit decoding happening in any relevant part of pygtk | 10:58 |
| cjwatson | but could be looking in the wrong place | 10:58 |
| cjwatson | woo. first stage of multiarch Packages/Translations enhancements to LP landing ... | 11:04 |
| bdmurray | cjwatson: I was looking at critical foundations bugs and discovered bug 562706 and was trying to figure what was wrong and if it is still relevant | 15:00 |
| ubot2 | Launchpad bug 562706 in tasksel "On i386, selecting the ubuntu studio packages fails to install" [Critical,Confirmed] https://launchpad.net/bugs/562706 | 15:00 |
| cjwatson | Launchpad doesn't know about ubuntustudio | 15:04 |
| cjwatson | we probably ought to fix that | 15:05 |
| bdmurray | cjwatson: could elaborate a wee bit? | 15:06 |
| cjwatson | done in the bug | 15:06 |
| cjwatson | pushing an LP branch for it now | 15:09 |
| bdmurray | cjwatson: got it thanks | 15:09 |
| bdmurray | additionally yesterday I'd mentioned memtest and redirecting some bugs to grub | 15:10 |
| cjwatson | I saw and replied | 15:10 |
| bdmurray | I had an apport branch merged for this but was wondering if there were more checks to be made | 15:10 |
| cjwatson | ("yes please" or words to that effect) | 15:10 |
| bdmurray | http://bazaar.launchpad.net/~ubuntu-core-dev/ubuntu/oneiric/apport/ubuntu/revision/1799 | 15:10 |
| bdmurray | specifically the 'grub_errors' list | 15:10 |
| cjwatson | maybe postrm.d as well as postinst.d | 15:13 |
| cjwatson | I suspect for other things you know better than me what's showing up in bugs :-) | 15:13 |
| cjwatson | you could probably ditch "error" from the last one; anything grub-probe whines about is likely to be a grub2 bug | 15:14 |
| bdmurray | sounds good thanks | 15:14 |
| cjwatson | oh, and also, the grub-probe case should be a grub2 bug even if update-grub is provided by grub | 15:14 |
| CIA-37 | ubiquity: evand * r4775 pygi/ (gui/gtk/stepLanguage.ui ubiquity/plugins/ubi-language.py): | 15:16 |
| CIA-37 | ubiquity: * Port Language GtkBuilder file to GTK+3. | 15:16 |
| CIA-37 | ubiquity: * Move language page sizing hacks to GTK+3. | 15:16 |
| CIA-37 | ubiquity: * Fix unicode handling in ubi-language. | 15:16 |
| CIA-37 | ubiquity: evand * r4776 pygi/ubiquity/frontend/gtk_components/keyboard_query.py: The resizable propery now sets whether a window's size can be modified. | 15:18 |
| bdmurray | looking at existing bugs I don't see any with postrm.d and zz-update-grub | 15:22 |
| cjwatson | possibly not - it is possible though | 15:23 |
| bdmurray | okay, its easy enough to add | 15:23 |
| cjwatson | as both scripts exist | 15:23 |
| kyleN | Hi, I have a question about oem-config. specifically, even though I have fully preseeded my custom image to zh_CN, and oem-config does launch in zh_CN, three of the four notification applets are still in English, even though LANGUAGE is zh_CN locale is same, and zh-hans language packs are installed. | 16:03 |
| kyleN | after oem-config runs, the applets display in zh_CN. so somehow they are not picking up the locale until oem-config is completed. | 16:03 |
| kyleN | ibus notification applet IS in zh_CN | 16:04 |
| kyleN | network-manager-applet, sound, and session are the three applets strangely in English | 16:05 |
| superm1 | the panel and those first three applets get started before zh_CN is applied to the environment I think | 16:07 |
| kyleN | superm1, yes | 16:08 |
| superm1 | ibus applet gets started later on after the U/I begins to come up | 16:08 |
| superm1 | and afte the preseed has been picked up | 16:08 |
| kyleN | superm1, so how do I preseed the panel to use my locale of choice? | 16:08 |
| superm1 | well i don't think there is a way right now | 16:09 |
| superm1 | a potential solution might be having ubiquity-dm try to read the the preseeded locale and set it up before it starts up the panel | 16:09 |
| kyleN | superm1, I don't think I need the panel. can I hide it? | 16:09 |
| superm1 | sure, just dpkg-divert it out of the way and it won't cause a problem | 16:09 |
| kyleN | superm1,I need the panel later, just not now. is there a gconf or other setting to hide it? i can reset it to show perhaps on first user login | 16:10 |
| superm1 | the ubiquity panel is not the same as the unity panel | 16:10 |
| superm1 | /usr/lib/ubiquity/panel is the ubiquity panel | 16:10 |
| kyleN | ah | 16:11 |
| kyleN | ok, that's a good lead superm1. thx | 16:11 |
| superm1 | sure. | 16:12 |
| kyleN | superm1, this looks like a panel bug to me. do you agree? | 16:14 |
| superm1 | it's a deficiency yes | 16:15 |
| superm1 | since it's spawned outside of the ubiquity process i'm not sure there is an easy interface to update the environment of those widgets | 16:15 |
| kyleN | ok. I'll look into filing a deficiency :) | 16:15 |
| superm1 | :) | 16:15 |
| kyleN | superm1, https://bugs.launchpad.net/ubuntu/+source/ubiquity/+bug/812988 | 16:33 |
| ubot2 | Ubuntu bug 812988 in ubiquity "panel does not pick up preseeded locale/translation" [Undecided,New] | 16:33 |
| superm1 | kyleN, does this fix the issue maybe? http://paste.ubuntu.com/647435/ | 16:40 |
| kyleN | superm1, I will check that after lunch and let you know. and thanks! | 16:42 |
| superm1 | ok cool, if not, then it's probably more involved and ev might have some ideas | 16:43 |
| === cr3_ is now known as cr3 | ||
| cornwall | Hi, is this the place for discussing and asking design ideas? | 17:02 |
| charlie-tca | That is usually #ayatana, but it might be a closed channel | 17:07 |
| kyleN | superm1, I added that text to /etc/oem-config, rebooted and it worked :) Except that ibus no longer shows in panel. | 17:11 |
| kyleN | sorry, to /etc/init/oem-config | 17:11 |
| superm1 | what triggers ibus launching? | 17:11 |
| kyleN | superm1, that may be sunspots and will do a more proper test | 17:11 |
| kyleN | superm1, I will try diverting the upstart file and install this one via another pkg. I'll reinstall the image, drop to tty before oem-config, install my pkg, reboot, and see what happens | 17:13 |
| superm1 | ok | 17:15 |
| kyleN | superm1, as to what triggers ibus. there are xinput variables: /etc/X11/xinit/xinput.d/ibus | 17:19 |
| cornwall | charlie-tca: I was asking that regarding ubiquity | 17:26 |
| kyleN | superm1, now that I think of it, since this is a conf file, i probably do not even need to divert, just install into place, eh? | 17:28 |
| ev | \o/ http://cdimage.ubuntu.com/wubi/current/ | 17:32 |
| ev | cjwatson: I added a bin/cron.wubi for that. Hacking it into bin/cron.daily-live,build-image-set,etc seemed like I'd be mostly #if 0'ing it for a very small bit of code, so I hope that's okay. | 17:36 |
| ev | I can refactor if it's going to be an issue | 17:36 |
| === davmor2_ is now known as davmor2 | ||
| kyleN | superm1, i was wrong about the divert not being necessary :) | 17:42 |
| stefanw_ | i did a clean re-install of 10.04 and now I am getting ADDRCONF(NETDEV_UP): eth1: link is not ready | 17:58 |
| stefanw_ | think it has something to do with the fact that we declined DHCP on install and am trying to get Static to work | 17:58 |
| james_w | hi, I'm getting "checksum mismatches" trying to write both natty and oneiric usb with usb-creator, any tips on debugging why? | 18:30 |
| james_w | ah, I should try oneiric directly too | 18:32 |
| james_w | that seemed to work better, it's booting the live image now | 18:38 |
| james_w | I'm happy to debug if that's desired | 18:38 |
| james_w | oh, this is a natty host by the way | 18:39 |
| superm1 | kyleN, hmm so any luck then after a full test with divert? | 18:39 |
| kyleN | superm1, it worked, but still I have a minor panel issue. no ibus and theme seems wrong. I am going for a more proper test though: install package at image build time rather than interrupting oem-config > tty > install pkg with dpkg > reboot | 18:40 |
| superm1 | the wrong theme tends to happen when changing languages previously too | 18:40 |
| kyleN | superm1, the wrong theme is that the panel is white, except where the notifications are, which is black | 18:40 |
| superm1 | that happened while ibus spawned previously i thought | 18:41 |
| superm1 | is ibus maybe spawning and then crashing or so? | 18:41 |
| kyleN | possibly. ibus-daemon certainly should spawn | 18:41 |
| kyleN | superm1, maybe ibus-daemon is not even being spawned properly any more. how can I tell if it ever ran? | 18:44 |
| superm1 | i'm not sure really | 18:45 |
| superm1 | does it log to syslog or anything? | 18:45 |
| kyleN | i've looked and have not found | 18:45 |
| kyleN | superm1, the language-selector's ImSwitch.py module does it, for one | 18:54 |
| kyleN | superm1, via the im-switch command | 18:55 |
| kyleN | superm1, and ubiquity does it via (src pkg): ubiquity/im_switch.py | 18:56 |
| superm1 | kyleN, ah right so that's exactly it. the LANG in os.environ matches what was preseeded so the im_switch.start_im() wouldn't get called | 19:01 |
| kyleN | aha | 19:01 |
| kyleN | superm1, so to get back to the start, why is the preseeded locale not being picked up? | 19:02 |
| superm1 | ubiquity-dm spawns the X session, followed by the panel, followed by the ubiquity process | 19:02 |
| kyleN | I suppose one could put a call to debconf into the code to get it | 19:02 |
| superm1 | the ubiquity process is what reads the preseeded locale | 19:02 |
| superm1 | so LANG hasn't been set by the time the panel is spawned. | 19:03 |
| kyleN | right | 19:03 |
| kyleN | can ubiquity-dm read the preseed? | 19:03 |
| superm1 | yes, but you achieved the same thing by reading /etc/default/locale in the upstart job | 19:04 |
| kyleN | well, except it doesn't quite work yet. | 19:04 |
| superm1 | I guess another way to solve it would be to unset LANG from the environment before ubiquity itself starts then | 19:04 |
| superm1 | that or modify the logic in ubiquity/i18n.py to start_im() on more situations | 19:04 |
| superm1 | ev, have you followed all above? which way would you lean? | 19:05 |
| kyleN | superm1, before that, can you check this: http://pastebin.ubuntu.com/647540/ | 19:05 |
| kyleN | there is an extra line in there that you didn't have but oem-config.conf did | 19:06 |
| kyleN | >/var/run/oem-config.upstarted | 19:06 |
| kyleN | seems find but I don't quite get a script line starting with ">" | 19:06 |
| kyleN | fine | 19:06 |
| superm1 | in oneiric that line is gone because of the /run transition | 19:06 |
| superm1 | i made the patch from trunk | 19:06 |
| kyleN | so you think it is not a problem? | 19:06 |
| superm1 | not related to this no | 19:07 |
| kyleN | superm1, I don't think we want to modify ubiquity to start_im() on more situations. that is, IM should only start for proper langs | 19:08 |
| kyleN | I would prefer a conf file solution so i don't need to patch ubiquity. | 19:09 |
| superm1 | well i think you'll need to patch ubiquity until this is properly fixed, i don't see the whole thing being fixable in files in /etc | 19:09 |
| kyleN | ok. not a big problem | 19:09 |
| kyleN | in this case, it is for a Chinese-only image. so if I could modify ubiquity to always start IM, that would not be a problem | 19:10 |
| superm1 | then ubiquity/i18n.py is where start_im() is spawned it looks, so just change that logic | 19:11 |
| kyleN | still, there would be the theme issue | 19:11 |
| kyleN | the panel theme | 19:11 |
| superm1 | that's been around a while, i don't think there is a solution for it yet | 19:11 |
| superm1 | maybe this whole thing would just be easier in your situation if you just didn't use the panel during oem-config | 19:12 |
| kyleN | i'd rather have the panel theme slightly wrong and have IM | 19:12 |
| kyleN | that is a possibility | 19:12 |
| kyleN | however it removes session. so user cannot reboot | 19:12 |
| kyleN | and network admin | 19:13 |
| superm1 | should the IM start on basically any non en_US language? | 19:14 |
| superm1 | it seems to me that calling start_im on any non en_US language should be safe | 19:17 |
| kyleN | in this image, there is only english and chinese/S. and english is really just for testing. so I would be willing to always start IM | 19:18 |
| kyleN | in the more general case (ubuntu), I don't think IM should start for all non-english langs | 19:19 |
| superm1 | well i'm talking more for the general case | 19:19 |
| kyleN | so im is simply not appropriate for non-english. | 19:19 |
| kyleN | for *all* non-english | 19:19 |
| superm1 | it looks like there is some logic to block it from starting in im_switch.py except in certain cases | 19:20 |
| superm1 | so for your case go ahead and just start it by modifying that logic, for the general case this will have to be investigated a little bit more | 19:20 |
| kyleN | superm1, first it is killed, then started if conditions merit | 19:21 |
| superm1 | it looks like it hunts for a conf file for LANG in /etc/X11/xinit/xinput.d first | 19:22 |
| superm1 | if it doesn't find one it won't spawn | 19:22 |
| kyleN | i will see if that conf file exists at the right time. If not, I can add one via another config pkg | 19:24 |
| superm1 | so i think this would be the solution then: http://paste.ubuntu.com/647559/ | 19:24 |
| kyleN | before looking at your suggestion, there are conf files in /etc/X11/xinit/xinput.d/ zh_CN amoung others. there is no "en" though. | 19:27 |
| superm1 | yeah it looks like it won't spawn on any language unless it's in that directory | 19:29 |
| kyleN | perhaps it is ok to hide the panel :) | 19:34 |
| kyleN | now I am thinking i may not need the IM during oem-config | 19:35 |
| kyleN | so the panel and apps are translated during oem config. but no IM. IM can only be used to enter Name anyway. this may not be required for Chinese users. If so, this would be a viable solutoin | 19:36 |
| kyleN | (panel and apps translated via your first suggestion, superm1, and thank you very much for that) | 19:36 |
| superm1 | ok sure no problem | 19:37 |
| superm1 | i'll add my proposed diff to that bug and wait for feedback from ev or cjwatson on which way they lean | 19:38 |
| kyleN | superm1, I am not sure why when I export LANG and LOCALE per your suggestion, then IM doesn't launch, whereas it did launch previously. Do you know why? | 19:47 |
| superm1 | yeah, it's because of that logic in ubiquity/i18n.py where it checks if LANG was previously set | 19:47 |
| superm1 | and that's why the diff in the second part of my suggestion should resolve that potentially | 19:48 |
| kyleN | ok. I'll check that. | 19:48 |
| kyleN | superm1, by modifying i18n.py, i can have my cake and pie: im starts, panel applets in chinse. | 20:06 |
| === kentb_ is now known as kentb-out | ||
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!