[12:04] is ubuntu in debconf high or critical? [12:07] srbaker: critical in the installer, high in the installed system [12:07] cool [01:00] hi all. [01:02] hrm.. no jdub, eh? === srbaker [~srbaker@blk-224-143-227.eastlink.ca] has left #ubuntu-devel ["Leaving"] === mdz [~mdz@69-167-148-207.vnnyca.adelphia.net] has joined #ubuntu-devel === wartylog [~warthylog@port1845.ds1-khk.adsl.cybercity.dk] has joined #ubuntu-devel === Topic for #ubuntu-devel: Ubuntu development -- general discussion on #ubuntu === Topic (#ubuntu-devel): set by mdz at Mon Sep 20 07:48:40 2004 [01:54] mdz: can we sync gftp with debian? -6 fixes RC bug for both us and them; the fix is in incoming; our bug # is 1806 [01:54] (we have -5 already) [01:54] Mithrandir: just saw that [01:54] yes, fine with me [01:54] please request it [01:55] willdo === sivang [~user1@80.179.82.108.forward.012.net.il] has joined #ubuntu-devel [01:55] night all [01:56] has anyone bumped into _very__very_ slow disk performance under inspiron 8200 laptop? I recently install a daily from about 4 days ago, and it just crewles.. [01:56] Is this more appropriate at #ubuntu ? :) === Mithrandir hopes he got the long addresses right and goes to bed. [01:58] night Mithrandir === lamont introduces his daughter to xlife. this is fun. [02:18] : [02:18] :) [02:18] any urgent bugs? === truk-away [~trukulo@docsis65-46.menta.net] has joined #ubuntu-devel === minghua [~minghua@danube.mems.rice.edu] has joined #ubuntu-devel === minghua [~minghua@danube.mems.rice.edu] has left #ubuntu-devel ["Leaving"] === runo [~runo@CM-lcon1-38-130.cm.vtr.net] has joined #ubuntu-devel === jamesh [~james@203-59-87-199.dyn.iinet.net.au] has joined #ubuntu-devel === jdub [~jdub@home.waugh.id.au] has joined #ubuntu-devel === tseng [~tseng@thegrebs.com] has joined #ubuntu-devel [04:40] justdave: happiness? [04:40] lamont: yep [04:40] what was it? [04:41] I accidently nuked a paren when I added the -f [04:41] ouch [04:41] because my stupid terminal was sending the wrong code for backspace :) [04:41] heh [04:43] guess postfix isn't as picky about that stuff as sendmail (or at least not by default) [04:43] sendmail makes you grant users permission to use -f [04:43] fehj [04:43] which is why I assumed I needed admin support :) [04:43] nc localhost 25 [04:44] yeah. then again, postfix's sendmail is mode 555 [04:44] and just writes a file into /var/spool/postfix/maildrop [04:44] one of these days soon, bugzilla will do smtp by default anyway [04:46] if you do, there's always sendmail -bs for those that can't run on port 25. But that's a gross hack from days gone by [04:47] back in 30 or so [04:47] later === npmccallum [~npmccallu@69-162-252-7.ironoh.adelphia.net] has joined #ubuntu-devel === lamont [~lamont@mix.mmjgroup.com] has joined #ubuntu-devel === lamont [~lamont@mix.mmjgroup.com] has joined #ubuntu-devel === lamont [~lamont@mix.mmjgroup.com] has joined #ubuntu-devel === doko [doko@dsl-084-057-030-252.arcor-ip.net] has joined #ubuntu-devel [06:44] morning guys [06:55] fabbione: morning [06:55] hey dani [06:56] daniels: i start to be a bit worried about Dec conf. [06:56] jane talks about end nov. mid dic [06:56] mark isn't sure... [06:56] personally i don't mind [07:18] oh man [07:18] the gnome-cups-manager test pages are all line-based, not font-based [07:18] ie. can't just s/XIMIAN DESKTOP/UBUNTU/g :) [07:19] ew! [07:20] jdub: btw, if you or anyone else can tell me how to preserve categories for contacts into evo2.0 and back, that'd be the last thing that's keeping me on jpilot... [07:20] lamont: to and from palm? [07:21] yes [07:21] the only use I have for evo is to sync my palm. :-) [07:21] heh [07:22] I _like_ the MUA I use, don't want some stupid gui MUA. :-) [07:22] lamont: so why do you sync with evolution at all? [07:22] atm I don't/ [07:23] but if I could preserve my address book categories, then I could drop jpilot. [07:23] the wifely one is using evo, because I didn't want to make her non-conformist... [07:23] ok, why do you want to sync with eovlution insetad? [07:23] oh [07:23] she discovered this evening that if you drop 5 concurrent events into evo, it segv's trying to print it... ;) [07:24] jdub: I'm _trying_ to be a good warthog/ [07:24] and it would be nice to share calendars in the house. === lamont admits to some usefulness to evo. I just don't want it for my MUA... === lamont looks at the clock, and realizes that he has to get up in about 5.5 hours. sigh. [07:25] daughter wants up at 5, instead of the normal 5:30. [07:25] I hate getting up early. [07:26] anyway, g'night [07:27] night [07:29] jdub: pstoedit? [07:31] mdz: thought about that, but doesn't convert to anything inkscape likes [07:31] but i could screw around with xfig [07:32] WWBJD? [07:33] jdub: there is a pstoedit SVG output plugin [07:33] maybe you could ask Ximian for whatever source format they have for the page [07:33] :) [07:34] it should probably be fixed in gnome cvs anyway [07:34] mdz: where's the svg output plugin? [07:34] http://www.pstoedit.net/plugins [07:34] eww, it's a binary [07:35] yeah [07:35] "Unless you get a key, these drivers will slightly distort the generated output. Colors will be changed and the letter e will be replaced by $." [07:36] oh hey [07:36] are you sure pstoedit doesn't support svg out of the box? [07:36] it looks like it can use the libplot svg output [07:37] pstoedit -f plot-svg xd2-testpage-a4.eps /tmp/test.svg [07:37] Interpreter finished. Return status 0 [07:39] mdz: it just gave me 0 length output\ [07:39] inkscape seems to like it fine [07:39] oh, now it's working [07:39] bong [07:40] ugh, the output is horrible [07:40] it's a little cracked out [07:40] might even be worth doing our own ;) [07:41] for some reason everything is covered by black circles [07:41] s/circles/shapes/ [07:41] if you delete that crap, the good stuff is underneath [07:41] maybe it's some weird printer magic [07:42] all of the objects are there, it's just the colours that are borked [07:43] mine are all misshapen [07:45] inkscape crashed on me [07:46] probably worth making our own :-) [07:46] I liked all the little test patterns they provided though [07:48] mmm === jdub has found a test page generator [07:48] will see if the output sucks or not ;) === pitti [~martin@195.227.105.180] has joined #ubuntu-devel [07:53] Hi guys! [07:53] pitti: good morning [07:53] mdz: Morning. Exhausting weekend... [08:11] http://www.gnome.org/~jdub/random/ubuntu-a4.eps or http://www.gnome.org/~jdub/random/ubuntu-letter.eps [08:11] anyone want to check those out? [08:14] hmm [08:14] that looks pretty good [08:23] jdub: ERROR 404: Not Found. [08:24] that makes ithard to test [08:25] mdz: 1808 what do you want me to do with this one? [08:25] jdub: ^^ [08:25] pitti: does this sound like it could be related to pmount? https://bugzilla.ubuntu.com/show_bug.cgi?id=1715 [08:26] pitti: I can easily fix up my gnome-vfs patch, but I'm wondering why the problem occured just recently. [08:27] fabbione: oh, that reminds me [08:27] fabbione: pipka's laptop only supports the full resolution of the lcd at 16bit [08:28] fabbione: which means the configuration fails :| === fabbione ask again... what should I do with it? [08:29] defaulting to 16 bits or let 24 bits? [08:29] one way or another we will get bug reports [08:29] do you mean in the general case? [08:29] "it doesn't work at 24 bits but only at 16" [08:29] http://www.gnome.org/~jdub/random/ubuntu-a4.eps or http://www.gnome.org/~jdub/random/ubuntu-letter.eps [08:29] "GRAVE BUG: my video card supports 24 bits and you only init at 16" [08:29] ^ try again :) [08:29] fabbione: ;-) [08:29] jamesh: this is indeed the same bug as #1599. [08:30] jdub: yes as a general case... it's either one or another [08:30] fabbione: if there is some way to auto-detect the situation, great [08:30] jamesh: this is not directly a pmount problem, it's a gvfs bug [08:30] fabbione: if not, close it [08:30] fabbione: either way, it's not important for the release (hence severity; enhancement) [08:30] mdz: hhhrrrmmm, it ends up being important [08:30] jamesh: seb recently uploaded a new gvfs, and my patch (which worked fine with the older one) does not play well with the new upstream version [08:30] jdub: what does? [08:30] jamesh: so I disabled it temporarily [08:30] pitti: I can probably get the computer:/// and network:/// patches to only move "connect to server" type entries. That would solve the problem, right? [08:30] mdz: well there is a way, but that will require xresprobe to return me the amount of ram on the video card and according to resolution set the depth [08:30] mdz: when the resolution chosen does not work at 24bit [08:31] jamesh: yes, this would solve both bugs. [08:31] mdz: but do you feel to implement it at this point in time? [08:31] fabbione: no, I really don't [08:31] jamesh: if you know the guts of gvfs well enough, it would be great! [08:31] mdz: i realize we did never discuss this problem before [08:31] mdz: so it's a point, but up to you [08:31] pitti: okay. I'll put together a patch, and we can see if it solves things. [08:31] this is only an issue with older graphics cards which do not have enough RAM, right? [08:31] fabbione: i'd say hoary [08:31] jamesh: I actually thought that I had to write new code to make the devices appear in Disks, but if that code already exists, so much the better [08:31] jamesh: thanks! [08:32] mdz: that's a lot of machines [08:32] mdz: not just older [08:32] pitti: my patch moves all GnomeVFSVolumes that don't have an associated GnomeVFSDisk object to network:/// [08:32] jdub: video cards sold today have more ram than most of my computers [08:32] jamesh: ah, maybe that was the reason why my previous patch just did not work any more [08:32] pitti: I'm not sure why things like the firewire and USB drives don't have a Disk object now though. [08:32] mdz: in ricer machines, sure [08:32] jdub: in something you buy from Dell or HP [08:33] pitti: s/GnomeVFSDisk/GnomeVFSDrive/ [08:33] mdz: it's a significant bug, but something that can probably wait until hoary [08:33] jamesh: upstream version only shows Disks which have an fstab entry [08:33] mdz: yes [08:33] jdub: I don't see that we can do anything intelligent in that situation; we would need to ask a question [08:33] jamesh: my previous patch "faked" such an entry by modifying the function which read fstab; it added the user-entries of mtab [08:33] "what's more important: colours or resolution?" [08:33] pitti: okay then :) [08:34] jamesh: but with the new upstream version, it became unreliable and breaked the trash [08:34] mdz: it's easy to determine from available video ram [08:34] jdub: you can't determine whether the user would prefer a higher resolution with fewer colours, or a lower resolution with more colours [08:35] that's the tradeoff, isn't it? [08:35] not on an lcd :-) [08:35] but it's something you can decide that they can reconfigure later [08:35] (on a crt) [08:36] that's what we do today :-) [08:36] they can't, because their system doesn't work [08:36] why not? [08:37] can the driver not detect that the mode is no good and fall back? [08:37] we don't do that today [08:37] when pipka installed her machine [08:37] it chose the right resolution for the lcd [08:37] and the wrong colour resolution [08:37] were she not a kung-fu master, she would not have got out alive [08:38] ah, ok [08:38] so the LCD was configured correctly, but the card can't drive it [08:38] if you can determine two useful configurations from the video ram [08:38] in the LCD case there is no tradeoff [08:38] you just need to drop the depth [08:38] then on a laptop you can choose the correct one for the lcd [08:39] on a crt you can make up some reason why you'd choose one over the other [08:39] and let the user configure it how they want later [08:39] #1808 is a different case [08:40] I think the emulator just sucks, but I wasn't sure [08:40] no, same case [08:40] it's just not hardware ;) [08:40] oh, you have a copy of microsoft virtual PC? [08:40] that'll come in handy for testing the fix [08:41] i don't, but i know it emulates an s3 [08:41] i think i can get access to a copy [08:41] I don't think the chipset is relevant [08:41] but probably don't need it, the fix is described above :) [08:42] the chipset isn't relevant, it's the 'hardware environment' [08:42] the 'hardware environment' in this case seems like a buggy emulator [08:42] which is a very differetn case from a card which is short on video ram [08:42] the amount of video ram can generally be probed [08:42] dude, it emulates a video card that can't do that [08:42] nothing wrong with the emulator [08:42] I _have_ an S3 card, and it _does_ to 24-bit colour [08:43] s/to/do/ [08:43] jdub: eps: maybe let the user know how many lines around the edges he is supposed to see? [08:43] apart from making a bad choice of hrdware setup [08:43] doko: hrm, that'd be a software issue ;) [08:43] doko: just replacing the poopie eps shipped with gnome-cups-manager [08:45] jdub: no, just print it somewhere on the page, "you should see ... if not, adjust your setup, help at ..." [08:46] fabbione: hmm, there is also a tradeoff with DRI [08:47] my mga card can only run 1600x1200 with DRI disabled due to lack of RAM [08:48] mdz: ok.. should we default to 16 bits? [08:48] fabbione: hell no :-) [08:48] fabbione: just another thing to think about when we decide to look at this issue in the future [08:48] mdz: we need to rewrite all the autodetection code anyway [08:48] and do a hell of a lot of a cleanup [08:49] s3s do do 24-bit, yes === pitti [~martin@195.227.105.180] has joined #ubuntu-devel [08:52] http://www.robertmoir.co.uk/win/vpcfaq/VPCFAQ7-KnownIssueswithVP.html#7.3 [08:52] http://vpc.visualwin.com/ [08:52] ^ mentions ubuntu ;) [08:53] hah, nice [08:55] Colin Barnhorst seems to have tried many different operating systems [08:55] hah, even knoppixmyth is on there [08:56] sick puppy ;) [08:57] but "does it worK? no" [08:59] hah, that page describes linux boot parameters as "cheat codes" [08:59] up-up-up-b-b-a-start-down-- WOOHOO! KERNEL 2.6!!!! [08:59] ^^vv<><>ABAB START [09:00] try to submit an entry for "Small Ubuntu Server", maybe it shows up in the front then ;) === mdz blinks at the PregnantPanda proposed release name [09:03] let's choose that name for the next release which requires new glibc, new gcc, new kernel, new X and new just about everything else [09:04] mdz: i like PokeyPenguin [09:04] that's Hoary, or Hoary+1 ? [09:05] hoary+2 is meant to be PerkyPenguin [09:10] ok bong-sippers [09:10] what do you think about those printer pages [09:10] ? [09:10] otherwise i'm just going to upload the suckers [09:11] ubuntu. [09:14] where are the printer pages? [09:15] ah, missed the 'try again' [09:15] http://www.gnome.org/~jdub/random/ubuntu-a4.eps or http://www.gnome.org/~jdub/random/ubuntu-letter.eps [09:16] jdub: definitely fixes the bug :-) [09:25] heh [09:25] you like? [09:25] fabbione: any idea about 1724? [09:26] pitti: I just bugzilla'd the new gnome-vfs patch. [09:26] jdub: I think it would be nice to have some test patterns and stuff, but for Warty purposes, what you have is more than sufficient [09:26] patterns rather than colours? [09:26] jamesh: thanks! Does it work now? [09:27] jamesh: I will take a look at it (out of curiosity :-) ) [09:27] jdub: text in a few different fonts, stipple pattern, parallel lines, etc. [09:27] mmm [09:27] pitti: it doesn't cause any problems here, and the patch to computer-method.c is a lot simpler [09:27] pitti: which means that it should behave more like upstream [09:28] jdub: I say ship it [09:28] SHIP IT! [09:29] pitti: I've only got a limited number of volume types to test on this box, which is why I'd like to know if it works well in other situations. [09:29] jamesh: still quite a big patch, but it looks as if it _could_ go upstream... [09:29] jamesh: I think there's only one possibility to find that out :-)) [09:29] pitti: it is a replacement for a patch we already have in gnome-vfs, and it is a bit smaller than the one it replaces. [09:30] pitti: the diffs between the two patches are quite small. [09:30] jamesh: shall I build and test the package before you ask mdz to approve it? [09:30] pitti: that would be useful, yes. [09:31] jamesh: do you already have a i386 deb somewhere? [09:31] no. [09:31] jamesh: okay, I build it myself. [09:36] this version of the patch could probably be sent upstream (the last version wasn't as clean) === azeem [~mbanck@socks-out.lrz-muenchen.de] has joined #ubuntu-devel === seb128 [~seb128@ANancy-111-1-3-129.w81-250.abo.wanadoo.fr] has joined #ubuntu-devel [09:50] Morning seb128! [09:50] morning [09:51] yo seb128 [09:51] grrr, 30 hours to fix the dsl connection, apparently dsl people don't work sunday [09:51] seb128: got my mail? [09:51] ouch [09:51] just back from 30 hours internet-less [09:51] my fetchmail/spamassassin/... is working :) [09:52] jamesh: patch works fine for me [09:53] pitti: great. [09:53] jamesh: this bug report was great; so far I thought that the new device would not appear in Nautilus at all [09:54] jamesh: I would have never come to the idea to look in "Network" for it :-) === seb128_ [~seb128@ANancy-111-1-18-15.w81-51.abo.wanadoo.fr] has joined #ubuntu-devel === trukulo [~mzarza@26.Red-81-45-239.pooles.rima-tde.net] has joined #ubuntu-devel [10:02] funny, I'm getting 2 weeks old mails for 3 days (WTF with my @debian) [10:04] jdub: the new hal 0.2.98 does not work with the current g-v-m, so I need to upload a new gvm [10:04] jdub: there is a new upstream version 1.0.2 of g-v-m which just adds some new translations [10:04] :) [10:04] seb128_: read planet.debian.net, there was a problem with mail server [10:04] jdub: may I upgrade g-v-m- while I'm at it? [10:05] pitti: yes thanks [10:06] http://www.extremetech.com/article2/0,1558,1651228,00.asp [10:06] ^ (p)review === elmo_ss [~james@host217-37-231-28.in-addr.btopenworld.com] has joined #ubuntu-devel === pitti [~martin@195.227.105.180] has joined #ubuntu-devel === pitti [~martin@195.227.105.180] has joined #ubuntu-devel === jdthood [~jdthood@aglu.demon.nl] has joined #ubuntu-devel === pitti [~martin@195.227.105.180] has joined #ubuntu-devel [11:30] seb128: the nautilus-cd-burner in warty does not yet use the new locking mechanism of hal to prevent further mounts; is this already upstream? [11:30] yes it's upstream [11:30] but need hal >= 0.2.98 [11:30] so it's not buildable for the moment [11:30] seb128: I just packaged crack-of-the day hal/dbus/gvm [11:30] seb128: now I test CD burning [11:31] ok [11:31] cool [11:31] seb128: I cannot find any trace of lockign in the n-c-b sources [11:31] it's in 2.8.3 [11:31] seb128: ah, so I'm going to package this as well [11:31] seb128: unless you want to do that... === azeem [~mbanck@socks-out.lrz-muenchen.de] has joined #ubuntu-devel [11:32] pitti: that's just a matter of dch -i, New upstream release and update the hal depends [11:32] pitti: so as you want :) [11:32] seb128: okay, since the other stuff is at my hd and not uploaded yet, I would do it [11:32] ok, thanks [11:32] seb128: so the guys can test the complete new stack from my archive without having it to upload first [11:33] yeah [11:33] jdub: [11:34] xfree86 (4.3.0.dfsg.1-6ubuntu22) warty; urgency=low [11:34] * Update 000_stolen_from_HEAD.diff: [11:34] + Include Xv/framebuffer fix for xf86xv.c. [11:34] * Update 030_Xserver_and_driver_region_primitive_fixups.diff: [11:34] + Fix REGION_EQUAL call in nv_driver.c. [11:34] * Update Czech debconf template translations (thanks, Miroslav Kure). [11:34] * Apply fix for Keycodes for PrintScreen and SysRq. (Closes: #1762) [11:34] -- Fabio M. Di Nitto Mon, 27 Sep 2004 07:58:55 +0200 [11:34] seb128: http://ftp.gnome.org/pub/GNOME/desktop/2.8/2.8.0/sources/ just has n-c-b 2.8.0 [11:34] seb128: where do you pull new sources from? [11:35] seb128: ah found it [11:35] seb128: sorry for disturbing [11:35] np [11:40] jdub: ok, changes done for the -fr list [11:47] seb128: hmm, ncb 2.8.3 still does not attempt to umount the CD-RW [11:48] seb128: how easy is it to add an _extra_ unmount option to the context menu? (By now you just have eject) [11:48] no idea [11:48] should ask jamesh about this, he had a look on that [11:50] seb128: I haven't looked at the exact nautilus code, but I could. [11:50] ok [11:50] jamesh,seb128: I think I can add the unmounting to n-c-b [11:51] jamesh: have you replied to my question about pygtk/gnome the other day ? [11:51] jamesh,seb128: but having a separate unmount option for CDs is a good idea nevertheless [11:51] pitti: why do you want to umount in n-c-b ? [11:52] seb128: by now you insert an RW, it is mounted automatically; if you want to burn on it, it is busy [11:52] seb128: however, there is no possibility to unmount it, it is ejected [11:52] you want to add an umount "option" ? [11:52] seb128: so we could pumount -l the device before burning [11:52] yes, but not an option [11:52] just umount it [11:52] seb128: okay, then without -l [11:53] seb128: by now, there is just a "TODO: unmount the device..." in the ncb code [11:53] I can ping an upstream about this === ross [~ross@82-44-126-41.cable.ubr03.croy.blueyonder.co.uk] has joined #ubuntu-devel [11:54] seb128: maybe a good idea; I don't know much about the code [11:54] seb128: I can add the umount there and try if it works [11:54] ok [11:58] seb128: I talked a bit with jdahlin about it. The fix is probably the right thing for Python < 2.3.5 [11:58] pitti: btw, the code for handling unmount/eject context menu item is in real_update_menus_volumes() in src/file-manager/fm-directory-view.c [11:59] pitti: it seems to use the same verb for both eject and unmount [11:59] jamesh: so the decision is taken at a lower level? [11:59] jamesh: that would mean it is nontrivial to add, isn't it? [12:02] pitti: the decision between eject and unmount is in the eject_for_type() function [12:02] pitti: which is set to show an eject option for cdroms, zip and jaz drives, and unmount for others [12:02] jamesh: I saw this once; it does not seem to be designed for just unmounting [12:03] jamesh: well, if its nontrivial to change and cd-burner supports automatic unmounting, it's not such a big deal, I suppose [12:03] pitti: unmount_volume_callback() uses the result of eject_for_type() to decide whether to eject or unmount [12:03] pitti: and the code for setting up the context menu uses eject_for_type() to decide whether to call the item eject or unmount. [12:04] it isn't a case of showing one item and hiding the other [12:04] jamesh: ugly... [12:04] jamesh: at least for having two commmands... [12:04] it is just a single menu item with a different label depending on the drive type. [12:05] pitti: well, other than the case you mentioned, when is a user going to want to unmount a cdrom or zip disk? [12:05] pitti: the correct fix would be to get n-c-b to do the unmount [12:05] jamesh: I would do it if I wanted to repartition my USB stick, ZIP drive or whatever [12:06] jamesh: but if I want to do that, I can as well type umount by hand [12:06] jamesh: I want to fix n-c-b anyway [12:06] pitti: actually, you probably couldn't umount by hand ... [12:06] if the stick has a trash dir [12:06] jamesh: these device locks, I suppose [12:07] jamesh: but that shouln't happen on CD-ROMs, no? [12:07] jamesh: seb128 suggested not to lazily unmount the CD, but do that normally [12:07] pitti: the Gnome vfs umount/eject stuff has a "pre unmount" signal so that apps like nautilus can stop monitoring directories [12:08] since fam/gamin's dnotify watches hold the device open [12:08] jamesh: hmm, so actually n-c-b should call this instead of umounting directly? === Sledge [~steve@80.46.37.1] has joined #ubuntu-devel [12:09] morning... [12:09] morning Sledge [12:10] pitti: yep. Get the GnomeVFSVolume or GnomeVFSDrive object for the cd drive and call its unmount() method. [12:10] jamesh: any idea how I can do that in n-c-b? I only have the device string (/dev/hdc or so) [12:11] pitti: gnome_vfs_volume_monitor_get_volume_for_path() might do it. [12:15] jamesh: do I have to call the pre_umount method myself or does GnomeVFSVolume's umount() method care for that? [12:15] pitti: the gnome_vfs_volume_unmount() routine handles it all for you. [12:15] jamesh: thanks a lot for your help! [12:19] pitti: if inotify existed a few years back, things wouldn't be so complicated ... === Capri [~makolb@mnch-d9ba4117.pool.mediaWays.net] has joined #ubuntu-devel [12:32] does somebody already had problems to boot the warty iso on a powerbook ? [12:34] seb128: define problems [12:34] the powerbook doesn't boot on the CD [12:34] it doesn't see it as a bootable one or something like that [12:34] seb128: hold 'c' as you power on [12:35] ok, not for me, I'm telling that to the guy ;) [12:35] thanks [12:35] sure [12:35] and for shared-mime-info, yes it was 0.15 :) [12:36] right, well, that bug is still occurring then :-/ [12:36] thom: 'c' doesn't work, the guy says it already does that [12:36] debian CDs boot ok [12:36] warty doesn't boot [12:36] s/it/he/ [12:41] seb128: btw, I did a patch for bug 1715 (firewire volume showing up under Network) [12:42] cool, thanks [12:42] seb128: pitti tested it, and it fixes 1599 too [12:42] ok, I'll do an upload soon [12:42] thanks [12:52] jdub, mdz: ping [01:05] seb128: I'd check that it burnt correctly; the CD build processes for Debian and Ubuntu are practically the same [01:06] the guy has burnt several time the debian and the ubuntu iso, the Debian always boot, the Ubuntu one not ... weird [01:15] very odd [01:15] is he certain that he downloaded the Ubuntu ISO correctly? [01:15] yes [01:15] rsync shows no changes? [01:16] make sure to run rsync with -c [01:17] I'm checking, but I think he downloaded the preview and one daily, and he can mount them without any problem [01:17] he deleted the iso and is downloading it again ... [01:19] the preview has certainly booted on lots of machines [01:48] jamesh: still here? [01:49] pitti: yep. [01:49] jamesh: gnome_vfs_volume_monitor_get_volume_for_path delivers a VFSVolume* with device path 'none' and umounting does not work [01:49] jamesh: do I need to initialize the monitor somehow? [01:50] jamesh: by now I just create an instance with GnomeVFSVolumeMonitor* gvm = gnome_vfs_get_volume_monitor() [01:50] jamesh: sorry for askign so much, but this stuff is not documented yet [01:50] pitti: hmm. I'm a bit new to the APIs myself. [01:51] jamesh: do you know what these _ref and _unref methods are for? [01:51] jamesh: maybe I need to call them [01:52] pitti: imaybe gnome_vfs_volume_monitor_get_volume_for_path() expects a path on the mounted volume [01:52] jamesh: argh; maybe. By now I give it the device node path [01:52] jamesh: I don't have the mount path in n-c-b... [01:53] pitti: the other thing to try would be to use gnome_vfs_volume_monitor_get_mounted_volumes() to get the list of mounted volumes, and gnome_vfs_volume_get_device_path() to find the device path for each mounted volume [01:53] find the volume that matches the device file name. [01:54] jamesh: sounds possible (although pretty much code for such a simple thing :-) ) [01:55] pitti: maybe that's why n-c-b doesn't currently do it :( [01:56] jamesh: :-) So, off to doing it [01:56] pitti: of course, the unmount may still fail if other apps have files open ... [01:56] jamesh: but I think this is reasonable in this case [01:57] pitti: but it will get nautilus to stop holding the volume open [01:57] jamesh: ncb will show a dialog "drive busy", then the user can close the app and click "OK" to try again === jamesh should get a cd burner to try these sorts of things out ... === T-Bone [~varenet@T-Bone.developer.debian] has joined #ubuntu-devel [02:00] seb128: 1762: who the screenshot stuff is supposed to work? [02:00] if i press print or alt+print, where is the screeshot saved? [02:02] fabbione: desktop or home, not sure [02:02] let me check [02:03] fabbione: it should open the screenshot dialog (same as in the computer menu) [02:03] there's a Sun guy working on adding printing support to the screenshooter dialog ... [02:03] seb128: ok thanks. than the fix doesn't work [02:05] seb128: does it need anything special installed? like gnome-foo-bar? [02:05] gnome-panel [02:05] that's all [02:05] it runs gnome-panel-screenshot [02:06] fabbione: pressing print or alt-print generally doesn't work in debian's X [02:06] ross: that's what i am trying to understand [02:06] ross: there is a patch missing [02:06] i've seen a patch to fix it [02:06] ross: that's the purpose of the bug report yes [02:07] aaah i see [02:07] ross: yes but the patch doesn't fix apparently === ross reads back [02:07] :) [02:07] ross: http://bugzilla.ubuntu.com/show_bug.cgi?id=1762 [02:08] xc/programs/Xserver/hw/xfree86/common/xf86Events.c @ 3.147 [02:08] 8. Fix for non-PC keyboard bug introduced by changes to make SysRq [02:08] generate the same keycode as PrtScrn (Ivan Pascal). [02:08] HMM [02:08] Is there any reason we don't modprobe apm on boot by default? [02:08] it would seriously suck for my laptop, for instance. [02:09] It won't do anything if acpi is already running [02:10] we should make sure acpi is loaded first, then. [02:10] Mithrandir: the initrd already loads thermal/fan [02:10] that should be enough, shouldn't it? [02:10] acpi is built into the kernel [02:11] Kamion: Does that happen in Debian, too? [02:11] Kamion: if you load apm on my system the modifier keys behave really badly, they get sticky. [02:12] pitti: i've forgotten pmount policy again -- would an external firewire hdd be mounted without a line in fstab? [02:12] ross: Should work, yes [02:12] hm, it isn't [02:13] Mithrandir: Really, apm won't do anything unless acpi is either implicitly or explicitly disabled [02:13] ross: does it appear in Device Manager= [02:13] There's no way to build modular acpi [02:13] ross: s/=/?/ [02:14] urgh. hald has hung in scsi_wait_req and is dead [02:14] mjg59: in d-i, yes; in the installed system, no [02:14] mjg59: (unless it was done somewhere different from where I was expecting) [02:14] ross: can you kill it and try again? [02:14] nope, its totally dead [02:17] Kamion: Bleah. We /still/ need to fix that. It does kill hardware. [02:17] ross: I packaged a newer hal, maybe it fixed the crash... [02:23] it looks to me that code is not even reached [02:23] either with the xfree86 patch or the x.org one [02:29] moof [02:29] moof, lamont [02:29] hey lamont [02:30] fabbione: which code, the radeon suspend shit, or the xkb stuff? [02:30] daniels: print and +print [02:31] ahr [02:31] we have both the fixed from xfree86 and x.org but alt+print still = SysRq [02:31] crap [02:31] it gives me the feeling that the part of the code is not used at all [02:31] let me try a gdb sessions [02:32] where for SURE it will work [02:32] therefor making the problem untraceable [02:49] hm, why is my numeric keypad not working at all [02:54] gar [02:54] how am i supposed to shoot people over lunch if my numeric keypad doesn't work [02:54] works at the console but not in X [02:59] seb128: 1762 it's a gnome problem :-))))) === pitti [~martin@195.227.105.180] has joined #ubuntu-devel [02:59] seb128: i just finished a gdb session on X and the keyboard mapping is working fine [02:59] seb128: meaning that something else (!=xfree86) is mangling the keyboard [02:59] and you know what is the best part...???? [02:59] ok [02:59] I CAN PROVE IT :P [03:00] fabbione: any idea why in X my numeric keypad wouldn't work at all? [03:00] (gdb) n [03:00] 579 if (scanCode == KEY_SysReqest) [03:00] (gdb) n [03:00] 580 scanCode = KEY_Print; [03:00] this is the patch to xEvents [03:00] it maps the SysReq to Key_Print and that works [03:00] (already the fact that is called means that it works) [03:01] fabbione: could you provide the evidences in http://bugzilla.gnome.org/show_bug.cgi?id=86506 ? :) [03:01] seb128: do i need an account to write there? [03:01] probably yes [03:01] seb128: humpf [03:01] :((( [03:01] just reassign 1762 to me with details [03:01] I'll follow upstream [03:02] ross: no.. keyboard? model? layout? laptop? desktop? [03:02] fabbione: standard UK keyboard, pc105, desktop [03:02] ah. when i press the arrow keys xev reports motionevents [03:04] ross: "Num lock"??? [03:05] moves the mouse cursor when its on and off === ross wonders what catches numpad events and turns them into mouse events [03:07] aha [03:07] i pressed shift-alt-numlock [03:07] obviously === pitti [~martin@195.227.105.180] has joined #ubuntu-devel [03:08] WTF === fabbione kicks crapzilla === lamont takes the car to the shop, where he will work on his laptop for a while. back in a few hours. === fabbione files a blocking bug on bugzilla for bugzilla [03:14] IT [03:14] IT'S UNUSEABLE [03:18] thom: 1221, the return mime for gnomevfs-info is still the same ? [03:19] seb128: MIME type : application/x-cd-image [03:23] Kamion: ping [03:24] i think fedora's bugzilla css is about as good as it gets [03:24] thom: aye. [03:25] ours is pants. [03:25] it makes me cry [03:25] so does planet.u.o [03:26] thom: could you try to patch with GNOME_VFS_FILE_INFO_FORCE_FAST_MIME_TYPE instead of SLOW and try again with gnomevfs-info ? [03:27] jdub: permission to upload 1821 === jdub attempts to say he's not at work. [03:29] fabbione: approved in comments [03:29] ok [03:34] seb128: i'll give it a try in a bit, sure [03:34] thom: thanks [03:34] jdub: OH MY EYES [03:34] (i just looked at planet) [03:34] uh huh [03:35] ARRRG [03:35] thom: seen our website recently? === thom drums fingers and waits for 400GB of sata love to arrive [03:35] jdub: planetplanet? [03:35] www.u.org [03:36] huh? a book shop? [03:37] ross: ubuntulinux [03:45] jdub: pong [03:45] jdub: 1787, please confirm ;-) [03:45] Kamion: how was oldentown? gothamburg? whereveryouwent? ;) [03:46] oldenburg [03:46] burg == castle, roughly === jdub spanks azeem [03:47] jdub: went well, got lots done, didn't *quite* get the graphical installer booting because mklibs hates me, otherwise good [03:47] Kamion: so there was some discussion a while back about switching off the hidden-menu grub setting if other OSes were detected... did that come to anything? [03:48] Kamion: noice. [03:48] jdub: I think it came to "good idea, put on Colin's to-do list" [03:48] ahr [03:48] (which is fine by me, will do it ...) [03:48] Kamion: did it have a milestone attached? === jdub buzzes off again :) [03:50] jdub: not sure it had a bug attached, even ... [03:50] PS several uploads waiting for approval [03:50] Kamion: What's the graphical installer based on at the moment? [03:50] mjg59: gtk [03:50] (er, not quite sure what you mean) [03:51] Just using the gtk debconf front-end, or something nicer? [03:51] And is this framebuffer or X? [03:51] just cdebconf-gtk for now, will have more bolted onto it [03:51] framebuffer [03:51] running AWAY from X [03:51] Hrm. Has anyone ported atk to the framebuffer? [03:51] It's quite X specific at the moment [03:52] It'd be nice to get the accessibility support for free [03:52] atk? the accessibility library? [03:52] dunno [03:52] Yeah [03:52] What's the main problem with X? [03:53] big pile more complexity that we don't need? [03:53] plus size; even with gdk-directfb we're going to be hitting initrd size limits on many systems unless we're very careful [03:54] powermac netboot already has no chance [03:54] Ah, ok - I'd assumed you'd be using a filesystem mounted off the media for graphical install [03:54] My (not very strongly founded) recollection is that RH doesn't do graphical installs when netbooted [03:55] we don't get to mount the installation media until after cdebconf has already displayed some questions [03:55] night all [03:55] www.directfb.org is probably the place to look for any accessibility stuff [03:55] there is XDirectFB, but I suspect daniels will consider it BONG; dunno [03:56] Looks like there's no support at the moment [03:56] That's a shame - an accessible installer would rock [03:56] in theory we could restart cdebconf with a different frontend, but it's not going to be trivial ... [03:57] has anyone tried to get the directfb frontend merged into gdk proper, I wonder? [03:57] mjg59: i'm going to do an accessible derivative [03:58] jdub: Rock [03:58] did you consider using plain GTK for the first stage, and just map the user directions to the debconf-database? [03:58] mjg59: there's a lot of bong requirements that sighted people will just baulk at [03:58] azeem: what do you mean? [03:58] jdub: Oh, most of your posts to ubuntu@ seem to appear twice [03:58] mjg59: i have a local testing team too ;) [03:58] mjg59: some; evo bug. :| [03:58] anyway [03:58] good night [03:59] Kamion: IMHO debconf-gtk looks pretty different from usual GNOME applications, so it might be more work to get it look right than just do it natively [03:59] haven't looked at it, of course [03:59] azeem: that would really, really suck [03:59] heh :) [03:59] azeem: you're talking about writing a new installer, not porting the existing one. [04:00] azeem: we need to use cdebconf, otherwise it is not maintainable [04:00] azeem: also, nobody's even touched the cdebconf gtk frontend since March (and not seriously for much longer, I think), so it's not surprising that it's raw [04:01] I did some work a few days ago to start cleaning it up, but it'll take a while [04:01] well, I meant just redoing the UI dialogs and using cdebconf to set the values, based on what the user does [04:01] azeem: we have some glade-related plans that will have that kind of effect, but it's not that simple [04:02] yeah, I guess [04:02] azeem: there is a lot of code in the installer that does things like setting the range of values presented to the user based on the answers to previous questions in the same udeb [04:02] jamesh: can I please ask you for help again? [04:03] azeem: if we bypass and duplicate that code, the result won't be maintainable, so we need to come up with a way to keep it (db_callback or similar) [04:03] jamesh: on www.piware.de/gvfs-unmount.c I put a version of my unmounting procedure [04:03] true [04:03] (I suggested db_callback to joeyh on Saturday and he shuddered ... :-)) [04:04] :) [04:07] Anybody here who knows about gnome-vfs programming? === pitti bites in his table [04:11] anyone know of anybody selling boxes with Ubuntu preinstalled yet? === pitti [~martin@195.227.105.180] has joined #ubuntu-devel [04:18] seb128: have you ever programmed some stuff with gnome-vfs2? Do you know how to correctly unmount a device? [04:18] yes and no [04:19] seb128: My function for unmounting the cd in the gvfs way worked after ten minutes [04:19] seb128: but it makes n-c-b crash when it starts to burn [04:19] utch [04:19] seb128: so I assume I have to free some resource or whatever to make it wok [04:19] work [04:19] seb128: www.piware.de/gvfs-unmount.c shows my current version [04:23] seb128: do you think it will hurt much if I just call "pumount /dev/whatever" instead of going through the gvfs calls? [04:23] seb128: respective pumount -l [04:23] do whatever you feel right [04:23] okay [04:23] I've > 100 bugs in my list now, I don't really have time to look on this [04:23] sorry [04:25] seb128: np, just asked [04:26] seb128: I don't know much about this gvfs stuff... [04:26] where are you testing this function ? [04:26] seb128: it's not documented at all [04:26] I know ... [04:26] seb128: where? on my pc? [04:26] in nautilus [04:26] or as a standalone soft ? [04:26] seb128: I call n-c-b /path/to/image [04:27] ok [04:28] speaking with a gnome-vfs guy, he says that you should start to test it out of nautilus-cd-burner === jdthood [~jdthood@aglu.demon.nl] has joined #ubuntu-devel [04:53] seb128: I think I have found my error [04:54] oh ? [04:54] seb128: can you tell me which function to call when I'm waiting for something? [04:54] seb128: while(...) sleep(1) is not good since it does not process callback functions [04:54] seb128: is there something like a gnome_sleep_but_process_events()? [04:56] you can use g_idle_add () [04:56] seb128: I call gnome_vfs_drive_unmount with a callback and must wait for the callback to be called [04:56] while (gtk_events_pending ()) gtk_main_iteration (); [04:56] might do what you want [04:57] yes, but that's ugly, isn't it ? [04:57] depends what you want to do [04:57] you should use async stuff in nautilus [04:57] seb128: actually I want to unmount the CD synchronously [04:57] seb128: that's why I have to wait for the callback to happen [04:57] you can probably block on a signal arriving [04:59] seb128: yeah, still no change with FAST [04:59] same mime in gnomevfs-info ? [05:00] yep [05:01] ok, thanks [05:04] l [05:14] gar. does cdbs really annoy anyone else? :( [05:14] thom: yep [05:15] thom: it's not that bad [05:15] i use it for all my xlibs packages, and it's used in dbus [05:15] what's wrong with it? [05:15] (then again, I have been known to like dbs) [05:16] it just feels totally over engineered, and makes it far too hard to work out what's going on under the hood [05:16] Kamion: oh good :-) [05:21] thom: how come acpi-support doesn't use dbs? [05:22] thom: bah, cdbs is cooool :) [05:24] daniels: shuttit,foo [05:24] seb128: you're french, though. i've seen french attempts at cool before ;-) [05:25] cdbs is yet another of those things that the maintainers of packages who uses it love, and anybody else who tries to understand what's going on in those packages hates === seb128 slaps thom [05:25] s/uses/use/ # I knew the grammar seemed funny when I wrote that [05:25] Kamion: you don't have to understand what's going on, the debian/rules has 3 lines most of the time [05:25] seb128: oh yes I do [05:25] when something breaks, as it does [05:26] or when I want to change something [05:26] I'm very very scared of any piece of software where somebody says "you don't have to understand what's going on" :-) [05:26] yes, cdbs is missing some documentation on the internals [05:26] yeah, but we use it for almost all the GNOME packages [05:26] because it usually means that when I need to understand it I won't be able to [05:26] and no problem at all [05:27] GNOME has a sane upstream, that's different ... [05:27] true === Kamion wonders how to test this debconf change [05:36] I guess I can hack the new package into my local Ubuntu archive and debootstrap === elmo_ss [~james@host217-37-231-28.in-addr.btopenworld.com] has joined #ubuntu-devel === mjg59 [mjg59@cavan.codon.org.uk] has joined #ubuntu-devel === npmccallum [~npmccallu@69-162-252-7.ironoh.adelphia.net] has joined #ubuntu-devel [06:18] l [06:18] sorry, wrong window... [06:31] mdz: already awake? === vuntz [~vuntz@fennas.vuntz.net] has joined #ubuntu-devel === pitti [~martin@195.227.105.180] has joined #ubuntu-devel [07:07] Kamion: can I get approval for 1516? [07:08] pitti: awake [07:08] mdz: Good morning! [07:08] mdz: The new utopia stack is ready [07:08] pitti: do you have i386 debs I can try? [07:08] mdz: but instead of typing it into IRC, I'm just writing an announcement to u-devel; might be better to test this on a broader basis [07:08] mdz: http://people.no-name-yet.com/~pitti/utopia/ (aptable) [07:09] mdz: The announcement will be on the list in a few minutes [07:09] ok [07:11] mdz: its now on the list [07:13] pitti: I did a dist-upgrade with your site in sources.list, it only upgraded dbus: no hal, etc [07:13] npmccallum: that's odd, let me look... [07:13] npmccallum: speling error :-) Packges.gt [07:14] npmccallum: can you please try it again= [07:14] npmccallum: s/=/?/ [07:15] nope, not yet [07:16] npmccallum: whoops, forgot about that. comment added. I want to look at that debconf encoding stuff before we blindly work around whatever the problem is. [07:16] pitti: I mean, I tried again, but its still not working [07:17] npmccallum: odd, http://people.no-name-yet.com/~pitti/utopia/Packages.gz now exists and shows all packages [07:17] npmccallum: did you apt-get update'd again? [07:17] yes [07:18] npmccallum: (because it might indicate some problem in debconf's internationalization support, and if it does then we *must* fix that) [07:18] ok, I removed it from sources.list, readded it, re apt-get updated, and its working now [07:18] but it wants to remove GVM!? [07:19] npmccallum: the new hal conflicts to the older gvm [07:20] npmccallum: you have to upgrade gvm, too [07:20] npmccallum: can't apt work this out automatically? [07:20] but its not upgrading it [07:20] pitti: it should, but its not [07:21] pitti: you have no GVM package in your ~pitti/utopia directory [07:21] npmccallum: bad. In fact the Conflicts: even existed before, I just bumped the version [07:21] npmccallum: good point [07:21] npmccallum: sorry, I upload it [07:23] npmccallum: uploaded [07:27] npmccallum: does it work now? === pitti [~martin@195.227.105.180] has left #ubuntu-devel [] === pitti [~martin@195.227.105.180] has joined #ubuntu-devel === pitti is still not used to the fact that ^W does _not_ delete the line in xchat :-( [07:29] heh [07:29] C-u does that [07:29] Mithrandir: tell that to a seasoned vim user :-) [07:30] pitti: everything is working well here, though I haven't done extensive testing [07:30] npmccallum: so apt-upgrading now works? Without any dpkg questions? === pitti is relieved [07:30] YES [07:30] YES [07:30] YES! [07:31] mdz: permission to upload X ubuntu22 [07:31] pitti: yes, everything now works [07:31] Xv extensions are working now [07:31] confirmed by the testers [07:31] pitti: the new version even picks up my usb key :) [07:31] pitti: the old hal didn't :) [07:31] Mithrandir: not for me, BTW. ^U does nothing [07:31] fabbione: nice catch :-) [07:31] npmccallum: has it partitions? [07:31] fabbione: any other changes? [07:31] mdz: ps i will upload tomorrow morning [07:31] pitti: yes [07:31] npmccallum: huh, what did hal-device-manager say to it? [07:32] npmccallum: anyway, great if it works now [07:32] mdz: xfree86 (4.3.0.dfsg.1-6ubuntu22) warty; urgency=low [07:32] * Update 000_stolen_from_HEAD.diff: [07:32] + Include Xv/framebuffer fix for xf86xv.c. [07:32] * Update 030_Xserver_and_driver_region_primitive_fixups.diff: [07:32] + Fix REGION_EQUAL call in nv_driver.c. [07:32] * Update Czech debconf template translations (thanks, Miroslav Kure). [07:32] * Apply fix for Keycodes for PrintScreen and SysRq. (Closes: #1762) [07:32] mdz: the first 2 are for the Xv extensions [07:32] fabbione: right, looks good, OK to upload [07:32] mdz: third one.. you guess it === pitti goes to find sth to eat [07:32] mdz: hold on.. the latter.. please read what i wrote on the bug [07:32] fabbione: I remember the bug [07:32] mdz: i added info [07:32] mdz: with gdb info [07:33] mdz: i applyed the patch from X.org but the problem is not in X [07:34] fabbione: so there is more than one bug, one of them not in X? or there is no bug in X (in that case, what does the patch do)? [07:34] mdz: the patch remaps SysRq (+print) to print [07:34] mdz: it is applied in 2 points of the code [07:34] mdz: but apparently applications still do not understand it [07:35] why should SysRq be remapped to print? [07:35] mdz: the gdb session shows it clearly [07:35] I thought PrintScrn was the key to take a screenshot [07:35] mdz: yes.. on i386 they are the same key [07:35] mdz: that's correct [07:36] so print screen should -> Print and alt+print screen should -> SysRq [07:36] or is there no keysym for sysrq? [07:36] mdz: that's what is actually happening [07:36] apparently SysRq is used for screenshot of a window [07:36] while print is for the entire screen [07:36] ahh, ok [07:36] in both cases gnome fails to detected it [07:37] so they bounced the crap down to xfree86 [07:37] xfree86 now has this lame remapping all over [07:37] but applications still fail to get it [07:37] anyway [07:37] i am off for today [07:38] mdz: someone needs to update linux-restricted [07:38] with the latest kernel [07:38] ok [07:38] good night [07:38] cya tomorrow [07:38] I don't see why X needs to remap that stuff, but if you think it is correct to include that patch, it's OK with me [07:39] goody [07:39] i will upload tomorrow === npmccallum [~npmccallu@69-162-252-7.ironoh.adelphia.net] has joined #ubuntu-devel [07:59] Kamion: ? [08:03] elmo_: yep? [08:15] kamion: linux-image bumped soname (or whatever) - can you let me know when I can remove the old ones? [08:15] (no rush) [08:18] npmccallum: the new default bash colours are really foul; when you pop up a terminal you get bright green on white for the username/hostname, which is bordering on painful to read [08:18] can we have some slightly more background-neutral ones? [08:19] Kamion: I'm not sure what you mean -- There are no colors on by default [08:20] npmccallum: I just installed from a daily build and got these [08:20] (amd64, but I'm sure it doesn't matter) [08:20] Kamion: it wasn't my doing then [08:20] npmccallum: ah, maybe doko's changes [08:21] I thought he talked to you about those [08:21] elmo_: ok, will do, need to rev linux-kernel-di-* [08:21] He mentioned that he found a typo (rogue colon) [08:21] but that was it [08:21] maybe he did some other stuff too, I have no idea [08:24] npmccallum: 1164 Sep 25 Matthias Klose ( 50) Accepted bash 2.05b-2-15ubuntu4 (source) [08:27] nevertheless, perhaps colours which look somewhat better with our default terminal would be a good idea [08:28] Kamion: I turned on colors for the terminal a while ago and almost got shot :) [08:28] yeah, I don't think colours-by-default is going to be too popular :) [08:28] its not with people coming from the debian world, but it is with people coming from gentoo and the like [08:29] do they really like eye-bleedingly painful colours? :-) [08:29] the colors that I originally put in are from Gentoo, though I don't know if they have been changed or not [08:29] (I wouldn't be too upset with darker ones) [08:29] bright green looks good with a black background [08:29] perhaps Gentoo's terminal has a black background by default [08:30] it does [08:30] well, not the gnome one [08:30] of course, I switch the terminal to a black background first chance I get, but ... [08:30] me too :) === Kamion wonders if there's a way to pick a readable colour based on the background colour [08:31] there isn't [08:31] its all ansi escape codes [08:31] didn't think so - oh well [08:31] well, you can retrieve the background colour using ANSI escape codes [08:31] it's just painful [08:32] elmo_: speaking of the kernel, there should be more NEW enjoyment for linux-meta now, and linux-restricted-modules shortly [08:33] I'm not sure that linux-meta is quite right yet [08:33] Kamion: we should just make gnome-terminal's colors white on black by default ;) (if you want to get shot that is) [08:33] Kamion: do you agree that it would be nice to have a single metapackage which would install matching versions of the kernel and restricted-modules, for d-i purposes? [08:33] mdz: definitely [08:34] mdz: that would fix bug #1835 [08:34] mdz: although I'm not sure how well it plays with people who want to remove the restricted-modules [08:34] Kamion: right [08:34] the question is whether we should have two [08:34] one which gives you the latest recommended kernel === Kamion is uploading incremented-ABI linux-kernel-di now [08:34] and one which gives you the kernel + restricted-modules [08:34] and if we should, what do we call them? [08:35] no good ideas on naming I'm afraid [08:36] currently we have linux- which depends on both image and modules [08:36] but it's in main, and depends on restricted [08:36] so it seems to want to be split for that reason as well [08:36] I can deal with installing two packages [08:36] well, we also need it for upgrades [08:37] otherwise the image gets upgraded and the modules never do [08:37] and they just vanish [08:37] you could make the restricted-modules metapackage depend on an exact version of the kernel metapackages [08:38] s/s$// [08:38] then you'd have to upgrade it or remove the metapackage, and if people remove the metapackage then we can't help them anyway [08:38] hmm [08:39] and since the two come from the same source package, there's no maintenance issue with having Depends: linux (= ${Source-Version}) [08:39] (or whatever) [08:39] well, the modules concrete package depends on an exact kernel version [08:39] so if we start installing the modules metapackage rather than the concrete, we should get reasonable upgrade behaviour [08:40] right [08:41] tying them together with a metapackage seems like the right thing to do, there is only this naming problem [08:41] Jane bailed us out of the last naming roadblock; I'll ask her [08:41] hm, she left [08:41] this is relatively easy to do in base-installer, but I was holding off fixing the bug that made it not work until I had a restricted-modules metapackage [08:41] kamion: my mistake, I left the color prompt enabled after testing. I'll have to make a new upload disablint it. [08:41] mdz: you don't get reasonable upgrade behaviour just from that, actually [08:42] mdz: since the old kernel concrete package stays installed [08:42] Kamion: aptitude will remove it [08:42] which is why I think metapackage dependencies are necessary [08:42] but arguably it should stay installed, in case the new one breaks [08:42] what? that's insanity [08:42] well, the install kernel should always stay, since that was installed explicitly [08:42] but things pulled in by metapackage dependencies will be removed by aptitude when the dependency goes away [08:42] won't the kernel package stay, since the prerm bails out if it's the currently running kernel? [08:43] hmm, possibly [08:43] I bet that pisses aptitude off [08:43] anyway, speaking of kernels, it's time for a couple of reboots to test if -3 fixes my issues [08:43] brb [08:44] elmo_: linux-kernel-di-* uploaded, let me know when you've NEW-processed them since I need to change debian-installer at that point [08:47] Kamion: hmm, don't see them? [08:47] oh, right, being built [08:48] yeah === mdz [~mdz@69-167-148-207.vnnyca.adelphia.net] has joined #ubuntu-devel [08:49] yay, fixes the APIC issue with my system [08:49] mdz: been trying to fix the debconf noninteractive/seen thing; turns out to be amusing to fix, my current attempt causes the aa_DJ locale to be generated by default on new installations [08:50] Kamion: aa_DJ and only aa_DJ? or all of them? === justdave [~dave@24.247.63.44.gha.mi.chartermi.net] has joined #ubuntu-devel [08:50] (I'm relatively sure we don't have enough Afar speakers to merit that :-)) === justdave [~dave@24.247.63.44.gha.mi.chartermi.net] has joined #ubuntu-devel [08:51] mdz: only aa_DJ; the default implementation of the noninteractive select element sets the value to the first item in the list [08:51] and evidently I've slightly changed the behaviour there, not so unexpectedly ... [08:52] linux-restricted-modules-2.6.8.1_2.6.8.1.1-6_i386.changes [08:52] SKIP (too new) [08:52] Rejected: nvidia-kernel-source_1.0.6111-1ubuntu2_i386.deb: old version (1.0.6111-1ubuntu2) in warty >= new version (1.0.6111-1ubuntu2) targeted at warty. [08:56] oh EW [08:57] need to manually update debian/rules for every new version [08:58] uploading -7 [08:59] off to karate, back in a couple of hours [09:09] hmm [09:09] is it just me, or are linux-meta and linux-source now building some of the same binaries? [09:09] hi. [09:09] elmo_: I suppose that's something that ought to be fixed in short order? [09:18] mdz: it's not disastrous from my POV? [09:18] elmo_: which one will end up in Packages? [09:18] whichever is newer [09:18] if they're the same one will be REJECTed [09:18] fascinating. that happens to be the right thing in this case [09:19] er, no it doesn't [09:19] -meta is at 2.6.8.1-5, and -source is at -8 [09:20] elmo_: if I remove the metapackages from linux-source, will that do the right thing and let the -meta ones come out, or will it disappear? [09:21] mm, people are wanting to go for food [09:21] mdz: nothing will disappear automatically [09:21] elmo_: so I should upload linux-meta with a higher version number, I suppose [09:22] mdz: yeah, sounds like a plan [09:22] elmo_: ok, don't let me keep you. a thumbs up/down on that plan would be great [09:22] ok [09:26] bah i can't sleep === thom lends fabbione a large baseball bat [09:32] or a copy of the X protocol specs [09:32] i guess either would work [09:33] sounds icky [09:33] thom: ? [09:38] thom: dude.. what's going on? === Keybuk [scott@descent.netsplit.com] has joined #ubuntu-devel === vuntz [~vuntz@fennas.vuntz.net] has joined #ubuntu-devel [09:51] mdz: as I need to disable the colorprompt anyway, should #1758 (lesspipe) really be enabled by default in the profile? [09:52] doko: you need to disable the prompt why? [09:53] mdz: the ugly color prompt I left enabled by default [09:53] doko: oh, did it become enabled in your last upload or something? [09:54] yes, I accidentally left it on after trying. [09:54] eek [09:54] I see, it is only in the skel [09:54] doko: do you see any harm in enabling lesspipe by default? [09:55] in that case I would like to alias more to less, so that the user has the same experience with both commands. [09:56] that doesn't make sense to me; they have different user interfaces [09:59] for a beginner, they are about the same, but "more changelog.gz" and "less changelog.gz" should have the same experience. [10:01] I think that enabling lesspipe in less and aliasing more to less are entirely separate propositions [10:02] welcome to the can of worms of user profiles ;-) I'll add the lesspipe suggestion, but I would prefer to leave it commented out. [10:02] some programs can view compressed files, and some cannot (e.g., vim/emacs vs. gedit), but they should remain distinct because users expect to get what they ask for [10:03] ok, if you have doubts, please raise the issue for discussion on ubuntu-devel === Keyb [scott@descent.netsplit.com] has joined #ubuntu-devel [10:15] doko: please don't wait to fix the colour prompt bug; that must be fixed for the next daily CD build [10:16] fine, and for now I added the lesspipe line as well (commented out). [10:25] fabbione: either one would help you sleep ;-) === thom yawns and stretches [10:26] thom: ehehe i got it.. very late ;) [10:26] just converted firefox to dpatch === Keybuk [scott@syndicate.netsplit.com] has joined #ubuntu-devel === Keybuk [scott@syndicate.netsplit.com] has joined #ubuntu-devel [10:44] Kamion: thanks for fixing the po-debconf bug :), you should probably report something on #1329 [11:14] npmccallum: it was waiting for a sync, hadn't yet read the mail that the sync was done [11:14] I did update the status whiteboard on #1329 earlier :) [11:16] npmccallum: oh, bugger, no I didn't, I had a mid-air collision in bugzilla and forgot to follow through. anyway, closing now === azeem [~mbanck@socks-out.lrz-muenchen.de] has joined #ubuntu-devel [11:58] Kamion: how long before the new kernel propagates to a daily CD?