[12:01] Keybuk: showed up on #d-release, I think. FWIW, I don't think libtool not being happy with ~ in upstream versions (the only part that appears in the directory name, after all) is a big deal for adoption, and I think it's already been claimed that the problem is fixed in current libtool. [12:01] rofl, it isn't fixed [12:01] keybuk: good question, on #d-d or #d-r last week? [12:01] jdub: hmm, what's esd getting replaced with? [12:02] Keybuk: ok, then fix it already. ;) [12:02] and wouldn't ~ be most common in the upstream version? [12:02] yikes, when did vorlon join :-) [12:02] thom: I have always been here. [12:02] hey dude [12:02] vorlon: can't, it's kinda core to the way libtool works [12:02] damn. [12:02] ;) [12:02] (libtool uses IFS="~" internally, basically) [12:02] *giggle* [12:02] vorlon: please don't start referring to mark as "the arrow" [12:02] that would just be disturbing [12:03] Keybuk: ah, well, whoever was noting it actually did suggest a patch, I apparently was under the impression it came from upstream [12:03] thom: [12:03] it'd be far easier to fix dpkg to use a less silly character [12:03] we're out of less silly characters. [12:04] though you're right about it being more likely to appear in upstream version #s (duh), but even so, that's a per-package implementation problem, hmm? :) [12:04] hmm, `=' doesn't seem to be used much. [12:05] doko: makes it more fun to parse version specs in apt-get install commands. :) [12:05] apt-get install ssh=4.0=rc5-3? [12:05] $version =~ m/[^-+:.0-9a-zA-Z~] / [12:05] ^ is the current regexp === AndyFitz [~andy@203-173-20-190.dyn.iinet.net.au] has joined #ubuntu-devel [12:07] make the version string utf8 encoded and we don't have the problem anymore. utf8 is a hoary goal anyway ;) [12:07] I still like the bug that proves nobody *really* uses the hurd port [12:07] (dpkg can't actually cope with "-" in arch names :p) === vorlon chuckles [12:07] let alone any of Millan's toy-of-the-week ports [12:08] keybuk: we should continue this discussion on #debian-glibc, jeff's currently there :) [12:09] vorlon: there's actually remarkably many silly characters left to be abused *shrug* [12:09] Keybuk: few that aren't already used somewhere else that would make it irritating to disambiguate and/or escape strings. [12:10] Keybuk: face it, it's libtool's fault. ;) [12:10] heh, half the archive tools don't even support ~ at the moment; so it's a bit of a dead duck usefulness wise [12:11] oh, I agree; I've waxed lyrical repeatedly that libtool should use IFS="\0" internally which would also fix the space problem [12:11] well, the archive tools don't need to support it until post-sarge anyway. [12:11] vorlon: the ones in sarge, like apt-ftparchive do [12:11] (though I think that's one of the few that does support it) [12:12] meh, it'd be nice if they would, but the only real requirement is that *our* archive be able to handle it before they're allowed into Debian. [12:12] (IMHO) [12:12] it would be nice if we could work out a way of removing this restriction that dpkg features need to be in a stable release first [12:13] it basically leaves amazingly critical things untested for too long [12:13] Solution: we use Ubuntu as a guinea pig for everything. [12:13] like the bz2 support, I could merge it in but it's just not tested enough for a stable release [12:13] we have the same problem [12:14] oh well, it was worth a try ;) [12:14] i don't mind being a guinnea pig [12:14] as long we we get it in early in the cycle [12:14] like, bz2 should go in immediately hoary opens up [12:15] well, the real solution would be to have an ubuntu derivative called "debian madness" or somesuch [12:15] ununtu? [12:15] and have all the daft ideas built in that on a weekly release cycle [12:15] hrm, I can't find that thread :-( [12:15] thom: you *know* that would be a magnet for Gentoo users... :) [12:15] vorlon: sure. and when they grow up, they use ubuntu [12:15] "Break My Buntu" === vorlon grins [12:15] it's a win alround :-) [12:16] because "break my ubuntu" is impossible for non-south-africans to say without dying of asphixiation in the middle [12:16] question is whether the Gentoo users would actually give useful feedback for sifting out the buggy from the non. :) [12:16] we should find some xhosa word with a nice glottal stop in [12:17] vorlon: that would be the problem === sabdfl [~mark@host217-37-231-28.in-addr.btopenworld.com] has left #ubuntu-devel [] [12:17] thom: and then we could use the glottal stop as a separation character in libtool. [12:17] sabdfl: do guinnea pigs have something to do with guinnes ;) [12:18] thom: http://www.ubuntuforums.org/viewtopic.php?t=442 [12:18] vorlon: some idiot would try to stick it in a filename [12:18] thom: gentoo is still the former distribution of choice for all the webforum people === vorlon scrubs the filenames in his homedir so Keybuk doesn't see [12:19] actually, I've thought of a bright side [12:19] vorlon: nice thought. i believe "xh" is the written form of a glotal stop for xhose tho [12:19] xhosa, even [12:19] daniels: ehh, no suse switcher at all ... [12:19] night, guys! [12:19] heh, some 'How did you find about Ubuntu?' thread [12:19] you don't *really* need to build packages in a directory named source-upstream [12:19] tempting to reply th that one [12:19] if you don't, dpkg-source will bitch, but make the .diff.gz anyway [12:19] but for everyone else, it'll still be unpacked in a directory named source-upstream [12:20] oh, wait, no === Keybuk derails his chain of thought [12:20] (was thinking it'd stop people shipping packages that ran libtoolize ... but it's libtool itself that's broken) === |trey| [~trey@ip68-230-75-109.ph.ph.cox.net] has joined #ubuntu-devel === thom ambles to bed [12:24] g'night === daniels blinks. [12:24] thom: er, night [12:25] daniels: why surprised? [12:25] thom: think of the kitt^Wtimezones [12:25] not surprised, it's just a little odd, is all [12:26] 23:25 [12:26] yeah === sabdfl [~mark@host217-37-231-28.in-addr.btopenworld.com] has joined #ubuntu-devel [12:38] elmo: around? [12:40] sabdfl: yeah [12:40] how big is the releases only dir of cdimage? [12:40] rsync://cdimage.ubuntu.com/cdimage/releases [12:40] gah, only 60kB/sec from auckland; still another hour to go for desktop [12:41] 1.7Gb [12:41] and with hoary that would double? [12:41] sabdfl: yeah [12:41] thanks elmo [12:42] vorlon: polypaudio [12:42] you've been fantastic through the push to release, hope you get some rest over the w/e! [12:42] sabdfl: thanks, me too [12:43] is there a way for them to mirror both packages and release install cd's and live cd with a single command? [12:43] (but pref not daily cd builds, just releases) [12:43] sabdfl: no, unfortunately not [12:43] +yet [12:44] ok, we'll work on that in due course [12:44] for the brave mirrors out there [12:45] gah, no clue where this bloody acpi bug came from === sabdfl [~mark@host217-37-231-28.in-addr.btopenworld.com] has left #ubuntu-devel [] === AndyFitz [~andy@203-173-20-190.dyn.iinet.net.au] has left #ubuntu-devel [] === hazmat [~hazmat@c-24-15-10-12.client.comcast.net] has joined #ubuntu-devel [01:13] daniels: do you sleep also times or are you 24/7 awake ;) === Riddell [jr@muse.19inch.net] has joined #ubuntu-devel [01:29] mdz: liveCD is progress, of sorts... [01:29] did y'all want X? :-( [01:30] lamont: please talk to alex about it [01:30] he's on jabber [01:30] awesome === lamont points mdz at the other window [01:31] then again, that's probably on the wiki [01:32] yep [01:37] err [01:37] lamont: I get a Morphix splash instead of an Ubuntu splash [01:43] that too === AndyFitz [~andy@203-173-20-190.dyn.iinet.net.au] has joined #ubuntu-devel [02:30] hehe === lamont [~lamont@mix.mmjgroup.com] has joined #ubuntu-devel [03:31] so why is it that I have to 'sudo nautius' to get a window? [03:31] nautilus, even === lamont goes in search of bandwidth === jamesh [~james@203-59-213-254.dyn.iinet.net.au] has joined #ubuntu-devel [03:35] lamont: I like that name, gotta find a use for it [04:16] heh, nautius [04:17] it was called naughtylus a lot when it was still 1.x ;) [04:18] ahh, eazel === lamont_r [~lamont@c-67-174-120-66.client.comcast.net] has joined #ubuntu-devel [04:23] mdz: the KernelHowto will have all the different options of building a kernel [04:23] I have gotten requests to show the different ways [04:23] hornbeck: great [04:24] mdz: what is there right now was what I got done at work tonight [04:24] hornbeck: I didn't intend criticism; I wasn't sure of your plans [04:25] I know, that is why I am letting you know === ..[topic/#ubuntu-devel:mdz] : Ubuntu development -- general discussion on #ubuntu | Happy Ubuntu Artwork week! | 5 (count 'em) major bugs [04:28] mdz: what do you think of the mono-on-buildhost situation [04:29] tseng: hard to say, considering I've no idea what that situation is :-) [04:29] mdz: well, most of the mono packages are failing [04:29] because mono-mcs has circular deps [04:29] I see [04:29] lamont: ping? [04:30] lamont should be able to take care of that [04:30] if he isn't around, send him an email [04:30] basically it needs the compiler to build parts of mono.. i installed them from sid to start out [04:30] 0.96 was problematic to bootstrap.. do we have sources somewhere to insert in to the archive? [04:30] there is a howto on bootstrapping [04:31] are we trying to bootstrap 0.96, or the new stuff from you tseng? [04:31] hm, no one is trying to bootstrap anything atm [04:31] but im looking at your build logs since we synced up my newer stuff [04:31] tseng: let me rephrase that... what are you trying to convince me to bootstrap... [04:32] so the new stuff is in our archive, and just needs to get helped through the binary circular dep-wait hell? [04:32] lamont_r: well, either bootstrap mono 1.0.1, or just install the bins [04:32] lamont_r: right. [04:32] mdz: ping [04:32] could you email lamont@canonical.com a pointer to the howto, and I'll do that in the morning. right now trying to focus on the *&%)^_ livecd [04:32] will do === warthylog [~warthylog@port49.ds1-van.adsl.cybercity.dk] has joined #ubuntu-devel === Topic for #ubuntu-devel: Ubuntu development -- general discussion on #ubuntu | Happy Ubuntu Artwork week! | 5 (count 'em) major bugs === Topic (#ubuntu-devel): set by mdz at Fri Oct 15 04:25:21 2004 [05:56] (mdz/#ubuntu-devel) it was 32C here today === Keybuk considers moving [05:58] you'll just have to unlearn a new timezone ;) [05:59] 21C here. [05:59] jdub: why should I care what the local timezone is? :p [05:59] so you can buy stuff [05:59] ;) [05:59] you don't have Internet Shopping? [05:59] then they have to wake you up to deliver it [05:59] some of the supermarkets used to do internet shopping [05:59] and 24-hour Wal-Mart ? [06:00] none of the convenience stores here are 24 hour, because it's "too dangerous" [06:00] here being a ten or so block radius [06:02] jdub: wtf? [06:02] STREETS OF REDFERN [06:03] see, you should move to Melbourne, because every Sevvo is 24h [06:03] and there are a fair few 24h supermarkets also [06:03] daniels: yeah, it's totally the redfern thing. everywhere else is fine. [06:03] jdub: wow, that's wack [06:03] i'm going to have to think about an air conditioner or something, if this is any indication of what summer's going to be like. :| === Keybuk wonders whether it's a name thing [06:04] most dodgy places seem to be [06:04] COCKFOSTERS -> not dangerous [06:05] so gold coast -> not dangerous? === AndyFitz will be in melbourne next week from 18th - 27th [06:06] monash uni is flying me down and back. what champs [06:06] jdub: see, this is why we need to rename Redfern to Cockfosters, remember? [06:06] no more negative stigma [06:06] Keybuk: try Eaglehawk, Bendigo, Ballarat, Melbourne :P [06:07] see, those names are just made up [06:07] like Wallabarralooloo [06:07] Keybuk: dangerous places around here are Frankston (Franger), Broadmeadows (Broady), and ... well, that's about it [06:07] so i was wandering around the afternoon of election day [06:07] Keybuk: do you mean Warrnambool or Woolloomooloo? [06:07] jdub: did you take a knife? [06:07] and this dude was saying, "haven't seen any of those f*ckin blacks out here!" [06:07] in Perth, people say that Northbridge is a bit dodgy. [06:07] jamesh: so do most geeks [06:08] which intrigued me, so i slowed down to listen [06:08] the bridge that northbridge is north of is quite interesting though. [06:08] must've been the only two old gizzers brave enough to be racist pigs in this area ;) [06:09] intel make binary graphics drivers too now? [06:09] the area I live in is primarily Asian [06:10] Keybuk: when you say asian, do you mean indian too? [06:10] jdub: yeah, Pakistani and Indian [06:11] mmm, that's always sounded weird to me [06:11] why? [06:11] asian, in .au, pretty much means CJK, SEasian, etc. [06:11] see, I'd call them Oriental [06:11] indians are called indians [06:12] definitely not indian (they're just called indians, much to the annoyance of the pakistanis) [06:12] uh, not exactly annoyance of Pakistanis ... I wouldn't *dare* risk calling a Pakistani an Indian [06:12] these are two countries that were nearly at nuclear war with each other only a year or two ago [06:12] people have been stabbed for less [06:12] a friend was booted out of a taxi for talking to the pakistani driver about indian stuff [06:13] indeed [06:13] at least with Asian, they don't kill you [06:13] but that's kind of a stand-out case here, as far as i'm aware [06:13] jdub: holy fuck, yeah, being racist in Redfern wouldn't last long [06:13] jdub: Britain, especially the midlands, has a very large Pakistani and Indian population [06:13] Keybuk: in .au, they'd seriously look at you oddly, unless they were UK "Asian" themselves ;) [06:13] largely from independance [06:14] Keybuk: asian -> SE Asian, North Asian/subcontinental -> Indian/Pakistani/Sri Lankan [06:14] Asian -> Indian/Pakistani; Oriental -> Chinese/Japanese/etc. [06:20] another interesting thing I noticed in the two weeks in London; when out in an Indian restaurant, I'll only eat with one hand [06:20] wherease the aussies just shovel in with both :p === fabbione [~fabbione@port49.ds1-van.adsl.cybercity.dk] has joined #ubuntu-devel [06:21] YES! === hornbeck [~hornbeck@adsl-68-92-227-78.dsl.okcyok.swbell.net] has joined #ubuntu-devel [06:32] fabbione: :) [06:32] Keybuk: huh? we use one for the knife and one for the fork ... [06:32] it's not like Aussies have some form of innate double-handed conveyor belt technique :P [06:32] heh [06:32] you'd be kicked out of my local balti for eating like that :p [06:36] which is good, because I don't eat like that :) [06:45] I mean for eating with a two hands in general :p === AndyFitz [~andy@203-173-20-190.dyn.iinet.net.au] has left #ubuntu-devel [] === lamont_r burns a test disk === lamont_r reboots === lamont_r [~lamont@c-67-174-120-66.client.comcast.net] has joined #ubuntu-devel === lamont__r [~lamont@c-67-174-120-66.client.comcast.net] has joined #ubuntu-devel [07:34] well. that was painful. [07:34] I think I'm going to go back home and send some email [07:35] fabbione: moved? === mdz reads other windows [07:36] fabbione: so I guess the new DSL is working [07:38] yeps [07:38] 100% [07:38] i am checking all the machines and stuff [07:39] mdz: anything special you want me to focus on for today? [07:44] fabbione: after a good night's sleep, and all of my high-priority issues fixed, I would like to revisit X [07:45] fabbione: the two autodetection fixes we talked about can go in [07:45] mdz: ok [07:45] mdz: anything particular that needs revisiting? [07:45] I am still unsure about xkb [07:46] fabbione: which autodetection fixes rae they? [07:46] daniels: the MAXRES and the res @60Hz problem [07:47] mdz: Mark also asked to change the default X background [07:47] daniels: do you have the patch ready? [07:47] fabbione: oh? [07:47] fabbione: i'm still unconvinced on not using the whitelist -- this will break stuff [07:47] fabbione: which patch, 099l? [07:47] mdz: it's to change the black/white points to something more sane [07:48] or the colour patch? [07:48] daniels: the color patch [07:48] I don't follow [07:48] mdz: when X starts up [07:48] you see the black and grey screen [07:48] currently it is black [07:48] fabbione: i still don't have the colour code [07:48] black.. it depends [07:48] mdz: the proposal is to change this to the base gdm colour [07:48] for me it takes longer and i can see the spots [07:49] interesting [07:49] we can change it to the default gdm background [07:49] that's Mark request [07:49] the patch is definetely safe [07:49] i don't think you should see the stipple with -br [07:49] so he #76848F [07:49] er [07:49] so he wants you to change it to #76848F? [07:49] that is the solid colour background used in gnome [07:50] altough it probably initialises the root window with MakeRootTile, then bzeros the pixels or something equally crackful [07:50] mdz: would you be ok with this? [07:50] mdz: no, #e1ca96 [07:50] jdub: what's that? [07:50] the gdm base colour [07:50] of the gdm theme [07:50] anyway, if Mark asked you to do it, who am I to argue? :-) [07:51] mdz: the Release manager? ;) [07:51] fabbione: er no, he's the CTO [07:51] jdub is the release manager [07:51] mdz: ok let's revisit the X changes again. one by one [07:51] fabbione: did he say "I want it in if it's OK with mdz" or "I want it in"? :-) [07:51] ok, I'll prepare a base tile patch [07:51] mdz: the latter [07:51] http://people.ubuntulinux.org/~fabbione/boom [07:52] it seems insane to me that we should need to patch X to change that, but... [07:52] mdz: it's hardcoded :( [07:52] mdz: insane != surprising [07:52] ok hold on [07:52] i need more coffee before we start [07:52] welcome to 1975, we have these things called "command-line options" [07:52] i woke several hours ago to move the office [07:52] mdz: either the root tile gets initialised to the grey/black stipple (hardcoded), or the pixels get initialised to zero (-br) [07:52] no, I'm not surprised either [07:53] the stipple is hardcoded [07:53] so someone went to the trouble of creating a command-line option [07:53] to select a different hardcoded pattern [07:53] YAY X === jdub doesn't really know why we need to choose something other than black, anyway. [07:53] mdz: not as heinous as ddxLoad, dude [07:54] black is great [07:54] means we won't need to patch it next time we change theme... [07:54] mdz: you're just hatin' on me 'cause I'm white [07:56] wtf [07:56] evo sans ldap support? [07:57] my brain just parsed "sans ldap" as a font [08:04] re [08:04] the electrician is giving me some extra plugs for the office [08:05] mdz, jdub: should we go item by item again? [08:05] fabbione: I do not want to update v4l at this time [08:06] autodetection stuff is OK [08:06] ok [08:07] the xkb changes are very difficult to judge because it is black magic [08:07] yes [08:07] i can't pass an opinion on the xkb stuff other than anecdotal reports from others of works vs doesn't work [08:11] so what should we do with XKB? [08:14] let me look at the diff [08:14] diff [08:14] yeah well people.u.o/~fabbione/diff [08:15] I already have a copy [08:15] ok [08:15] it is hard to read though because it is patching patches [08:15] not enough context [08:15] downloading source now [08:15] mdz: ok [08:15] it changes a ton of layouts [08:15] the diff doesn't contain the fix to the MAXRES -> $MAXRES [08:16] but i have it here locally [08:16] 099l_xxf86vm_increase_verbosity actually decreases the verbosity by increasing a macro called DEFAULT_XF86VIDMODE_VERBOSITY === mdz scratches his head [08:16] mdz: not quite [08:17] mdz: the verbosity level is 1 per default for console, 2 per default for logs (-verbose and -logverbose, respectively) [08:17] mdz: the way this does it (non-intrusive as possible) is it only prints that message via ErrorF() if the verbosity level is > DEFAULT_XF86VIDMODE_VERBOSITY [08:17] and apparently higher values mean less verbose [08:17] and it defaults to 3 [08:17] no, higher values mean more verbose [08:17] so this patch only prints the strings if verbosity > 3 [08:17] and since it is 1/2 per default ... [08:18] there is a better solution to this, but it's slightly more intrusive [08:18] so which part of what I said was incorrect? [08:18] it does make it less verbose ("decreases the verbosity") [08:19] yes [08:19] the macro is probably unfortunately namede [08:19] the correct fix is to use xf86DrvMsg() for everything, and use X_ERROR/X_WARNING/X_INFO/X_DEBUG as standard defined levels, but that's less intrusive [08:20] since it already has the if, just with stupid magic numbers (yay cruft) [08:20] mdz: become one with the crack [08:20] mdz: (it helps, seriously) [08:24] mdz: v4l update killed [08:30] 099j_xkb_new_layouts.diff seems to only change ca, hu, mn and vn [08:32] yes [08:32] 099z_xkb_fix_rules_xfree86.diff I do not understand [08:32] it fixes ca and ca_enanched [08:33] the latter is the one that makes some multilayout crap working [08:33] Kamion: 1659; I never got a reply from you. [08:33] 099z_xkb_level3_ralt_switch.diff I do not understand either [08:34] mdz: if i could explain that patch i would be able to understand XKB :P [08:34] isn't there a bug in our bugzilla about ca_enhanced or something? [08:35] mdz: yes. there was a note in a bug but i couldn't find it anymore [08:35] 1089 [08:35] ETOOMANY XKB reports [08:35] hm, ralt_switch_multikey vs ralt_switch [08:36] I don't see how this fixes ca_enhanced [08:36] the bug report says: Error: Can't find file "pc/ca_enhanced" for symbols include [08:36] Symlinking /etc/X11/xkb/symbols/pc/ca_enhanced to [08:36] /etc/X11/xkb/symbols/ca_enhanced fixes the problem. [08:36] diff -ruN xc-old/programs/xkbcomp/symbols/pc/us xc/programs/xkbcomp/symbols/pc/us [08:36] --- xc-old/programs/xkbcomp/symbols/pc/us2004-10-08 22:36:37.000000000 +0000 [08:36] @@ -1174,7 +1172,7 @@ [08:36] + key { [ slash, question, questiondown, dead_hook ] }; [08:36] + key { [ backslash, bar, notsign, brokenbar ] }; [08:36] + [08:36] -+ include "level3(ralt_switch_multikey)" [08:36] ++ include "level3(ralt_switch)" [08:36] mdz: the symlink is plain wrong [08:37] mdz: i did ask Denis in private and Branden confirmed that [08:37] the layout needs to be done properly [08:37] I can do setxkbmap ca_enhanced fine [08:38] with current Warty X [08:38] mdz: the problem is not when you load ONE layout [08:38] the problem is in multilayout environments [08:38] how can I test it on the command line? [08:38] no idea [08:38] hm [08:38] multikey is apparently compose-like behaviour [08:38] so I would posit that changes the behaviour of right alt from a level 3 switch to a multikey level 3 switch [08:39] e.g. altgr-' (or whatever it is) would no longer be @ on a gb keyboard [08:39] based on that and my tenuous understanding of XKB, I believe the patch to be correct [08:40] but I disclaim that belief as much as possible [08:42] so in #1089, Denis said basically that the status quo is correct [08:42] unless we convert things from old to new format, which we aren't [08:43] the problem is that we cannot verify anything [08:43] this patch could completely break the hu, vn and mn layouts and we would have no idea [08:43] unless some user reported it [08:43] yes and gnome people will say that is not true and that X is at fault [08:44] mdz: not really. this patch completes the layout [08:44] and considering that probably 5 people use those layouts, and 1 of them uses ubuntu, that seems unlikely [08:44] mdz: they were first imported in u23 [08:45] u23 which was uploaded october 4th [08:45] mdz: i think we should just throw our hands up and wontfix that one; there's really nothing we can do [08:45] or ubuntu24 [08:45] let me check [08:46] i am not a super robot :P [08:46] why not?? [08:46] mdz: u24 [08:46] ok, so it is fixing a layout that did not even exist until 1 week ago [08:47] it's completing the fix for a layout that was supposed to be there and it wasn't in the beginning [08:48] I'm OK with 099j_xkb_new_layouts.diff [08:48]