| somerville32 | bye | 00:40 |
|---|---|---|
| slangasek | superm1_: no | 01:19 |
| superm1_ | slangasek, i forget what exactly i messaged as i pinged you before. | 01:19 |
| superm1_ | slangasek, it was probably related to that mythtv sru | 01:19 |
| slangasek | superm1_: you asked me if I was still here | 01:36 |
| superm1_ | slangasek, ah okay | 01:36 |
| superm1_ | slangasek, well i was going to ask you about the sru then when you responded :) | 01:37 |
| === chuck__ is now known as zul | ||
| slangasek | and if it's about the sru, I'm really probably not here until tomorrow morning when I've had a chance to rest some. :) | 01:37 |
| superm1_ | slangasek, sounds good :) | 01:37 |
| superm1_ | slangasek, enjoy your evening | 01:37 |
| somerville32 | Hobbsee, I need some love. | 02:00 |
| somerville32 | Bug 160314 | 02:00 |
| ubotu | Launchpad bug 160314 in xfce4-session "xfce4-session: merge new Debian version" [Undecided,New] https://launchpad.net/bugs/160314 | 02:00 |
| Hobbsee | somerville32: oh? | 02:00 |
| somerville32 | :) | 02:00 |
| somerville32 | Hobbsee, Think you can help? :] | 02:03 |
| Hobbsee | somerville32: nope :) | 02:04 |
| somerville32 | Hobbsee, Why not? : *( | 02:04 |
| Hobbsee | because i'm doing uni stuff, and i'm only on irc to speak to some people | 02:04 |
| somerville32 | moogles. | 02:04 |
| BlaenkDenum | hey I'm wondering where I can propose something regarding the IRC channel, would the forums be a good place? | 02:32 |
| BlaenkDenum | I figured launchpad but then that's mainly for distribution problems I figure | 02:33 |
| Burgundavia | BlaenkDenum: irc channels are covered by the IRC council | 02:33 |
| Burgundavia | what is your issue? | 02:33 |
| PriceChild | BlaenkDenum, "the"... which channel? | 02:33 |
| BlaenkDenum | maybe the most popular, #ubuntu | 02:34 |
| BlaenkDenum | Burgundavia: oh haha there's an IRC council, where can I contact them? | 02:34 |
| Burgundavia | BlaenkDenum: what is your issue? | 02:34 |
| BlaenkDenum | Alright, well it's not really an issue I was just wondering if it's something to think about. | 02:34 |
| Burgundavia | instead of being obtuse, you could tell us what is on your mind | 02:35 |
| Hobbsee | #ubuntu-irc | 02:35 |
| Burgundavia | then we can help you a lot better | 02:35 |
| Hobbsee | Burgundavia: this still isnt the correct place :) | 02:35 |
| BlaenkDenum | I will, hold on I'm typing heh | 02:35 |
| BlaenkDenum | okay I'll go to #ubuntu-irc | 02:35 |
| PriceChild | Hobbsee, #ubuntu-ops you mean? ;) | 02:35 |
| Hobbsee | BlaenkDenum: use #ubuntu-irc. most of the develpoers are not ops. | 02:35 |
| Hobbsee | PriceChild: er, yes, that. | 02:35 |
| Burgundavia | Hobbsee: I have no idea what exactly he is saying yet | 02:35 |
| Hobbsee | Burgundavia: if ti's regarding the irc channel... | 02:35 |
| Burgundavia | Hobbsee: and he mentioned LP and the forums, thus I am still confused | 02:36 |
| Hobbsee | Burgundavia: as places to put his proposals, yeah. | 02:37 |
| * Hobbsee waits to see what it is | 02:37 | |
| somerville32 | Carefully BlaenkDenum, they might eat you alive :P | 02:40 |
| === bddebian2 is now known as bddebian | ||
| === mpt_ is now known as mpt | ||
| === elmo_ is now known as elmo | ||
| === TheMuso_ is now known as TheMuso | ||
| === shenki__ is now known as shenki | ||
| === popey_ is now known as popey | ||
| === blueyed_ is now known as blueyed | ||
| === tepsipak1i is now known as tepsipakki | ||
| viviersf | hi guys, does ubiquity copy files from the squashfs image only or the squash+unionfs ? | 12:01 |
| Mithrandir | viviersf: the squashfs, then a handpicked selection of items from the unionfs | 12:23 |
| Hobbsee | Mithrandir! | 12:23 |
| Mithrandir | hello little crazy Australian! | 12:24 |
| Hobbsee | hiya Mithrandir! | 12:24 |
| * Hobbsee is sad little crazy | 12:25 | |
| Hobbsee | Australian | 12:25 |
| Mithrandir | why are you sad? | 12:25 |
| Hobbsee | my boss is leaving :( | 12:25 |
| * Hobbsee notes that her underlings are *all* sad, and are saying she isnt allowed to go :P | 12:25 | |
| Mithrandir | heh | 12:26 |
| * Mithrandir ruffles Hobbsee | 12:26 | |
| * Hobbsee hugs Mithrandir | 12:26 | |
| Hobbsee | so we probably get a terror person, next | 12:26 |
| viviersf | Mithrandir, there a place one can define things to be copied form unionfs ? | 12:27 |
| Mithrandir | viviersf: you can set up a hook, sure. /usr/lib/ubiquity/target-config contains a list of hooks | 12:30 |
| viviersf | Mithrandir, heh okay, so i just put a scipt in that copies the stuff | 12:31 |
| Mithrandir | viviersf: that ought to work, yes. You might want to look at the bits there already to work out how it all fits together. | 12:31 |
| viviersf | Mithrandir, cool thanks | 12:32 |
| viviersf | Mithrandir, was this changed for gutsy ? in feisty it copied form all unionfs ? | 12:32 |
| Mithrandir | no, it's been the same way since dapper | 12:33 |
| viviersf | Mithrandir, in feisty i used to update packages from the live cd and install with the new versions. In gutsy it seems to revert everything | 12:44 |
| Hobbsee | morning cjwatson_ | 12:47 |
| Mithrandir | viviersf: if you got that, I think you would have dreamed it. | 12:47 |
| viviersf | Mithrandir, haha no i didnt. Maby there was a bug :P | 12:47 |
| Mithrandir | viviersf: somehow, I think we would have noticed. Just copying the unionfs would break a whole lot of things. | 12:48 |
| viviersf | Mithrandir, i dunno it worked. Its weird then | 12:49 |
| === zul_ is now known as zul | ||
| norsetto | stevenk: how is the battle against libtool going? | 13:15 |
| StevenK | norsetto: I will have my revenge! | 13:16 |
| Hobbsee | good luck with that. will you succeed? | 13:17 |
| norsetto | Hobbsee: the official score is libbtool 1 - stevenk 0 (see bug 139635 last comment) | 13:18 |
| ubotu | Launchpad bug 139635 in libgpg-error "[cryptsetup] library dependency in /sbin/cryptsetup" [Undecided,In progress] https://launchpad.net/bugs/139635 | 13:18 |
| Hobbsee | norsetto: pity. | 13:24 |
| manchicken | Hobbsee: Howdy :) | 13:56 |
| Hobbsee | hiya manchicken! | 13:57 |
| manchicken | How goes it/ | 13:57 |
| Hobbsee | manchicken: it goes, gnome style :) | 14:00 |
| === simira_ is now known as Simira | ||
| === \sh is now known as \sh_away | ||
| bddebian | Heya | 15:00 |
| tekteen | anyone here know what packages I can take off of the alt install cd for kubuntu? I need to make room. | 16:03 |
| BUGabundo | hya | 16:04 |
| tekteen | hi | 16:05 |
| BUGabundo | today I've add Hardy reps | 16:05 |
| BUGabundo | did a few (really few packages) updates | 16:05 |
| BUGabundo | found a little prob | 16:05 |
| BUGabundo | SoftwarePropertiesGtk didn't work! should I report it on LP? | 16:05 |
| warp10 | Hi all! | 16:05 |
| BUGabundo | is Michael Vogt here? | 16:06 |
| Hobbsee | BUGabundo: you want #ubuntu+1 | 16:09 |
| Hobbsee | and no, he's not | 16:09 |
| BUGabundo | thanks Hobbsee | 16:10 |
| highvoltage | Was anyone here at the Gobuntu sessions at UDS Boston? | 16:16 |
| popey | i was at one of them | 16:18 |
| === keescook_ is now known as keescook | ||
| === pedro is now known as pedro_ | ||
| fabbione | FYI: we are going to disable automatic package ACCEPT in launchpad for a few hours | 17:10 |
| === pedro is now known as pedro_ | ||
| Hobbsee | oh noes! | 17:10 |
| Mithrandir | fabbione_: as in, disable build-from-unpublished-source? | 17:18 |
| === chand is now known as chand[aw] | ||
| fabbione_ | Mithrandir: not sure.. ask pitti | 17:20 |
| fabbione_ | Mithrandir: we need to import the partner archive in LP | 17:20 |
| Mithrandir | ah | 17:20 |
| Mithrandir | pitti: ^^; are you disabling the publisher completely or just the build-unpublished-source bits? | 17:21 |
| pitti_ | Mithrandir: completely, just for safety | 17:22 |
| Mithrandir | ok | 17:22 |
| pitti_ | Mithrandir: b-f-a stays enabled for now | 17:23 |
| === pitti_ is now known as pitti | ||
| Riddell | carlos: spooky japanese https://translations.edge.launchpad.net/ubuntu/gutsy/+source/kde-guidance/+pots/guidance/nb/+translate | 17:26 |
| carlos | Riddell: https://bugs.edge.launchpad.net/rosetta/+bug/133315 | 17:31 |
| ubotu | Launchpad bug 133315 in rosetta "at least, four wrong language imports for gutsy" [High,Confirmed] | 17:31 |
| === fabbione_ is now known as fabbione | ||
| StevenK | pitti: libbsd-resource-perl should be okay to be promoted, it built everywhere | 17:51 |
| pitti | StevenK: it still needs at least a shallow review for MIR | 17:51 |
| StevenK | pitti: Yeah, okay, but that's you or me doing that? | 17:51 |
| StevenK | pitti: All I was planning on doing was crowbarring yada off of it | 17:51 |
| * StevenK beats bloody dexter with the crowbar | 17:52 | |
| warp10 | Hi all! | 17:54 |
| === rob1 is now known as rob | ||
| Joe_CoT | hey, the gutsy-security repo has an invalid signature. Has all morning. Anyone to report it to? | 18:17 |
| jdong | Joe_CoT: really? | 18:23 |
| jdong | I just applied security updates 1.5hrs ago | 18:23 |
| Joe_CoT | jdong, yeah. | 18:23 |
| Joe_CoT | W: GPG error: http://security.ubuntu.com gutsy-security Release: The following signatures were invalid: BADSIG 40976EAF437D05B5 Ubuntu Archive Automatic Signing Key <ftpmaster@ubuntu.com> | 18:23 |
| * jdong runs a 2nd update | 18:23 | |
| jdong | no error on apt-get update | 18:24 |
| Joe_CoT | really? hmm | 18:24 |
| jdong | sure you don't have an intercepting proxy or something upstream? | 18:24 |
| Kmos | i'm applying them without problem.. | 18:24 |
| Kmos | Joe_CoT: isn't a problem of your mirror? | 18:24 |
| jdong | Kmos: it shows sec.ubuntu.com | 18:25 |
| jdong | Kmos: somehow the traffic is getting corrupted on the way over to him.... | 18:25 |
| jdong | odd to say the least | 18:25 |
| Kmos | ah yeah =) is direct | 18:25 |
| === nicolai__ is now known as Kopfgeldjaeger | ||
| pitti | StevenK: ah, libbsd-r-p package looks so *delightfully* easy now | 18:48 |
| pitti | StevenK: mainified, thanks | 18:50 |
| pitti | StevenK: on a related note, gnutls13 FTBFSed again, probably because of the libtool-crackification in libgcrypt11 | 18:50 |
| StevenK | pitti: Thank you for the compliment. :-) | 18:54 |
| StevenK | pitti: I saw the FTBFS, I wanted to ask you about it. Shall I just fix and upload gcrypt11 and then we give back gnutls13 once more with feeling? | 18:55 |
| Chipzz | StevenK: weren't there plans to eliminate .la files from the packages several releases back? | 19:07 |
| slangasek | it's an uphill battle | 19:08 |
| Chipzz | looks more like a battle that never got out of the planning stage to me really ;) | 19:09 |
| StevenK | It's hard to kill .la files | 19:15 |
| StevenK | Since libtool is a pile of crap | 19:16 |
| geser | and some KDE apps need them | 19:16 |
| tekteen | can someone help me. My preseed file is not working. It does not seem to be answering any questions. I pasted it at http://paste.ubuntu-nl.org/43569/. | 19:16 |
| soren | tekteen: please don't ask questions and run away before you've given people a chance to answer. kthxbye | 19:20 |
| === pedro__ is now known as pedro_ | ||
| slangasek | soren: why? | 19:24 |
| * slangasek runs away! | 19:24 | |
| * soren grumbles | 19:25 | |
| * soren shakes his fist at slangasek | 19:26 | |
| * popey waves at soren | 19:40 | |
| soren | o/ | 19:40 |
| === ryu2 is now known as ryu | ||
| slangasek | mathiaz: http://lists.alioth.debian.org/mailman/listinfo/pkg-samba-maint | 19:48 |
| slangasek | bdmurray: are the Portland Cabal going to go in on Adilson's planning for airport transit? | 19:53 |
| StevenK | Heh heh. Portland Cabal | 19:53 |
| fabbione | FYI: uploads to LP will be processed automatically from now | 19:55 |
| bdmurray | slangasek: I've considered it | 20:06 |
| seb128 | keescook: the glib copy is pcre 7.2 | 20:09 |
| seb128 | keescook: the changelog has that | 20:10 |
| seb128 | * glib/gregex.c: define PCRE_ERROR_NULLWSLIMIT if it's not defined by | 20:10 |
| seb128 | PCRE, has PCRE 7.3 removed this definition. (#475474) | 20:10 |
| seb128 | not sure if that's a compatibility breakage or something not exported | 20:10 |
| slangasek | fabbione: hrm, what uploads to LP? | 20:10 |
| fabbione | slangasek: all of them.. we had the publisher in manual for a bit | 20:11 |
| fabbione | it's all good now | 20:11 |
| slangasek | oh | 20:11 |
| StevenK | pitti: Are you busy? | 20:13 |
| pitti | StevenK: currently typing a mail, but not particularly muc | 20:13 |
| pitti | h | 20:13 |
| StevenK | pitti: Shall I come find you and we can bitch at each other about gnutls and libtool? | 20:14 |
| slangasek | ugh, does gnutls still have .las? | 20:15 |
| StevenK | slangasek: It's worse than that. All three of us can talk about it if you want | 20:16 |
| slangasek | ok | 20:16 |
| StevenK | So as soon as pitti tells us where he is hiding, both of us can find him. | 20:16 |
| pitti | StevenK: main room | 20:17 |
| StevenK | Oh, duh | 20:17 |
| pitti | bdmurray: ok, mail is away | 20:17 |
| StevenK | I'm right behind you | 20:17 |
| StevenK | pitti: ^ | 20:17 |
| StevenK | slangasek: Come find us? | 20:18 |
| slangasek | StevenK: in a meeting | 20:18 |
| StevenK | slangasek: Ahh. pitti and I will discuss it. | 20:21 |
| slangasek | bdmurray: we probably need to get a plan together pretty soon, lest all the vehicles in town get filled up before Saturday? :) | 20:21 |
| broonie | c | 20:22 |
| broonie | Gah. | 20:22 |
| fabbione | is it normal that a stat(file, &sb) on a symlink that points to a dir will have S_ISLNK(sb.st_mode) set to FALSE and S_ISDIR set to TRUE? | 20:22 |
| zul | fabbione: gah? | 20:23 |
| fabbione | zul: stat(2) | 20:23 |
| zul | fabbione: ah.. | 20:23 |
| fabbione | zul: make a symlink pointing to a dir | 20:23 |
| fabbione | stat it | 20:23 |
| fabbione | check what it is via S_IS macros | 20:23 |
| fabbione | it tells me that it is a dir and not a symlink | 20:23 |
| fabbione | might be a bug or a feature.. just need to understand | 20:24 |
| lamego | fabbione, man fstat | 20:35 |
| fabbione | lamego: that's the same man page as stat.. | 20:36 |
| lamego | lstat() is identical to stat(), except that if path is a symbolic link, | 20:36 |
| lamego | then the link itself is stat-ed, not the file that it refers to. | 20:36 |
| lamego | no it is not | 20:36 |
| lamego | it has a clear statement regarding how slinks are treated :) | 20:36 |
| fabbione | lamego: read carefully what i said... man stat = man fstat | 20:36 |
| lamego | lstat() unlike stat() identifies links | 20:36 |
| fabbione | ok perfect | 20:37 |
| lamego | ok, whatever, it answers to your question :) | 20:37 |
| fabbione | that is not fstat.. i don't need that | 20:37 |
| lamego | fabbione, stat follows links, as per the developers man page | 20:37 |
| liw | violent agreement detected | 20:37 |
| fabbione | lamego: yeps right.. i missed lstat | 20:37 |
| fabbione | liw: no.. he is right for lstat.. we agree on that. | 20:38 |
| fabbione | liw: i don't agree that the man pages are different :) | 20:38 |
| lamego | you are correct | 20:38 |
| lamego | :P | 20:38 |
| fabbione | lamego: thanks tho | 20:38 |
| fabbione | :) | 20:38 |
| Riddell | bdmurray: bug 155784 | 20:43 |
| ubotu | Launchpad bug 155784 in qt4-x11 "[gutsy] /usr/lib/libssl.so missing" [Undecided,In progress] https://launchpad.net/bugs/155784 | 20:43 |
| === keir__ is now known as keir | ||
| pitti | seb128: new cdbs uploaded, works with libgcrypt11 | 21:17 |
| warp10 | pitti: Hi! Did you received my email? | 21:19 |
| seb128 | pitti: you rock! ;-) | 21:20 |
| pitti | warp10: no, seems I didn't; which mail you mean? | 21:21 |
| warp10 | pitti: I sent it 5 days ago, November 1st | 21:22 |
| norsetto | pitti: if I ask nicely, would you look at the fix for bug 155498 which is in gutsy-proposed since a month? | 21:33 |
| ubotu | Launchpad bug 155498 in rutilt "rutilt 0.15-0ubuntu5 crashes while applying a profile" [Medium,Confirmed] https://launchpad.net/bugs/155498 | 21:33 |
| pitti | norsetto: oh, in fact I looked at it yesterday, but there was something about it I didn't like | 21:33 |
| pitti | norsetto: give me a minute | 21:33 |
| norsetto | pitti: sure | 21:34 |
| pitti | norsetto: oh, right; it doesn't have approval from motu-sru | 21:38 |
| pitti | and it has a whole lot of changes in it, which don't look critical | 21:39 |
| pitti | so I want motu-sru to ack it | 21:39 |
| pitti | norsetto: also, it doesn't have a changelog which is useful | 21:39 |
| pitti | even less so for users, so it does not tell anything about the impact and severity | 21:39 |
| norsetto | pitti: ok, for the first point, that was motu-uvf at the time | 21:39 |
| pitti | oh, eww, I didn't know that | 21:40 |
| pitti | so that's replaced by motu-uvf team? | 21:40 |
| norsetto | pitti: but its ok, I subscribe motu-sru then | 21:40 |
| pitti | norsetto: no, please don't | 21:41 |
| pitti | norsetto: LP says that motu-sru is obsolete | 21:41 |
| pitti | norsetto: hang on a minute | 21:41 |
| pitti | norsetto: hm, seems that the MOTU SRU process recently changed to not actually require acks | 21:42 |
| pitti | norsetto: so, sorry, it's not your fault | 21:42 |
| pitti | but I don't like this at all | 21:42 |
| norsetto | pitti: I will make it only for that bug if necessary but its a pity to miss the other fixes, they really enhance the package | 21:42 |
| pitti | norsetto: but every change bears the risk of breaking it for current users | 21:43 |
| pitti | norsetto: only "severe regressions" or "loss of data" | 21:43 |
| norsetto | pitti: just between you and me, I'm using it since a month | 21:43 |
| pitti | norsetto: btw, with "I don't like this at all" I didn't refer to your patch, but to the process change of not needing acks from motu-sru any more | 21:44 |
| norsetto | pitti: sure :-) | 21:44 |
| pitti | the current process basically makes stable-universe fair game for breaking | 21:44 |
| StevenK | The point was more to trust MOTUs to make calls about what to fix. | 21:45 |
| pitti | well, the current mythtv in gutsy-proposed unapproved is a prime example of how *not* to do an SRU | 21:45 |
| StevenK | Who uploaded it? | 21:46 |
| pitti | "Rewrite the entire tv backend driver because it's better" | 21:46 |
| StevenK | Oh geez | 21:46 |
| pitti | and once we go down that route of never ack'ing uploads, I'm also afraid that too much effort will be spent fixing minor things in stables | 21:47 |
| pitti | which is really better spent on improving the current dev release | 21:47 |
| pitti | norsetto: ok, I let this through based on the fact that it matches the currnet policy | 21:49 |
| pitti | but I'll initiate a change to that policy | 21:49 |
| === d33p__ is now known as luisbg_ | ||
| norsetto | pitti: thanks, I really appreciate it | 21:49 |
| norsetto | norsetto hugs pitti | 21:49 |
| * pitti hugs norsetto | 21:50 | |
| * norsetto hugs stevenk too (he is in a hugging mood, so there) | 21:50 | |
| pitti | seb128: ok, we have fixed prepared for gnutls&friends | 21:55 |
| pitti | seb128: however, in the long run we need to convince Debian about this concept, or we need to use clean-la.mk by default in cdbs (right now it isn't) | 21:55 |
| pitti | seb128: I'm actually pondering doing the latter | 21:55 |
| seb128 | pitti: well, clean-la comes from Debian, so there is already people convinced there ;-) | 21:59 |
| seb128 | but right, should be discussed there | 21:59 |
| pitti | only amongst the gnome team, I guess | 21:59 |
| seb128 | I think Keybuk is not that happy with that and consider .la useful | 21:59 |
| pitti | what for? | 21:59 |
| seb128 | static building apparently | 22:00 |
| Keybuk | so, here's a theory for you ... | 22:00 |
| pitti | they just cause extra useless dependencies and other trouble | 22:00 |
| slangasek | seb128: the answer to that is to make sure a .pc file is provided when the .la is removed | 22:00 |
| Keybuk | if you don't want .la files | 22:00 |
| Keybuk | why do you compile with libtool at all? | 22:00 |
| seb128 | but I'm not sure that's true on linux nowadays | 22:00 |
| Keybuk | why not just use gcc? | 22:00 |
| pitti | Keybuk: because upstream build systems do? | 22:00 |
| Keybuk | upstream install the .la file | 22:00 |
| pitti | but TBH I think that dh_shlibdeps does a much better job | 22:00 |
| slangasek | no, libtool installs the .la file | 22:01 |
| seb128 | Keybuk: libtool install those | 22:01 |
| seb128 | that's a side effect | 22:01 |
| Keybuk | no, it's the primary effect | 22:01 |
| seb128 | upstream just wants to build their libs | 22:01 |
| Keybuk | the side effect is the installation of .so | 22:01 |
| pitti | and I'm not really concerned about having it working on a 20 year old VAX | 22:01 |
| Keybuk | libtool uses and manipulates .la files | 22:01 |
| slangasek | installing the .la file is an implementation detail of a poor implementation | 22:01 |
| Keybuk | so don't use the poor implementation | 22:01 |
| slangasek | don't have that choice | 22:01 |
| Keybuk | if you're using libtool, you should ship its .la files | 22:01 |
| slangasek | why? | 22:01 |
| pitti | right, we just hack around it ATM | 22:01 |
| Keybuk | if you don't want .la files, don't use libtool | 22:01 |
| slangasek | they're useless if you have a .pc file | 22:01 |
| Keybuk | it's easy not to | 22:02 |
| seb128 | Keybuk: that's not true | 22:02 |
| seb128 | what is the issue with not shipping those? | 22:02 |
| Keybuk | it breaks static linking | 22:02 |
| slangasek | seb128: none if you provide a .pc file instead | 22:02 |
| Keybuk | slangasek: not true | 22:02 |
| slangasek | .pc isn't broken; .la is | 22:02 |
| Keybuk | even with .pc files you need .la for static linking | 22:02 |
| slangasek | no, you don't. | 22:02 |
| Keybuk | yes, you do | 22:02 |
| slangasek | not if the .pc file isn't broken | 22:02 |
| Keybuk | the .pc file doesn't list the list of dependencies for the shared library | 22:02 |
| Keybuk | because if it did, you'd be removing those too | 22:03 |
| Keybuk | you need that list of dependencies for static linking | 22:03 |
| Keybuk | we fixed libtool years ago to ignore that list for dynamic linking | 22:03 |
| slangasek | er, no, you just need a version of pkg-config that supports the --static option | 22:03 |
| Keybuk | maybe the current Debian maintainer reverted my patches *shrug* | 22:03 |
| slangasek | Keybuk: except that libtool *still* has to recurse the actual .la files at build time | 22:03 |
| Chipzz | Keybuk: static linking tends to break for complex programs with lots of libs anyway | 22:03 |
| slangasek | so when the dependencies of your build-dependencies change, .la shows its true evil | 22:03 |
| Keybuk | slangasek: I have no issue with replacing libtool with something else | 22:03 |
| pitti | . o O { if that breaks static linking, so much the better :-P } | 22:03 |
| Keybuk | it's not difficult | 22:03 |
| seb128 | Keybuk: that's not something we want to do at a distro level | 22:04 |
| slangasek | and you told me that aspect of .la behavior couldn't be fixed without a compatibility break | 22:04 |
| Chipzz | back in the gnome 1.4 days I tried linking a program using orbit statically; broke horribly | 22:04 |
| Keybuk | Chipzz: yet we still ship .a files | 22:04 |
| Keybuk | seb128: we overwrite config.guess and config.sub for every package | 22:05 |
| Keybuk | adding a third file to that list seems trivial | 22:05 |
| slangasek | the only thing installation of .la files is legitimately needed for on Linux is static linking; there is a way to support static linking via .pc files without the side-effects of .la files. QED | 22:05 |
| Keybuk | slangasek: .pc has different problems ;) | 22:05 |
| Keybuk | the main one being they can only be installed in one place | 22:05 |
| Keybuk | so you can't (easily) parallel install different version | 22:05 |
| Keybuk | ...unless that's fixed, I'm not up to date | 22:06 |
| seb128 | Keybuk: how do other distros which don't ship .la do? that's not possible to do static linking on redhat? | 22:06 |
| Keybuk | seb128: ever tried static linking gtk? :p | 22:07 |
| pitti | Keybuk: but that's equally true of .la? /lib/libfoo.la doesn't work | 22:07 |
| Keybuk | *boom* | 22:07 |
| Keybuk | pitti: yes it does | 22:07 |
| Keybuk | assuming you compiled it for /lib | 22:07 |
| Chipzz | 23:01 < Keybuk> if you don't want .la files 23:01 < Keybuk> why do you compile with libtool at all? >> Several reasons actually, the prime one being "I didn't decide this as I didn't write the thing" | 22:07 |
| Keybuk | one of the usual complaints about libtool is precisely because it *does* support .la files by specific paths | 22:07 |
| Keybuk | Chipzz: the upstream maintainer almost certainly didn't decide either | 22:07 |
| Keybuk | most people only use libtool because automake demands it for shared libraries | 22:07 |
| Chipzz | second one: because it helps on other platforms than linux | 22:07 |
| pitti | Keybuk: we did, but libtool doesn't find them in /libs; we tried and it failed | 22:07 |
| Keybuk | Chipzz: we're Ubuntu *Linux* | 22:07 |
| Chipzz | Keybuk: I'm sure upstream cares about us being ubuntu linux ;) | 22:08 |
| Keybuk | like I said, upstream probably doesn't care about libtool | 22:08 |
| Keybuk | they got it as a side-effect of automake (which makes writing make files easier) | 22:08 |
| seb128 | right | 22:08 |
| seb128 | what upstream actually cares about is building and installing .so | 22:08 |
| Keybuk | yeah, and that's easy | 22:09 |
| Keybuk | just replace ltmain.sh with something that just calls gcc with the same arguments | 22:09 |
| Chipzz | Keybuk: anyway I think there's a difference between libtool being evil and .la files being evil | 22:09 |
| Keybuk | Chipzz: .la files are part and parcel about what libtool *does* | 22:09 |
| Chipzz | I'm not convinced of the former, but rather of the latter | 22:09 |
| Chipzz | Keybuk: libtool works fine without .la files | 22:09 |
| Keybuk | not true | 22:09 |
| Keybuk | I'm afraid | 22:09 |
| Keybuk | it works only for the commonest case | 22:09 |
| Keybuk | it's really a bit thick without them | 22:09 |
| Chipzz | try rm /usr/lib/*la and build something using libtool; it still builds fine | 22:10 |
| Keybuk | which is why I've continually resisted removing them | 22:10 |
| Keybuk | I *know* the bugs | 22:10 |
| Keybuk | actually | 22:10 |
| Keybuk | it doesn't | 22:10 |
| Keybuk | usual test | 22:10 |
| Keybuk | rm /usr/lib/*.la | 22:10 |
| Keybuk | have a package with a convenience library that links to something in /usr/lib | 22:10 |
| seb128 | do you have concrete example of things which doesn't work without those for people who don't know the bugs? | 22:10 |
| Keybuk | and an app that links to it | 22:10 |
| Keybuk | app will fail to link | 22:10 |
| Keybuk | (convenience libraries are arguably bugs in themselves, but people love them) | 22:11 |
| seb128 | what do you call "a convenience library"? | 22:11 |
| Keybuk | seb128: info libtool search for "convenience library" | 22:11 |
| seb128 | rm /usr/lib/*.la is the usual way jhbuild users get GNOME to build | 22:11 |
| Keybuk | what most people use when they spread code amongst multiple sub-directories in their package | 22:11 |
| Keybuk | yeah | 22:12 |
| Keybuk | GNOME are particularly well-behaved | 22:12 |
| Keybuk | it certainly never affects them | 22:12 |
| Keybuk | but it affects a random bunch of other crap | 22:12 |
| norsetto | pitti: err, I owe you an apology for bug 155431 ... can you pls. un-subscribe ubuntu-archive. Sorry :-( | 22:12 |
| Keybuk | I honestly can't remember, because I stopped seriously caring about this years ago ;) | 22:12 |
| ubotu | Launchpad bug 155431 in apr "documentation in /usr/share/doc/libapr1-dev missing" [Low,Confirmed] https://launchpad.net/bugs/155431 | 22:12 |
| seb128 | let's fix those then? | 22:12 |
| Keybuk | I came to the conclusion that the build chain support was utterly broken | 22:12 |
| Keybuk | and nobody had any desire to properly fix it | 22:12 |
| pitti | so, we could probably fix libgpg-error to ship the shlib in /lib *and* have a working .la file, but autoconf and libtool make this exceptionally hard | 22:12 |
| pitti | it seems that they actively make it hard to ship libs in /lib | 22:13 |
| Keybuk | I fully support replacing ltmain.sh with a very small shell script | 22:13 |
| Keybuk | so .la files cease to be a problem at all | 22:13 |
| Chipzz | Keybuk: one of the problems of .la files (which I'm sure you're aware of) is it totally defies any attempts of using -Wl,--as-needed | 22:13 |
| pitti | Keybuk: sounds good to me | 22:13 |
| seb128 | I don't know enough about the build system to have an opinion on that | 22:13 |
| Keybuk | Chipzz: patched in Debian years ago | 22:13 |
| seb128 | but if you are comfortable doing a such change | 22:13 |
| Keybuk | (when I maintained libtool) | 22:14 |
| pitti | norsetto: don't worry | 22:14 |
| Chipzz | Keybuk: yes, and if upstream doesn't use debians libtool and rolls a tarball you're still screwed | 22:14 |
| Keybuk | -- random tangent -- does anyone know how you tell OpenOffice what the individual text boxes on a slide master *mean* ? | 22:14 |
| Chipzz | (or you need to run relibtoolize) | 22:14 |
| Keybuk | Chipzz: autoreconf | 22:15 |
| pitti | (which tend to break more often than not) | 22:15 |
| === doko_ is now known as doko | ||
| Chipzz | autoreconf tends to be a chore to run, and it produces big patches which have to be reviewed | 22:16 |
| Keybuk | it only produces big patches if the upstream is done wrong | 22:17 |
| Chipzz | or with an older/different version than we use | 22:17 |
| seb128 | Chipzz: you can use the same version than upstream, just look to the Makefile.in | 22:18 |
| Keybuk | hell | 22:18 |
| Keybuk | just running "make" will always use the same version as upstream | 22:18 |
| Keybuk | (except where Debian deliberately change the upstream package to prevent all that useful functionality from working -- yay working around perceived problems) | 22:19 |
| seb128 | Keybuk: usually the libtoolize patch have an aclocal and an autoconf call | 22:19 |
| seb128 | anyway what do we argue about now? | 22:19 |
| Chipzz | Keybuk: we currently have autoconf2.13; what happens if for example upstream tarball was rolled with autoconf 2.12 ? | 22:20 |
| Keybuk | aclocal will just replace a file in m4 | 22:20 |
| seb128 | Chipzz: let's stop there | 22:20 |
| Keybuk | like I said | 22:20 |
| Keybuk | I don't really care about any of this | 22:20 |
| seb128 | most of us know the issues with the current system | 22:20 |
| Keybuk | it's all broken, and made worse by people doing workarounds rather than fixing the system | 22:20 |
| Keybuk | do what you want ;) | 22:20 |
| Keybuk | but don't complain to me when you break it <g> | 22:20 |
| seb128 | fair enough | 22:20 |
| slangasek | Keybuk: a convenience library referencing the .la files in /usr/lib is precisely a symptom of *why* we shouldn't ship the .la files :) | 22:21 |
| Chipzz | seb128: that actually was a genuine question; but whatever ;) | 22:21 |
| slangasek | if they weren't present when the convenience lib was built, the application will also link just fine | 22:21 |
| Keybuk | slangasek: no, it won't | 22:22 |
| slangasek | huh? | 22:23 |
| slangasek | if there was no .la file in /usr/lib, the convenience lib's .la will only reference the libraries, not the non-existent .la files | 22:23 |
| Keybuk | not true | 22:23 |
| Keybuk | but I can't be arsed to argue anymore | 22:23 |
| slangasek | you're claiming that the convenience lib's .la file *will* reference .la files that never existed? :P | 22:24 |
| Keybuk | no | 22:24 |
| slangasek | then what? you're saying that you can't have a convenience lib .la without having .la files for all the underlying libs? | 22:26 |
| gaspa | seb128: ping | 22:34 |
| gaspa | are you working on pygoocanvas? i should work with it, so i can take a look for the merge/sync of it. | 22:34 |
| seb128 | gaspa: no, I'm not, you are welcome do to the merging, feel free to ping me if you need sponsoring then | 22:36 |
| gaspa | seb128: ok, thank you ;D | 22:42 |
| seb128 | gaspa: no problem | 22:43 |
| LaserJock | seb128: are you generally ok with people merging gnome packages? | 22:44 |
| LaserJock | seb128: or are there ones that you need to handle yourself | 22:45 |
| seb128 | LaserJock: https://wiki.ubuntu.com/DesktopTeam/TODO | 22:49 |
| seb128 | LaserJock: feel free to claim any update or merge there | 22:49 |
| seb128 | LaserJock: just put a link to the .dsc or to a bug with the informations | 22:50 |
| === Fujitsu_ is now known as Fujitsu | ||
| === asac_ is now known as asac | ||
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!