IRC Logs for #crux-devel Tuesday, 2008-07-22

*** sepen has quit IRC02:03
*** mike_k has joined #crux-devel02:09
*** sepen has joined #crux-devel03:48
*** mike_k has quit IRC04:21
*** fetid has quit IRC04:36
*** Rotwang has joined #crux-devel06:53
*** aon has quit IRC07:47
*** aon has joined #crux-devel07:47
*** treach has joined #crux-devel08:04
*** fetid has joined #crux-devel08:31
*** treach has left #crux-devel08:48
*** Rotwang has quit IRC08:50
*** mike_k has joined #crux-devel10:14
*** sepen has quit IRC10:17
*** Rotwang has joined #crux-devel10:35
*** fetid has quit IRC10:59
*** treach has joined #crux-devel11:05
*** jue has joined #crux-devel12:06
cptnhey jue12:22
*** sepen has joined #crux-devel12:39
*** namenlos has joined #crux-devel12:45
cptnI guess we're complete for now12:58
cptnwelcome namenlos :-)12:58
cptnand sepen, for the first time in his new role as opt maintainer12:58
sepenyeha, many thanks guys12:59
namenloshi, sorry for being so absend the last time...12:59
cptnnamenlos: are you missing training now?12:59
namenlosmy actual working place is too far away from my training place.13:00
cptnwhen would you be back from training?13:01
cptnor are you too exhausted after that?13:01
cptnsince nipuL is also only joining around 22:30-23:0013:02
cptnalthough that's of cause even an hour later for aon13:02
namenlosif go on my motorbike i am at home at ~2145, if i take public transport, i am at home ~ 224513:04
namenlosor 230013:04
cptnokay; we can maybe discuss the time again some other time13:06
cptnlet's look at bugs:13:06
cptnthere's only one which is new I think:13:07
tilmanit's not a trivial fix unfortunately13:08
cptnthat's maybe something maintainers have to keep in the back of their minds for now13:08
tilmanbasically i think we'll have to drop ARCHIVE_UNLINK when extracting13:08
tilmanand unlink *files* ourselves13:08
tilmanwell, that's not rocket science of course, but i would have liked it to be a one liner :D13:08
cptnmight upstream care?13:08
juehmm, that's not a maintainer problem IMO13:09
cptnjue: no, but as long as there's no fix, maintainers should take extra care13:09
cptntilman: have you ever dealt with libarchive upstream?13:10
tilmanof course13:10
tilmankientzle is a nice guy, i'll ask him about this13:10
cptnokay, thanks13:10
cptnthen, on to
cptnnot sure if everyone saw it, but jue moved the FAQ to the Wiki already13:11
cptnthanks a lot13:11
juemom, #27013:11
tilmani didn't get to that last week13:12
juetilman: did you do something in your last udev commit13:12
tilmani just bumped udev13:12
* jue sees romster bumping as well13:13
sepenwell FS#323  the patch sent by Predrag can't be applied, we need nspr.pc13:13
tilmansepen: can you try that?13:14
juewe should file a bug at mozilla baout that13:14
tilmanjue: adding nspr.pc seems to be the right thing to do anyway. xulrunner.pc references it13:14
jueyep, that strange13:15
juethey don't supply one but need it13:15
jueit's kinda like the sqlite3.pc thing13:16
sepensee diferences between libxul.pc and libxul-embedding.pc, libxul requires it13:16
tilmansepen: can you try the patch i linked and see whether that makes mplayer-plugin compile?13:16
sepenyeah, need some time13:16
tilmancptn: can you create a wiki account for sepen?13:17
cptnI can try to :-)13:17
tilmanheh, thanks13:18
sepenthanks too13:18
tilmani'll announce him joining opt on the News page13:18
tilman(seeing the announcement stuff is in the agenda :D)13:18
cptnI was wondering whether we should send a mail to crux@13:18
tilmanprobably both :D13:18
cptnalso mentioning the other additions this year13:18
cptnplus the 2.5 outline13:19
sepentilman, I'll play sometime with flags13:19
sepenbut nspr.pc seems fine13:19
tilmancptn: dunno, it's kind of lame to announce new opt people now, five months (or so) after they joined13:20
cptnmmmh, not sure13:20
cptnthey deserve it, IMO13:20
tilmansure, just thought it might look strange13:20
tilmantechnically, hannes has +w in opt, too13:21
tilmanthough he never committed anything13:21
tilmanmaybe we should talk about 2.5 first :D13:21
cptnsure, why not13:21
tilmanimo: soon would be good13:22
tilmani don't think i can spend time on it before october though13:22
cptnalso, I'll be gone for two weeks in august13:22
cptnnot that you have to care :-)13:22
tilmani was scared by that bug report that said that gcc 4.3.1 miscompiles parts of the kernel (ext3 iirc)13:23
tilmanso i thought we'd better wait for 4.3.213:23
cptnsounds like a plan13:23
tilmanre. the locale generation13:23
cptnother than that, a regular update with current versions, and the usual bumps of major ports are enough13:23
tilmanlocale generation should go in definitely though13:23
tilmanas well as the other stuff we had on the list13:24
namenlosfor 2.5?13:24
tilman(new perl, /etc/ comes to mind)13:24
namenlosyes, i wanted to mention perl, too ...13:24
cptnwe have a list for this, right?13:24
tilmanit's in the bug tracker13:24
namenloshaven't seen that, yet13:25
tilmanlocale generation seems like a non-issue to me13:25
cptnso let's say we want to release 2.5 in Q413:25
tilmandocumenting it is trivial :p13:25
sepenjust only the handbook for releasing it too13:26
tilmancptn: what about it?13:26
cptnwould that be okay?13:26
juethe new yearly CRUX :-)13:26
tilmancptn: hehe, is misread that as 'qt4' ;D13:27
cptnI mean, you said "soon", but not before october13:27
tilmanlet's get started with it in october :)13:27
tilmanit will be delayed anyway13:27
cptnyeah, CRUX 2.5 will have a new pkgutils13:27
cptnbased on qt13:27
cptnsaves a lot of code on our side13:27
tilmanpkgutils6 will probably not be done again for 2.513:27
cptnno worries13:28
tilmandepends on whether i can go on a coding spree in fall13:28
cptnit's a drop in replacement anyway, so we can add it anytime13:28
* tilman flinches :D13:28
tilmanbut yeah, i don't think it's a problem either ;>13:28
cptnmaybe we can do one of the future meetings on 2.513:29
cptnso everyone can prepare his list of issues13:29
cptnbad idea?13:29
jueno, fine13:29
sepenor pseudo-oficial meeting for 2.513:29
tilmanaha, yeah13:30
tilmana non-casual meeting maybe13:30
cptnwell, given that we have a well defined agenda, that's implied :-)13:30
cptnthis is a quick on too: "allow posting from gmane "13:31
cptnany objections to that?13:31
tilmani don't know!13:31
cptnI have no experience with it, unfortunately13:31
tilmansame here13:31
cptnas far as I read, people will still have to subscribe to the list though13:31
cptnso it's not a meta address from gmane that's subscribed13:31
cptnbut the users have to subscribe first13:32
cptnand then they can post via the web13:32
cptnunless it mangles with message-ids, I see no problems13:32
tilmanwe can enable it and try it then13:32
cptnI can follow-up on crux@ and ask Colin if he has any experience with it13:33
cptnthen: "how to announce news, what to announce etc. "13:33
cptnwe quickly covered this13:33
cptnactually, we talked about this before13:33
tilmani seem to remember some blog talk :D13:33
cptnyep :-)13:33
tilmanwell, do you think we should go ahead with that?13:34
cptnit's probably worth a try13:35
cptnbut I'm not sure if it's worth the effort13:35
jueI don't think that we have a technical problem here, but we need someome how is able to add nice things to our HonePage13:36
juea "Chief editor"13:37
treachI heard eugenia quit osnews. ;)13:37
* treach takes cover13:37
cptntreach: :-)13:37
tilmancan/should we just sync the blog to the News page?13:38
cptnsounds good13:38
treachseems like a good idea, where should news be if not on the frontpage..?13:39
juewhich blog?13:39
cptnthe imaginary future crux dev blog13:39
juewhat's better with a blog than writing it into the wiki page?13:39
namenlosjue a blog is not version controlled.13:40
juegreat ;-)13:40
cptnwell, I guess a blog would be somewhat more personal13:41
namenlosin my opinion such things shouldn't be version controlled, but this is a matter of taste.13:41
cptnsince you usually have the authors coupled with a news item13:41
namenlosand people can comment the entry and give new input.13:41
juesorry, but on our main entry page we shouldn't have personal things13:42
jueand no comments from users13:42
cptnthat's not what I meant :-)13:42
cptnmore that e.g. the pkgutils news are from tilman13:42
cptnin any case, it's not a technical problem13:43
cptnso a blog won't help if no one posts13:43
jue.oO cptn for chief editor13:43
tilmanmaybe we should try to do more news update without a blog first13:44
tilmansee if there's interest at all :p13:44
cptnsounds good :-)13:45
aona friend was visiting so i'll quickly read the backlog13:45
tilmanbtwbug #17113:47
tilmaniirc we decided that utf8 configuration should be up to the user after all, right?13:48
tilmansince it's hard to get right from the start... see eg utf8 problems with man pages13:48
sepenyou mean localdef?13:48
aoni agree13:48
tilmanthis isn't strictly related to locale generation13:48
aoni hate how i have to disable utf8 everywhere else13:49
tilmanwhat do you mean?13:49
aonie. look for some random files to set things back to latin113:50
tilmanmmh, ok13:50
aonit's not like majority of our users need to write klingon or something :)13:50
tilmanother opinions?13:50
cptnI don't use an utf8 locale right now13:51
cptnand don't miss it13:51
juesame here13:52
namenlosi use it here13:52
namenlosi mean utf813:52
cptnwhat's involved to make utf8 the default, and what are the consequences?13:52
tilmansome/lots of man pages are misrendered13:52
tilmanif you tell man to use utf8 (/etc/man.conf)13:53
treachsince most other distros are utf8, shouldn't that be a passing problem?13:53
tilmanalso, users will have to tell their favourite editor to prefer utf8 etc13:53
tilmanand if your font doesn't support utf8 perfectly, you might be semi-forced to switch :(13:54
juethe user will have to convert docs if they contain chars > 12813:54
cptntilman: switch back, that is?13:54
tilmancptn: or switch to a better font ;D13:55
cptnah :-)13:55
juenamenlos: btw, would you please do s/รถ/oe/ on your ports13:56
juewe've decided last week to keep our ports tree 7-bit clean13:57
namenlosjue: ok13:57
juethanks, aon called it 'deumlautify' ;-)13:59
cptnI created a wiki account for sepen14:00
cptnand added him to Main/About14:00
cptnalso moved his profile from Wiki/ to Profiles/14:00
juetilman: <ot> you've seen the vendor_ruby dir with 1.8.7? </ot>14:01
cptnwhile we're at it, there appears to be a way to authenticate pmwiki users against an external database14:01
tilmanjue: yes. should we do anything with it? :p14:01
cptnaon found that14:01
cptnwould be nice if we could use the flyspray DB for that!14:01
cptn-> less user accounts and automatic registration14:02
jueI've changes my ports to use it14:02
juecptn: cool14:03
tilmanjue: you updated your ports to use vendor_ruby?14:03
tilmani'll add that to my .todo14:04
cptnshould we maybe start a porting best practices document?14:04
cptnwhich includes such things14:04
cptnso people creating their first ruby port can go check there?14:04
cptnsame for python (when using setuptools), or also the various build systems like cmake, *jam, scons etc.14:05
sepenand how to do after porting something (prtverify, tests, ...)14:05
tilmansounds good14:05
cptnso, back to utf8 as default locale14:06
tilmanuse C by default, and whip up a howto-use-utf8 document in the wiki14:07
cptnsounds good14:07
tilmanit's KISS :D14:07
tilmanpli would be proud ;)14:07
aonno cheating :)14:08
cptnand no anime smilies anymore14:08
cptnyou're an opt maintainer now!14:08
sepenhehen and Pkgfile and the others to utf8 too?14:09
tilmanlast week most of the guys wanted to stay with ascii in Pkgfiles14:09
juewe decided that14:09
sepenyeah, but imho patches and etc should be ascii too14:10
sepenI mean pre/post-install at least14:10
tilmansepen: i've never seen a non-ascii patch anyway :D14:10
teKI bet these exist .. in SuSE repos14:10
tilmancptn: what about linking your meeting minutes emails from MeetingAgenda?14:11
cptntilman: sounds good14:11
sepenhi nipuL14:12
cptnmorning nipuL14:12
tilman\o/ nipuL14:12
cptnis it like 5am in .au?14:12
nipuLwell, 20 past14:12
nipuLbrb, coffee is brewing14:12
cptnso then we still have the FAQ item14:13
cptnjue split it into multiple pages already14:13
cptnand it's merged into one on
juethe old one is still there if we don't like it14:14
cptnah, it's not merged then :-)14:14
juebut we can do it if we consider to go in that direction14:15
cptnI'd like to have a merged one too in the end, just that people can search for .footprint14:15
cptnbut in general, I prefer the split up pages14:15
cptnjust like for the news stuff, a chief editor would be great14:16
cptneither Eugenia or treach14:16
tilmansorry, is Wiki/Faq the new or the old one? :o14:16
cptnWiki/Faq is the old one14:16
cptnthe new one is on
jueto improve navigation in the splitted doc I've added a floating menu top right14:16
cptnfirst section14:16
treachcptn: I'm afraid that I'm too unreliable for that. ;)14:16
tilmanah, thanks :]14:16
cptnso, there's still the "devfs" FAQ in there14:17
cptnit's that old :-)14:17
tilmanvery nice jue14:17
cptnwe can hope that the users will help us to pick it up, but I'm not sure if that'll happen soon14:17
cptnjue: yeah, the menu is especially nice :-)14:17
treachnot really related, but I'd like to request having the "official" (core/opt/contrib) repos on an entirely separate page from the user contributed stuff.14:19
treachit's not good imo, if people get hit by something bad in the non-vetted stuff and then associates that to the project.14:20
tilmanwhich wiki page are you talking about? :o14:20
cptnportdb I think14:20
tilmanbtw, who's getting the Register mails these days?14:21
cptnViper I think14:21
cptnthere's a lot of things we could do with portdb14:21
tilmantreach: are you suggesting two port databases?14:21
nipuLyeah, i've been waiting for til to finish the xml feeds14:21
tilmani don't see that happening soon :>14:22
tilmannipuL: want to do it yourself?14:22
treachtilman: well, I'm not sure about how to solve it, but I don't think these things should be mixed.14:22
cptnyou're allowed to rewrite it :-)14:22
spraybotFS#175 - Support XML output for the portdb14:22
nipuLi have no problem with that14:23
tilmanoh btw14:23
tilmani added mail aliases for you guys14:23
tilmanguys = [nipul, sepen, namenlos]14:24
tilmannipuL: err, maybe you'd prefer lucas@?14:24
tilmannipuL: you should be able to ssh into now14:25
tilmanbbl, need to cook some food14:25
namenlosomg atm i am downloading with ~2500KB ..14:25
nipuLcheers, land yeah, i'd prefer lucas@ :)14:26
cptnI'll get myself some coffee then14:26
cptnI guess we could also affort two aliases :-)14:26
namenlos another quesion: now, since i am able to use my crux again, i decided to update/clean the openldap port. is anyone using the /usr/lib/{,,} links14:26
namenlosand i would like to use a release version of it (not stable). what do you think?14:27
cptnwhat means "not stable" here?14:28
cptna snapshot?14:28
namenlosno, openldap has release and stable14:28
namenlosit's some kind of confusing14:29
namenlosas i understood it until now: stable is the last version you should use before update14:30
cptnheh, hard to say then for outsides I guess14:30
cptn"Most users should use "release" (general use) versions."14:30
namenlosthe problem is, that there is a dos possibility in stable atm14:31
cptnwell, but then I think it's clear anyway, right?14:31
namenloscptn: this is why i would like to use release14:31
cptnupstream is recommending the use of "release"14:32
cptnand "stable" has a security issue14:32
cptnso I think 'release' is the way to go14:32
namenlosk, just wanted to be sure.14:32
namenlosshould i post a email on the ml for the links, so that people can check without stress?14:33
tilmannipuL: done14:34
nipuLcheers big ears14:34
tilmannamenlos: you could also post an updated port to the mailing list and ask for testers14:36
tilmannamenlos: people might not know whether they need the compatibility links :)14:36
namenlostrue. i will do that14:36
cptnanother random idea for the port db14:38
cptnapparently, some users do mainly "one time" ports14:39
cptni.e. they create them, but never update/maintain them14:39
cptnI do this occasionally too, but never upload these ports14:39
cptnif one could flag a repository as "one time porting", I might upload some of these14:40
cptnsaving others some work14:40
tilmanso that would be like another warning flag?14:40
nipuLthat was a good thing about when we used cvs14:40
cptnmaybe even a filter14:40
cptnso by default, you'd see the official repositories only14:41
cptnand then you could add the "active user" ones14:41
nipuLyou could see the age of a port14:41
tilmanmmh, that sounds like it might work14:41
cptnalso, it would be nice to include the "out of date" flagging, together with nipuL's bot, for core/opt/contrib14:42
cptnanyway, as I said, random ideas :-)14:42
nipuLshouldn't you be audting ports that you get from portdb anyway?14:42
cptnbut maybe you shouldn't expect a port from a "one time" repo to work14:43
cptnright now, some users might get the impression that some ports from portdb work, others don't14:43
cptnthey might not even see the difference between core/opt/contrib and $random, as treach mentions14:44
cptnif by default portdb would only search ports which are supposed to be maintained, the user experience might be better14:44
cptnbut it's just cosmetics in the end14:44
tilmanasking people to properly set the is-maintained flag might not work so well :)14:45
cptnthere's also the idea that maintainers could maintain their httpup entry themselves14:46
cptnand get a monthly reminder by mail14:46
cptnif they don't log in after that time, the repo is being marked as unmaintained14:46
cptnI'm sure some of the people that are listed on portdb don't use crux anymore14:46
namenlosi like this automatic approach14:47
namenlosotherwise i doubt (like tilman said), that people maintain their out-of-date-flag, if they don't even maintain their ports.14:48
sepentilman, mplayerplug-in builts ok with the nspr patch14:49
*** spraybot has quit IRC14:52
*** spraybot has joined #crux-devel14:52
tilmancptn: we should try to talk nipuL into writing a prototype for those features14:53
cptnwe should14:53
nipuLshould we?14:53
tilmanyes :p14:53
cptnbut we should define which features we want first14:53
tilmani thought we could make him go ahead and then kick the stupid stuff out ;D14:54
nipuLi'm not about to implement every action availble14:54
nipuLalthough it would be nice if we could hack feed.php to allow listing of an arbitrary number of items14:56
nipuLright now it only does 10 or 2014:56
tilmanwhat feed.php? :)14:57
nipuLin flyspray14:57
tilmani see15:00
*** mike_k has quit IRC15:00
nipuLah, found where all the web stuff is15:01
nipuLso are there any features anyone would like added to clispray?15:02
nipuLcrap i broke the bot15:03
cptncreating a new report15:03
tilmanmaybe the responses to those queries could include the bug's status15:04
*** spraybot has quit IRC15:04
*** spraybot has joined #crux-devel15:04
spraybotFS#279 - opt/sqlite3&#039;s pkgconfig file is wrong.15:04
tilmannipuL: displaying apostrophe's correctly would be cool ;)15:05
treachwhy, nobody uses them correctly anyway. ;)15:06
*** spraybot has quit IRC15:08
*** spraybot has joined #crux-devel15:09
nipuL!FS#279 status15:09
spraybotFS#279 - Status: Unconfirmed15:09
sepenofftopic question, when I should use [notify]? mplayerplug-in now depends of xulrunner instead of firefox, is this a reason for notify it?15:09
nipuLiirc, the way firefox works now is that firefox is essentually just an interface to xulrunner15:10
tilmanyep, so sepen's change shouldn't break anything15:10
tilman-> you don't need to use [notify]15:10
nipuLoh, i didnt read that properly15:10
tilmansepen: in 99% of all cases, [notify] is used for security notifications, and other very important stuff %)15:11
juenight tilman15:12
cptnnight tilman15:12
namenlosgn8 ;)15:12
nipuLdon't let the bed bugs bite....seriously15:12
juecptn: the combined FAQ is in place, was a bit tricky because I want a toc for the split pages as well15:14
juebut pmwiki is really cool, has a (:if:)(:ifend:) statement15:15
cptnlooks good!15:16
cptnI think once the FAQ is up to date, it should also be a top level item in the navigation15:16
jueyeah, needs a little bit more love, maybe we can force j^2 to take the job seriously :-)15:18
aon+1 :)15:20
*** spraybot has quit IRC15:29
*** spraybot has joined #crux-devel15:29
spraybotFS#279 - opt/sqlite3's pkgconfig file is wrong.15:29
juebtw, sqlite 3.6.0 still says:15:32
jue pkg-config --modversion sqlite315:33
* jue needs a chair with weels for the new huge monitor ;-)15:36
teKhow 'bout binoculars? :)15:37
nipuLdammit, portdb is written in php15:45
nipuLmy eye! the goggles do nothing!15:45
cptnit seemed like a good idea at that time15:45
Rotwangthats sad :{15:46
nipuLon the plus side it's only 272 lines15:46
nipuLalso, it's stuck in a v2.x sqlite db15:47
nipuLwhich is no longer available in ports15:47
nipuLsqlite3 doesn't appear to be backwards compatible15:49
jueno, you cannot open a version 3 db with 2.x15:49
nipuLoh hang on, sqlite is in opt, just not my opt15:52
jueI know ;-)15:53
nipuLnow to find out if v2 and v3 can coexist15:53
namenlosk, i got to go. gn8.15:54
*** namenlos has quit IRC15:54
nipuLlo, no wonder till got sick of adding xml to portdb16:06
cptnwhy is that?16:06
nipuLyou can tell the original script was quickly put together, he's had to put conditionals all the way through it16:10
nipuLi like to try to keep data collection functions seperate from display functions16:12
nipuLthis doesn't do that16:12
sepensounds pretty nice16:18
nipuLi think i might have a crack at a rewrite16:21
*** sepen has quit IRC17:31
*** jue has quit IRC17:37
*** treach has quit IRC18:00
*** Rotwang has quit IRC18:14
*** j^2 has quit IRC18:17
*** j^2 has joined #crux-devel18:17
*** fetid has joined #crux-devel20:15
*** pitillo has quit IRC21:31

Generated by 2.11.0 by Marius Gedminas - find it at!