IRC Logs for #crux-devel Tuesday, 2006-11-14

clbUpdate from opt: 14 Nov 09:32 - postgresql: changed mirror03:39
clbUpdate from opt: 14 Nov 12:36 - [notify] iproute2: updated to 2.6.18-061002, changed sbin location06:39
*** treach has joined #crux-devel07:05
*** treach has quit IRC07:59
*** jaeger has joined #crux-devel08:16
prologicaye jaeger08:25
jaegeraye? :)08:26
*** kieselsteini has joined #crux-devel09:18
*** kieselsteini has quit IRC09:26
*** j^2 has joined #crux-devel09:33
j^2morning guys ;)09:33
j^2how you doing jaeger09:38
jaegernot bad so far09:40
j^2late like always myself09:41
j^2class at night gets to me09:41
j^2dinged 49 last night09:44
j^2almost 1/2 to 5009:44
jaegernice :)09:44
j^2do you mind that we are still using your TS machine?09:44
jaegernot at all09:44
j^2awesome thanks09:46
j^2just checking09:46
*** kieselsteini has joined #crux-devel09:55
*** kieselsteini has quit IRC10:16
*** treach has joined #crux-devel10:36
*** jaeger has quit IRC10:56
*** jaeger has joined #crux-devel11:04
clbUpdate from opt: 14 Nov 17:07 - stunnel: update to 4.19 || 14 Nov 17:03 - dovecot: update to 1.0.rc14 || 14 Nov 17:00 - swig: update to 1.3.30 || 14 Nov 16:59 - screen: don't install screen suid root, suggested by Han Boetes11:10
*** clb has joined #crux-devel13:07
tilmanshould we start then?13:08
cptnI'd say so13:09
sipany news on the contrib/git woes?13:13
tilmannot that i know of13:15
sipwell, no_news=good_news in this case13:16
tilmannot sure whether eg jaeger tried to push again in the meantime13:17
*** bdev has joined #crux-devel13:17
tilmanshould we talk about the core/opt/contrib dupes issue now?13:18
sipI vote for kindly suggesting a rename of offending ports13:19
sipie pkgutils-han13:19
cptnhe actually added 'pkgutils' to contrib?13:20
tilmani vote for not allowing dupes ;)13:20
tilmancptn: and ports, and rpm2targz and openssh and lots of other stuff13:20
cptnwell, the rules state "No duplicates"13:20
cptnand to me, this should also be core/opt <-> contrib13:20
cptnnot just within contrib13:21
sepenand also -svn ports?13:21
tilmancptn: yes, it makes sense that way13:21
cptnsepen: would you consider svn port "latest stable"?13:21
juewe should avoid svn/cvs ports at all13:21
cptnsepen: considering that they're potentially untested in this revision?13:21
aon+1 for not allowing duplicates at all13:21
sepenIm according to this13:21
cptnViper_: around?13:22
cptnoh, away...13:23
cptndoes anyone know whether Viper_ talked to han already?13:23
tilmani doubt it13:24
tilmanhe wasn't around yesterday when we noticed the problem13:24
tilmanif there's no veto, i can fix up the wiki to clarify the dupe situation13:24
cptnsounds good13:25
aoni wonder if some clear guideline could be added for the maintainer email format13:25
aonit causes quite a lot of noise now if you for some strange reason run prtverify for whole contrib13:25
sipalright, we can also enforce that in the update hook if somebody insists with dupes13:25
tilmanwe could also fix prtverify to _not_ reject han's email adress13:26
tilmannot sure which i'd prefer13:26
jueaon: <- no enough ?13:26
aonwell, it is there, yeah13:27
tilmanmaybe it should be stressed that people should use this specific email format13:30
jueonly Han seems to have a problem with that, and prologic forgot the ',' :-)13:31
aontilman: yeah13:31
deus_excptn: wrt svn ports.What about openal, then?It doesn't have a current release(not even a svn snapshot), and old one is outdated/broken?13:33
deus_exI'll stfu now :)13:33
cptnjust create a tarball of a known and working svn revision13:34
tilmandoes everyone agree that we want one email format to rule them all in contrib, too?13:34
j^2+1 ;)13:35
tilmani just did a prtverify run on contrib13:36
tilmanit's annoying to see FATAL errors there13:37
sipwell, that's a reason why we have core/opt and contrib13:38
sipnonetheless we could help improve the situation (that was one of the aim of the new contrib)13:41
sipsuch as periodic tests and supervision (Viper_)13:41
tilman(contrib dupes)13:43
tilmanNote that this includes duplicates in contrib as well as duplicates in the other official CRUX repositories.13:43
tilmandoes that sound fine?13:43
tilman"official CRUX repos" would not only include core and opt, but also kde, xorg and eventually gnome13:44
sipjaeger, any comment on this?13:44
tilmanmy point is not about the inclusion of gnome atm13:45
tilmanit's rather whether core and opt are treated equally as kde/xorg/gnome ;)13:45
tilmani think they should, but i'd figure i ask for objections before going ahead ;p13:46
sipsorry, I missed the 'not' in 'not include'13:46
sipanyway, no dupe sounds good for now. shall we go on with the flyspray topic?13:49
tilmanwell, apparently "no dupes" wasn't very clear ;D13:50
sipheh. must turn off the tv.13:50
tilmaneven if "official crux repo" isn't well defined atm, it's probably better if people pop up and ask what that means13:50
tilmanrather than assuming it's only core and opt13:50
sipbetter replacing 'official' with ' hosted' imo13:52
cptnI'd rather go with official13:52
cptnsince we can then host a repo without making it relevant for contrib13:53
cptnand the we make it "official"13:53
cptnexternal -> hosted -> official13:53
sipa win-win ;)13:54
cptnand since all contrib post _should_ now have good email addresses13:55
cptngood meaning from their maintainer13:55
cptnrunning a script and informing the maintainer about dups should be easier than before13:55
jaegersorry, I was out to lunch and running errands, what am I commenting on?13:56
sipjaeger,  mainly dups in contrib and other repos13:57
jaegerpersonally, I'm not fond of dups but then again, we had them in the old contrib, too13:57
jaegeras long as the default order in prt-get.conf continues to list contrib AFTER core/opt, I'm fine with it13:57
tilmanwe voted on not allowing them. if that's not a huge concern to you?13:59
jaegerIt'll be ok for me either way14:00
jaegeron an unrelated note, I'm tempted to add distfiles mirror support to pkgmk14:01
sipwith a central ( mirror?14:01
jaegerthat'd be ideal, though I doubt Charlie would be keen on it14:02
jaegerbut I think it wouldn't be difficult to add14:03
sipyes, I mean entral mirror with all sources rather than different (ie soureforge) mirror options14:03
jaegeryes and no14:03
jaegersupport for an arbitrary number of mirrors14:03
jaegerbut central in that they would contain all the ports' sources14:03
jaegerfalling back on the source url in the Pkgfile14:04
sipsounds good to me14:04
sipshall we put it in the ToDo?14:04
jaegercouldn't hurt to have it written down, I suppose, though it's no more than idle chatter right now14:05
tilmanput it in 'future pkgutils plans' maybe14:05
cptnI think there are two patches floating around for that already14:05
jaegerI'm in the process of downloading all the core/opt/contrib sources now to see what kind of space requirement that'll entail14:05
jaegercptn: wouldn't surprise me, I'll have to search the ML14:06
tilmanspeaking of pkgutils... :)14:06
tilmancptn: do you want to put in han's sourceforge-mirror-html patch in pkgutils before we do the release?14:06
sipabout bandwidth, the mirror should be it only when there is some problem with the sources, so it won't be that bad14:06
cptntilman: sorry 'bout that14:07
jaegerpersonally, I'd rather pkgutils defaulted to the mirrors first14:07
jaegerlisted in preferential order in pkgmk.conf or something similar14:07
cptntilman: I wanted to answer today, but then I figured I'd be here anyway14:07
cptntilman: for me, the current situation is fine14:07
cptntilman: so I'd say release it14:07
jaegercptn: thanks, I'll have a look at that and test it14:07
tilmani'm tempted to quote the klingon programmer again14:07
tilmanbut i'll resist14:07
tilmancptn: i'll do the release then14:08
tilmantonight/tomorrow/the day after ;p14:08
cptntilman: good14:08
tilmanwhich leaves flyspray-and-contrib on my imaginary list14:09
sipalso 2.3 branching14:09
cptnmaybe Viper_ can send a mail to contrib talking about the three things?14:09
cptndups, email addresses and flyspray account?14:09
tilmanpoint 3 needs discussing _this_ first ;D14:10
tilmansip: do you have news/comments on the flyspray thing?14:10
sipI've done some basic test and I think we can move tasks to the contrib project14:10
sipthere seemed to be no technical issue14:10
tilmanvery cool14:10
tilmanthat would also mean we'd remove the 'ports - contrib' category in the mainline flyspray project, right?14:11
sipas I said yesterday it's more a user interface question: you must select the contrib before adding a task14:11
sipnot initially since it's kept with the move14:12
sipanyway, soon after the move we can rename/remove it14:12
tilmanwell, yes14:12
sipeverybody in favour of keeping the contrib a separate flyspray project?14:13
tilman+1 ;)14:13
cptn+1, but:14:14
cptnit seems a bit hard the way it is now14:15
cptnwe have a "CRUX" project, and a "Contrib ports" one14:15
cptncould we maybe make that difference a bit weaker? ;-)14:16
sipsuch as?14:16
tilman"CRUX contrib repository"14:16
sipwhat about makin it more general?14:16
sipin the sense that tomorrow there could be other contributions, not only ports14:17
cptn"CRUX contrib project (former CLC)"14:17
tilman"the flyspray project formerlly known as contrib (belongs to crux)"14:18
tilmanre. "other contributions, not only ports":14:21
tilmanthe current user pool for the "contrib" repo is only "port contributors"14:21
tilmani mean14:21
tilmanno, nevermind14:23
sipnames apart, I'll ajust the setup tomorrow and let Viper_ know when he can send the notes to contrib14:23
tilmani had a weak and silly argument in mind14:24
sipoh. that never stopped you :-P14:24
tilman2.3 branching?14:24
sipso, if we branch now and make a conservative release I think we can put 2.3 out before the end of the year14:27
jaegerwhat does a conservative release entail?14:27
jaegerrather, what does a conservative release entail that justifies a release at all? :)14:27
sipwell, xorg, setup+groups, same pkgutils with small improvements14:27
jaegersounds reasonable14:28
sipother than that, we can add other things time permitting14:28
cptnI'd also vote to include gcc 4.1.114:28
cptnI've been using it at work for a long time now14:28
tilmanhaving gcc 4.1 would be nice, yes14:28
jaegerI still see complaints every now and then about gcc 4.0.x so maybe that'd be a good idea14:28
sipsounds good for me, does not seem like a great jump14:29
cptnsome software needs additional patches, but the other distros have them all14:29
tilmando you plan to bump glibc?14:29
cptnno plan14:29
cptnbut mainly from a lack of time to test it14:30
tilmani wonder whether i could do it14:31
tilmanbut i'm not sure how to properly test glibc14:31
cptnwell, there are a couple of basic tests14:31
cptnthe most important one if a bootstrapped system works fine :-)14:31
cptnhowever, I think most other distros switched to 2.5, at least for testing14:32
jaegerIs there a big reason to bump glibc right now?14:32
sip'hello world' is the definitive test14:32
cptnso the step appears to be less critical than for other glibc updates14:32
tilmani'll try to read up on glibc development on the weekend14:32
treachsip: fwiw, I've been using cptn's 2.5 port quite a while. Haven't hit anything that seems related so far.14:33
tilmanjaeger: i think our kernel headers are broken wrt i2c. i'd like to see this fixed14:33
treachbut what do I know *shrug*14:33
tilmanjaeger: that's not really important though... just an annoyance14:33
cptntreach: ah, good to know14:33
cptnI have it on a test partition here14:33
jaegertilman: well, no reason to perpetuate brokenness, in my opinion14:33
cptnit would be nice to have a new glibc package though14:34
cptnusing 2.6.18's kernel headers14:34
cptnand we could therefore use 2.5 for our testing14:35
tilmanwhat about this:14:36
cptnand revert immediately for 2.3 if it's becoming critical for the release14:36
tilmancptn: you could move your gcc and glibc ports into 2.314:36
tilmanas they seem to work at least somewhat ;))14:36
cptnhere's another option:14:36
jaegerto be clear, we're talking about glibc 2.4 for crux 2.3, right? :)14:36
cptnwe branch 2.3, and use that for our own machines14:37
cptnand update the ports in 2.2 to get broad testing14:37
jaegeror is glibc 2.5 considered production-ready?14:37
tilmannot sure whether we're talking about 2.4 or 2.514:37
j^2i'd like to chime in and say....if you'll need testers for 2.3 i'd love to help at that place ;)14:37
jaegerI'm curious because I know people are using 2.5 but 2.4 is listed as the "current version" at
cptnjaeger: well, fc6 has glibc 2.514:38
tilmanis fc a good reference? %)14:38
cptnOTOH, glibc is developed mainly by fedora (or rather u. drepper), so that was to be expected14:39
jaegerat the risk of sounding rude, don't particularly care what fc uses... I'm curious if we want or need it for crux14:39
tilmani believe they like to put broken stuff in it to have the masses test it14:39
jaegerI've no objection to 2.5 if that is indeed the way to go14:39
cptnthere's a major reason to go for it:14:41
cptn"New Linux interfaces: splice, tee, sync_file_range, vmsplice."14:41
cptnnot sure if that's really critical for us though14:41
tilmand'oh, these are system calls?14:41
tilmani thought it was kernel internal stuff14:41
jaegerwell, if it's interesting, or useful, that answers part of the question for me... but the second half of it is this: is it stable enough? :)14:41
sepenwhat about qt4?14:41
tilmani don't really know14:43
tilmanneed to read up on the glibc mailing lists to be able to make a sane decision ;)14:43
jaegersepen: qt3 doesn't currently ship with the CRUX ISO, so not sure it's relevant at this time14:44
tilmani'd remove opt/x11 in 2.3 soon after it's branched14:44
tilmanif anyone can think of a reason to hold off on that, please speak up (not necessarily right now of course ;D)14:45
cptnIOW 2.5 is not sent out to the masses yet14:46
jaegerWarning: fsockopen(): unable to connect to deer:8010 in /home/httpd/gmane/php/lib.php on line 1814:46
siptilman, will you do the branching tomorrow?14:48
tilmancan do!14:48
sipfine. maybe do the git reset thing also for core-opt14:49
tilmancore.git and opt.git are also "dirty" on
tilman(ie, uncommitted changes)14:49
jaegerby the way, even though it's not announced, clb tracks the contrib rss, too14:50
jaegeras well as kernel.org14:50
tilmansip: please do it, then14:50
jaegercontrib, core, gnome, kernel, opt, pkgutils, xorg14:51
jaegerthe only ones announced in here are core and opt at this time14:51
sipas an addon for the iso setup, time permitting it would be nice to allow download of updated packages14:53
sipand kernel14:53
jaegerthat'd be nice14:53
tilmanmaybe we'll get the libarchive version of pkgutils ready for 2.314:54
sipI  suggest we start working on 2.3 now according to the plans and discuss/integrate addons gradually14:54
jaegerwould obsolete my updated CDs, which is fine14:54
sipthat also means somebody should provide the binaries ;)14:54
tilmansip: yes14:55
jaegerthe ones I update are currently provided in 3 different places14:55
sipgood :)14:55
tilmani think i'll do the pkgutils release now14:55
tilmandoes anyone know why core/pkgutils/pkgutils-largefiles.patch wasn't committed?15:00
jaegerdidn't really get much response from it15:01
jaegerbut I use it and it seems to have caused no trouble so far15:01
tilmansince we enable it in the port, i figure it should go in the tarball, too15:01
cptnsounds reasonable15:02
jaegerIt was my hope that it'd go in eventually, you won't hear argument from me15:02
tilmanpkgutils.git history seems a bit lacking15:03
tilmanis svnweb running anywhere still? %)15:03
sipno, but there's still the repo on crux.nu15:04
tilmandoes anyone have the path to pkgutils handy?15:05
tilmaniirc it was tools/pkgutils, but that doesn't work15:05
sipshould be15:06
sipthere's a chekout in /home/crux/temp/sip/repo15:07
tilmanthe oldest entry in pkgutils' changelog is from 2006-0415:08
tilmaneven these changes never went into a release :o15:08
tilmani'd like to remove ChangeLog from git and replace it with 'git log's output15:09
tilmanmaintaining ChangeLog manually is annoying work, prone to errors etc ;)15:09
cptnokay with me too15:09
jaegerok by me15:09
cptnalthough this means that typo's in ChangeLogs can't be removed easily15:10
jaegerare we going to replace pcmcia-cs with pcmciautils?15:10
jaegeralso, is someone willing to take over udev?15:12
sipthe system team ;-)15:12
cptnspeaking of which...15:12
cptnI'll have to drop some of my tasks in the future15:13
cptnor rather want to15:13
jaegerudev has seen quite a few updates since the current port release but I'm extremely sick of it15:13
cptnyeah, udev would probably need a dedicated maintainer15:13
juejaeger: actually the pcmcia-cs port was lost somewhere15:14
cptnOTOH, it has handy features, so there's probably little way around it15:14
jaegerjue: that's why I asked15:14
jaegerjue: I'm building another updated ISO and noticed it was gone15:15
tilmanshould we keep NEWS for pkgutils?15:16
tilmanprolly should.15:16
jaegerit's a bit more detailed than git's log15:16
tilmanit's nice to see the user visible changes quickly15:18
tilmancurrent pkgutils home page:
tilmani guess i'll contact per and ask about whether he wants to update it for new releases15:20
sipbetter change it to crux.nu15:21
siptogether with some other stuff, ie 'ports' sources15:21
sipso we don't have to bother him15:22
sipanyway, hearing from him is a good idea15:22
tilmani think i'll leave it as is for now15:23
juewas my thought as well15:24
tilmani think we have already, right?15:28
cptnwe could add for a files or dl. subdomain15:28
tilmanlet me see where it fits best15:28
cptnhow about we add dl1 and dl2 with RR DNS? ;-)15:28
jaegerand different locations, of which we only maintain 115:29
cptnerr, no15:29
cptnthey're two locations, and one is always down15:29
tilmancan someone put this somewhere?15:31
cptnsomeone create a real shell account for tilman15:31
jaegerjaeger@acheron(~)$ ssh crux....nu15:31 looks a bit chaotic15:31
* tilman hides15:31
jaegerthat's my usb fllllipping out15:31
jaegeragaiiin, hehe15:31
tilmanno idea where it should go really :P15:31
jaegerthis would be fun if it weren't annoying as hell15:31
cptntilman: okay, it's in /files15:32
jaegershould I put it in
jaegernm, guess it's done :)15:32
cptnwe also have distfiles/15:32
cptnand dist/15:32
tilmanand releases/15:32
jaegercould be cleaned up a bit :)15:33
sip /stuff15:37
sipgoing to bed. night guys15:40
cptnnight sip15:40
*** sip has quit IRC15:41
jaegercptn: so should we stay with glibc 2.4 or try 2.5? I saw some mention of missing linuxthreads with 2.4? or did I mix that up?15:42
cptnno, that's right15:42
cptnnot sure if this is really relevant for us though15:42
cptnwe want NPTL anyway, right?15:43
cptnI personally would go with 2.515:43
cptnit seems we are pretty close to start the testing of CRUX 2.315:43
tilmanmissing linuxthreads shouldn't be a problem at all15:43
jaegerI have no objections if it's stable15:43
jaegeraye, shouldn't be for us15:43
cptntilman: do you know what the status is of the setup?15:45
cptnWRT xorg15:45
cptnI think that's the only critical thing remaining, right?15:46
cptnwe'd then have:15:46
cptn- new gcc/glibc15:46
cptn- new kernel headers15:46
cptn- xorg 7.115:47
tilmandang, forgot to report a problem to sip15:47
cptn- beryl instead of windowmaker15:47
cptn- xyz instead of sendmail15:47
cptn- some changed ports15:47
cptn- s/ifconfig/ip/15:47
cptnI'd leave the texinfo changes out for 2.315:48
tilmangit produces ~/dead.letter on crux.nu15:48
tilmanalong with15:48
tilmanalong with fatal: protocol error: bad line length character15:48
tilmanafter a push15:49
tilmanshould i worry?15:49
cptndid I miss something?15:49
tilmani agree wrt texinfo :O15:49
tilmannot sure whether that's all15:50
tilmanah, there it is15:50
tilmanpkgutils unleashed, brace yourselves etc15:50
tilmanunless anyone finds any weirdnesses, i'll commit the libarchive changes tomorrow15:51
tilmanand branch off 2.3 afterwards15:51
cptnoh BTW:15:51
cptnit's too late now anyway, but some time ago I "missed" some include files15:52
cptnoh wait, nevermind15:52
tilmanin pkgutils?15:52
cptnnot related to pkgutils15:52
cptnanother question: I just saw we're down to eight project members now15:54
tilmanthat was a pretty nice and produce mini meeting15:54
cptnany candidates we could ask to join us?15:54
jaegertilman: aye, I think so15:54
jaegernot sure, honestly15:55
tilmanViper_: do you read your IRC backlog? :P15:56
tilmannight guys15:59
cptnnight tilman15:59
jaegernight, tilman15:59
clbUpdate from core: 14 Nov 21:46 - pkgutils: updated to 5.2116:10
jaeger913M    /usr/ports/srcfiles16:15
jaegernot done yet :)16:15
treachwhee. dbus 1.0 .. wonder if that cools things down..16:21
jaegerone can hope16:22
prologiccptn, I :)16:25
cptngood nigh16:36
cptnprologic: sure :-)16:36
*** cptn has quit IRC16:37
juenight all16:37
*** jue has quit IRC16:37
prologicbbl :)16:39
prologiccptn: let's talk more about this later :)16:39
treach23:37 Quit : cptn ..16:39
*** jaeger has quit IRC16:40
*** treach has quit IRC16:42
*** pitillo has quit IRC16:45
*** jaeger has joined #crux-devel16:58
*** j^2 has left #crux-devel16:59
*** j^2 has joined #crux-devel17:59
*** Romster has quit IRC18:03
*** Romster has joined #crux-devel18:11
*** jaeger- has joined #crux-devel19:03
*** jaeger has quit IRC19:20
*** jaeger- is now known as jaeger19:20
*** Romster has quit IRC20:32
*** Romster has joined #crux-devel20:36
*** jaeger has quit IRC22:08

Generated by 2.11.0 by Marius Gedminas - find it at!