IRC Logs for #crux-devel Tuesday, 2006-10-24

*** entreprenerd has joined #crux-devel00:08
*** pitillo has joined #crux-devel00:15
*** jaeger- has quit IRC00:28
*** rantolo_ has joined #crux-devel01:50
*** rantolo has quit IRC03:25
*** rantolo_ has quit IRC03:25
*** sepen has quit IRC03:40
*** sepen has joined #crux-devel04:04
*** Romster has joined #crux-devel05:19
*** Romster has quit IRC05:22
*** Romster has joined #crux-devel05:44
*** Romster has quit IRC05:45
*** Romster has joined #crux-devel05:46
*** deus_ex has quit IRC07:36
*** jaeger has joined #crux-devel08:02
jaeger doesn't show relevant changes in some cases... take the opt tree as an example, it doesn't track the 2.2 branch on the front page. looks odd08:37
Viper_you're right, that's a bit confusing.. :)08:41
jaegera minor issue but it does look weird to me08:43
*** j^2 has joined #crux-devel08:44
j^2morning lal08:45
j^2doing well today matt?08:45
jaegerok so far, you?08:46
j^2trying out something called Glucerna shake08:46
jaegersounds dangerous08:47
j^2it's nasty08:48
*** prologic has quit IRC09:33
*** jaeger has quit IRC09:48
*** jaeger has joined #crux-devel09:56
*** deus_ex has joined #crux-devel10:13
*** entreprenerd_ has joined #crux-devel10:20
*** prologic has joined #crux-devel10:22
*** sip has joined #crux-devel10:44
siphi tilman11:05
tilmani just saw this thread in my browser and wondered how the fuck i got there11:08
tilmanyay for instinctively clicking urls in irc11:08
sipbtw, tilman I think a new git version is out11:09
jaegerare the httpup versions of core, opt, and contrib being updated still? cd asking about it in #crux11:10
tilmanhaven't read all my mail yet11:10
sipI think so11:10
sipwell, not the contrib11:11
tilmansip: you're correct. i'll schedule a port update :D11:11
jaegerhe said all 3 are missing11:11
sipwhich was already broken via httpup AFAIR11:11
siplemme double-check11:12
tilmanplease limit your git commit comments to 72 chars per line11:13
tilmanto make 'git log' happy11:13
tilmans/make/keep/, there's no problem yet11:13
sipwhere are the 'official' .httpup files for core and opt?11:17
j^2sip,  iirc yep11:17
tilmanit's not a "yep/nope" question dude ;))11:17
j^2oh crap11:18
j^2my bad, i didnt read the "where"11:18
j^2i'll go back to my idling :-P11:18
jaegersip: might be a good place, though the ones there are out-of-date11:19
tilmanthey are linked from here:
sipaye, I followed those but that info seem to be outdated11:20
jaegerprobably best to remove the old ones to avoid confusion11:20
*** prologic has quit IRC11:20
jaegerand replace them with working ones11:20
sipI mean, replacing 2.1 with 2.2 works, but at some time we probably pushed something different to the masses11:21
* tilman has no idea11:21
jaegerwere they included in the httpup port?11:23
jaegerdoesn't look like it11:23
j^2yeah it was something you had to get off the site, "incase your firewall was 'too restricted'"11:23
*** prologic has joined #crux-devel11:32
*** Romster has quit IRC11:51
*** Romster has joined #crux-devel11:57
*** pitillo has quit IRC12:20
tilmansip: btw, there's one 'l' too many in the spelling of cptn's name in gitweb12:52
sipthanks, fixed12:53
*** cptn has joined #crux-devel12:59
*** ChanServ sets mode: +o cptn12:59
siphey cptn13:01
tilmanmeating time?13:02
jaegerhopefully not13:02
jaegerbut a meeting would be ok13:02
jaegerway to kill the room, tilman13:09
sipeverybody's scared now :)13:09
siptilman, please start the bbq with the proposal about patches13:10
tilmangive me 2 minutes to grab a pizza13:11
sipmeanwhile, a quick preview of the proposal for the homepage:
sipassuming we want the spotlight thing13:12
cptnso far, I got no single reply13:13
jaegerthat looks nice13:13
sipcould as well contain generic important communications13:13
cptnyeah, it definitely looks nice13:13
sipnow, we don't have any space for news item :/13:13
sipback on more serious issues, any note on the git migration so far?13:16
sipI've noticed we sorta stopped updating ports ;)13:16
cptnI thought that was only me :-)13:16
cptnseriously though, I haven't had time lately to look into it13:17
sipI'm working on a per-project default head for gitweb (see logs for #crux-devel)13:17
tilmansip: works like a charm. nice work there :)13:18
tilmansip: speaking of it. do you still plan to fix up xorg.git with a 2.2 branch etc? ;D13:18
*** jue has joined #crux-devel13:19
siphey jue13:19
tilmanhi jue13:19
sipnow we should prevent people of using a git driver for ports13:19
sepennice work sip13:19
sip(a already noticed a couple of requests on irc)13:19
jaegerwe should prevent it?13:20
sipwell, it's not meant for heavy traffic, also because the anonymous repos are via tcp_wrappers13:20
sipand I'm confident it'll choke with many incoming requests13:21
sipnot a priority now, just a note13:21
jaegerfair enough13:22
sipso you can advise against it if somebody asks13:22
sipanother option is to use djb stuff instead of tcp_wrappers, but this is infra-team matter ;)13:23
tilmanyou could also export the git repo via http13:23
tilmannot sure how cool that is for the clients though13:23
* sip launches a huge rock towards tilman13:23
tilmanmmh, i cannot find that mail to one of the lists13:25
tilmanwhere i suggested the patch handling thing13:25
sipthis should be anouch to introduce the topic13:27
jaegerI personally prefer not to put the version in the filename but it's really not a huge time-waster13:29
tilmanwhat's the _pros_ of doing it?13:29
jaegerwell, it's obvious which version it was *intended* to patch... not sure what else13:30
jueI handle that depending of the patch type, whether the patch is version specific or not13:30
tilmani find it annoying that we put the same (usually/almost) file in the repo repeatedly13:30
juee.g. dovecot-conf.patch and python-2.4.4.patch13:31
tilmansounds sensible13:31
sipAs I wrote before I have no clear preference, in any case if we decide for the version-less option I suggest we change it in 2.313:34
jaegerok by me13:34
tilmani suggest to adopt jue's scheme then13:35
tilmanalthought it might not always be clear whether to put the version there or not13:35
cptnI can't think of an advantage to justify the renaming in git13:36
cptnin any VCS, for that matter13:36
cptnOTOH, I don't mind having the version in there and hardcoding it in Pkgfile either13:36
cptnwe could leave it to the user13:40
cptnand recommend to _not_ user versions if it's a patch which will potentially stay valid13:40
tilmanyeah, i feel a bit uneasy to call for one scheme to use seeing that most of us don't really care13:41
tilmancptn: alright13:41
tilmanlet's do it that way then13:41
tilmanobjections? :]13:41
tilmancptn: played with the libarchive patch yet?13:42
cptntilman: unfortunately not13:42
tilmanmy plan[tm] is to give it a beating, and if it seems to work, tag the current tip of master as 'pre-libarchive' or something and put it in13:43
cptnsounds good to me13:43
tilmanoh btw13:43
sipwill you compile a static binary as with libtar?13:43
cptnrelease it :-)13:43
tilmanwhat about making a stable release for the INSTALL stuff & html-fix?13:43
tilman"Ship it! Ship it and let them flee like the dogs they are!"13:44
tilmansip: yeah, i think so13:44
sipcptn, ypu forgot to tag prt-get13:44
cptnsip: forgot? ;-)13:44
cptnwho said we had to tag anything...13:44
cptntilman: I think we should just decide on how to promote INSTALL13:45
cptnand write that down somewhere13:45
tilmanwanna do that? :]13:45
sipI like seeing the visual tag in gitk :)13:45
tilmani think we should tag releases13:45
cptntilman: well, I can write it, if we agree on something13:46
tilmani'm fine with your proposal from some weeks ago13:47
cptnmy idea was that we recommend to not do thing on per port level13:47
cptnokay, I'll write a draft13:47
sipsystem team, any idea on what to do with the iso dirs from svn?13:48
jaegeranything wrong with a system repo?13:48
jaegeror perhaps iso repo13:48
tilmanspeaking of teams13:48
tilmandid we ever define some team membership thing?13:49
cptntilman: don't think so13:49
tilmandefining all that team business would be good13:49
cptnwe kinda did for infra, and ports is implied13:49
cptndefinitely agreed13:49
sipyes, also for the users13:49
cptn"go talk to Viper_ for that damnit" ;-)13:49
sipexactly what I was thinking of :)13:50
sipI added a todo item13:54
tilmanwhat's the status of contrib.git?13:54
tilmanwhat's left to do to get the ball rolling? ;)13:54
* tilman pokes Viper_ 13:54
sipOnly a couple of members requested access so far13:54
sipthe repo itself is already exported, but since there are few ports I kept a static version of the old one for now13:55
juesip: btw, could you add me to the contrib group on, please13:55
sipjue, done13:56
jaegerI'll start using it if it's ready13:56
sipyou'r gonna beat some world record if you plan to put ports also in contrib :)13:57
sipjaeger, as a user or maintainer?13:57
jaegerI got a mail from Viper_ about it but wasn't sure from the wording if it was ready13:57
jaegeras a maintainer. I'd simply import my old .sync'd contrib ports13:58
jaegerunless I need to go through the lint process again :)13:58
sipadded jaeger and jw to the contrib group13:59
cptnnow that was easy14:00
sipnp, I'm on 24/7 lately14:00
jaegerdo we have post-commit hooks with git? for the commits list14:00
sipyes, and a volunteer is needed :)14:00
sipwe have to adapt a bit the default update script to trigger mail for security related commits14:01
sipI worked on it a bit but I'm very far from getting a grasp of the hundred git commands available14:02
* tilman thinks14:03
tilmansip: let's talk about this tomorrow, maybe i can help14:03
sipgood, thanks14:04
sipso, is system/iso.git ok for the iso stuff? If so I think we finally have all the stuff needed for branching 2.314:04
jaegerIt seems fine to me14:05
sipI mapped the svn references in flyspray to git commits and added a markup14:07
sipwhick sucks compared to svn :)14:07
sipie ports/sip/40d96d2f46031b0a8249e13cb386475cc69345d5 maps to http://localhost/gitweb/?p=ports/sip.git;a=commit;h=40d96d2f46031b0a8249e13cb386475cc69345d514:08
sipsorry for the small terminal guys ;)14:08
sipwell, apart from the localhost thingie14:08
tilmanwe got the idea :)14:08
sipnow, small fixes apart, do we want to define a date for branching 2.3 and a target period for the release?14:11
tilmani have no idea how much work it is to update glibc and gcc14:12
tilmanso it's hard to estimate how long our 2.3 time window should be14:13
cptnfrom my initial tests, it should be pretty straight forward14:13
sipI'm more worried of xorg :)14:13
cptnhowever, I likely won't have time in the next couple of weeks14:13
sipbtw, tilman: renamed master in 2.2, it's now exported httpup/rsync:
tilmanre. xorg...14:15
cptnsip: it's mainly in installer issue, right?14:15
tilmandepends on how content you are with the set of ports that i have etc14:15
tilmanor whether we agree to restructure things ;D14:15
sipcptn, yes.14:15
cptnsip: how about we just have a magic 'xorg' entry14:15
cptnwhich - if selected - installs the whole set of packages?14:16
cptnsure, it's kinda besider the point of modular X14:16
sipwell, the group patches should be pk to handle this14:16
jaegerand it should be made clear that that's a metapackage14:16
jaegeror metaentry, if you prefer14:17
tilmanis it a problem to make the default set of packages to install the ports that 'xorg' depends upon?14:18
sipwell, with dependency tracking one could leave out xorg and all the (minimal) deps are automatically added if something like fierofx is selected14:21
sipthat was firefox, obviously :)14:22
tilmanyou mean iceweasel? ;)14:22
sipthat one :)14:22
tilmanwe could also only put the minimal xorg packages on the iso14:22
tilmanand always install all of them14:23
tilman+ maybe the most popular drivers14:23
tilmanto not piss people off too much ;p14:23
cptnsounds sensible14:23
sipis it much bigger than R6?14:23
sipif not, I'm also fine with including all the packages14:24
tilmani don't even know, never built all of the ports at the same time i think14:24
sip <- this handles groups and dependencies, if anybody plan to work on the setup14:25
tilmanso essentiall what's needed for 2.3 is setup work for the dreaded xorg, glibc and gcc update and testing?14:25
tilman(+ the damned texinfo stuff, which i haven't worked on since the last meeting)14:26
sipyes, I think it's more a question of choosing the way of doing things rather than do them14:26
sipunless we want some deeper change for 2.3, but I'd leave them for 3.014:27
sipwhatch out Debian, we're catching up! ;)14:27
tilmani'd prefer getting 2.3 rather soonish14:27
*** entreprenerd has quit IRC14:29
sipyes, there is some stuff as R7 that's almost mainstream now and should be updated....unless we're *really* after debian footstep14:30
juelooking at DevVison I see one point that requires some work too: non-root modes for pkgmk and ports14:31
*** sepen has quit IRC14:31
jaegerIs there any true solution yet? fakeroot seems the best to me but still fails on some things14:33
sipcan't we just move that some line below?14:34
*** sepen has joined #crux-devel14:34
sipI mean, just adding a section in the official handbook for fakeroot building of packages14:34
sipwithout forcing it as the standard way (yet)14:34
jaegerI would prefer that method, myself14:35
jaegerI'd prefer it never be default, personally. It's easy to set up.14:35
sipfine form me, regarding 2.3 at least14:36
juecptn ?14:37
Romsteri've jsut got the hang of chroot that could be used to make a clean safe enveroriment.14:38
cptnmmmh, it would be a nice feature :-)14:38
sipgah, you ruined it all ;)14:38
cptnseriously though, I don't mind if we add it to the handbook14:39
tilmansip: agreed14:39
sipwell, we don't necessarily have to take a decision now, this can be worked on as 2.3 gets developed14:39
cptnit's certainly better than doing some special case handling for those ports that fail14:39
juecomments on my suggestion to use exim as the sendmail replacement ?14:42
jaegermakes no difference to me, I remove sendmail in favor of postfix anyway14:43
sipsame here, qith qmail14:44
cptn!sendmail sounds good14:44
tilmanmay we ship qmail binaries?14:45
sipI think so, if the ports is in line with djb layout14:46
sipI meant the package14:46
sepenhi! we are develop a web for crux-es, now we need a logo but Im in doubt with clasical crux logo or this what are your opinions?14:46
cptnsounds off topic here14:49
juetilman: my suggestion is mainly based on the fact, that exim works without any post-install script14:49
sipsince we're OT i like this:
tilmanjue: ah, good point14:49
juein the same way as sendmail does14:49
jaegersepen: the text alignment in the table of contents seems broken but otherwise looks ok to me14:50
cptnyeah, I'm not sure if the current postfix port could be simplified14:50
sepenthanks jaeger14:50
sipI vote for exim, for it has fewer deps14:51
tilmani'm fine with anything that's not sendmail14:52
*** Romster has quit IRC14:52
*** Romster has joined #crux-devel14:53
*** treach has joined #crux-devel14:55
sipsleeping time for me. see you soon.14:56
jaegertake care14:56
cptngood night14:56
jue night14:56
*** sip has quit IRC14:56
*** pitillo has joined #crux-devel15:08
*** Romster has quit IRC15:18
*** Romster has joined #crux-devel15:20
*** Romster has quit IRC15:30
*** Romster has joined #crux-devel15:33
*** treach has quit IRC15:37
*** pitillo has quit IRC15:44
*** cptn has quit IRC16:13
*** jue has quit IRC16:18
*** jaeger has quit IRC16:46
*** jaeger has joined #crux-devel17:05
*** j^2 has quit IRC17:08
*** deus_ex has quit IRC18:08
*** sepen is now known as se18:44
*** se is now known as sepen18:44
*** sepen is now known as is18:45
*** is is now known as sepen18:46
*** entreprenerd has joined #crux-devel18:52
*** jmvr has joined #crux-devel19:36

Generated by 2.11.0 by Marius Gedminas - find it at!