IRC Logs for #crux-devel Wednesday, 2008-05-07

*** pitillo has joined #crux-devel00:07
*** tilman has joined #crux-devel01:21
*** namenlos has joined #crux-devel01:24
cptntilman: debian.as?01:32
tilmanexodus.xmms.se isn't "a good host" anymore01:33
tilmanso i had to switch shells ;)01:33
tilmandebian.as is owned by another xmms2 person01:33
tilmanbbl01:36
*** sepen has quit IRC02:09
*** sepen has joined #crux-devel02:09
cptnokay, enough with the crux-devel@ flooding :-)02:49
pitilloxD02:51
pitilloI wish all flooding could be like that....02:52
sepenxD02:53
aonjejeje xDD02:55
sepenbut maybe it should be better if @daily job like prtverify and checkurls would be moved to another ML (imho)02:55
cptnI'm also not sure if they have to run daily02:56
cptnat least the prtverify one02:56
cptnI wonder if it could be triggered by post-commit hooks, to only check the newly committed port...02:57
aonwasn't there a crux-commits@ at some point?02:57
sepenyeah, crux-commits sounds a good place for02:58
cptnmmmh02:58
cptnbut crux-commits is very verbose02:59
cptnso if one's only interested in those reports, crux-commits has a bad S/N ration02:59
cptn*ratio02:59
aonah right, it still exists, i even subscribe to it and read it daily :)02:59
sepencrux-commits shows the same output as the gitweb atm03:00
sepenerr, gitweb for core/opt/xorg03:00
aonyeah, seems a bit overkill to found a new list for those03:02
aonafter all, you can filter them to a separate mailbox if necessary03:02
sepennot when reading from the web03:03
cptnideally, we'd make the scripts silent in case there's no problem03:03
cptnand fix all the existing problems03:03
cptncheckurls reports a number of false positives03:04
sepenimho ML's subscriptors should be only persons and not a box with crontab jobs03:04
sepenspecially crux-devel03:04
* sepen needs 10min for a cofee03:05
cptnalso, the scripts could send their output to the maintainers listed in the Pkgfiles03:05
cptninstead of the list03:05
pitillomay be it can be good to solve/fix little problems faster (if maintainers don't take a look to ML, which is quite strange)03:38
pitillobut how much effort can it be and which advantadges give to CRUX? I mean if it's good to expend that efforts03:39
aonsepen: don't read it from the web03:39
sepennow here03:44
*** DarkNekros has joined #crux-devel04:03
*** deus_ex has quit IRC04:14
namenloscptn: why would i need the -R swith for git-diff?04:21
cptnbecause otherwise it's the wrong way around04:35
cptnif you do a checkout04:36
cptnand then remove a port04:36
cptnports -u should get that port again04:36
cptnif you run diff, it's show that port as D04:36
cptni.e. "will be deleted if you commit"04:38
cptnbut what we want to show is "will be added if you reset"04:38
cptn-> -R04:38
namenlosstupid question: what if i do the reset before the diff? i think i didn't get it...04:57
*** mike_k has joined #crux-devel05:03
cptnnamenlos: if you do the reset before, you don't know what you resetted05:06
cptnso you'd do:05:07
cptnreset05:07
cptnerr, wait05:07
cptndiff05:07
cptnreset05:07
cptnfetch05:07
cptndiff05:07
cptncheckout05:07
cptn(that's for an update)05:09
cptnclone05:09
cptndiff05:09
cptncheckout05:09
cptn(for the initial fetch)05:09
mike_kcan anyone take a look at http://www.archlinux.org/news/205/  Looks like it can be implemented in 2.5 together with UTF8 transition06:16
cptnI think predatorfreak wanted to look into that06:22
cptnnot that link, but subject06:22
mike_knice06:23
sepenhttp://crux.nu/bugs/index.php?do=details&task_id=17106:24
sepen(FS#171 - utf8 locale)06:24
mike_kyep, that is a bit related06:24
cptnor maybe it was tilman06:25
sepenwas tilman x)06:25
*** teK has quit IRC07:03
*** DarkNekros_ has joined #crux-devel07:04
*** DarkNekros has quit IRC07:05
*** teK has joined #crux-devel07:19
cptnI wonder if ports(8) should run as a dedicated user07:22
cptnor rather could07:22
*** DarkNekros_ has quit IRC07:28
*** DarkNekros_ has joined #crux-devel07:29
mike_kmaybe it should require root privileges, but drop them for the actual filesystem operations07:32
cptnyeah07:32
*** predatorfreak has joined #crux-devel07:36
Romstercptn, yeah i'm thinking of patching ports so it checks it the directory is writable so i can use it as pkgmk user07:44
*** mwansa has quit IRC07:46
*** DarkNekros_ is now known as DarkNekros07:46
*** mwansa has joined #crux-devel07:48
*** mwansa has quit IRC08:16
*** mwansa has joined #crux-devel08:45
*** treach has joined #crux-devel09:26
*** mwansa has quit IRC09:32
*** mwansa has joined #crux-devel09:38
*** DarkNekros has quit IRC09:48
tilmanwow, spam on crux-devel10:13
tilmanoh wait, it's just the git ports driver thread ;)10:14
*** namenlos has quit IRC10:52
cptnheh10:58
*** jtnl has joined #crux-devel10:58
*** namenlos has joined #crux-devel11:00
*** sepen has quit IRC11:06
cptntilman: could be an alternative to the rsync driver11:08
cptnthis way, you could get rid of the syncing alltogether11:09
tilmanyeah. i'm not sure about the traffic though11:09
tilmannot sure whether i want to have a real git repo lying around11:09
tilmangrabbing the whole repo everytime is kind of dumb though11:09
cptndoes clone --depth 1 still grab the whole repo?11:10
cptnthat would indeed suck11:10
tilmanwell, it grabs the latest revision of every file11:10
*** namenlos has quit IRC11:56
*** sepen has joined #crux-devel12:00
*** jue has joined #crux-devel12:29
*** jtnl has quit IRC12:33
sepenhey jue12:44
sepenxD12:44
juehi sepen12:44
sepenyour php split idea just rocks12:45
aonevening12:45
mike_kjue: hi12:50
juesepen: thx, I'm on the way to commit the ports12:50
*** pitillo has quit IRC12:51
*** pitillo_ has joined #crux-devel12:51
mike_kjue: I'd like to ask if it's intended that new (php-apache/mod_php) does not read any of the php.ini-dist/php.ini-recommended ?12:53
*** pitillo_ has quit IRC12:53
*** pitillo_ has joined #crux-devel12:54
*** pitillo_ is now known as pitillo12:56
*** jue_ has joined #crux-devel13:00
jue_mike_k: sorry, you got my answer?13:02
sepennop13:02
jue_ php.ini-dist contains just the default values if no php.ini is13:03
jue_found, so installation is not needed13:03
sepenjue_, did you read my openbsd attachment?13:03
jue_But, the user should cp php.ini-recommended to php.ini anyway13:03
jue_sepen: yes, but not to the last detail13:04
jue_mike_k: objections?13:05
sepenjue_, hmm just to take an idea about installations for extensions13:05
sepenobsd provides phpxs command, imho could be interesting for include something similar inside the php port13:06
jue_I think we should keep the things still simple, with the new structure it's easy to add extensions13:08
mike_kjue: hmm, cp php.ini-recommended to php.ini could be documented/advised explicitly13:08
sepenjue_, yeah just KISS13:09
sepenonly was for getting more ideas13:09
mike_kjue_: yep, I guess easy enough. I'll start switching to this structure on a few boxes after making ports for some extensions.13:10
jue_fine, suggestions are welcome ;-)13:10
mike_khmm, and what anout those few lines in php.ini. User supposed to merge php.ini  and php.ini-recommended...13:11
jue_ah, I have changes something in that area:13:12
mike_kisn't it ok to put those few lines (just with extension=foo) to conf.d/basic_extensions.ini13:12
jue_we have now a extensions.ini in /etc/php/conf.d13:12
mike_kright13:12
jue_an no php.ini in /etc/php13:13
jue_that's what I want to commit13:13
mike_kare you going to commit to 2.4/opt or somewhere else?13:13
jue_yep, to opt13:14
*** jue has quit IRC13:14
sepenjue_, how many extensions?13:15
jue_?13:15
sepenphp-mysql only?13:15
sepenin opt/ I mean13:16
mike_kyes, that is important since adding --enable-foo is faster that building separate port for anyone going to modify mod_php13:16
mike_kmaybe to wait for some extensions to arrive13:16
jue_yep, at least for now from my side13:17
sepennice13:17
mike_kjue_: have you tried spawn-php from lighttpd? it's nice enough to replace simple php rc script13:19
jue_yes I know it, used it prior. But nowadays I use the -b option of php-fcgi to spawn fcgi processes.13:22
mike_kI've used that a bit. any complains so far?13:26
jue_mike_k: I've written two wiki pages for mathopd about that13:26
jue_http://www.mathopd.org/wiki?PhpFastCGI13:26
jue_and13:26
jue_http://www.mathopd.org/wiki?SpawnFcgi13:26
jue_no, but I don't see any advantage if you can use a TCP port13:27
mike_kok, I have to explore the features of each before choosing the right for me13:29
*** jue_ is now known as jue13:30
*** predatorfreak has quit IRC13:35
*** treach has quit IRC13:56
juehmm, seems that [notify] does not work14:18
tilmanwonder whether my new post-update hook broke it14:26
aon:-(14:27
juetilman: wrt php.ini location, I'm unsure here. Both mod_php and php-fcgi are more server than user-software so /etc seems right, php (the port) is comparable to every other interpreter so /usr/etc might be better ...15:13
*** predatorfreak has joined #crux-devel15:13
jueopinions?15:28
predatorfreakjue: On what? XD15:34
tilmanjue: no idea15:34
predatorfreakjue: Both don't really get touched frequently.15:36
predatorfreakBut the generally-used location is /etc/php.15:36
predatorfreakAs far as I've seen.15:36
mike_k /etc(/php) is very common15:39
jueok thx, keeping /etc/php for now15:42
*** jue has quit IRC16:49
*** mike_k has quit IRC17:21
*** pitillo has quit IRC17:50
*** tilman has quit IRC20:15
*** tilman has joined #crux-devel20:16
*** sepen has quit IRC20:58
*** predatorfreak has quit IRC22:09

Generated by irclog2html.py 2.11.0 by Marius Gedminas - find it at mg.pov.lt!