IRC Logs for #crux-devel Wednesday, 2014-09-03

*** mavrick61 has quit IRC02:53
*** mavrick61 has joined #crux-devel02:54
*** novak has quit IRC03:30
*** Workster has joined #crux-devel06:09
*** Workster has quit IRC06:09
*** Workster has joined #crux-devel06:09
*** mike_k has joined #crux-devel06:16
*** sepen has joined #crux-devel07:18
*** sepen_ has joined #crux-devel07:23
*** sepen has quit IRC07:23
*** mike_k has quit IRC08:25
prologicno md5sum mismatch here :(08:42
nrxtxprologic: can you remove the downloaded files from your directory?08:47
nrxtxits the shell script mismatching08:47
nrxtxif not i have to try that again08:48
Romsterthe docker file also mismatches i never reported it.08:48
prologiclemme double check again08:48
prologicI did pkgmk -c :)08:48
Romsterhttps://github.com/dotcloud/docker/archive/v1.2.0.tar.gz cea0b533c0fae5d2aaef6b3abd92af5d08:50
Romsteractually i think i know what it is.08:50
Romsteri have another file of v1.2.0.tar.gz because github sucks at supplying names.08:51
prologicyeah sorry no mismatches here guys08:52
Romster361f364342af357ed78ae4eabe465da8  /var/ports/distfiles/v1.2.0.tar.gz08:52
prologiccleaned everything and re-tested08:52
Romstergithub god damn sucks....08:52
nrxtxmaybe they generate them on the fly on different servers08:53
prologicPEBKAC :)08:53
Romsternot my fault i use a common distfiles directory and no names on files causes name colisions.08:54
Romsterpkgmk should save them with a name08:54
Romsteri used to use a patch and save them as $name/... but that sucked when i had more than one port that used the same sourcefile08:55
Romsterwhy i started doing this. http://202.168.100.149/distfiles/hash/v1.2.0.tar.gz/md5/08:57
Romsterlike fedora does on my distfiles mirror and then symlink the ones that match the md5sums.08:57
Romsteri should make pkgmk aware of this scheme.08:58
nrxtxi think nicest would be if you have the possibility to supply a output filename for a source file09:07
prologiclike:09:12
prologicsource=(http://foo.com/file.tar.gz#file=bar.tar.gz)?09:13
prologicthat way it's both optional and backwardsa compatible09:13
prologicand easy to parse out in pkgmk09:13
Romstersomething like that09:28
RomsterYou have been unsubscribed from the crux64 mailing list09:29
Romsterdid that recently too.09:29
RomsterteK____, any chance of removing that mailing list? crux6409:29
nrxtxmaybe just wait for a response to that of one of the core developers09:29
Romsteri don't see it as being useful anymore.09:29
nrxtxprologic: that would be a way but keep it optional at any time09:30
Romsternrxtx, think there was a feature request or if not we been talking about it on and off.09:30
nrxtxRomster: a way like prologic proposed?09:30
Romsterpersonally if we are gonna do that i'd like git hg svn support in source=() as well09:30
Romsternrxtx, yeah09:30
nrxtxwhereas all of them kill md5sums09:34
frinnstyeah relying on github tarballs isnt going to work09:34
frinnstthe crux64 mailinglist still exists? why? :)09:35
nrxtxfrinnst: what do you think about supplying an output filename to "source" or using git directly in "source"?09:36
Romsteror screw github host the tarball with a name09:40
frinnstyeah10:10
frinnstIm not a fan of git stuff in source10:11
*** mike_k has joined #crux-devel10:31
*** Romster has quit IRC10:45
*** Romster has joined #crux-devel11:28
jueme neither11:40
jaegerthe ability to force the filename would be nice, though I also agree about avoiding vcs/rcs stuff in the source array12:11
jaegermany projects have stupidly named tarballs on github, though, so that would be a handy thing12:11
frinnstI think github automagically does that when a release is tagged12:14
frinnstjust zips up the source and names it v$version12:14
juejaeger: FYI, update of my old i686 laptop with your repo was fine, no problems12:21
*** prologic has left #crux-devel ("Leaving...")12:30
jaegerglad to hear it. :)12:36
jaegerfrinnst: yeah, I think so. It causes a problem when you're building something like razor-qt which has several packages with the same version number12:37
nrxtxjaeger: do you think it might be a problem adding such a parameter to source as prologic proposed or do you have another way?13:32
nrxtx(define a output filename=13:32
nrxtx)13:32
jaegerwon't know without some testing but I imagine it could be implemented without trouble13:33
Romsterit'll brake other tools though. but oh well.14:06
Romsterhttp://202.168.100.149/distfiles/hash/v1.2.0.tar.gz/md5/ i avoid it by using a sum for my distfile mirror locally. but probably not ideal for pkgmk todo that.14:07
*** mike_k has quit IRC14:33
*** mike_k has joined #crux-devel15:55
*** dwts has quit IRC16:08
*** dwts has joined #crux-devel16:13
*** sepen_ has quit IRC18:12
frinnstwow, the util linux changelog and release notes differs quite a bit18:42
jaegerbuilding an i686 ISO if anyone wants to try it18:58
*** mike_k has quit IRC20:23
jaegerwell, it will boot and install, at least20:53
*** prologic has joined #crux-devel21:08
jaegerhttp://jaeger.morpheus.net/linux/crux/files/i686/21:24
*** novak has joined #crux-devel22:56

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