IRC Logs for #crux-devel Friday, 2012-06-15

juejaeger: wrt gdk-pixbuf, thanks for the pointer, dunno why I've ignored the missing libs, but never had problems with it over the last months ;)02:21
*** mike_k has joined #crux-devel02:44
*** horrorStruck has joined #crux-devel03:37
horrorStruckjue: you may know this already but new glib needs python at build time03:47
Romsterusr/bin/env: python2.5: No such file or directory03:51
Romsteri didn't notice that before...03:51
horrorStruckhi Romster03:52
horrorStruckxf86-input-synaptics 1.6.2 is out ;)03:52
Romsterthe way linux is going it's driving me crazy03:52
Romsterhi horrorStruck03:52
horrorStruckit's just for some tests03:52
Romsterk i'll bump that, i did check xorg yesterday.03:52
Romsterprobably could be disabled.03:53
horrorStruckthere's no --disable-tests at least03:53
Romsteryeah i don't see one either.03:56
juehorrorStruck: what do you mean with 'needs', builds fine for me even without an installed python03:56
Romsteri do see use internal pcre wonder why it's not defaulting to the system one.03:56
jueoh no, I'm wrong, sorry03:57
Romsteri just found that03:57
Romsterwhich i did not notice before.03:58
horrorStrucksomeone is not using a clean chroot :P03:59
Romsteri didn't build it after safe-build clean04:00
Romsteri assumed jue did all his checks on that.04:00
Romsterso i only tested my stuff with what's on the depends on line only.04:00
horrorStrucki'm using jue's glib since always but just discovered that while updating my chroot04:01
Romstermoral of the story always check other peoples work and don't take anything for granted.04:03
Romsterprobably could patch glib to not run them tests.04:04
juewell, it's no a big issue currently, because most/all users probably have python installed, but it's a bug we have to fix of course04:05
Romsterpython2 (optional) - for gdbus-codegen04:07
Romsterbut i don't see a way to disable that if it's so called optional.04:08
jueRomster: where do you see this?04:08
horrorStruckRomster: check their PKGBUILD, python2 is a makedep04:10
Romsteryeah i see that...04:10
horrorStruckthis optional dep is at runtime i guess04:10
Romsterit'll be the build time dependency not run time.04:10
Romsterrun time optional as horrorStruck suggests.04:11
horrorStruckmakedepends=('pkg-config' 'python2')04:11
teK_would anybody be so kind and try to rebuild chromium? It wont work for me:
Romsteri do see that horrorStruck04:13
teK_should berelated to thenew glib, the latest release fails at link time -_-04:13
RomsterteK_, sure but it'll take some time (understatement)04:14
teK_a ccached installation took  ~2m until linking *g*04:16
teK_but you have to get there, first :o04:16
Romsterun i got to bump the port first i imagine04:16
Romsteroh your on 18 still04:17
teK_the glib problem occurs with 18.xxx04:18
teK_the linking problem with 21.xx04:18
teK_21.0.1174.1 to be precise04:18
Romster$ df -h /var/cache/ccache04:18
RomsterFilesystem               Size  Used Avail Use% Mounted on04:18
Romster/dev/mapper/data-ccache   20G   17G  3.5G  83% /var/cache/ccache04:18
Romsterall my chroots and system point to the same cache that sits on it's own LV04:19
Romsterthat is striped over a few disks.04:19
jueteK_: look at my gtk-engines port, a similar sed line should fix the glib error04:19
Romsternss bypasses the CC so it wont work with ccache04:20
teK_jue I will, thanks04:20
teK_I want to get rid of the red broder in chromium :O04:21
prologicI saw tilman's farewell email04:23
prologicsad to see him go - he's done quite a lot for us all04:23
Romsteryeah i'm sad too04:23
prologicwhere is crux heading in the future guys?04:23
Romsterthe same as it always has?04:23
prologicwe seem to be lacking steam04:23
prologicand I'm of no help :/04:23
Romsteryou would be if you pickled up a few ports in contrib..04:24
prologicperhaps I should04:24
prologicbut I have such limited time04:24
teK_  /o\04:24
prologicwith work, family, etc04:24
Romsteri don't have much time either.04:24
prologicif I were able to get some decent processes in place04:24
prologicI might be able to afford an hour or two updating, fixing ports :)04:24
Romsterper a week :D04:25
teK_you may laugh but I spare quite sometyping by calling ccommit instead of git commit:04:25
teK_function ccommit() { # {{{04:25
teK_old=$(git diff Pkgfile | sed -n 's/^-version=//p')04:25
teK_new=$(git diff Pkgfile | sed -n 's/^+version=//p')04:25
teK_echo git commit . -m "'$(basename $PWD): $old -> $new'"04:25
teK_git commit . -m "$(basename $PWD): $old -> $new"04:25
teK_} # }}}04:25
teK_you may laugh = you are not allowed to laugh!04:26
prologicthe most significant ports I do maintain are mythtv04:26
Romsteri thought about doing $(basename $PWD) but was too lazy to bother.04:26
prologicanything else in my personal repo I can't gurentee04:26
Romsteri might do that now that teK_ has saved me some effort.04:26
prologicI wish we'd switch to mercurial :)04:26
prologicand just bloody well make it a bitbucket repo :)04:26
Romsteri have a C app that reads the diff off git and commits but i can't fine where i put the source for that.04:26
Romstergit is not hard to use i can use it and i'm less of a programmer than you prologic04:27
teK_lucas' (?) idea about github was not bad either..04:27
prologicwell even github yeah04:27
prologicI don't mind particularly04:27
prologicbut I am particulrly fond of mercurial04:27
Romsterif it's that painful, you use hg push to some repo i setup here then i can merge that into git. if that is any help.04:28
prologicoh no it's quite alright04:28
prologicI can actually manage git repos with mercurial04:28
prologicI suppose I just have to commit the time to get back into contrib again04:29
prologicand tidy up all my personal ports04:29
prologicI did have a small system/process working for me once when I had heaps of time :)04:29
prologicihmo, I think crux needs a new website and a new domain - we need to relaunch crux to the public :)04:29
horrorStrucki like my underground crux04:30
Romstera part of me died when changed to not as underground.04:31
horrorStrucknot being in distrowatch top 100 makes me feel so 1337 :P04:31
Romsteralso sad too that we do need more coverage. i was thinking of getting some crux shirts made up as i know the business here that can do embroidery04:32
Romsterbut posing around in this town wouldn't get me many users if any.04:32
prologictbh I do like the underground-ness too04:32
horrorStruckdon't we have any idea about the userbase?04:32
prologicthe neich (spelling error?)04:32
prologicactually we don't04:32
prologicwe know very little about crux's user base04:32
prologicwe used to have a map, but I can almostly certainly gurentees it's useless04:33
Romsteri'd like ot know how many unique ips there are for core.rsync fetching.04:33
horrorStruckfrom rsync maybe?04:33
prologicthat might not be indicative of crux users though04:33
prologiccould be mirrors, bots, etc04:33
teK_I'm fine with our domain/website but suggestions are always welcome04:33
prologicI've always thought (at the very least) we should try to acquire something like or cruxlinux.org04:34
prologic(not checked those)04:34
teK_I wanted to do a script collecting ports+arch for statistics04:34
Romsterproper wiki would be nice for the site other than that it's fine as it is.04:34
teK_what's improper?04:34
Romsterneed one that has authentication anyone can edit it currently.04:34
prologiceven if we just redesign the website a bit - with a more modern design04:34
teK_that is nottrue04:34
jue.oO the whole glib/python thing is a bigger issue than I thought first, gdbus-codegen is a python script with libs in /usr/lib/gdbus-2.0 and is a kind of swig for dbus stuff04:34
Romsteroh gawd04:35
jueRomster: what's wrong with our wiki?04:35
prologicI think he means the structure and content could do with a cleanup and update :)04:35
teK_it's a wiki, go ahead (or send diffs :})04:35
Romsterproper user account creating and logging in authentication. i don't believe anyone has to even register to edit the user wiki, the dev wiki is protected i know that.04:36
prologicbut hardly anyone does04:36
Romsteri might attack the wiki some more with text when i got time.04:36
prologictbh though (all in all)04:36
prologicI'm actually still very happy with CRUx in general as a Desktop and Server OS04:36
prologicand I don't see myself ever using anything else really04:36
Romster/usr/include/glib-2.0/glib/gutils.h:28:2: error: #error "Only <glib.h> can be included directly."04:37
RomsterIn file included from /usr/include/glib-2.0/glib/gutils.h:34:0,04:37
Romster                 from ui/base/l10n/
teK_I did:04:37
Romsteri think i've seen that error before teK_04:37
teK_  sed -i 's|<glib/.*\.h>|<glib.h>|' ui/base/l10n/l10n_util.cc04:38
teK_  make chrome chrome_sandbox BUILDTYPE=Release04:38
teK_  install -m 0755 -D out/Release/chrome $PKG/usr/lib/chromium/chromium04:38
prologichere's an idea though04:38
teK_as jue suggested04:38
Romsterbut from a different port.04:38
teK_seems to build fine (until now)04:38
prologicperhaps we could adopt a more distrbuted approach to maintaining crux and it's core/opt/contrib repos04:38
teK_how would that look like?04:38
prologicwe'd have a github repo04:38
prologicusers are free to clone/fork and submit pull requests04:38
prologicthe core crux team manages what they pull in04:39
prologicrather than having to go through a process of maintaining a personal repo for some time, then becoming a contrib maintainer, then later on joining core, etc04:39
prologiclet's make it easy for others to make crux better (at our disgression)04:39
teK_as I said before, this sounds not bad. The question is: is the higher effort for core (using/learning github stuff) "better" than our users creating patches by using git send-email?04:41
prologicpersonally I think the effort is worth it04:42
prologicbecause of the openness and better flexibility we can achieve with a more open crux04:42
Romster  sed -i 's|glib/gutils.h|glib.h|' ui/base/l10n/l10n_util.cc04:42
teK_Romster: yeah I already told you *g*04:42
prologicI believe we've (till this day) maintained too much structure/hierachy04:42
teK_I have to checkout the github process before stating an opinion04:43
Romsteryeah i was looking at it and the 's|<glib/.*\.h>|<glib.h>|' looked wrong to me.04:43
teK_Romster: with this sed call chromium will build again, just finished04:43
prologicwell as you know I'm a bitbucket/mercurial fan04:44
prologicbut the same can be acheived with github/git04:44
jueRomster: why looks it wrong?04:44
teK_rebuilding chromium did not make the red borders go away04:44
jueRomster: why looks it wrong?04:44
Romsteroh right your regex line would work ok, looked weird to me.04:45
horrorStruckteK_: there's no learning curve wrt github, it's super easy to setup, i'm sure it'd work great for that purpose04:47
Romsterrelying on github do we want to do that?04:47
teK_I bet you can keep the repos on and only pull from github04:48
teK_so github would be the write-only part of that :}04:48
horrorStruckjust mirror the repo on github, people fork it and send pull requests04:49
Romsterso someone has to look at and review everything all the time than the current contrib04:50
horrorStruckexcept if you want to give write access to someone, you can set this up easily04:51
Romstergitolite is good for that too.04:52
RomsterteK_, isn't chromium 21.... beta?04:54
Romsteri'd assume that chromium-20.0.1133.0.tar.bz2 is the most current stable verison04:54
horrorStruckcurrent is 19.x i think04:54
horrorStruck20 is beta and 21 is dev04:55
teK_Romster: maybe.04:55
Romster yeah 21 is dev 20 is beta 19 is stable.04:56
Romsterthough i'd tink dev is more alpha or no one uses alpha anymore.04:57
jueplease give it a test: a new version of pkg-config, which uses a internal glib, is in my private repo06:10
jue -> httpup sync pkg-config06:11
Romster=======> Building '/var/ports/packages/chromium#19.0.1085.0-1.pkg.tar.gz' succeeded.07:41
Romsterfinally and wrong channel...07:41
Romsteri'm testing pkg-config now07:41
horrorStruck^ got it installed too, will report if any issue08:01
Romsteri just thought of a alternative nick for you horrorStruck.... wait for it struckHorror08:04
Romsteris that what you based your name off08:22
Romstergeez that's bad...08:26
*** joe9 has joined #crux-devel08:40
horrorStruckjust tried the pull request process on github, this is *really* nicely done. any dev willing to setup an account and mirror the repo so random users like me can start forking it and send pull requests, it takes seconds to open an account. it seems no one is really hostile to this idea, do you devs think we could give it a try?08:42
teK_we'd have to adjust our repos on or each developer's local workflow08:43
horrorStruckit shouldn't modify your workflow that much, there's just a need for someone to keep the repo in sync and devs who agree to accept/reject pull requets. then pull from github master and push to crux.nu08:48
rmullWhat about using gitolite like
Romsteranother words more of us reviewing.08:50
Romsteri'm using gitolite here and i absolutely love it08:50
horrorStruckRomster: are you saying this as good thing or a bad thing?08:51
Romsterjust a statement. needs someone to review ports.08:51
Romstera whole group would be better but we need guidelines it would fall under the whos going to look after xorg debate, if it's a group we need guidelines.08:52
Romstersadly we haven't got tons of man power.08:52
Romsterso anything that's fast efferent and takes little time is a plus.08:52
*** mike_k has quit IRC15:06
*** horrorSt1uck has joined #crux-devel15:07
*** horrorStruck has quit IRC15:10
*** mavrick61 has quit IRC21:48
*** mavrick61 has joined #crux-devel21:50
*** joe9 has quit IRC22:43
*** joe9 has joined #crux-devel23:59

Generated by 2.11.0 by Marius Gedminas - find it at!