IRC Logs for #crux Sunday, 2012-07-01

Romsterjoe9, i use a seperate ccache logical volume that's striped over many disks to gain extra speed, a SSD would be even better.00:35
horrorStruckglibc 2.16+tzdata/binutils 2.22/gcc 4.7.1 I FEEL LIKE FUCKING JAMES BOND NOW01:00
Romsterno issues?01:23
Romstercare to point me to your Pkgfiles :D01:24
horrorStruckso far not really, just starting to rebuild some ports to see how much it breaks stuff01:24
Romsteri wouldn't mind to test in a chroot on what ports i need to fix01:24
RomsterteK_, i'm poking at llvm some more.01:25
horrorStrucksure np, let me updload this somewhere01:25
Romsterlibdrm 37 already out O_O01:26
Romsteri'll bump that shortly in xorg R7.701:26
horrorStrucklet's see if it builds :P01:26
horrorStruck=======> Building '/home/pkgmk/packages/libdrm#2.4.37-1.pkg.tar.gz' succeeded.01:28
Romster16:29:51 up 9 days, 16:04,  1 user,  load average: 7.73, 6.88, 4.5401:29
nogagplz_did you go over those ports yet Romster01:37
nogagplz_I'm looking forward to the report on my desk first thing tomorrow morning01:38
horrorStruckRomster: sha1 7ed04e6580d38ec5a9bd062512945e34db0ed911 romster.tar.bz2 hopefully i didnt forget anything01:42
Romsteri'll let you know if anything is missing.01:48
Romsteranything to watch out for?01:48
Romsterthis is i686 testing i see.01:50
horrorStruckyes i68601:51
Romstercool i'll start that up in another chroot.01:52
Romstercloog ppl changes too?01:52
horrorStruckdont forget i'm one who asked *yesterday* on how to properly build a toolchain :P01:52
Romsteryeah lol01:52
horrorStruckdo we even have them in repos?01:53
horrorStruckno we dont01:53
Romsteroh new deps i see.01:53
horrorStruckenables graphite branch in gcc01:53
horrorStruck(and maybe doing more stuff)01:54
Romsterlast time i tested graphite in 4.6.x it was horribly broken.01:54
Romsteri'd like elfutils in core for -flto support on gcc.01:54
Romsterbut not enabled by default, let the user decide if they want to enable that on some ports.01:55
Romsterworks on the majority of ports.01:55
horrorStrucki use graphite all the time without any issue. flto is nice but i have quite a lot of build failures01:55
Romsteralso gold linker i have never tested that might be worth a try01:55
horrorStruck-fuse-linker-plugin doesnt work for me for some reason01:56
Romsteryou can mix -flto and non -flto ports together without any issues.01:56
Romsteras i was doing, but i plan to hack pkgmk a little to have a /etc/pkgmk.d/portname file for fine tuning flags, than my older hvl attempt of PKGMK_LIBDVDNAV_CFLAGS='-O3' PKGMK_FLAC_CFLAGS='-O3' style.01:58
Romsterthat picked off the name= and used that as the uppercase var.01:58
horrorStrucki like the idea, i see many case where this would be useful01:59
Romsterthat way i can also strip out -flto and have that on be default.01:59
Romsterstrip out -flto where it brakes a port.01:59
Romsteryeah users can share files of there settings that way too. or ports can even install there own defaults if that was a nice idea.01:59
Romsterit's a bit ricing though but users can choose there own settings.02:00
*** vaddi has joined #crux02:00
horrorStruckhave to leave right now but i do have some questions about this, 'later if you're still around02:02
tilmanhorrorStruck: so either you're fucking james bond or this shit isnt as much voodoo and black magic as you thought it would be :D02:06
Romsternot so bad it's when you cross compile is when you hit the hard stuff.02:18
joe9romster: ghc build segfault on x86_6403:11
joe9maybe it is the LDFLAGS that I have. I will try without them and keep you posted.03:11
Romsterlast time i tested was a few weeks ago built on both arches ok03:12
joe9Romster: no --enable-udev in libdrm anymore..03:18
Romsteroh damn it...03:18
Romsterjust finished bumping gtk-32 and friends.03:18
joe9ok, thanks.03:20
Romsterwhat was that ruby thing about wxgtk? or was that tk03:23
joe9i don't know. it built fine now.03:24
Romsterpthreads or something?03:25
joe9yes, there is some message about pthreads03:28
joe9romster, does gcc-fortran build on x86-64? it is not for me.03:29
joe9not sure what needs it. but, I cannot get it to build.03:29
Romsterno it's not 64bit i can make a 64bit version if it's needed.03:39
joe9Romster: yes, please. I need the 64 bit version of the gcc-fortran03:40
joe9I need libf2c which is dependent on that.03:40
Romsterk give me some time i'm doing a ton of compiling as it is now.03:40
joe9footprint update..03:42
Romsterare you on multilib or pure 64bit?03:45
joe9on pure 64 bit.03:47
joe9Romster: is python-stdlib-extensions still relevant.03:47
joe9the files in the .footprint are being installed from python package itself.03:47
Romstermight have been merged now perhaps?03:48
Romsterdo they all conflict or only some?03:48
joe9all, I think.03:48
Romsterxorg-font-inconsolata-dz fixed.04:01
Romsteralso missing dep added fontconfig for post-install04:01
joe9cool, thanks.04:01
Romsterhmm time to nuke that port python-stdlib-extensions04:02
Romstermy romster repo doesn't get alot of love.04:04
Romsterwith everything else i do.04:05
joe9ok, thanks.04:05
Romsterghc ok? gcc-foprtran compiling now.04:19
joe9ghc, compiling now.04:22
joe9without LDFLAGS.04:22
Romster19:22:08 up 9 days, 18:56,  1 user,  load average: 11.43, 8.45, 6.3304:22
Romsteri haven't investigated if ccache can use ghc04:22
joe9that is what I am doing. running ghc with ccache.04:25
joe9ok, I will remove it, if this does not compile.04:25
Romsterit might be broken with the libffi update and i designed it to bootstrap off installed ghc04:26
Romsterif found04:26
joe9i do not have an installed ghc. fresh install.04:27
Romstergcc-fortran failed needs -fPIC figures recompiling.04:28
Romsterhow did you get ccache to use ghc?04:29
joe9oh, sorry I meant. ghc to use ccache.04:31
joe9where do you put the -fPIC?04:31
Romsterin CFLAGS04:31
Romsteror CXXFLAGS or whatever language it's using.04:31
joe9 ghc segfaulting. no cflags, cxxflags or ldflags in its' build.04:33
Romsteri'll test ghc shortly.04:34
joe9ok, thanks.04:35
joe9 is the Pkgfile04:35
joe9similar to yours, except for the unset on cflags, cxxflags and ldflags.04:35
joe9Romster: what is surprising about the ghc build segfault is that it failed at a different place this time.04:39
joe9i am thinking it might be due to something else with the compiler or space or speed on my system.04:39
Romstersome race condition perhaps04:44
Romsternot using distcc with that?04:44
Romsterdisable distcc04:44
joe9no, not using distcc. but using ccache though.04:44
Romsterccache shouldn't be an issue. ghc hardly uses that afaik04:45
teK_I had corruption issues with ccache lately04:45
Romsterwonder if i can push to contrib-x86_64 lst time i tried git on had a fit with permissions on some files.04:45
Romsteroh what exactly?04:46
teK_format error(?) for random object files04:46
teK_disabled it -> worked fine04:46
teK_Romster: try it and  let me know04:46
Romsteron what ports?04:48
Romsteroops wrong channel04:49
Romsterah no i'm on right channel teK_ is jumping channels04:50
Romsterwas a very long time ago i had ccache issues.04:50
RomsterteK_, flush the cache and retest you probably got corruption04:54
*** vaddi has quit IRC04:57
teK_did not help05:34
* Romster shrugs05:46
Romsteri can never push to x86_64
Romsteralways some file permission error.06:14
RomsterteK_, maybe you can help06:14
horrorStrucktilman: yeah true, not so much voodoo involved, it is quite a smooth upgrade (so far at least). at first when i googled "upgrading glibc" there was nothing but big fat warnings on how the sky will fall on my shoes. anyway, this makes me think that CRUX could be 100% rolling release, no?06:40
Romsteryou had a good day, it's not always good, but that could be a possibility.06:41
horrorStruckRomster: did you have time to build the toolchain yet?06:42
*** jue has joined #crux06:43
*** ChanServ sets mode: +o jue06:43
Romsternot yet horrorStruck i just finished with ghc and gcc-fortran06:43
Romsterwill soon06:43
horrorStruckok, just curious06:44
horrorStruckhi jue06:44
frinnsthorrorStruck: glibc is usually pretty painless06:44
frinnstbut with some releases they introduce compability issues06:45
frinnstif you use 2.16, try going back to 2.11 or something and enjoy the horror :)06:45
horrorStruckfrinnst: looks like but at first i really read only scary things about that. maybe that was true few years ago, dunno06:45
horrorStruckah ok that's why06:45
Romsterstarted chroot how did you build it horrorStruck ? some order or did oyu just do a pkgmk -r -i06:45
Romsterhi jue06:45
horrorStruckRomster: glibc -> binutils -> gcc -> binutils -> glibc06:46
horrorStruckbut you need cloog before building gcc06:46
frinnsthorrorStruck: would be nice if you could append the list at:
horrorStruckand ppl06:46
horrorStruckand isl before cloog06:47
horrorStruckfrinnst: will do06:47
Romsterprt-get depinst cloog ; prt-get update glibc .... should do it.06:47
horrorStruckRomster: i think my Depends are OK but i didnt use them, just pkgmk. worth a recheck06:48
horrorStrucknot sure how many times i forgot my flyspray password :(06:53
Romsterwrite it down in a notebook07:02
Romsterwouldn't a wiki page be better than flyspray?07:04
Romstertzdata for glibc too..07:11
frinnstyeah that's already on the list07:22
horrorStruckRomster: btw,"my"  tzdata is just a shameless ripoff of arch's package07:48
Romsterit works that's all there is to it. some of my stuff is a ripoff07:48
RomsterhorrorStruck,  i see what youve done minimal gcc to then rebuild the main gcc compiler? as you've only enabled gcc and no g++08:06
joe9romster: ghc and gcc-fortran build ok for you?08:09
Romsteryes but i can't push gcc-fortran to contrib-x86_6408:10
joe9just updating /usr/ports/romster now.08:10
joe9can you please paste it?08:10
joe9Romster:  what is your change to ghc/Pkgfile? I do not see any changes?08:12
joe9or, are your changes to a different repo?08:12
Romsteri haven't changed anything to ghc it works fine as is.08:13
Romsterghc will be here shortly
Romster5% to go08:14
joe9how about to gcc-fortran08:16
Romstersame patches as in core-x86_64/gcc/08:17
horrorStruckRomster: --enable-languages=c,c++,objc \  ??08:20
Romsteroh i just saw the $name-$version.tar.bz2 and assumed you removed the individual $name-{core,g++,objc}-$version.tar.bz208:21
horrorStruckah no that's new archiving style upstream08:21
joe9romster, looks like you bumped the version on gcc-fortran too08:21
Romsterjoe9, yes to match current gcc08:22
Romsteri should do the same to i686 contrib too.08:22
Romsterdon't tell me that's too new for your needs?08:23
nogagplz_gcc 2.95 ought to be enough for anybody08:23
Romster</witty remark>08:24
Romsterraid kitchen time bbs08:25
joe9Romster: how do you know that you have 5% left on the ghc build?08:28
joe9do you use a different make system that can show you the progress?08:28
joe9Romster: in gcc-fortran, why are you doing this?08:39
joe9shouldn't that be left to the pkgmk.conf variables?08:39
joe9make -j 1 DESTDIR=$PKG install08:40
joe9i like the above line.08:40
joe9if a packages will not build with -j (>1), it would be good to have that line instead.08:40
joe9frinnst: i am noticing more than a few packages which are failing with -j >1 value.08:41
frinnstlots of packages fucks up the install when running lots of jobs08:41
frinnstif they are in core/opt/contrib/xorg, let us know08:41
joe9i gave up trying to fix such packages, I removed the -j flag from my makeflags instead.08:42
frinnstsince it's a race you might not always hit the error08:42
joe9i am seeing such race with my ghc install too. it always fails somewhere else.08:43
*** mike_k has joined #crux09:01
joe9Romster: i think it is a good idea to add the make -j 1 to ghc too, given that it fails at a different place for me each time. I will paste the pkgfile after adding it.09:06
Romsterjoe9,  on the upload i meant it was on 95%09:07
joe9oh, ok.09:07
Romsterpkgadd that09:08
Romsterit's built in a chroot with x86_6409:08
joe9Romster: ghc pkgfile:
joe9ok, let me try the build one more time. before pkgadd'ing that09:08
joe9Romster: thanks for the pkg file.09:09
Romsterghc never seems to fail for me09:09
Romsterif it builds with -j1 then there is a obvious race i'm not seeing09:10
joe9i am trying that.09:10
joe9i think it is a race because it fails at a different location each time.09:10
Romsteri add to -j1 to any port i find broken of mine due to a race ghc builds faultlessly on -j409:11
joe9 qt fail09:15
Romsterqt4 also works for me i submitted bug reports to sepen to get that working... and it works here what on earth are you doing?09:16
joe9Romster: what is your glib version?09:18
joe9this seems to be a known error/bug09:19
Romsterthe one in opt09:19
Romsterprt-get sysup if oyu haven't already and make sure you've installed libffi09:19
joe9Romster: gcc-fortran built fine.09:20
joe9romster: libffi version ?
Romstergcc-fortran should it builds for me too.09:22
joe9Romster: regarding the libffi. what version do you have?09:26
joe9yes, gcc-fortran builds fine for me too.09:26
Romsterbuilding qt4 now09:26
Romstereverything in opt up to date is what i have09:26
joe9romster, i am throwing in the towel with the ghc. I am using your package.09:33
Romsternot working with -j1 even?09:34
joe9no, it failed even with that. trying one last time with -j 1 again.09:37
joe9actually, with no -j09:37
joe9 if it does not build, I will add your package and then build it from there. so, it does not have to worry with the bootstrapping.09:38
Romsterhave you tried a clean chroot?09:38
Romsteri have several for testing...09:38
joe9no -j flags gets farther along the build than having a -j 1 oslt.09:38
joe9anyway, how is the qt4 build coming along?09:39
Romsterstill going09:39
joe9Romster: do you have webkit installed?09:40
Romsterqt4 built uploading09:49
joe9Romster: thanks for checking on that.09:50
jaegerThe two of you run into more problems than I can imagine09:50
Romsternot me it's all building fine09:52
Romsterwell i do have a ton more ports and i do tweak with stuff too.09:52
Romsterto be expected09:52
joe9i think it is me because I am mucking around with cflags, cxxflags and ldflags against romster's advice.09:53
Romsteragainst my advice?09:54
Romsteri did tell you the safe things you can do from experience.09:54
joe9Romster: yes, you did. you warned me about ricer stuff too.09:54
joe9i just liked how fast firefox was running after some of the flags, hence experimenting with them.09:55
Romsterwhat the fat ed is on tv09:55
horrorStruckRomster: new version of xorg-libpixman#0.26.2-1.pkg.tar.gz09:59
joe9romster, I think qt4 needs this patch
RomsterhorrorStruck, pushed10:10
Romsterjoe9, then why does it build for me without that?10:10
joe9Romster: that is what I do not get.10:10
joe9romster : this file /usr/include/glib-2.0/glib/gthread.h10:12
joe9typedef union  _GMutex          GMutex;10:13
joe9do you have that line or is it a struct?10:13
Romstertypedef union  _GMutex          GMutex;10:14
Romstersame as yours10:15
Romsterwhy would that change10:15
joe9qt-everywhere-opensource-src-4.8.1/src/3rdparty/webkit/Source/JavaScriptCore/wtf/gobject/GOwnPtr.h : in this file, do you have glib.h10:16
joe9 -- I am trying to check if you have this already in your build10:16
joe9for some reason.10:16
joe9or, if it decides that it does not need this as webkit is already installed and has this patch.10:17
joe9romster, I cannot explain why I have the issue and not you.10:19
Romsterthe webkit i am using is the one in contrib/webkit10:21
Romsterlook there10:21
joe9let me try after installing webkit10:22
joe9Romster: on a different note, ghc build still fails:
joe9i will just use your package.10:23
Romsteri can't explain it sorry10:24
*** nogagplz_ has quit IRC10:33
*** nogagplz_ has joined #crux10:34
joe9can anyone bump the version of imlib2 from 1.4.4 to 1.4.5?11:17
joe9romster, the version has been bumped in your opt branch.11:19
joe9by someone else before me. not me.11:19
*** joacim has quit IRC11:43
*** tilman has quit IRC12:12
*** tilman has joined #crux12:13
*** ChanServ sets mode: +o tilman12:13
*** Rotwang has joined #crux12:24
*** vaddi has joined #crux12:31
Rotwangobama is half white lol13:06
Rotwangcontast betwen white folk to the left and that footboler obama in  the middle13:07
Rotwang<it wasn't meant to be racist>13:19
Rotwangwhere can I watch euro final streaming?14:12
teK_ ?14:13
teK_dunno if this works outside germanyu14:13
RotwangteK_: there is some error, I bet I have to find polish one14:14
teK_Romster: retry (git push)14:15
Rotwangdon't care who wins14:20
Rotwangdon't even know why I am watching this14:20
teK_my son wants to watch.14:22
jseSports. lol.14:23
Rotwangjse: don't like football?14:24
teK_sitting on my couch is quite sporty14:28
jseIt's more entertaining to play football yourself than watch others.14:29
Rotwangjse: indeed, but it is much easier to watch others playing it14:29
teK_I stopped playing world class football some years ago :(14:31
Rotwanghow do you call number 10**12 in english?14:45
tilman10 to the 12th power14:46
tilmanor 10 to ther power of 1214:47
RotwangI've hoped that there is a name for it14:47
Rotwangwe call it bilion14:48
tilmanherp derp14:49
tilmanRotwang: in english it's trillion14:51
Rotwang. .. pron15:00
niklaswehmm any idé what i have forgot when my raid doesnt "auto" active when booting up my system15:35
Rotwangtotal destruction15:36
Rotwangall I can say about the final15:36
niklaswethe right team won :)15:37
*** hawksbill has quit IRC15:38
Rotwangniklaswe: no idé15:40
Rotwangworks here15:40
RotwangI just assemblet it once, and it works from then on15:40
niklasweRotwang: can you paste your mdadm.conf?15:40
Rotwangniklaswe: I have generic one15:40
niklaswe I have this in my mdadm.conf  = ARRAY /dev/md0 UUID=4ed8e941:8b3bc232:eefd2a84:4bb366bd15:40
Rotwangok, I don't have one15:41
frinnstniklaswe: udev tries to be clever with mdadm-arrays. you might need to deactivate that rule and do it manually15:41
Rotwangfrinnst: doesn't kernel auto assemble it?15:42
Rotwang4:0 incredible15:42
Rotwangsuch a score in a final15:42
*** hawksbill has joined #crux15:44
frinnstno, only if you use ancient metadata15:45
frinnstits done in userspace these days with metadata v1.215:45
frinnstI had issues when running a hot spare, udev trying to be smart and assembled it in degraded mode with every reboot15:46
*** vaddi has quit IRC16:04
*** joacim has joined #crux16:19
joe9n #ghc16:44
*** mike_k has quit IRC16:49
joe9Romster: ghc version bump to 7.4.216:54
*** Rotwang has quit IRC16:55
joe9romster, I could not get ghc to build. and am checking the arch version and I think the flags are wrong.17:57
*** nthwyatt has quit IRC18:00
*** vee has joined #crux18:10
veehey guys18:10
veedid anyone here make the cruxex cd?18:31
joe9 romster, new
joe9romster: new Pkgfile:
joe9romster, just added this line:    cp $SRC/ mk/build.mk18:51
joe9Romster: i am trying with -H128m. I will keep you posted if it works.18:52
joe9Romster: also bumped version to 7.4.218:52
jaegervee: cruxex?19:08
jaegerah, the one at exton.net19:09
jaegerno idea who made that19:09
veeits pretty nifty. a crux live cd that you can install as well. its an easy way out to crux :P19:14
*** nogagplz_ has quit IRC21:01
*** mavrick61 has quit IRC21:38
*** mavrick61 has joined #crux21:40
*** vee has quit IRC22:10
*** nthwyatt has joined #crux22:49
*** zhvtar has joined #crux22:51
*** zhvtar has left #crux22:52
*** SiFuh has joined #crux22:55
*** SiFuh has quit IRC23:24
*** nogagplz has joined #crux23:41
*** nogagplz has quit IRC23:57

Generated by 2.11.0 by Marius Gedminas - find it at!