IRC Logs for #crux-devel Thursday, 2012-10-18

*** pedja has quit IRC00:42
*** deus_ex has joined #crux-devel00:57
Romsterty jue01:52
*** mike_k_ has joined #crux-devel01:53
*** horrorSt1uck has quit IRC01:59
*** c0x has joined #crux-devel02:35
*** c0x` has quit IRC02:36
*** Nomius has quit IRC05:22
*** deus_ex has quit IRC05:42
*** deus_ex has joined #crux-devel05:58
*** deus_ex has quit IRC06:38
*** horrorStruck has joined #crux-devel06:38
*** deus_ex has joined #crux-devel06:46
*** sepen has joined #crux-devel07:04
*** sepen has quit IRC07:09
*** sepen has joined #crux-devel07:10
*** sepen has quit IRC07:16
*** horrorSt1uck has joined #crux-devel07:18
*** sepen has joined #crux-devel07:19
*** horrorStruck has quit IRC07:20
*** sepen has quit IRC07:25
*** sepen has joined #crux-devel07:27
juesepen: I have no use for swig anymore, would you like to takeover the port?07:40
sepenyep since its a dep for subversion-{python,perl}07:47
jueRomster: oh, just saw that pyopengl dpends on swig as well, so you might be an candidate too?07:47
juesepen: ok, thanks07:48
frinnstcleaning out opt? :)07:48
sepenI've installed another rc3 and worked just fine07:48
juefrinnst: only a little bit :)07:49
frinnstmight clean out contrib as well, atleast the port thats been bugging me: vpnclient07:49
frinnstseems it has slipped through every release since per left :)07:49
frinnstuntouched since 2004 atleast07:50
Romsterthe same maintainer also has pekwm and a bunch of others last i looked he was maintaining his ports.07:51
frinnstsorry. untouched since 200707:52
Romsteror has that guy just silently left all his ports to rot?07:52
frinnsti have no idea who it is07:52
RomsterRene Lengwinat, rugek at dirtyhack dot net07:52
frinnstbefore my time07:52
Romstercan't remember his nick on irc or if he was only contactable by flyspray email and then he was hard to catch up with too.07:53
juehe left us without any notice, IIRC07:54
Romsterhe has 8 ports in contrib07:54
Romsterif that's the case i may as well pick up pekwm since i am using that.07:54
Romsternot sure if i even use anything else of his.07:55
frinnstsomeone is already bumping pekwm according to git07:55
Romsterwhy wasn't his stuff marked as unmaintained in the wiki ages ago.07:55
frinnstbut the last commit he did seems to have been in 200807:55
sepenmore ports to Orphaned list maybe, you can ask him via email07:55
juethere's a first attempt for 2.8 release notes, please have a look ->
sepenjue: are you the original packager for swig?07:58
Romsternope just pekwm is all i'm using out of that lot.07:58
Romsterwhos going to email the guy?07:58
juesepen: yes07:59
Romsterdoes he even have a key in git for contrib?07:59
sepenjue: just one thing, I'd like to have one point about --enable-kernel requeriment for the toolchain in 2.807:59
juesepen: good point, will add something08:01
Romsterugh now my firefox is acting up...08:01
sepenjue: at least its useful for chroot environments (I used to run chroot crux with safe-crux on my ubuntu laptop)08:02
frinnstwhats wrong with firefox, Romster ?08:02
Romsterdunno i might have a broken dep perhaps... or having two firefoxes 32bit and 64 bit on the same profile was a bad idea.08:03
sepenjue: and what about the udev decision? should we have something in rnotes?08:03
frinnstsepen: +108:03
*** deus_ex is now known as pedja08:03
frinnstthe fact that we will never use systemd should probably be mentioned08:03
frinnstmake crux a safe haven for redhat refugees :)08:04
Romsterwe oppose all forces of systemd and pulseaudio :D08:04
juewell, I don't think that for such a statement the release notes are the right place ;)08:04
frinnstwell in the announcement then :)08:05
sepenkill with power!!!!08:05
Romsterweird now firefox loads up... took long enough08:05
frinnstwhats the reason you have a 32bit firefox?08:05
sepenjue: ok, so where should be the right place for you?08:06
sepenjue: I've no experience with that :P08:06
Romstersome deprecated video driver has been removed from xorg s/driver/drivers08:06
Romsterxcb has been split up? or is that been like that for awhile now.08:07
Romsterwhat about glu separate from mesa3d08:07
sepenjue: would be ok to add something about the migration from libungif to giflib?08:09
sepenjue: its only a port rename since both are ABI compatible IIRC08:10
sepenahh no, sorry is not part of opt iso packages08:10
sepenmy bad08:10
frinnstwhen is 2.8 expected? i've been a bit afk for a couple of days08:11
frinnstshould probably get a x86_64 iso rolling08:11
sepenjue: did you see pitillo comments about bash the other day?08:12
jaegerjue: didn't get a chance to write the message last night, sorry. will do it this morning08:16
frinnstjaeger: when do you expect to start building a 2.8 final iso?08:17
jueRomster: thx, glu and xcb added somehow08:18
jaegerwell, about to write the rc3 announce so people can test it but I think we're pretty close08:18
jaegerthe changelog, release notes, and handbook had to be updated but I doubt the package content will change08:18
jaegerunless someone updates a port that needs to go on the ISO08:18
Romsteri can't think of anything else.08:18
jueuntil now there's nothing important08:19
Romsteri haven't been that involved in 2.8 release.08:19
Romsterand i still have a ton of ports to unbreak.08:19
juesepen: yes, I've read pitilo's comment, do you know what kind of problem he has with bash?08:20
sepenjue: yeah, he talked about to use --without-bash-malloc and use glibc's instead which seems better08:20
juewhat is better?08:20
sepenjue: nowadays we're crosscompiling and seems he found something interesting on bash08:20
juenever had any issue with bash08:20
sepencan't remember now, maybe that bash's malloc in not maintained anymore08:21
sepenor obsolete08:21
sepencan't rememeber all sorry08:21
jaegershould we save that change for after 2.8's release or no?08:21
frinnstI've had lots of crashes with bash.. especially when rebuilding everything: prt-get update -fr $(prt-get listinst)08:21
juewell, it's still the default in bash08:21
frinnstthat's why i use dash as /bin/sh08:21
frinnstwell, thats the only time i've had crashes :)08:22
sepenjue: but seems that glibc can handle that and better08:22
juejaeger: don't see a reason to do such a change now08:23
jaegernor do I, really08:25
jaegerput it on the 3.0 todo list?08:25
frinnstyeah let it be for 2.808:25
* frinnst builds bash --without-bash-malloc08:26
juewell, I'd do the change only if it really fixes something08:26
pitillohey, sorry. That comment was just for know a bit jue08:26
sepenjue: bash could be provided via 2.8 ports08:27
pitilloI was reading about that option, and I was out of curisioty about it, and asked here to see if someone knew about it08:28
Romsteri haven't seen bash crash ever.08:28
Romsteri do have kernel lockups sometimes though with binary nvidia video driver.08:28
pitillono intention to make a change in bash really, just out of curiosity08:28
Romsterjust stating i've never seen any bash issues08:29
pitilloI've never experienced problems with bash neither, but reading lfs about it, seems they prefer to relay in glibc malloc functions instead of bash's malloc calls08:29
sepenjue: +1 for rnotes now08:30
Romstermakes sense. but i haven't read that LFS for awhile.08:30
sepenpitillo: this problems appears in lfs doc from years ago (bash-2, ...)08:31
juewhat I've read is that the bash malloc is faster08:32
pitillogood to know then jue08:34
sepennah, thanks anyways08:34
pitillosepen: it isn't from years ago SVN-20121015 (bash 4.2 with this patchset bash-4.2-fixes-9.patch)08:35
*** horrorStruck has joined #crux-devel08:36
pitillobtw, it was just out of curiosity, as I don't know so much about its development. Nothing important :)08:37
sepenpitillo: sorry I read that ^^ which is 6yo08:38
*** horrorSt1uck has quit IRC08:38
sepenjue: why you dropped webfs?08:40
pitillowell, there is an explanation and it explains a bit the reasons, about using or not using it. Arguments from both sides. That's why I asked here about what does CRUX currently08:40
juepitillo: the fact the LFS is using that option isn't a real reason for use to do the same08:41
jueI'm subscribed to the bash mailing list since a long time and never heard, IIRC, complains about the internal malloc08:42
jueat all, bash is one of the well maintained programs, IMHO08:43
pitilloyes jue, but their agument is a bit "hard" that's why I asked here (and highlighted you as I saw your port maintenance) to get maintainer decision/explanation, just curious :)08:44
sepenpitillo: you done well, IMHO, that's the point of trying to improve CRUX08:45
sepenall info is always welcome, later research and decisions ;D08:46
pitillosure, improve CRUX and a little bit my small knowledge08:46
Romsteragreed sepen08:47
*** mike_k_ has quit IRC08:48
sepenjue: can I resurrect webfs to opt? I really like this port (100% of my crux boxes have it)08:50
juesepen: sure08:50
juepitillo: I've just looked at debian, they are using --without-bash-malloc only for building a static bash, not for the 'normal' version08:53
juepitillo: anyway, thanks for the info, it's always interessting to see what others are doing and maybe frinnst comes up with new insights ;)08:56
sepenwe could just move some tickets to 3.008:56
jaegertime to sacrifice the work laptop back to ubuntuland08:56
sepenjaeger: I've ubuntustudio for a year or so ;D08:57
jaegerI've been running crux on it for a long while but need ubuntu for it now08:57
sepenjaeger: another possibility is to run crux on a chroot08:57
pitillojue: great, I've seen it and I've found some comments about static linking (in sepen's link there is note about it too). Thank you for the explanation too :)08:58
jaegerhonestly I don't need more crux machines :)08:58
jaegergot plenty of them already, between real hardware and VMs08:58
sepenwell chrooted environments on new safe-crux I'm preparing worked really fine08:58
jaegerI'm still running crux on my main workstation here at work anyway, just gonna switch the laptop08:58
jueI guess #780 is already implemented?08:59
sepenat least to build/test some ports its helpful08:59
juejaeger: right (FS#780)?08:59
jaegerjue: yes, I was planning to close it with some comments when 2.8 is released08:59
jaegerthough there's no reason it needs to wait until 2.809:00
jueok, great09:00
sepen#743 and #780 to 3.0?09:00
sepenwell all the others too09:00
jaegerI'll close #78009:00
jue.oO FS#603 reminds me that we need a new maintainer for pkgutils :(09:02
jaegerpkgutils and prt-get both, probably09:02
jueFS#603 is easy to implement, but generally09:02
*** sammi` has left #crux-devel09:03
jaeger#769 won't be an issue in 2.8+09:03
jaegerIt wouldn't hurt to fix the documentation for 2.7, though09:03
sepenwe should edit at least tickets marked as 'due to 2.8'09:05
sepenjaeger: I'll file a ticket about rootfs changes in iso.git we talked09:06
jaegerthere are quite a few updates since rc3 =/09:14
jaegerI hate releasing like that09:15
jueoops, you mean iputils and man-pages?09:15
jaegeriputils, man-pages, glib, libpng, xterm, xorg-xf86-video-intel09:16
juewell, what's the problem if we ship the old versions?09:17
jaegerThere's no technical problem, I just don't like it :P09:17
jueok, agreed :)09:18
jaegerIn the end it doesn't matter whether the time from release to updates is 0s or days/weeks09:18
sepenhow we could avoid that?09:19
jaegerNo useful way, really, unless we did a freeze. No point in that, it's not worth it just for my opinion09:19
jaegerIf it were a technical issue it might be09:19
sependo a freeze is not easy imho09:19
sepenwhat about to use another way to work in branches09:20
jaegerI was just answering the question, I don't think it's something we should do09:20
sepenI mean, instead of create 2.8 from 2.7 we can just create 2.8-rc1 from 2.709:20
sepenand increase the rcX until 2.8 be released09:20
sepenwork in rcX branches instead of doing tags for every rcX09:21
sepennot sure if I explained that correctly09:21
sepenimho the problem is to work directly into 2.8 and use git-to-rsync since the first commit in 2.8, isn't that right?09:22
jaegerWell, we do start work in a new branch when the release is close, what would be the advantage of having all those tags for every rc?09:22
sepentag is a snapshot in time IIRC09:23
sepenjaeger: did you mean we should start to work in 2.8 ports after released?09:26
sepensorry I'm a bit lost09:26
jaegerWe already work in 2.809:26
jaegerIt hasn't been released yet but the 2.8 branch is where 2.8 development is happening09:26
jaegerI don't understand what you're asking, I guess, or suggesting09:26
juesepen: there's no real solution to the main non-problem we have, we need some time between rcX and final and in the meantime upstream releases new versions nevertheless09:26
sepenso can we improve our work to avoid the problem?09:27
jaegerI don't think so, at least not in a way that's reasonable for us09:27
jaegerIt's not a real problem, anyway09:27
sepenhmm sorry just I was trying to get the point09:27
sepensometimes I'm a bit lost with my poor english ;D09:27
jaegerThe point was that I was complaining, not that there's a real issue :)09:28
juethe point is that jaeger just announced rc3 and we have already newer versions of some progs in our ports tree ;)09:29
sepenI assume that if we were a company with many employees would develop differently, but we are few ;D09:29
jaegeryeah. I think the way we do it is probably the best we can for the size of the group09:29
sepenat the end is not a problem to download the iso, and see that people is working all days to improve ports, if you have some ports to update its a good new to me!09:30
jaegerI'll continue to make the update ISOs regardless09:30
sepen2.7.1 was the updated one09:32
jaegerthat was an official interim release, though, I'm talking about my updated ISOs09:32
jueI think it's not a matter of our group size but a matter of our goal to have always latest version available09:32
sepenyep, as described in the homepage09:33
juebut jaeger said it already in nice words:09:34
jueIn the end it doesn't matter whether the time from release to updates is 0s or days/weeks09:34
sepenI know some old co-workers that used to download fedora4 and wait until fedora5 be released to update packages ;)09:36
sepennah!, you are doing a good job and that spreads :)09:38
jaegerawesome... out of the box the amd video driver installer in ubuntu 12.04 fails :P09:52
jaegerworked after a reboot but that's not a reassuring start09:54
*** horrorSt1uck has joined #crux-devel10:21
*** horrorStruck has quit IRC10:24
*** sepen has quit IRC10:34
*** sepen has joined #crux-devel10:36
frinnstaussie kids-tv seems a lot more entertaining than swedish10:38
frinnstoh, wrong chan.. sorry10:39
frinnsthmm, my bash problems seems to remain:
frinnstback to dash10:48
juehow do you trigger this?10:53
frinnstby rebuilding everything: prt-get update -fr $(prt-get listinst)10:53
frinnstday to day use works just fine10:54
*** io has joined #crux-devel10:58
*** io has left #crux-devel10:58
juefrinnst: ever tried to debug this a bit deeper?12:04
*** sammi` has joined #crux-devel12:18
jaegerhow odd12:45
jaegerdo the faults always happen at the same times?12:45
frinnstjue: nah, never bothered13:32
frinnstjaeger: not sure.. but there are always a handful of segfaults when i rebuild everything13:32
*** sammi` has left #crux-devel15:38
frinnstjue: btrfs-progs was tagged with 0.20-rc1 a few weeks ago. Also there has been some fixes recently. I'm using a snapshot from oct 415:52
*** Nomius has joined #crux-devel16:35
pitillocan someone check pciutils md5sum?18:19
*** sepen has quit IRC19:34
*** sepen has joined #crux-devel19:36
Romsterpitillo, straight off the site cdb1e5a95ab7ac67349f10c8f77a6d3c  pciutils-3.1.10.tar.gz20:41
Romsterwitch matches the .md5sum file.20:42
pitilloRomster: yes, and if you have old sources to check vs them... I didn't see any change here20:42
pitilloah, it matches in your case20:42
Romstereven one in my distfiles matches
pitilloRomster: thank you for confirmation, I downloaded it and I got different md5sum and same sources20:44
pitilloI did something wrong here. I got it again and all seems right, sorry20:44
Romsterprobably corrupt download.20:45
pitillostrange thing as diff didn't found any differences between them... may be it's related to these hours. Sorry for the noise20:46
Romsterhappens to the best of us that double/triple check and still mess up.20:47
pitilloit's clear... I was getting sources from another site... my bad20:56
Romsteroh interesting...20:59
Romsterother site must have a old copy or hacked.20:59
Romsteror corrupt.20:59
pitilloI couldn't see differences between them
*** __mavrick61 has quit IRC21:49
*** __mavrick61 has joined #crux-devel21:50
*** sepen has quit IRC22:13

Generated by 2.11.0 by Marius Gedminas - find it at!