IRC Logs for #crux-devel Thursday, 2010-09-16

*** nthwyatt has quit IRC00:14
*** nthwyatt has joined #crux-devel00:18
*** nthwyatt has quit IRC00:24
*** nthwyatt has joined #crux-devel00:29
*** nthwyatt has quit IRC06:36
*** nthwyatt has joined #crux-devel06:40
*** kreed has joined #crux-devel06:51
jaegerok, there's an rc4 iso uploaded but I'm still going to test a slightly different kernel and see if I can avoid the lenovo bug09:27
jaegerhrmm... no luck on the lenovo issue, no idea what's causing it10:08
jaegerupdated the BIOS and no luck10:48
*** jue has joined #crux-devel11:13
juejaeger: thanks for rc4, downloading now ...11:17
jaegerno problem. still working on the lenovo issue11:18
juewhat do you think, should I wait for that?11:21
jaegerI think we should for release but rc4 at least has the updated ports changes11:22
jaegerOn the other hand, I have *zero* idea what the problem is11:22
jaegerI cannot find anything useful online because thousands of people get the error message I'm seeing for lots of different problems11:23
jaegerIt seems very specific to lenovo hardware, though11:24
juehmm, odd11:25
jueand it works with 2.6?11:27
jaegerNo11:27
jaegersame problem11:27
jueno sure if we should delay the release because of that11:31
jaegerI wish I could say I'll know soon11:31
jaegerooooh, I think I have something11:33
jaegernot a solution but at least it's different11:33
jaegerneed someone else with the problem to test11:33
jaegerThere's part of the message that I wasn't considering before, it was something about a buggy MP table11:34
jaegerbooting with the "nosmp" kernel param allows it to work, though SMP is obviously disabled11:35
jaegerwhich isn't an issue most of the time for installs, I guess11:35
jueagreed, that shouldn't be a problem11:41
jaegerMy inclination is to leave it alone and if someone comes into #crux asking about it we tell them to use the "nosmp" boot option11:45
jaeger(assuming that I get some other lenovo users to verify it works)11:45
jaegerI'm also doing some rewrites of the init script right now to clean up module loading11:45
jaegerand remove the deprecated IDE stuff11:45
jueso you are going to replace rc4 with a newer version soon?11:47
jaegerI will upload it alongside rc4 so we can test both but yes, should be soon11:47
jueok11:47
*** mhi has joined #crux-devel11:56
mhi@devs: Would you mind if I would bashify some CRUX shell scripts? /usr/bin/setup etc.11:56
mhiThe shebang says bash, but it's often just regular sh syntax.11:56
juewhat's the idea behind, to improve the scripts?11:59
mhiNo functional changes, just improving the syntax, yea. (== making them more elegant)12:00
mhiOf course I could also try to functional improve them, but most scripts looked pretty okay to me.12:02
jueno objection wrt setup and other stuff we are using on the ISO, if we are talking about system init scripts most people would like to see it the other way around, meaning no bash syntax but plain sh12:05
mhiOkay, that's understandable.12:06
jaegerhrmm... running into a problem with busybox sed again14:00
jaegerI think I ran into this once before, can't remember now14:00
*** jue has quit IRC14:42
*** jue has joined #crux-devel14:44
*** sip has joined #crux-devel14:58
jueheyo sip15:13
jaegerAnyone remember how I fixed that last time?15:14
juejaeger: sorry no, but I remember that there was something with sed ...15:15
juebtw, got a 'illegal instruction' after/while the media search with my netbook15:16
juebut manual mounting works15:16
juethis is with rc4, worked with rc315:17
jaegerthink I found it again15:18
jaegerjue: that's the sed crash I'm referring to15:18
jaegerhttp://crux.morpheus.net/irc/?channel=crux-devel&date=08Jul2010#11:21:2715:18
jaegerhrmm, I guess that isn't it, it's already built with -O015:20
sipheya15:21
jaegerhey15:22
juesip: have you seen ticket #646?15:24
jaegerjue: when it fails there can you run the grep and sed commands manually?15:24
jaegerlines 46 and 47 of init15:25
sipjue, yeah it's been there for ages :-S15:25
jaegerok, I'll temporarily work around this with awk instead15:33
jaegeractually, awk will give cleaner output anyway15:33
jaegerjue: ok, I've got a fix I'll upload now15:36
mhiAre you working with a bash script by any chance? You can often use bash's parameter expansion to save using external commands like sed, awk, basename etc.15:36
jaegermhi: busybox ash in this case15:37
mhiah, oki :>15:37
jaegermhi: http://crux.nu/gitweb/?p=system/iso.git;a=blob;f=initramfs/init;h=2898d4808a972af806afc7f5f51413b6ca60bdf9;hb=2.715:38
jaegerjue: ok, fixed rc4 is up if you wouldn't mind testing it15:40
jaegerjue: I'd recommend making a copy of rc4 and rsyncing over it, save time downloading :)15:40
jaegermorpheus.net::crux27rc/crux-2.7-rc4-fix.{iso,md5}15:41
juejaeger: sed dosn't work15:42
juethanks, syncing the new one now15:42
prologicsorry guys but late with this one maybe15:47
prologicjaeger, what's this lenovo issue ?15:47
jaegerif you have it on a usb stick I can upload just the initramfs15:47
jaegerprologic: the AHCI controller in a few lenovo laptops we've tried causes a hang and failure when booting the install media15:48
prologicoh15:48
prologicbummer15:48
prologicI don't have or use lenova hw :/15:48
prologicI'm guessing it's specific to this particular controller ?15:49
jaegerseems to be15:49
jaegerjue: as for the sed issue and my init rewrites, http://jaeger.morpheus.net/linux/crux/files/tmp/init.diff.txt15:49
jaegerThe comments are still in, which is visually messy, but the script is working now on all my test hardware15:50
*** sip has quit IRC15:50
prologicjaeger, would a newer kernel have fixed this bug ?15:51
prologic2.6.35.4 perhaps15:51
jaegerprologic: nope, that's what I'm using15:54
prologicdamn15:54
jaegerI've tried a few, it even exists back in the kernel that 2.6 uses15:54
prologicwoe to those that use lenovo then15:54
jaegerIt can be worked around by booting with the 'nosmp' command line option15:55
prologicI saw15:55
prologicbut what if you have this AHCI controller and an SMP system15:55
prologicthen what ? :)15:55
juejaeger: yeah, works again :)15:56
mhiprologic: The thing is, it only seems to happen with the CRUX kernel, self-compiled kernels work. Yesterday I had to disable the AHCI support in the BIOS, install CRUX, compile a new kernel, reboot with AHCI enabled and it works. (Thanks again, jaeger!)15:59
prologicwtf16:07
prologicwhat do you mean CRUX kernels ?16:08
prologicthere's no such thing mate :)16:08
tilmancrux iso kernels probably16:08
prologicyes but my point is still valid16:08
prologicthere is no such thing16:08
mhiUhm, yea, that was misworded. I meant the kernel used to booting the, in my case, the usb image.16:08
prologiccrux kernels are vanilla linux kernels16:08
prologicright ?16:08
* mhi nods16:08
prologicahh ic16:08
prologicI gotcha16:08
prologicwell I fail to see how the ISO/USB kernel might be any different to what you'll hand-compile16:09
juejaeger: booted my laptop/netbook successfully from cd resp. usb-stick16:09
juewill try a full install/update tomorrow16:10
jaegerprologic: presumably there's some other option enabled that causes it but I haven't been able to track it down yet16:10
jaegerjue: awesome, thanks :)16:10
prologicjaeger, well try a clean kernel config (make ..)16:11
juethank you :)16:11
prologiccan't remember the option16:11
prologicand let's bisect it :)16:11
jaegerjue: that rc-fix has the deprecated IDE stuff entirely removed16:11
jaegerer, rc4-fix16:11
prologichttp://linux.slashdot.org/story/10/09/16/0340226/Adobe-Releases-New-64-bit-Flash-Plugin-For-Linux?from=rss16:13
prologicrofl!16:14
prologicwhat did I tell you!16:14
prologictalk about it enough and it'll happen :)16:14
jaegerheh16:14
prologicnow we have no excuse :)16:14
prologicwe just have to wait for wine to port to 64bit too16:14
prologicso, CRUX 2.8 64bit only ? :)16:14
prologicdrop 32bit support and leave 1-2 maintainers to maintain that16:14
jaegerheh16:15
jaegermaybe so16:15
prologicseems a good idea to me16:15
prologicthis means we need 2 sets of iso/usb maintainers in the clc16:15
jueprologic: if 2 maintainers are working on 32bit there's no one left for 64bit16:16
prologicbut a main focus on the 64bit arch16:16
prologicerr16:16
prologicyeah :)16:16
prologicso we roadmap this for crux 2.816:16
prologicwho (other than jaeger) maintains and builds iso ?16:17
prologiciso(s)16:17
prologichttp://rss.slashdot.org/~r/Slashdot/slashdot/~3/8c2JU1-mpE4/story01.htm <-- geez christ16:19
prologictime to change careers me thinks :)16:19
tilmanfrinnst: dude, you need to add yourself to the Main/About page in the wiki16:48
prologicI'll add myself as an Ambassador and Endorser of CRUX :)16:50
prologichaha16:50
frinnstdo i even have an account?16:55
juefrinnst: I've added you already16:56
frinnstthanks16:56
juenp, thank you16:56
jueyou mean a wiki account?16:57
frinnsti noticed i was added to the page, but i dont think i have a wiki-account16:57
juelooks like you have one, name Fredrik16:58
frinnstthen i have no idea what the pw is :>16:59
mhiBTW, I have problems with 2 ports. w3m and ctags both download a HTML file instead of the actual tar.gz. (both opt)17:00
tilmanasdf17:00
frinnsthttp://crux.nu/bugs/index.php?do=details&task_id=64917:00
tilmanty17:01
prologicdo we arelly need w3m in opt ?17:01
prologicisn't that the w3c browser ?17:01
prologicteK_'s the maintainer for that17:02
mhiteK_: Please correct w3m's wrong download path :>17:04
mhitilman: and you ctags ;) The diff seems right, but the actual package is wrong.17:05
tilmanyeah, will do17:08
juewell, sooner or later we have to fix all sourceforge ports17:12
prologicsed them all ?17:13
juebetter would be to adjust them to the new sourceforge layout17:15
juethat's at least that what I'll do for core17:16
mhiI'd do for other ports, if you need the help.17:16
juemhi: have you seen http://crux.nu/Wiki/FileDownloading17:17
prologicI was thinking we could write a script that'll do a blanket change to the new layout17:17
prologicis the mirrors supportin pkgutils part of the standard build17:18
prologicor a patch built by Danny ?17:18
jueno that will not work, because the new url follows not always the same pattern17:18
mhijue: Ah, thanks for the pointer.17:18
mhi(Same SF error with libid3tag, which is needed for feh, teK_)17:21
juemhi: you'll run in a lot of such failures, I'd suggest to use 216.34.181.129 as the IP for all sourceforge hosts17:24
juefrinnst: ping17:24
mhiOkidoki. Thought the maintainer should know, but you're right.. automatically correcting those Pkgfiles seems the saner way.17:27
frinnstzzz17:32
prologicmhi, can't help sf mirror issues :/17:33
prologicsf is a right royal PITA17:33
*** jue has quit IRC17:34
mhiwell, it works pretty well with the /etc/hosts trick; still annoying nevertheless :)17:37
prologicof course it's annoying17:37
prologicbut what can ya do huh :)17:37
prologicmaybe crux should ship with a known working mirror included in /etc/hosts :)17:37
prologichaha17:37
mhiBTW, would it be against CRUX guidelines to 'ln -s /usr/bin/zsh /bin/zsh' like most distros do nowadays? Most modern zsh scripts use that in the shebang, too.17:38
mhiHeh, yea ;]17:38
prologiczsh binary is normally installed in /usr/bin/  ?17:44
prologicI personally would leave it alone myself17:46
prologicCRUX never really has been in the business of "doing everything"17:46
prologicif someone wants to use zsh, they'll know the binary is in /usr/bin/zsh and set shell (on their user account) appropriatly17:47
prologicor symlink it themselves on their system17:47
prologicie: /bin/sh -> /usr/bin/zsh17:47
mhiYea, I'll just use the symlink from above.17:48
prologicyeah I mean symlink it yoruself once it's instaleld17:48
prologicno need to do so in the Pkgfile17:48
prologicbut then I don't really see the point anyway17:48
prologic/usr/bin/ is _usually_ in the $PATH17:48
mhiHmm, the zsh devs (from #zsh) stronly suggested to use /bin/zsh as shebang.. I forgot why, but I'll ask later :>17:50
prologicthat's just their personal perference and subjective opinion17:51
prologicthere is no good reason to have it in /bin as opposed to /usr/bin17:51
prologic:)17:51
prologicyour shebang line could be #!/foo/bar/bang/zsh for all anyone cares :)17:51
mhiNoted. Apparently I was just naive and thought: "They're the devs, they must have a reason." ;p17:52
prologicno :)17:53
prologiceven devs have subjective opinions and personal perferences :)17:53
prologicI should know - I have them!17:53
mhiHah!17:53
prologicI dunno what's so great about zsh anyway lol17:54
prologicso many *nix colelages of mine both past and present have always loved it17:54
prologicgah :)17:54
prologicbash is good enough for me17:54
prologichaha17:54
mhiHmm, guess the main argument would be the awesome completion. :>17:55
mhiBut yea, I could live with bash, too.17:55
prologicfair enough17:55
prologicalthough I seem to get by with bash's completeion17:55
prologic*meh* :)17:55
prologicthat's why open source is so great!17:55
mhiI remember I always had problems with Debian's bash-completion package.. ;]17:56
prologicI have a very simple philosophy17:56
prologicif it's too much to read, I don't use it17:57
prologicif it doesn't compile easily, I don't use it17:57
prologicand if it's too compilcated, I don't use it :)17:57
prologic#1 is why I don't use Ruby or Ruby on Rails17:57
prologicread a few pages of the tutorials and went "screw this" and stopped reading17:58
mhiWell, I left CRUX back then (around 2.1), because my computers were too slow to compile things like gcc, webkit etc.. I used Debian since then and still quite like it. But CRUX' simplicity is just uber ;]17:58
mhi(Was there even webkit around that time? :P)17:59
prologicwell you know ...18:03
prologicthat's why you get others to compile things for you :)18:03
prologicI used to publish large archives of x86 (32bit) crux packages for most of core/opt and contrib18:04
mhi:)18:08
jaegerargh, firefox/xulrunner updates21:41
*** mavrick61 has quit IRC21:42
*** mavrick61 has joined #crux-devel21:43
*** prologic has quit IRC21:45

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