IRC Logs for #crux Wednesday, 2006-09-06

*** pitillo has joined #crux00:09
*** koefz has joined #crux01:05
*** brian|lfs has joined #crux01:34
*** brian|lfs has quit IRC01:53
Romstertilman, the cd-rw DMA issue i've had it in the past, no other devices on the same IDE cable? also try a 80 conductor cable might solve it, if its not a 80 conductor already.02:11
RomsternipuL, capacitor < nope thats how its spelled in my australian english text books.02:13
*** richi_aut has joined #crux02:30
tilmanRomster: i've got a cd-rw and a dvd-rom connected to ide102:38
tilmannot sure whether it's a 40 or a 80 pin cable :P02:38
Romster40 wire cable is alot more ridged looking, the 80 has a smoother and alot more wires on it (in the same width as a 40 wire cable)02:48
Romstertry unpluging one drive then use dma on it.02:48
Romsteryou might find one drive is slowing the connection speed down too.02:48
tilmanthanks, that's a good idea02:49
Romstersee what mode both drives run at when on there own and when on the same cable.02:49
Romsterthats what i've done in the past02:50
tilmani should also try swapping the drives02:51
Romsterthe ide bus will only go as fast as the slowist drive attached to it.02:52
tilmandidn't do that yet either02:52
tilmantrue, but i don't really care in this case02:52
Romsternow time for dinner :)02:52
tilmani expect them to be roughly equally fast anyway02:52
Romsternot sure if that'll do much but its worth a try02:55
Romsterif here roughtly the same age yeah, my problem arised when i had a say 4yo drive with a newer one on the same cable02:56
tilmani don't like messing w/ hardware03:00
tilmaneven swapping graphics cards annoys me ;)03:01
*** rxi has joined #crux03:02
Romsteri got sick of puting the cover on my pc so now i leave it off03:04
*** destruct has joined #Crux03:30
*** destruct_ has quit IRC03:47
richi_autOpenSSL 0.9.8c is out. Security Advisory:
cptnrichi_aut: already testing, but thanks :-)04:07
*** MrX has quit IRC04:07
cptnrichi_aut: update will hit core within the next five minutes04:12
richi_autcptn: ok, thanks04:12
cptnrehabdoll: just for the record, openssl seems to loop forever when using -j5 :-)04:19
*** richi_aut has quit IRC05:05
cptnrehabdoll: fix committed, so it should work now05:10
*** Romster has quit IRC05:20
*** Romster has joined #crux05:23
*** treach has joined #crux05:28
*** gaqzi has quit IRC05:51
*** gaqzi has joined #crux05:53
*** gaqzi has left #crux05:56
*** gaqzi has joined #crux05:56
*** Romster has quit IRC06:07
*** Romster has joined #crux06:10
*** mrpk has joined #crux06:16
*** richi_aut has joined #crux06:39
tilmanmaro: hey, do you read dberkholz' (aka spyderous) blog?06:39
*** bd2 has quit IRC06:44
*** MrX has joined #crux07:08
*** richi_aut has quit IRC07:16
*** richi_aut has joined #crux07:17
*** mike_k has joined #crux07:24
pitillourxvt-lite seems to use less resources, nice optimized port Han07:47
HanI don't maintain that port.07:54
pitilloHan, well, at info seems to be packaged by other and you like maintainer07:55
tilmangood that it's in contrib then!07:55
cptnit comes from 'vico'07:57
cptnhe apparently missunderstood the meaning of packager and maintainer here07:57
tilmanah, the rcs id is telling07:57
pitillocptn, who?07:59
cptnhave a look at the Pkgfile08:00
cptnthe "packager" field is set correctly08:00
pitillocptn, true08:01
HanI made the port, but after vico proposed three patches I gave him the maintainance of the port. :-)08:03
brointhemixeeem, what about vsftpd 2.0.5?08:04
brointhemixit's officially out08:04
tilmancute desert fox08:04
brointhemixwill the port be updated or should we update by hand?08:04
tilmana whole army of desert foxes08:05
cptnbrointhemix: you'll have to run 'ports -u' yourself08:05
cptnwe can maybe find someone to hold your hand08:05
tilmanor a desert fox08:05
tilmanyou can nurse08:05
* rxi nominates tilman 08:05
brointhemixcptn: thanks, i'll manage :)08:06
tilmanrxi: i'm not a fox, i only play one on tv ._o08:06
HanThere is a special sauce for that:
brointhemixlol, i got vsftpd updated already O_O08:08
brointhemixguess i started typing 'prt-get sysup' mindlessly some time ago ;)08:09
brointhemixmany thanks for your patience cptn :)08:09
treachmmh, that rxvt-unicode port might be decently "optimized", too bad it didn't carry over to the readme. :/08:13
mike_kregarding latest openssl update: since it is in core, I assume all ports(at least core,opt,contrib) listing it as a dependency are statically linked and should be rebuilded. Can I rely on that?08:23
tilmancore doesn't imply that apps link it statically08:23
tilmanthat sentence sucked, but you prolly understand anyway ;)08:24
mike_ktilman: I mean there is a rule regarding listing deps:
mike_koh, I am wrong.08:26
mike_kthat rule makes it harder to update all affected(by that openssl issue) apps. no easy way.08:29
cptnit says that you _must_ list dynamically linked08:30
cptneven if they are in core08:30
cptni.e. core/curl must list core/openssl08:30
mike_kbut core/curl might be leaved as is, while others(statically linked to openssl) must be rebuilded08:31
cptnbut which ones are statically linked?08:32
tilmanany port that links openssl statically is broken i think08:32
mike_kand what is the suggested way to find those statically linked ones08:32
cptnyou could probably use nm to spot the symbols08:38
j^2hey guys08:38
cptnhey j^208:38
cptnmike_k: if you find any in core, please submit bug reports08:39
*** Romster has quit IRC08:39
tilmancptn: what about _not_ building libssl.a?08:39
mike_kcptn: ok. thats a question of 'general usability' afterall08:39
cptnyes, "generally don't use static linkage"08:40
cptnmike_k: what would you suggest as alternative?08:40
cptnintroduce a special dependency notation?08:41
*** Romster has joined #crux08:41
mike_kof coarse not. no real solytion so far (08:42
*** Romster has quit IRC08:44
mike_kreally? I should be banned.08:45
cptntilman: a bit restrictive maybe, since there are legitimate uses08:46
cptni.e. keeping a statically linked sshd around for binary incompatible ssl upgrades :-)08:47
treachmike_k: fyi coarse -> crude, unpolished, rude.08:47
cptnmmmh, I guess 'nm' doesn't work for statically linked stuff08:48
mike_kthanks. but that way souds like of course(of track/ of road).08:49
mike_kI'd like to avoid nm anyway08:49
cptnalso, the idea behind omitting build dependencies wasn't really to not list statically linked libraries08:50
cptnso we could change that easily in the guideline doc08:50
cptnthen you could at least go through all ports depending on $lib, check whether they're linked dynamically and if not rebuild08:51
*** Romster has joined #crux08:52
mike_khmm, that rule already sounds complicated... but if its the only reliable way08:59
pitillowhen creating the pkgmk user to build ports like an unprivileged user, home directory is created. What is that purpose?09:01
mike_kkeeping sources and binary packages f.e.09:06
richi_autmike_k: so if i setup a special folder for the sorces and packages (I think this is done in the /etc/pkgmk.conf), I could delete this folder?09:09
pitillomike_k, well, in documentation keep them in /usr/ports/{}. I think that may be isn't necesary. Thanks :)09:09
mike_knp, just keeping all that in home is easier to set up sane permissions09:10
pitilloI think permisions are sane just doing what documentation says. But thinking in that too :) Is a nice way to separate ports and sources09:13
*** Husio has joined #crux09:21
*** Brzi has joined #crux09:30
*** Husio has quit IRC09:33
jjpk"irc quickies" :p09:36
*** mrks_ is now known as mrks09:36
*** richi_aut has left #crux09:39
RyoSis there a way to identify the font firefox uses to display "special symbols"?09:49
marowhy is "make install_root=$PKG localedata/install-locales" run in the glibc Pkgfile?09:49
maroto me it looks like it only generates the locales that we delete afterwards09:50
*** sip has joined #crux09:54
*** ChanServ sets mode: +o sip09:54
*** Brzi has quit IRC09:58
*** MzOzD_ has quit IRC10:22
*** Husio has joined #crux10:24
*** Husio has quit IRC10:43
*** rxi has quit IRC10:43
*** davl has quit IRC10:45
*** rxi has joined #crux10:46
*** MzOzD has joined #crux11:01
*** mjeq has joined #crux11:15
*** thrice` has joined #crux11:17
*** morlenxus has quit IRC11:19
*** MzOzD_ has joined #crux11:24
aoncdrkit didn't build out of the box :)11:24
*** morlenxus has joined #crux11:25
aonapparently some libcap headers missing11:25
*** Husio has joined #crux11:26
tilmantreach: for the 2nd time, the openbox struck me at 18:30 :D11:30
tilmanthis must be a coincidence (the exact time), but it's funny anyway ;p11:30
*** MzOzD has quit IRC11:37
treachmm. I begin to suspect openbox is coded by the spanish inquisition.11:39
treachI got bitten by that bug thrice, then I went back to fvwm... hopefully it has existed long enough for them to get the worst bugs sorted out.. :p11:40
*** Husio has quit IRC11:42
*** morlenxus has quit IRC11:44
*** morlenxus has joined #crux11:44
jjpkSpanish inquisition moved to coding? I'm amazed.11:45
treachA small joke for the bored people out there: - What's the opposite to firewall? - Watergate  :D11:46
jjpkWhat's next? The Illuminati taking over the church? :>11:46
jjpk:D :D11:46
treachjjpk: you know, you never expect the spanish inquisition etc etc.11:46
*** thrice` has quit IRC11:47
treachand you certainly don't expect something as common as a wm to completely fuck up your system at any random (or apparently according to tilman, not so random) time.11:47
tilmanit's not random11:47
tilmanit happens when rdate sets my system clock, and the difference is > MAGIC_DELTA11:48
treachbah. I've had enough of that "computers are determinitic" crap.11:48
treachafaic they are as quirqy old cars.11:48
tilmanYES, translucent windows in openbox11:48
treachand *deterministic as well. :/11:48 about a working keyboard.. :>11:49
tilmanstupid gtk checks for a composite manager atom that xcompmgr doesn't set yet11:49
tilmantreach: yesyes, i'll look into that next week i think11:49
tilmanexa and compositing is more fun11:49
treachanything that *isn't* more fun than hunting obscure bugs?11:50
*** mike_k has quit IRC11:51
tilmanunderstanding matrox card specs/driver interface specs11:53
treachheh, gotta admitt I never tried that one.11:54
tilmanbut i guess it was worth the pain11:55
treachI guess I won't be digging out my old mystiqe either way. :p11:57
tilmanthe current render accel code is G400+ only11:58
aonat least openbox changed the themes to something awful :)12:01
aonneeded to get blah41 from an old tarball12:01
tilmanuse Blah4212:02
tilmanit's a bit better :P12:02
aonand where can it be found?12:03
tilmanhang on12:03
tilmansomewhere on that site12:03
*** bd2 has joined #crux12:05
treachhuh? Mr Ratpoison wants a ob theme? ;)12:06
j^2where's jaeger? i miss him :(12:08
tilmankoefz: is #crux-bsd still alive? :P12:12
*** predatorfreak has joined #crux12:14
koefztilman, err.12:20
marocomposite, hah, that's so 2004 ;)12:21
tilmanfast xrender certainly isn't12:22
cohanpuh, my crux devel support site is now available, featuring isdn ports for the start12:33
aontilman: didn't find it there :|12:33
cohanand rsync/svn/wiki-infrastructure to easily expand the group of contributing users/developers12:33
aonanyway, ->12:34
tilmanaon: i'll dig it up12:34
treach1995 just rang and asked for it's terminal-adapter.12:34
cohanis it okay to just use the blue crux penguin for such a site? or at least to modify it?12:34
aontilman: thanks12:34
treach*its (dammit)12:35
tilmancohan: do it12:35
tilmantreach: isdn is still widely used in germany at least...12:36
treachtilman: Yeah, I suspected that. It never became a hit here since our stupid telco monopoly back then decided to essentially price it to death.12:37
treachI know it had much wider use in germany though. (Lots of radio commercials in the bus to C-BIT -97 :P )12:38
tilmanaon: did you check ?12:43
tilmanmaro: compiz should work otob with a recent xorg-server (aiglx), right?12:55
marotilman: you'll need mesa > 6.5 too and patches for compiz and the xorg-server12:59
maroI'm not sure if all krh's aiglx patches for xorg-server are applied in git13:00
maro(I'm using 1.1.1 + patches)13:00
tilmanok, thanks13:01
tilmancohan: what's this "crux devel support site"?13:01
*** koefz has quit IRC13:03
*** mike_k has joined #crux13:05
tilmanmaro: i think i'll be lazy and wait until the patches hit git...13:06
tilmanwow, there's a german xgl users community!13:08
tilmanthat's really awesome13:08
tilmangoes well with compiz.net13:08
treachwell. In 80 M people, there bound to be a few interested ones. :p13:09
tilman90% of the guys that visit those sites are gentoo users13:09
tilmani'm sure13:09
treachhehe, no doubt13:09
treachthe debian users will arrive sometime shortly after Duke Nukem 4 Ever has been release. :)13:10
tilman"In diesem Bolg werden [...]"13:11
tilmanBOLG sounds much better than 'blog' anyway13:11
tilman(i mean the blog entries)13:11
tilman((if you know german, i guess))13:12
predatorfreaktreach: I agree :)13:12
treachtilman: usually, in my experience, blogs are more interesting if you don't know the language they are written in :P13:15
marotilman: hmm, I don't use xgl13:15
maroit's way too hacky for my liking13:16
maroaiglx provides the same eyecandy, but without the hacks :)13:16
tilmanyes, but i  wasn't serious anyway13:16
predatorfreakmaro: I go for the performance improvements, not the eye-candy.13:19
predatorfreakOther than that I agre.13:20
maroI don't have performance issues with regular 2D X...13:20
predatorfreakmaro: I more-or-less mean 3D improvements.13:21
maro3D isn't faster with AIGLX than regular X13:21
cohantilman: i started "layouting"... and playing with apache config... if i am done, i'll post the link13:21
marohm, wonder why cdrkit wants sys/capability.h13:21
cohanto the "crux devel site"13:21
predatorfreakmaro: For indirect stuff it is ;(13:25
predatorfreak-( +)13:26
predatorfreakThat's the entire point of AIGLX.13:26
mike_kcptn: I can not think of a reliable way to find that apps linked to openssl statically ((13:26
cptnmike_k: well, in similar cases we usually increase the release numbers of affected ports13:27
cptnto enforce a rebuild13:28
mike_kI see13:28
cptnsince the maintainers would need to know that they're linking statically13:28
cptns/would need to/should/13:29
cptnI'm just not aware that we're doing this in core and opt13:29
marothe "build in clean chroot" approach would fix that :P13:29
mike_kmaybe that kind of rule should be pointed out in PortGuidelines ?13:29
cptnmaro: how so?13:30
marocombined with some reverse rebuild script that is13:30
marosince you'd know exactly what packages could be linked against openssl13:30
marowhat apps *need* a static openssl btw?13:30
cptnmaro: I don't really see how you'd address that problem13:32
cptnother by generating lists of affected ports13:32
mike_kmaro: that is the queastion!13:32
marocptn: yes, and so rebuild those13:33
maromike_k: I'm wondering if there are any13:33
mike_kmaro: even if none, similar things can happen in the future (shit happens)13:33
marousually I rebuild my entire system once a month or so13:34
maroit only takes a night anyway, and makes me aware of any inter-port problems13:35
bd2mike_k, "gcc c3.c -static -lssl -lcrypto -ldl && nm a.out | grep openssl_ | wc -l" => 213:35
maroe.g. the new autoconf mandir13:35
bd2c3.c is sample openssl app13:36
cptnbd2: what if you strip it?13:36
bd2cptn, good catch ;-)13:36
cptnmike_k: I really thing the better solution would be to list even static deps13:37
cptnmike_k: and then write a very small checker script13:37
cptnsimilar like revdep rebuild, checking static binaries13:37
bd2but anyway, "strings a.out  | grep -i openssl | wc -l" => 3813:37
marocptn: er, shouldn't those already be listed?13:37
maro(since you can't really build the port without them)13:38
cptnmaro: they're not explicitely mentioned13:38
*** nullpointer has joined #crux13:38
marobut they're _dependencies_13:38
cptnmaro: but the rule says "no run time deps from core"13:38
marocptn: that rule is silly, IMO13:38
marothere are lots of stuff in core that one can safely remove13:39
cptnI'm sure MarOS will do a lot better in all those regards :-)13:39
maroit's BASIX, dammit ;)13:39
marocptn: personally I think the toolchain and the like are always assumed to be dependencies13:40
mike_kbd2: seems 'strings' is not very reliable way (13:41
marobut things like curl (static or not) aren't13:41
bd2mike_k, agree13:42
mike_kbd2: there should be another way. It's unix afterall13:43
aonaon@beech:~$ tar tzvf openbox-themes.tar.gz|grep -i blah13:43
marocptn: is there a reason why the requested package name is shown for "prt-get dependent"?13:44
cptnmaro: yes, that's an ugly ugly bug13:44
maroah :)13:44
cptnactually, it's a silly bug13:44
cptnbut fixed now :-)13:45
marodude, you rock13:45
cptnthat's what my fortune cookie says too13:45
predatorfreakBloody hack idea, but you could just parse through the depend info of all ports and rebuild all the ones that depend on OpenSSL.13:45
predatorfreakLike I said, bloody hack.13:45
maroreally love your idea about making prt-get dumber and using scripts for the cool stuff13:46
maromuch more flexible :)13:46
maropredatorfreak: problem is, the "port guidelines" tells not to mention runtime deps in core13:47
maroso really the ports that depend on openssl are being naughty, if you take that rule seriously13:47
maroor was it compile time? or both? :)13:48
predatorfreakmaro: Well I say that's silly ;)13:48
maropredatorfreak: me too, but say that to the rule author ;)13:48
mike_k...runtime deps in core that are not dynamically linked13:48
predatorfreakTo whomever decided that: Your rule is silly.13:48
maromike_k: huh?13:49
maros/dynamically/statically/? :)13:49
cptnmaro: you got that wrong13:49
cptnyou have to list runtime deps13:50
cptnif they are dynamically linked13:50
cptnso if openssl is updated, you can easily rebuild it13:50
cptnyou don't list non-dynamically linked runtime deps though13:50
cptnlike grep, or wc, or all those tools13:50
marosure :)13:50
cptnnow, the static linking case is special13:50
cptnbut was not considered when writing those rules13:51
nullpointergotta go, bye13:51
*** nullpointer has left #crux13:51
cptnI personally would consider that listing them would be okay13:51
marome too13:52
cptnalthough you'd rebuild too much i.e. for a binary incompatible update of openssl13:52
cptnsince static binaries don't need a rebuild13:52
maropersonally I think a short list of packages which shouldn't be listed would be much more useful than "stuff in core"13:52
cptnOTOH not listing them leaves you with no way to know which binary linked in an old version13:52
cptni.e. a potentially vulnerable one13:52
maroe.g. bash, glibc, coreutils, gcc, ...13:52
cptnwell, chances are that future core is much like this13:52
cptni.e. less httpup kind of things13:52
marosounds good :)13:53
marorock on!13:53
marocptn: I _think_ the second make target in glibc can be removed13:54
maro(the one that runs localedef)13:54
maroat least I haven't found any problems yet, but I've only been using that package for a day now13:55
marothe only difference is that there's no /usr/lib/locale/locale-archive (since localedef hasn't been run) and that locale -a shows only C and POSIX (oh yeah! :))13:56
maroand it cuts off like half of the build time13:56
cptncan you send me the Pkgfile?13:57
cptnI'll have a look sometimes this weekend13:57
maromine is heavily modified13:57
marobut you can just comment out the "make install_root=$PKG localedata/install-locales"13:57
maroI think it's only needed to pass some of the tests (which we don't run anyway, any longer)13:58
cptnah, sounds reasonable13:59
marobut perhaps we (well, you) should probe per why he included it in the first place - perhaps there's some other reason I just haven't found13:59
cptnyeah, I thought the same :-)14:00
*** Husio has joined #crux14:02
*** lasso has joined #crux14:05
*** predatorfreak has quit IRC14:12
*** Brzi has joined #crux14:21
*** predatorfreak has joined #crux14:24
*** Husio has quit IRC14:29
*** mjeq has quit IRC14:42
*** JackMacOKC has joined #crux14:52
*** JackMacOKC has left #crux14:52
*** sen has quit IRC15:02
*** sen has joined #crux15:02
*** pitillo has quit IRC15:27
*** Brzi has quit IRC15:44
*** mike_k has quit IRC16:04
Auge^sh: error while loading shared libraries: cannot open shared object file: No such file or directory16:10
Auge^lrwxrwxrwx 1 root root 14 Aug 23 00:28 /lib/ -> libdl-2.3.6.so16:11
Auge^-rwxr-xr-x 1 root root 9944 Apr  8 20:31 /lib/libdl-2.3.6.so16:11
Auge^mh, any idea where i have to search for the problem?16:11
treachhave you tried a fsck?16:12
treachyour symptoms feel uncomfortably familiar.16:13
*** lasso has quit IRC16:14
Auge^fsck means "clean"16:15
tilmanactually, it means 'file system check'16:15
*** bd2 has quit IRC16:16
*** bd2 has joined #crux16:19
*** predatorfreak has quit IRC16:31
*** Romster has quit IRC16:40
*** sip has quit IRC16:44
*** the-ruediger has joined #crux16:54
*** mrpk has quit IRC16:55
*** bd2 has quit IRC17:11
maroAuge^: using xfs? :P17:15
tilman"it means clean"17:15
tilmandid you want to say:17:15
tilman"fsck says: 'clean'"?17:15
maroAuge^: what did you do for that to happen?17:15
Auge^fsck.ext3 /dev/hdXY says clean17:16
Auge^maro: just start dvdrip17:16
tilmanhint: "it means clean" == "es bedeutet 'clean'"17:16
marotry running ldconfig17:16
Auge^=) of course17:16
Auge^goliath# ldconfig -p | grep libdl (libc6, OS ABI: Linux 2.6.0) => /lib/ (libc6, OS ABI: Linux 2.6.0) => /usr/lib/libdl.so17:17
maroyou aren't doing anything fancy at all?17:18
Auge^a clean and fresh install...17:19
treachI had errors like that, which turned out to be a failing hd.17:19
treachls found the file, but I couldn't open it, and if I tried to remove it the system crashed. :-/17:20
marotreach: were you able to ldconfig that?17:20
treachit wasn't a lib file.17:20
treachbut the beaviour was similar.17:21
Auge^there is no problem to open it manually17:21
maroAuge^: try to update your glibc package17:21
Auge^just the dvdrip-script have a problem...17:21
maroupdate as in reinstall17:21
marowhat script is that?17:21
marohave you read it?17:22
Auge^dvdrip is just a perl-script... i think i should strip it down till the error is gone...17:23
*** treach has quit IRC17:36
Auge^so... good n817:36
*** copwork has quit IRC17:57
*** copworker has joined #crux17:57
*** thrice` has joined #crux18:12
*** copworker has quit IRC18:52
*** cowpork has joined #crux18:52
*** predatorfreak has joined #crux19:11
*** the-ruediger has quit IRC19:26
*** scree has joined #crux21:08
nipuL!seen jaeger21:40
-MelOne- nipuL, I last saw jaeger in this channel (quit saying "Read error: 104 (Connection reset by peer)") 1 year, 105 days, 23 hours, 43 minutes ago21:40
nipuLhas harry homeowner's home eaten him or something?21:41
nipuL"screw maintaining gnome, i'm going to repaint the vestibule"21:42
*** vertah has quit IRC22:17
*** Romster has joined #crux22:30
*** mrks_ has joined #crux22:53
thrice`@seen jaeger22:53
clbthrice`: jaeger was last seen in #crux 4 weeks, 0 days, 8 hours, 45 minutes, and 27 seconds ago: <jaeger> don't know KDE, myself22:53
*** predatorfreak has quit IRC22:56
*** laod has quit IRC22:59
*** morlenxus has quit IRC23:05
*** morlenxus has joined #crux23:05
*** mrks has quit IRC23:08
*** thrice` has quit IRC23:19
*** thrice` has joined #crux23:20
*** thrice`_ has joined #crux23:21
*** thrice` has quit IRC23:21
*** thrice`_ has quit IRC23:43

Generated by 2.11.0 by Marius Gedminas - find it at!