IRC Logs for #crux Tuesday, 2011-10-04

*** crshd has joined #crux00:14
pitillogood morning01:09
Romsterrmull, for contrib access you need to make a few ports not just one and maintian them for some time before applyig for contrib. so we can see the quality of your ports and maintaining.01:54
Romsteralso register your ports on portdb
Romsteralso if oyu can change the mime type of Pkgfile from binary to text.01:57
*** joe9 has quit IRC01:58
Romsterrmull, there is a tar.lzma archive of it than the tar.gz one to save some space and download time.01:59
Romsterthe foo: metatags should have 1 space after them not tabbed or multiple spaces.02:00
Romster# URL: http://www.graphicsmagick.org02:00
Romsterlike that.02:00
Romsternot # URL:http://www.graphicsmagick.org02:00
*** lasso|qt has joined #crux02:01
Romster# Depends on:02:02
Romstercan be removed if you have no dependencies to list.02:02
*** jkramer has left #crux02:03
*** jue has joined #crux02:08
*** ChanServ sets mode: +o jue02:08
juegood morning02:09
Romstergraphicsmagick depends on zlib and bzip2 it links to them so should be listed as dependent rmull. other than that a README of other optional formats and ports needed for them be nice but not required.02:15
Romsterhi jue teK_02:15
cruxbot[opt.git/2.7]: fakeroot: update to 1.18.102:24
cruxbot[opt.git/2.7]: hpijs: update to 3.11.1002:24
cruxbot[opt.git/2.7]: libevent: update to 2.0.1402:24
cruxbot[opt.git/2.7]: unbound: update to ldns 1.6.1102:24
cruxbot[opt.git/2.7]: nfs-utils: update to 1.2.502:24
cruxbot[opt.git/2.7]: nvidia: update to 285.05.0902:24
cruxbot[opt.git/2.7]: mc: update to
cruxbot[core.git/2.7]: pciutils: update to 3.1.802:24
*** feature has joined #crux02:25
Romsterhi feature if you haven't i repled to your firefox ran as daemon & and you didn't detatch it from the terminal the other day.02:31
featurethx, touche on lazy :)02:32
Romstereh i speak my mind sort of vent my days work when i get home a little.02:38
Romsteri sure hope you do search before asking. you ask alot of questions and then you don't stay around long enough for someone to reply.02:38
Romsteri should try to refrain from such comments. just lately we been getting more users that ask a lot of questions some of which act troll like. (not saying your one).02:40
Romsterthat vee/mint i'm undecided so far.02:41
*** deep42thought has joined #crux02:44
Romsterjue, /etc/rc.d/unbound stop does not stop unbound i tried to restart it after updating unbound and it said socket already in use.03:37
*** feature has quit IRC03:38
Romstersudo killall /usr/sbin/unbound03:41
Romster/usr/sbin/unbound: no process found03:41
Romster$ pgrep unbound03:41
Romsterwonder why that wont kill it.03:41
jueRomster: that's the usual behaviour we have for all daemons that are stopped with 'killall <path_to_binary>', you have to stop the service before the update03:42
teK_may it ignores SIGTERM?03:42
Romsterbefore and not after?03:42
juebefore, than do the update and restart03:43
Romsterkillall unbound works03:43
Romsterit don't work with the path of /usr/sbin/unbound03:43
Romsterah ok so it must match some sum or something and since i restarted after updating it wont match... my bad.03:45
Romsterdidn't think much of it as linux can run that program from ram while the disk gets the new version and not have it complain.03:46
Romsterand restart the service to get the new library into ram.03:46
Romsterguess i should be stopping before i update.03:47 up again o/04:03
cruxbot[opt.git/2.7]: Revert "gstreamer: updated to 0.11.1"05:16
cruxbot[opt.git/2.7]: Revert "gst-plugins-base: updated to 0.11.1"05:16
*** ardo has joined #crux05:53
deep42thoughtDoes it mean anything that 3.0.3 sources vanished from (I'm using this one)05:54
*** ardo has quit IRC06:10
*** ardo has joined #crux06:11
frinnstonly the git repos are online i think06:18
deep42thoughtbut it's not in the listing anymore06:32
Romsterit said the git trees will be online for each developer once they get access again.06:34
Romsterthank you for the gstreamer revert, guess now i can close the bug report applied to webkit.06:34
rmullRomster: Thanks for the tips re: my repo. I'll get those changes made.06:58
Romsteryou can have a look at prtverify too handy tool.06:59
*** ThePub has quit IRC07:06
rmullRomster: Made the changes, thanks07:22
rmullExcept for the mimetype switch07:25
rmullso the Pkgfile has no extension, so I'd have to set a default mimetype of text/plain07:26
rmullBut I thought that was bad practice07:26
rmullI'll do a little research and get back to you07:27
Romsteri have all files set to text/plain that have no extension.07:28
Romstermakes it easy to view the Pkgfile's on the browerser07:28
rmullYeah, I'm gonna have to find where I read that I shouldn't do that. It's not obvious to me why I shouldn't, but I've gotta run to work ATM so I'll check in a little bit07:29
Romsterk no rush07:30
frinnsti hate repos with some wierd mimetype for Pkgfiles07:37
*** ardo has quit IRC07:49
*** ardo has joined #crux07:50
*** ardo has quit IRC07:55
*** ThePub has joined #crux08:22
*** jdolan has quit IRC08:25
*** ardo has joined #crux08:41
*** ardo has quit IRC08:44
*** ardo has joined #crux08:45
*** jdolan has joined #crux08:53
*** ChanServ sets mode: +o jdolan08:53
*** ThePub has quit IRC09:18
*** joe9 has joined #crux09:25
*** joe9 has joined #crux09:28
*** sepen has joined #crux09:49
sepenPkgfile.docx would be ok to me :D09:49
joe9is there a command that can do both prt-get search && prt-get isinst on the search packages automatically?09:51
joe9prt-get search gettext | xargs prt-get isinst09:51
joe9seems to do it09:51
joe9but, it gives this error:, though it seems to work.09:52
rmullRomster: Set Pkgfile, .footprint, .md5sum, and README mimes to text/plain and stopped hiding dotfiles for the repo09:53
rmullSo I think I'm good to go now?09:54
*** deep42thought has quit IRC09:57
Romsterrmull, yeah providing you didn't unhide any .htaccess files in the process.09:58
Romsterrmull, be sure to do a pkgmk -um and upload the new .md5sum it's got the gz sum still.09:59
Romsterare you using a content manager like git or hg for yourself? if not it might be good to learn git since crux uses git and it's handy to go back in history if you need to review something you did ages ago.10:01
Romsterother than the sum being now wrong your on your way for some good ports.10:03
rmullI use git for my personal stuff, yes10:05
rmullAnd oh, sorry about the md5sum, will fix right away10:05
Romstersometimes happens.10:06
Romsterfar easier to click on a file and see the text in it now. if oyu ever have to patch anything you can add the patch to the port too and be sure .patch files are also text/plain10:07
Romsterwhich they should already be.10:07
rmullOkay sum is updated10:08
Romsterif you happen to have a really large patch .patch.gz it but only if it's a couple killobytes or more. this is extremely rare except for ports that haven't been maintained upstream for ages.10:08
Romsterof course you can browse the many Pkgfiles already for ideas and how others have done it.10:09
Romsterit becomes second nature after a little while.10:09
rmullI'm sure of it10:10
rmullSo I should load my ports dir into a git repo?10:10
Romsteryeah i have done that, it's upto you if you do or not.10:11
Romsterone git tree and many directories for each port you make.10:11
Romsteri see your on a edu so will this only be a temporary place until you finish uni and then have to move hosting to some other site?10:12
rmullOkay, and that git tree is for my own tracking purposes? I can't integrate that into the crux infrastructure unless I'm in contrib, correct?10:12
rmullRomster: No, I'm already out of uni10:12
rmullI ran the LUG there for three years and now admin that machine10:13
Romsterah sweet10:13
rmullIT says it's okay10:13
Romsteryou may want to add a header to that page like the nickname your using for the portsdb entry.10:13
rmullOkay, will do10:14
rmullBut not right away :) gotta do some work first10:14
rmullthanks a ton for all the handholding, I know that sort of thing gets tiring10:15
Romsterportpage --header=FILE i got lazy and did a sed on mine.10:15
Romsteryeah no rush just a suggestion makes it easier to associate page with portsdb entry since your url doesn't give your rmull nick name away.10:16
Romsterthey should be OK with open source GNU/Linux contributions :)10:17
*** ardo has quit IRC10:23
*** ardo has joined #crux10:27
*** ardo has quit IRC10:29
Romsterok i so better get to bed or i wont be awake for work later.10:30
*** vee has joined #crux10:35
veesup fellas10:35
*** vee has quit IRC10:41
*** Romster has quit IRC11:00
*** Evil_Bob has joined #crux11:27
*** lasso|qt has quit IRC11:30
*** lasso has joined #crux11:49
*** lasso has quit IRC12:10
*** Rotwang has joined #crux12:13
*** treach has joined #crux12:15
*** Romster has joined #crux12:34
frinnsthead like a fucking orange13:12
Rotwangany c person around14:07
RotwangI need to know how to copy file ;f14:07
rmullI know some C, what's up14:07
Rotwangcan't find anything appropriate on google14:07
Rotwangthere is no standard function for doing that in ansi c14:07
Rotwangbut I hope there is one specified by posix or something similar14:08
rmullOpen both files, read out of one and into a buffer, copy that buffer into the new file - here's a relevant snippet:
Rotwangand I can't use third party libs14:09
Rotwangrmull: yeah, thats what I've found14:09
rmullWhich part of this are you trying to avoid?14:09
Rotwangbut I've hoped that there is some nicer way14:09
rmullThere's no like, cp(2) or anything14:09
Rotwangso thats how `cp' work basically?14:10
tilmanmy copy of man just broke14:10
rmullI haven't looked at the source, but it's probable14:10
rmullThey might use a buffer wider than a byte, though14:10
tilmangnu's cp is horribly more complicated14:10
RotwangI've thought so14:11
tilmanRotwang: if this is POSIX, i'd use open(2) instead of fopen(3)14:11
tilman(and read(2), write(2), close(2))14:11
tilmanalso, use:14:11
tilman  char buf[4096]; // or 8192, or BUFSIZ14:12
Rotwangthanks [:14:13
tilmanuntested, but should be close14:19
rmulltilman: Question for you - why use for(;;) when while(1) compiles to the same assembly?14:22
rmullI've seen this a couple of times14:22
rmullBut i'm not sure why14:22
tilmanrmull: taste. personal preference.14:24
tilmanno technical reason14:24
rmullOkay, just curious, thanks.14:24
tilmanRotwang: with the code i pasted, try to copy a _large_ file and then play with the size of the buffer. figure out how the buffer size affects performance14:25
rmullYou can get the system page size with sysconf()14:29
rmullIf you're feeling fancy14:30
tilmantrue, though in practice you'll just use 4k or 8k usually ;D14:30
*** crshd has quit IRC14:31
* Rotwang looks into tilman's code14:34
RotwangI don't care about performance that much14:34
Rotwangin this case14:34
tilmanwow, prologic owns an ipad14:35
tilmanteK_: your signature delimiter is broken (missing space :P)14:36
teK_btw.. anyone knows where I can get a hp touchpad (from the US)14:36
jaegernowhere, from what I've heard14:36
teK_tilman: i first replied only to romster, then copy and pasted (and mis-checked the signature)14:37
tilmanaha :]14:37
Rotwangrmull: I prefer for(;;) to while(1) because it is less ambiguous in some way (doesn't contain any "magic" numbers)14:43
teK_#define FOREVER 114:43
tilmanRotwang: using 1/0 for truth values isn't magic.14:43
tilmanRotwang: but there's always stdbool.h and true/false14:44
*** vee has joined #crux14:44
RotwangI know, but for(;;) has it for free14:44
veegudev fails during xfce install14:44
treachit's gudev's way of telling you to install fvwm.14:45
tilmanfv-what? :P14:45
rmullRotwang: Interesting perspective14:45
sepenvee: gudev? or sources?14:46
rmullBut then again while(1) makes more sense linguistically than for(;;)14:46
rmullIMHO, of course14:46
sepenvee: do you have a paste?14:46
Rotwangtilman: by magic number I meant a number without an explantation14:46
tilmanrmull: don't even try to make sense of for(;;) linguistically :D14:46
veesepen give me one sec14:46
Rotwangit's when some values appear in code without #defines or const's14:47
teK_tilman: it's probably a compiler-sid macro that expands to while(1)14:47
sepenhmm, Romster announce his application in crux-general, iirc they were normally made in crux-devel14:48
tilmanteK_: no. have a look at the c standard and how for loops work :P14:48
sepenvee: you need to download the source manually14:48
teK_do you allude to asm-level code?14:48
Rotwangsince is down [;14:48
teK_it's not!14:48
sepenvee: unfortunatelly is down14:48
rmullIt's back14:49
teK_11:03 < teK_> up again o/14:49
teK_read the backlog :>14:49
veei figured  it was14:49
Rotwangok, I stand corrected14:50
tilmanteK_: no. for with empty init/check/post-iteration statements == endless loop14:50
tilmanteK_: no need for whatever kind of 'macro'14:51
teK_it was supposed to be a joke14:52
Rotwangtilman: one last question, why open is better in this case than fopen?14:56
RotwangIt seems that fopen is defined by c standard, no?14:57
tilmanstdio (fopen etc) doesn't really buy you anything in this case. it's just overhead14:57
tilmanstdio implements buffering, but in this case you can easily take care of that yourself14:58
veethanks sepen, ill look into it in a bit14:58
*** sepen has quit IRC15:01
veeanother question guys. on the home page, it says crux is optimized for i686 cpu's, but, its not just limited to that, is it?15:02
Rotwangvee: no you can pass your owv cflags15:03
veeonce my payment for my tires go through, im going to start building a new rig.15:04
rmullI'm running crux on x86_64 and there is also a crux-arm project.15:05
treachthere's also frinnst's unofficial 64bit release for the daring.15:05
jseOfficially crux is limited to i686 cpus.15:06
RotwangI think vee asked if he could optimize for specific i686 compatible processor15:07
jseRotwang: ok.15:08
jsevee: was that your point?15:09
veebut wouldn't running crux on a computer that has 4+ gigs render the rest of the ram useless since its not a 64 bit?15:09
veeno, i wanted to know if crux could be installed on something like an i7 without having too many issues15:09
jseI thought it was possible to make use of it with PAE (or whatever it is called).15:10
Rotwangvee: you can use unofficial 64bit iso15:10
teK_you can adress more than 4GiB RAM with 32Bit15:10
treachvee: I guess it would work better on that i7 than an old p2 233. :>15:10
treach(building ff on the latter would be.. interesting)15:11
Rotwanganyway x86 processors can use offset (8 bit?) while addressing ram15:11
veetek, i remember with windows, if you had lets say 4 gigs of ram, and were running 32 bit, only 3 gigs would be useable15:11
veethat might be windows though, i shouldn't compare15:12
Rotwangbecause windows is crap(tm)15:12
teK_that depends on the CPU (/board?), too15:12
teK_look up PAE on wikipedia. :-()15:12
frinnsti run on an i7 with 16gig15:12
teK_me too :>15:12
veeim jelly15:12
frinnstpae is _UGLY_15:12
teK_(but on 32bit)15:12
teK_i meant.15:12
veeoh yea? pentium 4 here. xD15:12
frinnstyeah, me too (who would have guessed?)15:12
tilmani probably wouldn't put a 32 bit system on hardware that has >4 gig15:13
rmullIMO the CRUX project should consider embracing x86-64 a little more fully15:13
teK_it is. It's not officially embraced. That's all :)15:13
rmullThat's what I mean15:13
frinnstrmull: buy tilman and jue 64bit cpus with lots of ram :)15:13
* tilman rolls eyes15:13
teK_frinnst: we could spare them some ssh-accounts *g*15:13
rmullOn the front page (from a marketing standpoint solely) it seems that it's 32-bit or nothing15:14
treachfrinnst: why? I bet they could find those in the nearest dumpster.. :>15:14
veewell, you guys who do have the i7, is it worth the extra cash? i have a friend who has one and goes on and on about how awesome it is, and one that doesn't (that could be why he hates it) talks about how you never would really need to use the extra cores15:14
frinnstit is exactly as awesome as you think it is15:14
tilman? :D15:14
rmullvee: I have a little single core Via chip that's low wattage and 64-bit15:14
teK_vee: if performance does not mean everything, get one of these 20W TDP Xeon CPUs :>15:14
* treach considers going for fusion :>15:15
veei have my eyes set ont he i7's, but, i feel like when i buy it, it'll become obsolete xD15:15
veenew stuff coming outby the hour :(15:15
teK_rmull: sorry, we have to (geek)discuss this hardware issue first :P15:15
veegive me one sec gaise, gotta get the laundry15:16
frinnstivy bridge will still be a while15:16
rmullI like the sandy bridge stuff because the GPU is integrated with the CPU and you can use fully open source drivers for h.264 acceleration15:16
teK_vee: then don't buy a new box and *stay* obsolete15:16
rmullBut I can't justify buying new hardware15:16
frinnstdont you miss that new-hardware smell?15:19
tilmanburnt cpus?15:19
veeexcellent advise, teK_ lol15:19
rmullSo when is the next cruxcon?15:21
frinnstcan we stay at your house?15:21
teK_depends on where is 'your house'15:21
rmullI live in a studio, no space15:22
rmullBoston area though, I see you guys had one here before15:22
veei have a spare bedroom O.o but its filled with alcoholic beverages15:22
rmullAre you sure that's not your refrigerator?15:22
teK_I can provide 94qm space+ patio and parking lots :>15:22
tilmanvee: hahaha :D15:22
treachI'm sure people would help you fix that15:23
veedont really fridge wine15:23
veethe oldest wine i have15:23
veeis from the mother land15:23
vee30 years old15:23
veea dab of that, and we're all best friends15:23
jseWith that you would definitely get the 64-bit version made official. :D15:24
teK_we need a (wine) sponsoring bar on crux.nu15:24
teK_and a picture of tilman or jue15:25
teK_just as wikipedia has15:25
tilmancrux has an article in the german wp o_O15:25
tilmanthat picture + "send patches, *****"15:26
teK_and WINE15:26
teK_Die Verbreitung von Crux ist gering.15:27
teK_angezweifelt :>15:27
veeso crux con is a no go?15:27
teK_Simone Rota stimmt auch nciht mehr so ganz15:28
tilmanvee: it says "annual", but it only took place twice in 2004/2005 i think15:28
* teK_ votes for a crux con in germany ;)15:28
veeguess if you guys decide to have it near la15:29
veei might be able to accompany a few of ya15:29
veein exchange for your i7's15:29
teK_look at him.15:30
tilmanla? doubt it :D15:30
veeits near la, nto really la15:31
teK_wrong continent15:31
veeouch. but im a good cook ;)15:31
rmullIs wine really that hard to cook?15:32
veeits hard to cook WITH wine15:32
veeill even let you guys watch jersey shore here15:33
jseYou sir drive a hard bargain.15:34
* tilman stopped reading at 'mtv' and 'reality show'15:34
veeill reduce your iq's, get you drunk off some nice wine, and cook you delicious food15:34
veeall in exchange for your i7's15:34
veeseems legit15:35
teK_i.e. make you fat :>15:35
teK_dumb, drunk, fat. \o/15:35
jseWho in their right mind could say "no" to that!15:35
veei make ridiculously good bbq15:35
veeill give you that for free ;)15:35
teK_in your studio?!15:35
veeby ridiculously good, i do mean reallllllllllllly good15:35
teK_how'd that work ou15:35
veei guess you'll just have to find out15:36
veei'd send some out, but, im not sure what it'd taste lik ewhen it gets there15:36
rmullteK_: That's me with the studio15:39
rmullMy bbq isn't as good15:40
*** Evil_Bob has quit IRC16:20
*** jdolan has quit IRC16:21
*** jdolan has joined #crux16:55
*** ChanServ sets mode: +o jdolan16:55
*** treach has quit IRC17:46
*** Rotwang has quit IRC18:10
*** Ovim|afk_nA has joined #crux18:29
*** Ovim-Obscurum has quit IRC18:32
*** vee has quit IRC18:32
rmullI don't need to list "gcc" as a dependency in my Pkgfiles, right? :P18:51
thrice`right, usually anything from core is assumed18:52
rmullthrice`: thanks.18:53
thrice`sure :>18:54
*** aarchvile has quit IRC19:07
*** aarchvile has joined #crux19:07
enteI think that's not entirely right19:18
entegcc and glibc are assumed, anything from core that isn't linked against is also assumed19:18
enteor something like that19:18
entethere was something about it in the wiki19:19
thrice`"build dependencies provided by core are not listed in the dependency header "19:19
thrice`   , yeah, I guess you are kinda right :p19:19
enteah, right, not the wiki19:20
rmullSo... we leave in /usr/share/{applications,icons,mime} and that cruft because people can opt out with an INSTALL rule, correct?20:01
rmullBut we take out /u/s/doc if it's full of html and png help files?20:02
enteactually all plaintext in /u/s/doc is taken out20:07
rmullOh, wow20:07
rmullOh right, just kill the whole /u/s/doc, gotcha20:07
rmullWhy should users be allowed to opt out of all those crufty icon/pixmap dirs, but not the doc dirs?20:08
*** vee has joined #crux20:21
thrice`docs are just killed explicitely, as the idea is "if you want docs, open up a web browser" :)  the icons and desktop files are used by eg. XDG-complient menu systems, like xfce, etc20:34
Romsterrmull, also i can't find it but if it links to it with ldd ( i use finddeps portname) then list those but not glibc binutils glibc tool chain if it uses perl but does not link to it don't list it. if it links to zlib list it.20:46
Romsterrmull, then after that and i cp it to the /usr/ports/foo/portname/Pkgfile i run findredundantdeps -s portname then clean up the Depends on line.20:47
Romsterbest to do that in a chroot that's only got the depends on line ports installed.20:47
Romsterthere is safe-build and safeenv (i think it was called safe-env) former is my work and i use it to setup a chroot and chrot back in safe-build clean to remove ports prt-get depinst portname then clean the depends on line and .footprint20:49
Romsteri can't remember where the details in the manual or wiki that say if it links to a core port but not the tool chain it should be listed are i'm sure jue or someone will point the url out.20:50
*** Dudde has quit IRC21:28
*** Dudde has joined #crux21:29
*** vee has quit IRC22:55
rmullRomster: I uploaded another port (geda - I'm going to do the whole suite of electronic cad tools probably)23:19
rmullGot a weird overlap of mime files when I went to install23:19
rmullhad to do pkgadd -f23:22
rmullGonna compile it in a clean chroot and see what happens23:24
rmullgood night23:24
*** ThePub has joined #crux23:24
*** ThePub has quit IRC23:42

Generated by 2.11.0 by Marius Gedminas - find it at!