IRC Logs for #crux Wednesday, 2008-04-09

*** pitillo has joined #crux00:07
*** Romster has joined #crux00:22
*** Romster has quit IRC00:29
*** maxus has quit IRC00:37
*** maxus has joined #crux00:46
*** joacim has joined #crux00:58
pitillogood morning01:14
racerI need to go to bed01:45
rehabdolli need to go to work :(01:45
racerI'm still at work... almost 2:00am here01:45
racerhave to be back at least by 8:00am, if not sooner01:46
rehabdollouch :)01:46
*** sepen has joined #crux01:50
*** Romster has joined #crux02:07
tilmanrehabdoll: bah, just call in sick02:11
cruxbot[core.git]: rsync: update to 3.0.202:16
*** namenlos has joined #crux02:35
cruxbot[opt.git]: stunnel: update to 4.2202:46
cruxbot[opt.git]: libxml2-python: update to 2.6.3202:46
cruxbot[opt.git]: libxml2: update to 2.6.3202:46
cruxbot[opt.git]: mpg123: update to 1.4.102:46
cruxbot[opt.git]: whois: update to 4.7.2602:46
cruxbot[opt.git]: swig: update to 1.3.3502:49
*** mxs has joined #crux02:54
*** maxus has quit IRC03:12
*** mike_k has joined #crux03:28
cruxbot[contrib.git]: swfdec: 0.6.2 -> 0.6.404:07
cruxbot[core.git]: autoconf: update to 2.6204:40
cruxbot[opt.git]: git: updated to
cruxbot[opt.git]: scons: updated to
pitillolet's see how E builds with this new autoconf04:49
*** RedShift has joined #crux04:57
tilmancptn: slim's crux-smooth theme (the one that ships with the port) looks silly on my system (the string "Default" reaches into the username text box). is that because of my X settings, or did some freetype change cause that?05:00
cruxbot[contrib.git]: inadyn-mt: Updated 02.11.87 -> 02.11.9605:02
sepentilman, same here with Mode 1280x76805:04
tilmani'm using 1280x800 on this laptop fwiw05:04
sepenerr, s/768/102405:06
pitilloummmm I can test 1280x800 here05:09
pitilloI can't reproduce it, working fine with that resolution05:10
cptntilman: no idea, I'll have a look though05:10
tilmanhi cptn05:10
tilmanit's probably freetype, i noticed that problem some years ago or so ;p05:11
sepenhmm indeed I've another box running also 1280x1024 without any problem and it appears fine05:12
pitillosepen, with your res (1024x768) is working fine too05:13
cptncould it be a dpi issue?05:13
cptnhhhe, where's this "Default" string?05:15
sepencrux-smooth is set to Verdana fonts05:16
pitilloseems that E have been built fine with this autoconf version05:28
cruxbot[opt.git]: opera: updated to 9.2705:29
cruxbot[contrib.git]: doxygen: Updated 1.5.4 -> 1.5.505:59
tilmancptn: i think it's the name of the session06:21
tilmani have slim run ~/.xinitrc, so it just displays 'Default'06:21
cptnmmmh, it's just that I don't get any string matches for 'Default' in the source06:25
cptnhowever the string in front of the username text box should be "Username:", not "Default"06:27
howlcptn, what line of work are you doing these days?06:28
cptnconsumer electronics06:29
cptnI'm in a team that writes firmware for semi-professional audio interfaces06:30
cptnthink: multi channel, low latency etc.06:30
howlAhh, interesting.06:31
cptnit's pretty nice, especially since most guys here use the devices privately as well06:34
howlSo you get to suit the devices for your own needs. :P06:35
cptnyeah, at least to some extent06:36
cptntilman: could you try this slim.theme
cptnjust copy it to /usr/share/slim/themes/crux-smooth06:37
cptnyou can run slim - /usr/share/slim/themes/crux-smooth to test06:37
howlObviously only if you can persuade the managers that this and this feature will produce more-sale. :)06:37
cptner, slim -p /usr/share/slim/themes/crux-smooth to test06:37
cptnhowl: sure, but in our case the developers are part of the target audience06:38
cptnplus, you can always build a firmware just for you :-)06:39
tilmancptn: i suck. it said 'Username', not 'Default' ._o06:41
cptnthe slim.theme I uploaded should address that06:41
tilmanyeah, it does06:42
cptnoh, I actually hast that crux-smooth theme...06:42
*** jdolan_ has quit IRC06:44
cptntilman: I uploaded a fixed tarballed called slim-crux-smooth2.tar.gz (to avoid an md5 error)06:46
tilmancool, thanks06:46
cptnshould I send a mail to sip?06:47
tilmanyes please :p06:54
cptnwat wat07:04
tilmanwat wat wat07:30
jesse_wat wat wat wat07:30
tilmanbleh, i was going for fibonacci :(07:31
*** brendan` has joined #crux07:48
*** brendan` has quit IRC07:51
*** jdolan_ has joined #crux07:56
*** sahko has joined #crux08:15
*** Rotwang has joined #crux09:06
cruxbot[contrib.git]: leafpad: Fixed maintainer line.09:16
cruxbot[contrib.git]: $(dirname $i): Fixed maintainer line.09:23
cruxbot[contrib.git]: $(dirname $i): Fixed maintainer line.09:24
cruxbot[contrib.git]: $(dirname $i): Fixed maintainer line.09:24
cruxbot[contrib.git]: $(dirname $i): Fixed maintainer line.09:24
sepenhh sit09:24
cruxbot[contrib.git]: sgmixer: Fixed maintainer line.09:26
cruxbot[contrib.git]: `dirname $i`: Fixed maintainer line.09:26
cruxbot[contrib.git]: sunbird: Fixed maintainer line.09:27
cruxbot[contrib.git]: xalan-c: Fixed maintainer line.09:31
sepenpffff bad day, sorry guys09:31
teKlaze bastard *g*09:35
Romstergitk --all, look before you git push... what i always get told09:36
Romsteralso you could of done all the maintainer lines in one commit...09:36
Romsterlike i did.09:36
*** treach has joined #crux09:41
tilmansepen: ffs, run 'git log' before you push09:42
*** Rotwang has quit IRC09:47
treachyeah, think of the children. :>09:48
jesse_Don't forget to ring unicef and the endless brigades of mothers against/for xyz ad infinitum. :D09:53
treach"mothers against bad commits" -- sounds plausible. :p09:54
tilmansepen: did your script run 'git commit one, git push, git commit two, git push, ...'?09:54
tilmansepen: if it did, it's fucking retarded. do 'git commit one', 'git commit two' etc and push *once*09:55
jesse_Master git inspector tilman. :D09:58
cruxbot[core.git]: [notify] libtool: updated to
cruxbot[contrib.git]: wgetpaste: 2.10 -> 2.1109:58
tilman"i'm the fucking git master"09:58
treachThe Git MoC09:59
tilmanthe git git09:59
treach"damn someone made a bad commit, git the git git." :D09:59
sepengit commit git -git10:00
sepentilman, reading10:00
sepen<tilman> sepen: did your script run 'git commit one, git push, git commit two, git push, ...'? yep10:01
sepenI think was cause by quotation marks '$()'10:01
tilmanthe primary problem is that you never had a look at the commits you generated >:10:01
sepenpffff Ill never repeat this10:02
Romsterplus  ' ' is strong quoting when you need weak quoting "" with variables.10:02
sepenyeah, that does the trick10:03
sepensorry one more time people10:03
sepenIm fixed my various and erroneous maintainer emails, see yours
sepenI updated the about-page-ticket10:05
sepeni.e: Thomas Penteker, tek at serverop dot de || Thomas Penteker, tek at serverop dot de10:05
sepenisn't clair this, but i.e this another one:  Matt Housh, jaeger at crux dot nu || Matt Housh, jaeger at morpheus dot net10:07
Romsteri noticed that ages ago.10:08
tilmanhe probably didn't care much because both addresses work ...10:08
j^2it seems crux isnt dead :P10:09
Romsterwell i don't see anything wrong with that..10:09
Romsterother than not being consistant.10:09
sepenyeah, what are waiting for to close this ticket?10:12
sepenalso I found more errors in the website
sepenwebsite is unmaintained actaully imho10:13
Romsterfound web site issues too didn't bother to report.10:16
Romsterexcept a few of the bug tracker but nope no change as far as i can tell.10:16
sepenwell more people could complete this first info10:17
tilmaninstead of file bug reports with a random list of names and emails you could start by providing an understandable problem description etc10:18
tilmanyou could even ask to get a wiki account to fix things yourself10:18
Romsterya great help i've asked and didn't get anywhere <<10:19
tilmanyour english sucks too much to let you edit the wiki10:19
Romsteroh wow and sepens doesn't....10:19
tilmanit probably does10:19
treachnobody have said he'd get that account. :>10:19
Romsteri can use a dictionary..10:20
treachRomster: tilman only said he could have asked10:20
tilmansepen: btw, jaeger isn't an opt maintainer anymore...10:20
tilmanfwiw, i fixed that typo you found10:20
tilmanand i'm going to fix the fucking about page now10:20
tilmansepen: it's annoying to get that 'what are you waiting for to close the bug' response just because i'm pointing out that jaeger's email addresses both work >:10:22
Romsterat least i provide better descriptions on the bug tracker. but who cares, i'll just go do my own stuff and not waste my time where it's not wanted.10:22
sepen<Romster> oh wow and sepens doesn't....10:24
sepenhehehe my english is the worst of this channel10:24
treachno bickering please, that's hardly constructive.10:24
Romsterno offense but your not natie englich speaker and tilman said you could apply and i'm a english speaker and tilman tells me off, that's very encouraging.10:25
Romsterbad typing there.10:25
sepenhmm <tilman> sepen: it's annoying to get that 'what are you waiting ...10:25
tilman    treach> Romster: tilman only said he could have asked10:25
sepentilman, sorry, but isn't relative to this question10:25
sepenI mean the 'about' page10:25
Romsteranyways have fun..10:25
tilmangah, whatever10:25
*** tilman has left #crux10:25
treachRomster: stop that.10:26
sepensounds strange tilman is out after that?10:27
treachyou are putting way too much into what people say, and then you go off venting on all kinds of injustices based on that interpretation.10:27
*** tilman has joined #crux10:27
*** ChanServ sets mode: +o tilman10:27
sepentilman, any problem?10:27
cptnsepen: well, it's just annoying if people keep repeating the same problems10:27
cptnI mean it's not like bugs are ignored because devs don't care10:28
sepencptn, then I should shut up and come to my house?10:28
cptnit's about time and priorities10:28
cptnit's a good thing that you submit the bug report10:28
sepencptn, its due to the number of sysdevels imho10:28
cptnbut as tilman said it would help to describe the problem10:28
sepenonly 3 sysdevels can't liderate a big project imho10:29
sepenmany work10:29
cptnyeah, then why put additional pressure on them here?10:29
tilmanokay, who wants to maintain
tilmangood written english required10:29
sepenmeeting, ML's10:29
tilmanbeing on freenode a plus10:29
cptnI'm sure you meant no harm, but repeating bugs which are on the bug tracker always sounds like "why are you so lazy?"10:30
sepencptn, sure10:30
cptnand if you're on the recieving end, this is very annoying10:30
cptnand the more you get of this, the more you want to stop being a dev10:30
sepenso, which is the problem? the number of crux devels┬┐?10:31
cptnwell, depends10:32
cptnI have no problems ATM10:32
cptnmy CRUX systems run fine10:32
sepenwe have the whole infrastructure to get a big and stable, secure linux distro10:32
cptnmaybe some links are dead on the website, but TBH that doesn't bother me much10:32
tilmanthe "not caring" part is definitely true10:32
cptnsepen: let me give you an example:10:33
sepencptn, my crux's too10:33
cptna while ago, there was this thread on the ML10:33
cptnby you or pitillo10:33
cptnregarding "too few devs"10:33
tilmanif links to "CRUX on a dell inspiron 5150" which is dead, people at least have something to google for10:33
tilmani could add "These links might be outdated, go cry me a river."10:33
cptnto me, that sounded like a complaint10:33
sepenimho (and only imho) the core system should have more people10:34
sepencptn, as Per days10:34
cptnbut that's not a problem per se, is it?10:34
cptn"too few"10:34
cptnoutdated ports would be, compilation failures etc.10:34
sepenbut the system now is not the same10:34
cptnnow, a while later, some guy had an httpup/proxy problem10:35
pitillocptn, was me, not only talking about devs, add a little note about organization, btw, I saw the point the last time we talked, the mail is there and there is no need on both10:35
cptnno one who previously complained in the other thread helped that guy10:35
cptnso why is the community not helping users, but expects devs to do more work?10:35
cptnwhy is it easier to complain about devs than helping out other users?10:35
sepencptn, how many SYS retired memeber we have?10:36
tilman2 or 3?10:36
cptnand think about this: if you were a developer, how would you find motivation to do that work if all you hear is complaining?10:36
cptnit's not like I often read "thank you" on the ML10:36
sepen22 noted10:37
cptnsepen: yeah, and how many of those left because they couldn't motivate themselves anymore?10:37
cptnit's not like crux couldn't move faster10:37
cptnit probably could with more people10:37
sepenbut in all comunities there are people who 'dead' and others who 'born'10:38
cptnbut the current pace seems just fine too10:38
cptnwell, check sourceforge... how many dead projects are there?10:38
cptnbut sure, I'm not saying it's all good10:38
cptnmaybe things could be better10:38
cptnbut those still working on it are doing their best10:39
cptnand deserve better than getting reminders of things they still haven't done10:39
cptnIMHO, of course10:39
sepenI think there are a lot of contributors which are working hard, like Romster, and imho they should be part of the core devel10:40
cptnthat's your opinion10:40
cptnjoining the team was never only about hard work10:40
cptnbut also about compatibility10:40
cptnI remember when I still was a developer Romster did a number of things in a way I wouldn't agree10:41
sepensure, Im not saying 'goods', only people who can work10:41
cptnnot wrong, just different10:41
sepencptn, you remember Han's style?10:41
sepen*do you10:41
sepenlike this case┬┐10:41
cptnhe didn't really get along with the devs either?10:41
cptnforget that question mark...10:42
tilmansepen: did you ask simone to bump your flyspray privileges?10:42
sepenyep, but actually there are only 3 devs10:42
tilmanit would help if you could assign bugs to people10:42
cptnsepen: it used to be only one10:42
cptnsepen: and it worked too10:42
sepentilman, sure10:43
tilmanthanks for pointing that out10:43
tilmanin the old days, NOBODY complained about anything10:43
sepenIm doing some efforts trying to report bugs or whatever strange10:43
tilmansepen: initially, bugs aren't assigned to anyone. if they were, people would see them more quickly :P10:43
cptnsepen: you know, if there was man power so somehow coordinate the efforts, like your bug reports, or Romster's ports work, that would be grea10:44
cptnbut it's not there right now, so everyone should just try to make the best out of it10:44
sepen1 sec, Im reading10:44
cptnwhich might just mean that the bugs don't get fixed for a while10:44
mike_ksepen: yes, please go on assigning bugs if you have spare time. It would be really nice.10:45
sepen+1 for me10:45
cptnanyway, I'm on my way home10:45
sepenbut also I'll happy if people allow this kinds of reports and tries to fix them, Ill try don't be repetitive10:46
sependid you see this?
sepenare there some tool plained for to verify all oficial/pseudoficial repositories┬┐?10:47
tilmanno plans, that is10:47
sepenI known actually prt-verify10:47
mike_ksepen: nice page anyway10:50
sepenstill with some erros checking url sources10:51
sepenbut could be some other guy can fix or develop a more complete log report system10:51
mike_kafterall, it's up to people themselves to make valid ports, checking them, etc10:51
mike_kand the most ugly things should became bugreports10:52
sepenthat's only an idea which was proposed by me at ML's10:52
*** predatorfreak has joined #crux10:56
sepenwell, time to go home. later10:56
*** sepen has quit IRC10:57
mike_kpredatorfreak: scroll logs if you feel bored10:58
predatorfreakI'm not :D10:58
predatorfreakI just woke up, so I'm not too boredness yet10:59
mike_kthen skip that10:59
tilmanfucking awesome, new autoconf breaks the xserver build11:00
predatorfreaktilman: \o/11:01
*** js has quit IRC11:02
predatorfreaktilman: BTW, I just sent my complaint about merging the wget ignore-SSL-certificates patch, it's purely nit-picking at the code though11:02
mike_khmm, some paranoid people would prefer a variable (though, maybe turned on by default)11:05
mike_klike ones with SHA hashes =)11:06
predatorfreakmike_k: if [ BLAH ]; then; alias wget="wget --no-check-certificate"; fi11:06
mike_kyep. and /etc/pkgmk.conf: BLAH=true11:07
predatorfreakEither way, I think an alias is better if every-dang-time you intend to call wget you're going to use that.11:07
predatorfreakand I can make a version of that on-request.11:07
raceris there a place that the default configuration files for crux are available?11:09
mike_ka pkgmk.conf setting along with command line switch is preferable here (while I don't really care)11:10
racerfor example, I've modified /etc/rc.d/net, but I need to know what it looked like originally so I can create a set of instructions on how to modify it -- what's the easiest way to get the original file?11:10
racermike_k: thanks11:10
predatorfreakerr11:11 isn't in rc.11:11
predatorfreakI forget where it comes from XD11:11
racerit's in iproute211:13
racerjust in case you wanted to know :)11:13
predatorfreakYou know, it would be very nice if shell supported stuff like export BLAH=BLAH11:16
predatorfreakthen export BLAH+="BLAH"11:16
predatorfreakWhich would append onto the variable..11:16
* predatorfreak e-mails off his version of the SSL patch11:19
tilmanwhy would you not want to use $CHECK_CERT?11:23
predatorfreaktilman: If you keep on using that.11:24
predatorfreakWe'll eventually have a billion variables for one command11:24
predatorfreakIt's infinitely more readable to just use alias wget="wget OPTIONS" and say "Look there."11:25
predatorfreakPlus, even if it has to be in a variable.11:25
predatorfreakIf it's getting called every time11:25
predatorfreakput it in DOWNLOAD_OPTS11:25
predatorfreakWhere it properly belongs.11:25
tilmanwhy does it have to be configurable?11:25
mike_ktilman: some people might try to avoid self-signed sites11:26
predatorfreaktilman: Some of us might prefer to leave it on.11:26
tilmanW H Y11:26
predatorfreaktilman: Because I don't want to download via https if I'm giving up the damn benefits.11:26
tilmanmike_k: we have md5sums to check the integratity of our downloads11:26
predatorfreakIf I just let any tom, dick and harry put shit on https then I might as well stop using it and go use http.11:27
predatorfreakWhich doesn't have the encryption overhead.11:27
mike_kpredatorfreak: and md5sums are br0k3n, aren't they?11:27
tilmansuppose i stick a port in my repo that uses https because http isn't available11:27
predatorfreakI haven't encountered a single god-dang port in core/opt/contrib that uses https11:28
tilmanwould be pretty not-so-awesome if everybody then had to edit pkgmk.conf to make the port work11:28
predatorfreaktilman: Enable it by default11:28
predatorfreakAlthough I just realised a slight bug in my version11:28
predatorfreakmike_k: Yes.11:28
predatorfreakI've already voiced that concern and no one seems to care enough.11:28
tilmanmike_k: s/md5sum/$MECHANISM_TO_CHECK_FILE_INTEGRITY/ my point still remains ;)11:29
mike_ktilman: only if md5sums were more reliable.11:30
predatorfreaktilman: Well, if our mechanism to check file integrity is borked.11:30
* mike_k kidding11:30
predatorfreakI'd like some level of backup.11:30
tilmanif $mechanism is broken, using https urls won't help11:30
predatorfreakPlus, I don't know of a single host who puts files on https exclusively.11:30
tilmanpredatorfreak: i do, and i already said so11:30
mike_kanyway, configurable and ON by default does not seem to bad?11:30
predatorfreaktilman: You can't spoof an ssl certificate.11:31
predatorfreakYou can spoof md5sum.11:31
tilmanyou missed my point11:31
tilmanand i'm already tired of this pointless debate11:31
* predatorfreak just forks pkgutils and says fuck it11:31
predatorfreakI'm sick of pointlessly submitting proper patches only to have them get rejected or ignored <_<11:32
tilmanwho says it's rejected?11:32
predatorfreaktilman: By the sound of it, you don't like it.11:32
predatorfreakand you, jue and Simone control pkgutils.11:33
tilmani don't like the fucking discussion11:33
tilmanif md5sum sucks, .md5sum needs to be fixed. using some https crap doesn't fix the problem11:33
tilmaneod ffs11:33
predatorfreaktilman: and no one seems to care about my sha256sum patch which I submitted months ago.11:34
tilmani already explained how we can tackle that problem11:35
predatorfreaktilman: Regardless:
tilmanand it still applies11:35
predatorfreakIf you actually intend to use my version, that's the one without the bug.11:35
predatorfreakSo it's properly disabled by default.11:35
mike_kpredatorfreak: if I ever had a vote, I'd say yes to sha256. but it's not like that.11:36
mike_ktilman: anyway, what's with that SHA stuff? Is it really rejected or up to discussion on the next developer's meeting?11:37
predatorfreaktilman: I'm annoyed by the fact it's taking so long to even take a couple minutes, get jue and Simone in #crux-devel and quickly debate it.11:37
tilmanmike_k: it's a pretty major change, we cannot just slip it in :|11:37
predatorfreaktilman: I wouldn't say it's major, it's done in a purely incremental way.11:38
predatorfreakIt doesn't break a single thing.11:38
tilmanmike_k: also i don't want to just override jue's strong objections against sha256.. and he *does* a point, too. noone ever tried to pull such a job (providing forged tarballs)11:38
predatorfreaktilman: ... by that logic.11:38
predatorfreakWe don't need md5sum11:39
jesse_You seem to forget every single port has to be updated and IF it were changed, you need to constantly slap users in the face.11:39
predatorfreakand should scrap that too.11:39
predatorfreakjesse_: You don't need to update every-single-port.11:39
mike_ktilman: I hope predatorfreak will beat the objectors =)11:39
predatorfreakJust do for port in /usr/pors/core/*; do; pushd $i; pkgmk -um; popd; done11:39
predatorfreakThen commit the changes to core/opt with a message of "Add .sha256sum files for ports."11:40
cptnso what's the u in '-um' other than update?11:40
cptndon't forget the -d11:40
predatorfreakcptn: -um is update sum.11:41
predatorfreakIt doesn't even rebuild the ports.11:41
cptnso you're updating every single port11:41
cptnone file for every port11:41
predatorfreakcptn: No, you're adding one file.11:41
predatorfreakThere's a big difference between an "update" (i.e. version bump) and adding one file11:41
*** namenlos has quit IRC11:42
predatorfreaknot to mention, it can fallback to md5sum if sha256sum sum's aren't present.11:42
mike_knoone will have to rebuild (or prt-get won't do a sysup)11:42
predatorfreakI covered all the bases for a proper transitional route.11:42
jesse_It still is an update because you are changing an aspect in each port: the crude verification that your source code is not b0rked.11:42
cptnso if it's backward compatible, it's still as insecure as md511:42
predatorfreakcptn: No.11:42
cptnproviding a false sense of safety11:42
predatorfreakif .sha256sum is present11:43
predatorfreakIt uses that11:43
predatorfreakand skips md5sum.11:43
predatorfreakIt ONLY falls back to md5sum if there's no other option.11:43
predatorfreakcptn: Plus, right now, md5sum is a false sense of security.11:43
*** mike_k has quit IRC11:44
*** Romster has quit IRC11:44
jesse_Why would you preserve an md5sum hash when you often bring up how insecure it is in the end?11:45
predatorfreakI swear people love to gang up on my patch even when they obviously didn't read it <_<11:45
predatorfreakjesse_: It's a TRANSITIONAL patch.11:45
cptnwho does?11:46
predatorfreakI don't want to break shit for old users.11:46
predatorfreakcptn: Well, I'm not attacking you here, but you didn't read my patch.11:46
predatorfreakNor did jue when he slammed it.11:46
cptnI'm just saying as long as it's impossible the users will think they're safe ("I have sha-pkgutils"), but the ports tree with only md5sums will work just fine11:46
predatorfreakcptn: The idea is to not break shit and hope people aren't idiots.11:47
predatorfreakI think number 2 should be covered by the CRUX userbae.11:47
jesse_It simply does not make sense to keep a legacy function.11:47
cptnwell, some people are still on 2.3 trees11:47
jesse_It will disappear when you force it out and replace it with a so-called better alternative.11:48
predatorfreakChrist first jue attacks me without reading my patch for breaking shit.11:48
predatorfreakNow I'm getting attacked for not breaking shit.11:48
* predatorfreak smacks head against desk11:48
cptnI'm not attacking YOU11:48
tilmanwho's attacking you?11:48
cptnjust the transition path outlined11:48
predatorfreakjesse_: There's it completely skips md5sum if sha256sum's are present.11:48
predatorfreakPresumably, before sending the new pkgutils version into core.11:49
predatorfreakUpdating core/opt to have them should be simple.11:49
predatorfreakThen merely telling the contrib maintainers to do pkgmk -um should take care of contrib.11:49
predatorfreakThen, after awhile, I can provide a patch to totally drop md5.11:50
predatorfreakBut in order to facilitate a proper transition, I think it's best to provide backwards compatibility.11:50
predatorfreakcptn, tilman: I don't like getting criticised once for doing something I didn't, then getting criticised for the same damn thing I got criticised for not doing before <_<11:51
tilmanif the .foosum is missing, pkgmk will just create it11:51
tilmanand if pkgmk believes that md5 is worthless (which it probably does, since it uses sha256 then) why should it read .md5sum11:51
tilmani think cptn hinted at  that earlier, i'm not sure11:51
predatorfreaktilman: Read above.11:51
predatorfreakI'm sick of repeating it.11:52
tilmanpredatorfreak: you wanted a discussion, now you got it11:52
jesse_My point being WHY do you want to keep md5sum around when it is failing to do its job? Get rid of it, be decisive and avoid sending mixed messages.11:52
tilmani did read it11:52
tilmanmaybe you should give us some credibility, too11:52
predatorfreaktilman: I already answered that ten damn times over.11:52
predatorfreakIf people really want the breaky approach.11:52
predatorfreakThen sure, I can provide a version for that as well.11:53
tilmanpredatorfreak: if you wish to convince jue that it doesn't all suck, you should have a nice cup of tea before that meeting so you don't go all omfg-read-the-patch on him ;)11:55
predatorfreakI think I'd need a god-damn sea of tea.11:58
predatorfreakand very, very angry music11:58
treachmh. being angry is rarely a good way to persuade anyone. :>12:10
treachthat said, I don't really understand what jue have against sha256.12:11
predatorfreaktreach: I don't like getting told to do something which I already did, then being told to not do that <_<12:14
predatorfreakCall me crazy, but I like some level of uniformness to getting criticised.12:14
cptnI was discussing a feature, not a patch12:15
cptnif jue was doing the same no wonder there was no agreement there12:16
treachpredatorfreak: well, if you always get criticized whatever you do, isn't that uniform enough? ;)12:16
predatorfreakcptn: As far as I'm concerned, they're still the same thing.12:16
predatorfreakIt's an out-of-tree sha256 feature/patch ;)12:16
predatorfreaktreach: No :D12:17
cptnpredatorfreak: yeah, I think that's the problem12:18
*** joacim has quit IRC12:18
cptnpredatorfreak: you're trying to push a patch/implementation12:18
predatorfreakcptn: Because it desperately needs to go in.12:18
*** joacim has joined #crux12:19
cptnwell, no12:19
cptnthe feature has to12:19
predatorfreakOtherwise we should just drop sums period, because our current implementation is broken.12:19
cptnbut not necessary your implementation12:19
predatorfreakcptn: I don't see any other implementations.12:19
predatorfreakI have the only working option available right now.12:19
cptnare you saying it's rocket science?12:20
predatorfreakI'm saying no one else seems to care.12:20
predatorfreakI'm the only person to come forth with a working, complete solution and I'm willing to adapt it to whatever the conses is on how SHA256 should be integrated.12:21
cptnokay, sounds fine12:21
predatorfreakFucking hell.12:31
predatorfreaknew libtool bork libgphoto12:31
predatorfreaktilman: There should really be a policy about pushing autoconf/libtool updates into core like this12:31
tilmandid you see the notification mail on the ml?12:32
predatorfreakYes, I did.12:32
predatorfreakIt breaks BUILDING libgphoto2.12:33
predatorfreak"configure: error: libgphoto2 requires the ltdl library, included with libtool"12:33
tilmancan't be too hard to fix :D12:35
predatorfreak"checking for lt_dlcaller_register in -lltdl... no"12:35
predatorfreakWhy do we have a policy against dangerous toolchain updates12:35
predatorfreakif we can randomly update other parts of core that ARE breaky-type-things12:35
tilmanbecause they were well-tested12:36
predatorfreaktilman: ... not really.12:36
tilmanalso libtool isn't in the same league as gcc or binutils12:36
predatorfreaklibtool is already breaking shit.12:36
predatorfreakand binutils updates rarely break anything for me.12:36
treachpredatorfreak: I gett the feeling you're blaming the wrong end here12:37
predatorfreakI get the feeling we shouldn't just update libtool, ESPECIALLY when it's as scary as "binary incompatible" releases12:37
predatorfreakWithout asking for some testing first.12:37
treach"source incompatible"12:38
predatorfreakI asked for testing for something as simple as FIREFOX12:38
predatorfreakSurely it's not impossible to get some testers to rebuild ports with libtool.12:38
predatorfreaktreach: Well, yes, it is source incompatible by the looks of it12:38
treachbut is ltdl really a part of libtool?12:39
predatorfreakFucking hell.12:40
predatorfreakNot only is libtool breaking libgphoto2.12:40
predatorfreakBut the new autoconf breaks it too.12:41
predatorfreakI can't even fix the idiotic libtool problem because of it.12:41
treachwhee! teh end of boredom. :D12:41
* predatorfreak gets out his whip12:42
predatorfreaktilman: What was that about libtool/autoconf updates not being dangerous? <_<12:44
*** jue has joined #crux12:44
*** ChanServ sets mode: +o jue12:44
tilmanthat's not what i said, read again, yada yada yada12:44
tilmanhi jue12:45
cptnhi jue12:45
predatorfreaktilman: It's still quite annoying.12:45
tilmanyes, let's crucify someone12:45
predatorfreakI'm reverting to autoconf 2.61 for the moment so I can actually fix problems caused by libtool...12:45
tilmanor, we could stay calm and try to fix things12:45
predatorfreaktilman: You can't do autoreconf in libgphoto because autoconf 2.62 broke it.12:46
predatorfreakand you can't fix the libtool source-level incompatibilities without fixing the m4 autoconf files.12:47
tilmanyes,i  got that12:47
predatorfreak... great, fixing the libtool problem generates a 7MB patch12:56
predatorfreak.. screw it, I'm using the Gentoo hack, commiting a 7MB patch is insane <_<12:57
juetilman: no objections12:59
tilmanjue: we should probably revert the autoconf bump13:01
rehabdolloh how i loooooooooooove sourceforges mirror-system13:01
tilmanAC_SUBST is b0rked in 2.62, or at least code that used to work isn't accepted anymore13:01
juesure, didn't find any problems myself and upcoming coreutils are build with 2.62 so I though that it is save to update ...13:04
predatorfreaktilman: oh and, version without my alias changes:
predatorfreakRather than roll one option into it's own variable, it seems better to just add it to DOWNLOAD_OPTS13:05
cruxbot[contrib.git]: libgphoto2: Update to 2.4.1, fix libtool problems.13:07
tilmanpredatorfreak: i'll ponder the reasoning for making it configurable13:12
predatorfreaktilman: Oh, yeah, that one isn't configurable either.13:12
predatorfreakIt just adds --no-check-certificate to DOWNLOAD_OPTS13:12
predatorfreakIt's the extreme minimal changes version.13:12
*** sepen has joined #crux13:13
*** ^racer^ has joined #crux13:31
*** racer has quit IRC13:31
*** sahko has quit IRC13:35
*** mrks has joined #crux13:42
*** mrks_ has quit IRC13:58
sepenwup wup14:01
*** Rotwang has joined #crux14:02
rehabdollhey, its mr $(dirname $i)14:02
ryuosed 's|stonecrest|troll|'14:02
sepenI sent a mail to crux-contrib ML's but unfornately it have some problems, thats the aim point of it
tilmansepen: our mail was accepted by the MTA, and delivered to mailman, and tehree's no pending requests in the admin interface :|14:04
tilmani have no idea what's up there14:04
ryuoit ate your mail.14:04
sepenwell Ill try later in both cases14:04
sepenryuo, ?14:04
ryuoMTA: Burp.14:05
sepenO_o ?14:05
ryuoIt likes to eat mail from sepen. :->14:05
tilmanhe's kidding ;)14:05
ryuoor maybe it got uploaded to /dev/null14:06
* ryuo shrugs.14:06
treachcontrib, isnt't there where everyone contributes their offers of p3n1z enlargenment etc?:>14:06
ryuoaka, the linux blackhole :->14:06
tilmantreach: no, we fixed that...14:06
treachhah, ok.14:06
ryuoanything that goes in never comes back14:06
* sepen echo "hello crux" | mail -s "subject" 2>&1>/dev/null14:07
treachI think I've recived like 5 mails from it that wasn't spam :P14:07
* ryuo sends his email to /dev/null.14:08
tilmani don't get why everyone ignored that problem14:08
sepenwell I updated the unmaintained port list14:08
sepenfor any contrib devel who want14:08
treachtilman: nobody was responsible for it.14:08
cptnwell, but those mail were so obviously spam that any decent filter should catch them14:09
tilmanviper should have felt responsible, being the contrib guy ;)14:09
ryuoi think contributors who dont update their ports after a certain time when they need an update should 'expire' and be allowed to be updated by other contributors14:09
thrice`I don't think my filters (gmail) will catch it.  I send ML stuff right to archives14:09
sepenwell, Im also propose you to create a dedicated perl repository for placing all perl libs on it14:09
ryuoperhaps 6 months to be fair?14:09
sepenI found at least over 500 ports14:09
rehabdollsplitting up repos is great in principle.. the trouble starts when a port depends on something thats not in the same repo or in core/opt/contrib14:12
treachget duplicates then.14:13
sepenbut in this case only perl libraries14:13
treachit's easier than having some freaking  huge mesh of interdependent repos..14:13
sepenis not the same as if we have  perl oriented applications14:13
tilmanthe idea doesn't sound that good to me either14:15
tilmanbut i've had my share of #crux discussions for today :P14:16
*** jue has quit IRC14:18
thrice`er, caps14:20
*** will has joined #crux14:24
jesse_Sure, go ahead and split the ports. You'll see a lovely mess and let the whining games begin. :]14:34
*** mxs has quit IRC14:38
*** maxus has joined #crux14:38
*** maxus has quit IRC14:42
cruxbot[contrib.git]: flasplayer: release 3 -> 414:42
*** maxus has joined #crux14:44
*** maxus has quit IRC14:47
predatorfreaktilman: I've had enough enough discussions too, I'm done fucking with CRUX for today :D14:51
predatorfreakTime to rape avisynth some more14:51
*** maxus has joined #crux15:43
*** jnengland77 has joined #crux16:04
*** maxus has quit IRC16:21
*** Rotwang has quit IRC16:26
*** jdolan_ has quit IRC16:32
*** will has quit IRC16:37
*** jnengland77 has quit IRC16:45
cruxbot[contrib.git]: flock: Updated -> 1.117:03
*** jnengland77 has joined #crux17:04
*** pitillo has quit IRC17:12
*** RedShift has quit IRC17:22
*** destruct1 has joined #Crux17:22
*** destruct has quit IRC17:33
*** andarius has joined #crux17:36
*** destruct1 is now known as destruct17:37
andariusgreetings and salutations17:37
jesse_hi andarius.17:46
andariusperhaps he should be like the others then and release it before its done ?17:54
andariussorry, wrong data for the channel. i am having auto-scroll issues :(17:54
treachheh, better late than never they say, but I wonder in some cases. :p17:58
cruxbot[contrib.git]: flock: Replaced init script by symlink. Also removed one EMPTY file17:59
*** DarkNekros has joined #crux18:00
*** DarkNekros has quit IRC18:01
*** jnengland77 has quit IRC18:35
*** sepen has quit IRC18:52
*** joacim has quit IRC19:02
*** Romster has joined #crux19:37
*** treach has quit IRC20:28
*** DivineAnt has joined #crux21:15
DivineAnthi, i just installed crux, and is having some truble figuring out the ports and prt-get system21:16
DivineAnti added the contrib dir but prt-get doesnt search it21:16
DivineAntnevemind, found prtget.conf21:21
DivineAnthow do i get prt-get to use distcc? ive added it to hosts.allow and PATH21:43
predatorfreakRomster: distcc is your thing21:49
*** andarius has left #crux21:49
Romsteryou don't need to add it to hosts allow.21:52
Romsterand there is a README in the distcc directory prt-get readme distcc21:52
Romsterbut i'll pastebin what i got so far.21:52
DivineAntlets se what happens now, I exported makeflags21:56
Romsteryou'll have to edit it some still.21:59
DivineAntill give it a spin, thanks22:00
Romsteri'm about to cruse so if you want any more questions ansered i'll wait 5-10 minutes then i'll be out for upto 2 hours.22:00
Romstermy setup is a little diferent as i build as pkgmk and not as root.22:01
DivineAntok, im used to arch linux, but my via epia doesnt like 686 so i just wanted something kinda similar22:02
DivineAntbut i think im good for now22:02
Romsterk wwell if your building for jsut that pc use what you can find in man gcc22:03
Romsterand i'd use -mtune= for slightly faster code but spefic to your cpu or -march= for more generic code and it's -march=i68622:04
Romsterthat is default in crux22:04
DivineAntthere is a 586 port available whis i installed22:05
Romsterthat should work fine but i just fine tune my CFLAGS and CXXFLAGS22:05
DivineAntso that not a problem22:05
Romsteroh your on a i586?22:05
Romsteri have one box here that is a -mtune=k6-222:05
Romsteror -march=i58622:05
Romstertake ya pick mtune si more cpu spefic and less portable.22:06
DivineAntjep, via epia C3, its been hell finding a proper distro for it22:06
Romsterbut a litter quicker22:06
Romstercrux should do fine with the right flags22:06
predatorfreakRomster: if you're building for an Architecture, -march=blah22:06
predatorfreakWill specify -mtune=blah22:06
Romsterpredatorfreak, yeah22:06
predatorfreakBut, yeah, if you want generic support.22:07
Romsteri build per a cpu though.22:07
predatorfreak-march=generic -mtune=blah22:07
predatorfreakis best22:07
Romsternot with distcc22:07
predatorfreak(generic is an march)22:07
Romstergeneric or native messes it uo22:07
predatorfreakRomster: Generic isn't like native.22:07
DivineAntdefault in the 586 build is march=i586 and it seems to be working fine22:07
predatorfreakIt's akin to specifying -march=i686 or whatever as a baseline.22:07
predatorfreakOnly it's designed with all x86 processors in mind22:07
Romsterah k22:08
predatorfreakBasically it won't bork any single processor.22:08
Romsteris that a c3 or a c3-2 ?22:09
DivineAntmodel name: VIA Samuel 222:09
Romster-mtune=c3 would give you the best results then.22:10
Romsterif your not looking for portability.22:10
DivineAntyou sure, because i think i read somewhere that not all C3 are the same22:11
Romster-O2 -mtune=athlon-tbird -fomit-frame-pointer, is what i use.22:11
predatorfreakRomster: In fact, I think I should propose, when the time for 2.5 rolls around, that -march=generic -mtune=blah becomes the default, as it will defeat the need for i586 builds and shit.22:11
Romsterthats why there is a C3-222:11
Romstercat /proc/cpuinfo22:11
Romsterpredatorfreak, maybe22:12
Romsterthere isn't alot of i586's about really.22:12
Romsterand jue has a iso for i58622:12
predatorfreakRomster: I know.22:12
predatorfreakI'm saying jue wouldn't need to roll a separate ISO anymore.22:12
predatorfreakWe'd roll one ISO and be done with it.22:12
DivineAntno its just a c3, it doesnt have sse22:13
DivineAntflags: fpu de tsc msr cx8 mtrr pge mmx 3dnow22:13
Romsterbut it wont be so optomised *shrugs*22:13
predatorfreakRomster: Compiler rarely affects your performance that god-damn much22:13
predatorfreakESPECIALLY -march.22:13
RomsterDivineAnt, ok so C3 will do for you what you maybe read about is that some use the C3-2 cpu but yours does not.22:13
predatorfreakand most users change their march after installation anyway22:13
predatorfreakI know I did.22:13
Romsterwhich is why i use -mtune <<22:13
Romsterand don't set march22:14
Romstermarch is generic for a arch type like all 586 and up or 686 and up22:14
predatorfreakRomster: I think it's best as a "generic" thing for the ISOs are shit.22:14
predatorfreakso that we can stop wasting time rolling separate ISOs for shit like that.22:14
Romsterguess we should have a march=i586 and that will work on a 686 too.22:15
predatorfreakPlus there's very few applications which benefit significantly from compiler optimilisations.22:15
Romsteranyways i'm off if you need mroe help DivineAnt give a hoy in a couple of hours.22:15
predatorfreakNamely, video codecs.22:15
predatorfreakEverything else it's ricery <_<22:15
Romsteri'm just a whore at optomisations where i can.22:16
Romsterbut i dont' go insane like a ricer does.22:16
predatorfreakx264, in particular, benefits like fuck from a good compiler.22:16
DivineAntRomster, i think ill be good for now, soon off to bed. But thanks for the help22:16
predatorfreakxvid to a lesser extent, they've got well-optimilisation0red assembly.22:16
Romsterk drop in some more DivineAnt :)22:16
predatorfreaklibavcodec encoders are pretty optimised too :)22:16
predatorfreak"BTW, did anyone note yet that this thread started about HD. Now we've reached VHS."22:24
*** morlenxus has quit IRC23:05
*** morlenxus has joined #crux23:05
*** DivineAnt has quit IRC23:33
*** ryuo has left #crux23:50

Generated by 2.11.0 by Marius Gedminas - find it at!