IRC Logs for #crux-devel Sunday, 2012-12-23

Romsterminor issue though as i was going to enable it anyways00:00
Romsteroly got it in a chroot atm00:00
Romsterdo i have access to compat-32?00:00
jaegerlooks like you do00:01
Romsteri take it jue or tek are the only ones that can edit rsync.conf00:03
jaegerI can as well but it shouldn't need any changes00:03
jaegerat least not for compat-3200:04
Romsterit needs the one for comtrib 3.0 ^00:04
Romsteri made the branch earlier today00:04
jaegerah, so we probably need to update a post-update hook, not change rsyncd.conf00:05
jaegerthere's only one repo in the rsync config, ports00:05
Romsteri got some hollidays coming up so i can get to more fixing and testing.00:05
Romsteroh i see00:05
Romsterthat makes sense.00:06
Romsteralso hi i didn't say hi earlier :)00:06
jaegerok, added 3.0 to the post-update hook00:06
Romstercan you run it manually00:08
jaegerI think it'll run with the wrong umask but if it does I can fix that manually00:09
jaegerran it, can you sync?00:11
jaegerlooks fine on my end, for what that's worth00:11
Romsteryeah it's fine now, thanks.00:16
Romsteri went outside for a bit, really warm here today00:16
Romsterhumid too00:16
jaegerI think the plan for compat-32.rsync was to make it inactive but I can't remember for sure. I think we talked about doing it that way, need to search my logs00:19
Romsteri'm pretty sure it was meant to be inactive and only those that need it to enable it.00:20
Romsterbbs grocery shopping00:21
jaegerI think so, we can always change it back if needed, though I suspect we won't need to00:21
jaegerok, fixed up ports and prt-get. now I need to get some sleep, catch you guys after a few hours00:27
*** mike_k has joined #crux-devel04:20
Romsterwern't we going to remove libpthread-stubs not just bump it's version?06:13
Romsterguess xorg-libxcb wont work without it even though it's a stub06:18
jueRomster: we tried to remove it, but other stuff is looking for pthread-stubs.pc and errors out if not found06:18
Romsteri just tested and that's what i came up with too.06:19
jueit's not worth the trouble to patch other ports because of that06:19
Romsteralso damn i've been nuking the libpthread-stubs out of the ports i've been updating.06:19
Romstermeaning removing it off the depends on line06:20
juewell, it's still a build-time dep for some ports e.g. libxcb06:21
Romsteri would of thought that port would be optional not required. i guess it breaks too many things for xorg to remove that yet.06:22
Romsterreally only 3 ports seem to depend on it xorg-libxcb xorg-xf86-video-nouveau libdrm06:23
Romsterbut i guess it has to stay as even LFS has it.06:24
pitillowin 206:25
pitillosorry :)06:25
pitillowin 206:25
pitilloummm no slah from the phone, my bad :)06:26
jue /win 106:26
Romsterno forwared slash on phone?06:26
pitilloummmm nop, I'm connecting from vnc to the laptop and there are problems with the slash this morning06:27
pitilloummm last e17 update made something06:28
Romsterjue, i'm not so sure wine in opt is right as that flavour depends on compat-32, or is compat-32 meant to be on by default and not disabled.06:29
Romsteron the 3.0 branch06:29
Romsterwine can work as 64bit only but it's not much use unless the user only has a 64bit PE to run. but not sure if opt/wine should be 64bit only and compat-32/wine be the multilib one... but then compat-32 needs to be above opt. or just leave stuff as is.06:31
Romsterpitillo, broke something?06:31
Romsterheck what's the permissions on compat-32? any compat-32 user can update all ports like the xorg repo is. or is each port associated with the owner of the 64bit version... I'm really out of the loop on what's going on.06:33
pitilloyep, seems I'm making something wrong wiyh kbd06:33
Romsterthat is always a pain, figuring out what broke it.06:34
jueI'd say compat-32 is a group-effort like xorg, but I've no strong opinion on that06:36
Romsteralso depends on jaeger's opinion since he started it i just continued it.06:37
pitillostrange and possibly a casuality... setxkbmap again instead of doing it from xinit reloaded kbd without problems06:37
Romsterpitillo, is it because your using a different keymap than english?06:37
jueRomster: wrt wine, that's a good question, but at a first thought I'd say that wine should stay in opt, even if it depends on stuff in compat-3206:39
Romsterreally 64bit wine is not that much use so keeping only one multilib wine is prefferble. but i did say to jaeger that i thought compat-32 was meant to be disabled by default as not everyone needs 32bit, but then that will break wine in opt.06:39
jueyeah, compat-32 should be inactive by default06:40
Romsterhaving compat-32 on by default probably isn't a bad idea then.06:40
pitilloyes Romster, I think it's picking english by default since e17 stable release removed old configs06:40
Romsteror is it, wine in opt wnt build without that.06:40
Romsterso really the rules for opt ports are they can depend on core xorg and compat-32 only.06:41
pitilloinstead of adding compat32 a README FILE in wine's port explaining those needs could be enought06:41
RomsterI guess I could add to the wine README that it needs compat-32 enabled to work.06:42
Romstermakes sense.06:42
Romsterwont be many 32bit ports in opt, wine is just a specal case.06:42
Romsterbeen wondering if wine would be better in contrib.06:43
juewhat's better than?06:44
Romsteri like other devs opinions on such complex and confusing decisions as i don't always have the best answers, or fail to see some fault.06:44
pitillothat's another option too, or a wiki page... lot's of options and all good06:44
jueit still depends on other stuff in opt and compat-32?06:45
Romsterwell there is other ports that wine can build with and there in contrib already, but regardless still need compat-32, just thought keeping wine off opt that there by no other ports in opt that would require compat-32 to be enabled.06:46
pitillofrom my pov to keep consistency would be better to move it to contrib and add a readme file for it (opt will stay as it should be with only one problematic port)06:46
Romsterit does yes but removes the compat-32 being needed at all on opt.06:46
juesorry, what's better if we move it to contrib?06:46
Romsterunless that's not an issue at all then no point shifting wine to contrib.06:46
pitilloare there more ports in opt which will need compat32?06:47
Romsteri was thinking of the compat-32 required for wine in opt. wine is the only port in opt that requires compat-32 afaik. while there could be more in contrib at a later porint from the emulators ports tree.06:47
Romsternot what i'm aware of pitillo06:47
juewell, I don't get it ;)06:48
jueeven if wine in contrib you have to enable compat-32?06:48
Romsternevermind then it's probably not worth the shifting.06:48
Romsterthe requirment of compat-32 would be on contrib only instead of opt is all i'm getting at.06:49
Romstershifting wine to contrib would remove compat-32 off opt and put it for contrib only.06:50
pitillokeep things as more homogeneus as possible. opt will work as always... if you want wine, mpup the port if you son't want the entire collection06:50
jueI don't see a requirement of opt on compat-32, it's wine that needs it06:50
Romsteri'm thinking fo the rules of ports in general using wine as an example.06:50
Romsteropt currently has a core xorg only rule.06:51
Romsterhaving wine in opt would mean addign the compat-32 to that rule too.06:51
pitillosorry if I tell something wrong, long night and breakfasting computer on sunday is a bad combination :)06:52
Romsterthat's fine pitillo06:52
Romsterif i still make no sense it probably does not matter and i suck at trying to explain myself. no loss either way.06:52
Romsterit's probably not a problem and if there was a problem jue would certantly tell me about it.06:53
Romsterso as it is now jaeger has disabled compat-32 by default as it is enabled in rc106:54
Romsterand wine stays in opt, everything is good.06:54
Romsteri'll add to the wine README on requiring compat-3206:54
jueguys, don't make the things more complicated as they are ;)06:57
jueif we introduce a new 'official' repo like compat-32, it's fine for me if it is on the same level than opt/xorg06:58
Romsteri was seeing if tings could be made less complicated :)06:59
jueand having a note in the README of wine should be enough06:59
Romsterand done.06:59
Romsteri just wanted to pass the ideas/opinions as to how it should be.07:00
Romsterconsidering this is a group project.07:00
jueyeah, sure, we definitely should discuss such things07:01
Romsteri'd prefer to, than make some change and make everyone angery -_-07:01
pitilloand that's really the way to go, talking and discussing things gives more than a solution of a topic07:01
Romsterlike you guys know i've changed over the years on this project.07:02
Romsteralthough i do have less time than i would like to devote here.07:02
Romstersome time is better than no time though.07:02
Romsterwhile where here, is there any other things that need taking care of?07:03
Romsteri'll likely goto bed shortly.07:03
juethe discussion shows me that we have different views on the contrib repo, basically contrib is a community effort and therefor keep in mind that moving a port to contrib means that we move it outside of the official project07:06
pitilloRomster: changing while learning, bit by bit :)07:06
pitillojue: yes, your explanations (the new "official" repo sentence) explained the differences between repositories and the side effect of moving an opt port to contrib :)07:07
jueok :)07:08
Romsteri'm aware of the contrib repo being a comunity effort.07:10
Romsterpersonally contrib is a semi-offical status but ports in there should be at the same calibra as opt or perhaps a bit lower on testing quality.07:11
jueof course, that only my point of view, if others have different opinions about the status of compat-32 we need an other solution for wine and the like07:20
jaegerRomster: I understand what you were saying regarding moving wine to contrib but I don't have any objections to it staying in opt. A note in the README would be fine with me. On the other hand if you WANT to move it to contrib I don't have any objections. Up to you.07:44
Romsteri just updated README about compat-3207:51
Romsterrest is fine as is07:51
frinnstwhy not just move wine to compat-32?07:54
frinnsta pure 64bit wine is more or less useless, no?07:54
Romsterunless user has only 64bit PEs yes07:58
Romsteri thought about that too.07:58
frinnstwine is the only port we ship that really needs 32bit stuff, right?07:59
Romsterand stuff in emulators08:00
Romstermight be other stuff i'm not aware of too.08:04
jaeger /wave08:59
juebad news, sed is unmaintained now ->
jaegerwow, never expected that09:06
jaegerI imagine someone will pick it up09:11
jueyeah, hopefully09:14
juebtw, there are a few new feature requests, namely 869 and 871-874, in our tracker, opinions?09:18
jaegerwill go look09:19
jaegerheh, #869 is something I suggested like 4 years ago and nobody liked09:20
jaegeron #871 I like the noop for non-rotational disks but not everyone uses deadline, we shouldn't force that09:21
jaeger#872 doesn't appeal to me because while having a firewall is good, everyone does their iptables scripts differently09:22
jaegerRegarding #873 and #874 I have no strong feeling, either is fine09:23
jaegerUnrelated to those, I think #190 can be closed. If someone wants to maintain a texlive port it can go into contrib09:25
juewrt #873, you read my comment in #826?09:26
teK_hi folks09:27
teK_a fresh install of 3.0-rc1 went 100% smoothly on my thinkpad09:28
jaegerheyo, teK_09:28
jaegerglad to hear that. :)09:28
juehi teK_09:28
teK_building my usual tool chain right onw09:28
teK_chromium (aka the kraken), cmus, xfce-Terminal etc.09:29
teK_not binutils and firends ;)09:29
jaegerah :)09:29
teK_is there anything specific I'm ought to test BESIDES the 32-compat glue?09:29
jaegerI would say just do your normal things and report if anything goes wrong09:30
teK_maybe I get to try and build steam and wine wrt 32bit09:30
jaegerthe wine port in 3.0 is already multi-arch so that's an easy one to test09:30
juere #871: I'm asking myself what our policy is here, adding our own stuff to udev or going mostly with that what upstream delivers09:30
jaegerjue: my preference would be to leave upstream alone, honestly. We could always put that stuff into a wiki page (or the guy who made the ticket can)09:31
jaegerI just put 'echo "noop" > /sys/block/sdX/queue/scheduler' into my /etc/rc.local :)09:31
jueagreed, I'd prefer this as well09:31
jaegerWe do have an 81-crux.rules file for fuse so it's not 100% upstream, I guess09:33
jaegerbut I'd rather not add things when we don't need to09:33
teK_ meant steam and skye ;)09:33
jaegerteK_: ah09:33
jue#872 is a no-go for me as well, at least I wouldn't maintain such a firewall-script09:33
jaegerI view it as a subjective thing. Some people like to put that into /etc/rc.d/net, or /etc/rc.d/firewall, or /etc/iptables-save, or any other thing09:34
jaegerWith a small team we have to let users do some of their own stuff :)09:35
juemy main concern is again the security POV, if we officialy ship such a script we are responsible for it ;)09:36
jaegerfair enough09:37
juere #190, we have a texlive port in opt, maintained by Danny :)09:40
jaegerok, problem solved :)09:44
jueobjections if I reject #873?09:59
jaegernone from me10:14
jueok, closed that and added comments to #871 and #87210:17
juelsscsi seems handy, even more nowadays where nearly every disk is ata/scsci10:19
frinnstnah, "no" on #871. "meh/no" on #86910:36
frinnstlsscsi might be useful10:38
frinnst#873 might be a good idea10:39
frinnstbut ah, i see we need to move libs around10:40
frinnststupid backlog :)10:40
juewhy? I think it's useless for a distro like CRUX where most users have only a very limited number of modules10:40
frinnstyeah, i thought there was no "downside" to it.10:41
frinnstafk again10:42
jue#871 closed too10:45
*** clb has joined #crux-devel14:30
frinnstgeez, building firefox is becoming more and more difficult with each version15:02
teK_what are they at .. like 20 or s?15:52
teK_chromium took 2 hours on this i7 laptop15:52
frinnstyay, the alsa-ftp is broken again16:02
*** __mavrick61 has quit IRC21:51
*** __mavrick61 has joined #crux-devel21:52
*** pedja has quit IRC23:09

Generated by 2.11.0 by Marius Gedminas - find it at!