IRC Logs for #crux-devel Wednesday, 2015-11-18

*** ___mavric61 has quit IRC03:27
*** ___mavric61 has joined #crux-devel03:29
*** Workster has quit IRC04:07
*** fktkfidbjmhautpa has joined #crux-devel04:07
*** Workster has joined #crux-devel04:07
*** bfxzzvxshgtijrdi has joined #crux-devel04:11
*** Workster has quit IRC04:11
*** Workster has joined #crux-devel04:11
*** Workster has quit IRC04:12
*** rgrxoursmswfnuwl has joined #crux-devel04:12
*** Workster has joined #crux-devel04:12
jaegerfrc5 seems to work ok here, I don't notice any immediate issues04:21
*** sudobaal has joined #crux-devel06:18
*** Romster has quit IRC07:53
*** Romster has joined #crux-devel08:57
*** ziggypup has joined #crux-devel11:25
*** ziggypup has left #crux-devel ()11:25
juejust finished the move of my desktop to 3.2, still no issues13:08
jueI'd say we release rc5 as final ;)13:09
Romsteri'm still working on ports but there not part of the iso. i'm sure i'll get reports soon enough.13:12
Romsteri'm not sure i like ncurses why are we messing with the libraries? don't we want everything to use the wide libs only?13:14
frinnstread the commit message13:18
frinnstor the release notes13:18
frinnstanyways, somewhere13:19
Romster heh13:23
frinnstFS#1210 was the ncurses bug i was thinking of13:32
jueRomster: did you run in any problem with ncurses?13:36
jueI think there's no reason to not use the wide library13:38
Romsterno just when i went to update ncurses-32 i looked over core/ncurses and i looked at some other bug reports that i don't recall where about only wanting to link to the wide character and i haven't looked at lynx yet in crux 3.213:38
Romsterwhich is supposedly broken due to ncurses, bu ti haven't checked that yet13:39
juelynx is broken because of a 404 ;)13:40
Romsteri see gentoo uses --enable-symlinks13:46
Romsteri already bumped lynx13:46
Romsterso it's not 404 now.13:47
Romsterunless that broke again13:47
jue--enable-syslinks for what?13:47
Romsternevermind i see the ncurses Pkgfile and it looks ugly. it's probably ok.13:47
juewhy looks it ugly?13:48
jueif you have something to improve it, give me a patch13:48
frinnstwell even touching libs looks ugly. i agree13:49
frinnstbut ncurses is a real bitch to package so13:49
Romsteri don't know yet but i've only just started looking at it, and looked for bug reports on it.13:50
juefrinnst: sorry, I don't understand, what do you mean with "touching libs"?13:51
Romsteri understand the move to /lib and then symlinking it, since core requires it. the rest i dunno much about yet. i'll do some research on it.13:51
juewell, the only ugly thing I can see are the sysmlinks for version 513:53
RomsterMany applications still expect the linker to be able to find non-wide-character Ncurses libraries. Trick such applications into linking with wide-character libraries by means of symlinks and linker scripts: ...
juebut we can remove them later on13:53
Romsterdon't worry about it now but i'll look into it.13:53
Romstermight be a case of many programs rely on them symlinks.13:54
Romsterinstead of using the pkg-config files.13:54
Romster(which still gets me why are projects having libtool and pkg-config?)13:55
jueyeah, our port is mostly a copy of LFS, I said this already13:55
Romsteri missed that point to LFS sorry. clearly LFS knows what is best. and they are doing it so little to no breakage in that.13:56
Romsteri jsut saw it a jungle of symlinks to libs and headers, and went omg...13:57
jueRomster: but lynx is still 404 for me ;)13:57
Romsterit's fine.13:57
Romsteri'll have to look at that /again/13:57
jueah, sorry, it works now, I forgot to pull13:58
Romsterah :D13:58
Romsteri'm atm fixing up 3.2 compat-3213:58
Romsterand started to build all my ports in contrib 3.2 and see what is broken13:59
frinnstjue: moving, symlinking etc. its ugly. upstreams makefiles should fix all that14:01
frinnstideally atleast14:01
Romsterwonder if upstream has a bug open for that.14:02
Romstersomething to look for.14:03
frinnstprobably not worth the effort14:03
juewell, it's not the only port where we move libs form /usr/lib to /lib14:04
Romsterit works as is though.14:04
Romsteryeah i am not fussed with the moving to /lib it's all the other stuff.14:04
Romsterbut then LFS is doing that so it's ok.14:05
jueit's easier to do it that way as to install with --prefix=/ and move devel-stuff to /usr/lib14:05
juemost progs do not support the split into /lib and /usr/lib and in times of systemd this will become more and more uncommon14:07
jueinteresting case is udev/eudev which supports it ;)14:10
Romsterwont systemd just end up with /lib and that's it no /usr/lib14:10
jueno, with systemd everything is in /usr14:10
Romsterspeaking of gudev there is a new library i need to find that again. since it's now stripped out of eudev.14:10
Romsterwas mentioned in here awhile ago.14:11
Romsterta that's it.14:15
juefrinnst: what do you think about rc5? are we ready for the final release?14:16
frinnsti'd say so. I havent tried it though14:16
Romsteri haven't had time to test but if you lot are happy with it then so am i.14:16
frinnstbut the issues I hit were fixed in rc3.5 (that I rolled myself)14:16
Romstersorry i haven't had more time for this.14:17
Romsteri am testing compiling of ports though.14:17
Romstercompat-32 for crux 3.2 should all be good now.14:18
frinnstRomster: ipset looks a bit outdated15:34
frinnstlatest version is 6.2715:34
frinnstjue: I have one crux install that is still on 3.1, I'll upgrade that with rc5 tonight15:41
juefrinnst: good, let's see how it goes15:50
*** sudobaal has quit IRC17:05
frinnstno issues17:17
frinnstother than my block device is slooooooooooow17:17
frinnstold usb stick that I use in case of emergency :)17:17
frinnsti'd say we're more or less good to do17:18
juefine, let's wait for a comment from jaeger20:13
jueI've updated the ChangeLog today, so mostly everything is done20:14
jaegerI'm currently busy with a work project but I'll try to get the UEFI updates into the handbook tonight20:27
frinnsthm, what happened to git?21:16
frinnst3.1 lists 3.2 changes21:17
rmullI think 3.2 stuff got pushed to 3.121:22
rmullMy p5-xml-parser footprint on 3.1 is looking weird - the footprint looks more appropriate for 3.221:22
frinnstyeah, wonder if it was something I did21:28
frinnstI dont see how:
frinnstline 1 is where I update btrfs-progs in core/3.121:33
frinnstthen cd into opt/3.2 and you can take it from there21:33
frinnstgit reset --hard c83ade7bbfc939ebb14527740298926c4d2c9941 seems to fix it21:39
juethat's my fault, I did a wrong merge which I've interrupted21:39
juebut now I see that it has happend partly already21:40
frinnstyeah my merge then really fucked it up I think :>21:40
frinnstwhats the best way to fix it i wonder?21:41
juehmm, good question21:43
Romsterif it's just one port just cherrypick it back21:48
Romsteri don't have time to look now21:48
Romsteri make sure with git branch before i go and commit stuff21:48
jue.oO what a mess, my apologies21:54
jueno idea how to fix it22:28
juefrinnst: if we do the hard reset, we lost some of the commits to 3.1 from today22:33
Worksterso what got merged from opt 3.2 to 3.1 ?22:37
jueyeah, looks like22:37
Worksteroh geez this is going back nearly a page of commits22:39
Worksteri think, because i see my 3.2 icu update in 3.1 tree22:40
jueyeah, a wrong merge is a disaster22:41
Worksterthis is why i make extra sure i am on the correct branch22:42
Worksteri see many "Merge branch '3.1' into 3.2" not once have a spotted a "Merge branch '3.2' into 3.1"22:42
juesure, normally I do as well, but shit happens22:43
Worksterfirst off we need to find what is wrong in 3.1 now. then we can git revert or cherrypick the older changes. or manually alter it in 3.122:43
Worksteri can't see where the wrong merge took place at what commit22:44
juewell, for me it looks like 3.1 and 3.2 are the same currently22:44
jueI have a working opt repo as of 2015-11-16, as a last option we can use that22:56
Worksterdon't think it's that bad22:56
Worksterit's not clear who did the merge. we can just git revert some of the commits that were only meant for 3.2. in the 3.1 branch?22:59
Worksterwont be many22:59
juethat works with the hard-reset frinnst posted above23:03
jueand we lose only the commits starting from wine until gtk323:06
Worksterhmm ok guess do that then. i'll leve it to you i got too much on here23:08
Worksteri can't o much until after work in like ~7 hours time23:08
jueWorkster: ok, thanks anyway23:09
juewell, I did the hard reset to commit c83ade on 3.1 and cherry-picked the lost commits from the 3.2 branch23:23
juehope everything is back in a good condition now23:25
Worksteris icu reverted or at the new version?23:30
Worksterand how many would of updated and broke half there system as a result.23:30
Worksterof ABI changes23:30
jueicu is at 52.1 on 3.1 and at 56.1 on 3.223:34
Worksterah phew that's okay then23:35
jueyeah, I think we are back again, at least a 'ports -u opt' showed nothing, which is a good sign23:41
jueI'm going to sleep now, cu23:46

Generated by 2.11.0 by Marius Gedminas - find it at!