IRC Logs for #crux-devel Tuesday, 2013-07-02

*** rmull has joined #crux-devel00:05
*** kris has joined #crux-devel02:19
*** nrxtx has quit IRC02:22
*** __mavrick61 has quit IRC02:30
*** __mavrick61 has joined #crux-devel02:31
*** kris1 has joined #crux-devel05:35
*** kris has quit IRC05:37
juegood morning06:56
jueRomster: you know that your icu update breaks a lot of other stuff?06:57
frinnstouch, what broke?07:02
frinnsticu infects everything it touches :)07:02
juewell, it's not ABI compatible, so-name changed from .49 to .5107:04
juethanks good I've remove icu from my system as soon as you removed it as dep from harfbuzz ;)07:06
frinnstit is removed07:06
frinnst:)07:06
jues/good/god/07:06
frinnstyeah that harfbuzz mess was a real fredrik-mess ;)07:07
juea nice quote from jaeger -> My laptop is also somehow infested with this icu plague... :)07:09
prologicthgoutougly test before eyou push :)07:15
jueRomster: we should do such sensible library updates only if _really_ necessary and only together with a special information for our users, like a notify commit or a mail to the ML. Most often it's better to postpone such updates and keep the old version of the library, like we do it for libpng etc. see TODO31.07:17
prologicif something is going to horribly break large parts of a system07:18
prologicI'd rather keep that for a point release of crux07:18
prologicwith related release notes07:18
prologicI think users pay more attention to release notes07:18
prologicthan random mailing lists emails or notifications from cruxbot or such07:19
frinnstto be fair to romster, i run icu here and dont notice anything07:20
jueprologic: that's the reason why I said '_really_ necessary' above07:20
juefrinnst: even after the icu update?07:21
frinnstyeah07:21
juehmm07:22
frinnstrunning a revdep07:23
juewhat do you get with a "ldd <something_linked_against_icu>" ?07:23
juenot something like /usr/lib/libicudata.so.51 ?07:24
frinnsti cant think of something thats linked against it :)07:24
frinnstuhm07:24
frinnstsilly brain07:24
frinnstrevdep seems to only report the usual suspects07:29
juejust did a test with harfbuzz-icu, it is linked against the versioned icu libs -> http://c204bf79c5973ae6.paste.se/07:48
Romsterjue, i tested icu on all ports that mentions icu on there depends on line.... my system is fine i rebuilt everything else that needed rebuilding08:02
juewhat do you mean with 'everything else' ?08:03
Romsterprt-get dependent --all icu08:04
Romsteri tested all those.08:04
teK_I rebuilt pango and gtk and everything went back to normal08:04
Romstersame08:04
jueyeah, that's the point ...08:04
Romsterpango gtk harfbuzz-icu webkit boost couple small other ports.08:04
Romsteris that too much effort and i wasn't expecting a ton of other ports to also break.08:05
jueyou have to rebuild a couple of other things to keep your system alive08:05
Romsteralso i didn't think many even used half of my ports. including icu seems to be a pauge like dbus08:05
Romstershould i revert it and bring more chaos or leave it b?08:06
Romsterbe*08:06
Romsterseriously how often do i break stuff... and when i do you all go at me like a ton of bricks.08:06
jueit's too late, I'd say08:06
Romstercooking food replies may be slow. no one told me sooner.08:07
Romsterwell the few ports i had to rebuild not an entire system, i thought it would be ok.08:07
Romsternot like it's libpng that infects the entire system08:07
teK_well .. revdep is mandatory after each sysup so I guess you did not exactly break things :-)08:07
juewell you know that we always have postpone sections in our TODO* and you know why we have it ;)08:08
Romsteralso if other maintainers would list icu in there Pkgfile i'd of tested those first.08:08
Romsteri test in a minimal chroot then on my system08:08
juewell, the point is not that the other ports work with the new version of icu, but breakage of other ports08:08
Romsteri have minimal time that thought never crossed my mind about icu being postboned.08:08
Romsterso really where do we draw the line on updates, do we postbone everything that causes other ports to break or do we draw a line at x amount of broken ports or what?08:09
juemy question: was the update really necessary or would it be possible to postpone until 3.1?08:09
Romsterthe rule is really not clear.08:09
Romsterit could of been postboned.08:10
jueok08:10
Romsterbut the time i bumped it i was going though opt and i notced i've never bumped that for /ages/08:10
Romsterso i tested about 4 or so ports broke ok not that bad...08:10
Romsterwe are experts we should know about how to fix that. but then i might be wrong *shrugs*08:11
juefor me stability and trouble-free operation is more important that to have always the latest version of a library08:11
Romsterso how many ports broken if any considers postboning a not urgent bump?08:11
jues/that/than/08:11
Romsteri go for stability than bleeding edge version08:11
Romstericu 15.2 has been out for awhile now.08:12
juehow ofter do you have a library update with a ABI change (= so name change)?08:12
Romsteri really don't know. i see a bunch of version changes on alot of ports. though i find not many break any other ports.08:13
jueI never would commit a libray which breaks ABI commit without special need and special notifications08:13
Romsterso i screwed up sorry08:13
jueRomster: we are talking about libraries not ports in general08:13
Romsterk well some ports have libraries... i'm interchanging words here sorry.08:14
Romsteri really haven't seen that many other than boost libsoup but i needed a libsoup bump for webkit in the past.08:15
Romsterpossibly a few others not many and now icu.08:15
Romsteri'll need ot be more warry of what i do change and if it's a ABI changed library weigh up if it can be postboned.08:16
jueyep08:17
Romsteragain sorry, my standards are usually high and i prefer stability over bleeding edge.08:18
jueteK_: you will almost never see a commit from me that requires a revdep, if you need one it was a [notify] commit to warn our users08:21
jueteK_: the situation is of course different after a version change of CRUX08:21
teK_ok08:22
jueteK_: I'd consider everything else as a bug ;)08:22
teK_bug.. imagine my looks when, after sysup, (xfce) terminal wouldn't start up anymore :p08:23
teK_rebuilding it did not help08:23
juewhat was the reason finally?08:23
teK_ldd printed the .48. something soname.. so I rebuilt terminal and vte08:24
teK_didn't help.. finally I found out that it took deps from pango/gtk08:24
teK_so after I rebuilt these two terminal worked without another rebuild08:24
teK_I was quite puzzled as I did not think of recursive deps.. pkg-config  printed 51-.. as version, too so everything seemed right except the deps in terminal08:25
jueyeah, every library used by gtk and friends is sensible because a lot of other stuff depends on it08:27
teK_heck even qemu broke08:28
teK_--> ef37bb5af982bfc67e297685fc73272c47fec245 ;)08:28
jueguess that's the --disable-gtk commit? ;)08:29
teK_it is08:30
Romsteroh08:31
jueanyway, I'll be offline for the rest of the week, take care guys :)08:31
teK_have fun :)08:31
teK_Romster: don't worry!08:31
juethanks and bye08:31
Romsterlike i said i did a revdep on my systema and i expeareanced no broken stuff after prt-get depinst foo for everything that mentioned in prt-get dependent --all08:32
Romsteri'm not too fussed teK_08:32
teK_bye jue08:32
Romsteranoied at myself dong something that stupid...08:32
teK_good08:32
teK_c08:32
*** jue has quit IRC08:32
Romsteri shuld write a mail i guess explaining if you don't want to rebuild half your system to prt-get lock icu08:33
teK_why not08:34
Romsterguess i'm more a novice than experienced or i just made a mistake.08:34
teK_shift happens08:35
teK_gotta lern for my exams NOW. afk. :)08:36
Romsterk08:36
teK_oh and gtk3 apps like gtk-spice need a rebuild of gtk3, of course08:42
teK_now. AFK!108:42
teK_:)08:42
Romsteryeah i broke alot of stuff...08:54
*** sepen has joined #crux-devel08:56
Romsterprologic> thgoutougly test before eyou push :) <- you really can't talk since when do you ever push stuff <_<09:00
frinnsthey, ive done that! :)09:22
*** sepen has quit IRC09:59
*** sepen has joined #crux-devel10:00
Romster:)10:27
teK_310:57
teK_err10:57
teK_http://forums.freebsd.org/showpost.php?p=120324&postcount=310:57
Romstereeps that really is /not/ a good idea.11:00
frinnstheh11:01
frinnstALL: ALL11:01
frinnstall my hosts.allow look like that ;)11:02
Romsterewww11:03
Romsterthen again if you got iptables blocking it what's an extra layer that's bypassed.11:03
Romsteri prefer the belt and suspenders approach11:04
frinnstits either public or physically isolated in my case11:07
teK_ALL: 192.168.1. for me .. my laptop is a moving target :)11:32
Romsterhttp://www.linuxfromscratch.org/lfs/view/systemd/11:41
Romsterewww11:41
frinnstxml-parser, libdbus, dbus and systemd are the new deps?11:58
Romsterseems so11:58
jaegerAnyone else notice font weirdness in firefox lately? Some lines or parts of lines of text get randomly bolded12:20
jaegerIf you highlight them they return to normal but if the page refreshes or you scroll they get messed up again12:20
Romsteryeah i have seen that.12:20
Romsteri don't know why.12:20
jaegerExtremely annoying visually12:28
Romsterat first i thought i was going crazy12:29
Romsterah crap i made firefox crash12:30
frinnsthm, probably some firefox-cairo bug12:37
Romsteri'm finding others had issues with type 1 fonts but no real specific leads as of yet12:38
RomsterWhy is all the text on Facebook bold suddenly?12:41
RomsterEDIT: I found the answer to this question, it was because I had accidentally uninstalled the system font Tahoma from my system and so it reverted to the next best thing, Tahoma Bold...12:41
Romsterlol12:41
jaegerhaha, nice12:45
Romsterhttps://bbs.archlinux.org/viewtopic.php?id=136771 that's nto it but it hits at freetype12:53
Romstercould be cairo or xcb or .... too ugh12:55
Romsterhwo to find this sucker.12:55
Romstersupose i could try using the cairo that comes with firefox and see if that fixes it.12:57
*** horrorStruck has joined #crux-devel12:58
Romsterarch still has cairo in firefox than system12:58
Romstercompiling13:01
Romsterprobably not cached in my ccache off to fire up my distcc cluster.13:02
Romstermean while jaeger has probably already recmpiled firefox and solved it :D13:23
jaegerIt's still building, heh13:24
Romstersame13:24
jaegerI had to focus on work stuff for a while this morning13:24
Romsterthat happens13:25
jaeger24m56.278s to build, just finished13:31
jaegerthat did indeed fix the issue13:31
horrorStruckanyone else having build failures of subversion 1.8 without adding -j1?13:35
Romstermine's done  didn't time it though and i was in the kitchen so it finished before i got back13:37
jaegerhorrorStruck: don't have it installed currently but I can test it in a bit13:38
Romsteri havne't updated that yet13:38
Romsterbut i got my distcc cluster running so now is a good time to try13:39
Romsteronly using -j12 though13:39
horrorStruckjaeger: thanks13:39
horrorStruckonly -j12 :)13:40
Romsteranyone got a known bad site for bold?13:40
Romstergoing though my history none yet13:40
Romsteryeah hehe wait till i get more nodes horrorStruck13:41
jaegeranything with lots of text on it13:41
jaegerI tested it with slashdot13:41
Romstergood idea13:41
jaegerman, I have a lot of ports that make revdep squawk these days... I should rewrite the patch that searched ports for their own reverse links13:42
jaegerjre, libreoffice, samba, syslinux, thunderbird, vagrant13:42
Romstergot a few myself13:42
horrorStruckif you guys want to test your fonts: http://www.infinality.net/files/font.html13:42
Romsterawesome13:43
jaegeryes, subversion does fail for me13:46
horrorStruckjaeger: thanks for testing and confirming13:46
horrorStrucksepen: ^ :)13:47
sepenhorrorStruck: thats my fault13:49
sepenI removed -j113:49
sepenI'll revert that, thanks13:49
Romsternot too blow my own horn but this paste of mine does the bold lines in places https://gist.github.com/anonymous/5876469/raw/a08c232ff08bd3e041b7161c241443e8845aab23/stdin.txt13:50
Romsterseems to move about on scrolling though how weird13:51
Romsternow to test on inbuilt cairo13:51
Romsterinbuilt cairo doens't do it system cairo does.13:52
sepenhorrorStruck: anyways I had -j4 as makeflags and the built worked13:53
Romstertesting subversion13:53
sepenjaeger, horrorStruck, Romster thanks ;D13:54
horrorStrucksepen: ah strange, it built fine only after adding -j1 on my machine13:54
jaegernp13:54
Romsteryeah system cairo is smoother to scroll but makes stuff bold.13:54
Romsteri haven't finished testing subversin yet13:55
jaegerI'd rather have the scroll flicker than the bold, myself13:55
Romsterlibtool: install: warning: `/usr/ports/work/subversion/src/subversion-1.8.0/subversion/libsvn_ra_serf/libsvn_ra_serf-1.la' has not been installed in `/usr/lib'13:57
horrorStruckyeah same stuff13:58
Romstersepen, yeah -j1 on make install only13:59
Romstercompiles fine on multiple jobs.13:59
Romster-j12 here13:59
Romsterthey obviously didn't make make install SMP safe.13:59
Romsterconfirmed chroot does the exact same sepen14:04
Romstersame jaeger damn cairo what hapend to it.14:09
*** deus_ex has joined #crux-devel15:26
*** nrxtx has joined #crux-devel15:30
frinnstwhat sites are you seeing this on?15:44
frinnstim unable to reproduce it15:44
jaegerAny site with a lot of text15:45
jaegerslashdot, for example, or that infinality.net link horrorStruck posted15:45
Romstergetting tired i'm of to bed, and i just noticed a mesa3d 9.1.4 update waiting, i'll do that after i slept if no one gets to it first hint hint, gnight.15:55
frinnstdo you need to autoscroll or something?16:15
frinnstah yeah. managed to reproduce it16:16
frinnstwhen firefox loses focus it reverts16:17
frinnstdamn, forgot to test the new intel driver at work today :(16:19
frinnstbtw, keeping the bundled cairo might cause other bugs: https://bugzilla.mozilla.org/show_bug.cgi?id=67130216:21
jaegerIt wasn't related to focus for me, just happened to random lines any time the contents of the main window moved or changed17:54
jaegereven if it was one click of a scrollbar or mousewheel17:54
*** pitillo has quit IRC20:38
*** pitillo has joined #crux-devel20:38
nrxtxis there already something similiar or interest in? http://pastebin.ca/241506922:17
prologicRomster:  I cannot find the time to be an active quality maintainer :)22:46
jaegerweird... all of a sudden I can't get usb storage support working on my updated ISO which makes installing from USB pretty goddamn hard23:01
jaegernrxtx: looks useful23:02
jaegerhrmm... maybe 3.9.7 is just broken, I can't see any problem23:24

Generated by irclog2html.py 2.11.0 by Marius Gedminas - find it at mg.pov.lt!