IRC Logs for #crux-devel Thursday, 2014-08-28

jaegerI built some updated i686 stuff while I was working on a slow project at work00:27
*** Workster has joined #crux-devel01:54
Workstersupose we could do a i686 iso? seems we get a few asking but it's just that much effort when i'm all x86_64 now01:56
jaegerI've considered it, just takes someone doing the work02:17
*** mavrick61 has quit IRC02:40
*** mavrick61 has joined #crux-devel02:41
*** c0x` has joined #crux-devel05:57
*** `c0x has quit IRC06:01
teK___how old has your machine to be to not support 64bit hardware-wise?08:15
nrxtxintel builts again some for low power devices08:33
nrxtxand many 3-4 year old netbooks still have those non-64bit hardware08:34
nrxtxno idea why08:35
juejaeger: no sure what I'll do, my old i686 laptop is still in use ...12:29
jaegerjue: I'll share up the ports I hacked last night, then. Might save you some time12:32
juejaeger: thanks, do you have a deeper interest in keeping the i686 overlay up-to-date?12:42
frinnstI bought an eeepc 2009 with a 32bit atom cpu12:43
jaegerI'm not sure yet, to be honest.12:43
jaegerI think that it might not be a lot of work once it's ready but we'll see12:43
jaegerIt's just that we still have a few people asking about it now and then12:43
jueok, if so we should probably share the github repo12:44
juejaeger: btw, I did it yesterday ;)12:45
jaegerdid what?12:45
jueordered an update for this one -> :)12:46
jaegernice :)12:47
juewell, the current one is almost 4 years old, so I though it couldn't harm12:48
juebut I have to wait until beginning of november for it12:49
jaegerit'll be worth it :)12:50
jaegerIf you prefer to merge the updates into your repo that's fine. Your repo has some more things in it that I haven't messed with yet13:00
jaegeroops, I left a build log in gcc :)13:06
jaegerIf I do end up building ISOs from the overlay it would help to split it into core and opt, probably13:08
juejaeger: thanks for sharing the repo, will update my latop with it the next days :)13:14
jaegerno problem13:14
jaegerI'll be doing some cleanup in it, too. removing build.log, multilib patches, etc.13:14
jaegerI wonder how much RAM it takes to link 32-bit firefox now15:22
jaeger2G RAM + 1G swap was enough, surprisingly16:18
*** Workster has quit IRC17:16
frinnstany objections to me pushing the glibc fix for CVE-2014-5119 ?17:42
frinnsti've tested it on 3 machines without issue. the patch is in glibc master too17:43
frinnstpushing in 317:45
jue frinnst: no, works for me as well on 3 boxes :)17:59
nrxtxRomster: do you still plan to use port tagging? might need some testing later18:01
jaegerfrinnst: none from me18:31
nrxtxteK___: you are maintaining ports in repo "alan" and "prologic"? ;)19:10
teK___wtf? no :p19:12
nrxtxat least when you look for maintainer lines :D19:13
teK___give me a real hint please19:13
nrxtxcyrus-sasl and libgcrypt1519:14
teK___% grep Maintainer opt/cyrus-sasl/Pkgfile19:15
teK___# Maintainer:  Thomas Penteker, tek at serverop dot de19:15
teK___I dont get it..19:15
teK___that's one form of giving me credit :P19:16
jaegerJust means they copied the port without changing it, probably19:17
teK___btw! frinnst jue jaeger Romster: I am going to be away from Saturday to Saturday, feel free to trim important ports I maintain; I'd be grateful for a short notice (preferably by mail), just in case :)19:18
jaegertrim as in remove them? :)19:19
teK___trim also means ausruesten, dekorieren or einordnen! :)19:20
jaegerduly noted :P19:20
teK___I have the skillz!19:20
teK___wrt the glib post on the ML..19:36
teK___  g_atomic_int_inc (&source->ref_count);19:36
teK___this makes gsettings segfault19:36
teK___but I lack the debugging skills finding out what's wrong with that19:41
teK___uh oh19:48
teK___this goes on to __sync_fetch_and_add19:48
teK___which is a G_GNUC_EXTENSION19:49
teK___so either glibc or gcc b0rkage? ._.19:49
nrxtxthe reason why it fails is because "source" var is null19:53
teK___too easy =)19:54
teK___could have spotted that in:19:54
teK___#0  g_settings_schema_source_ref (source=0x0) at gsettingsschema.c:20319:54
nrxtxaltough it should never happen except in one case...20:00
nrxtxthis happens when he is unable to find a single schema source directory20:01
nrxtxwhen iterating through none he sets initialized as true even though no valid source is there20:02
nrxtxhe is looking for XDG_DATA_DIRS, /usr/local/share/ and /usr/share/  each appends "glib-2.0/schemas"20:07
teK___yeah but this is clearly a bug in glib?20:08
teK___write to the ML already, Thorsten is getting nervous20:09
nrxtxit seems like it looks for /usr/share/glib-2.0/schemas/gschemas.compiled20:10
nrxtxwhereas the port supplies gschema.dtd Oo20:10
nrxtxand there we go20:13
nrxtxplease do not forget to execute "glib-compile-schemas /usr/share/glib-2.0/schemas/"20:14
nrxtxafter installing the port20:14
teK___google uses .aspx pages for rating their interview process, ZOMG!!1120:14
nrxtxfrinnst: that's something for post-install20:15
teK___nice work nrxtx20:15
nrxtxand you might need a readme20:15
jaegerthis is where pre- and post-install scripts being on by default would be handy20:19
nrxtxyeah but the post-install script for glib does not exist yet20:20
jaegerthat could easily be remedied20:21
nrxtxdepends if frinnst is still active or if he is out of coffee after drinking that much this morning20:26
teK___he's probably drunk right now, it's Thursday after al20:27
frinnstglib cant use post-install because its shipped on the iso20:28
frinnstso even with a post-install script for glib - a fresh crux install will still segfault20:29
nrxtxthen you might call it in build()20:29
nrxtxwith absolute oaths but don't know if that works20:29
teK___and we could fix it in the setup script, too20:29
frinnststill same problem when installing a binary package20:29
frinnstan ugly workaround was to do this with gtk3 post-install20:30
nrxtxfrinnst: what about doing it in build()?20:30
nrxtxcould result in problems when cross-compiling ports but don't know if this is done20:31
frinnsthm, yeah. something to play with this weekend i guess20:31
jaegera runonce type script could do it for the ISO if needed20:33
nrxtxteK___: does it work for you thorsten on the ml gets errors?20:34
teK___that is not a sentence20:35
teK___I can reproduce the problem, if you meant that?20:35
nrxtxrunning the glib-compile-schemas command20:36
teK___it solves the problem, yes20:36
nrxtxhave a look into the mail he wrote, can't see why theres a problem running this command20:37
teK___strange :)20:38
jaegerteK___: do you have any schemas in the schemas dir?20:38
teK___-rw-r--r-- 1 root root  2839 [2014/07/15] gschema.dtd20:38
teK___hm not the file, I guess :o20:39
jaegerwould be xml files like org.mate.terminal.gschema.xml20:39
teK___yeah, plenty20:39
jaegerI was just curious if not having any installed caused the problem, but it seems not20:39
teK___% pkginfo -o 2.0/schemas | grep -c xml20:39
teK___but none is part of glib20:40
teK___gtk3 and virt-manager put their stuff there20:40
jaegerwell, that makes sense, then, I'd guess. gsettings doesn't do much without schemas, I bet20:40
teK___what about an empty .xml file in the directory?20:40
teK___will test that20:40
nrxtxwon't work20:41
jaegerI wonder what he's trying to accomplish with gsettings anyway20:41
nrxtxit needs the gschemas.compiled20:41
nrxtxhe doesn't look for the xml directly20:41
nrxtxonly for that file20:41
jaegerbut if you have no schemas it will not create that file20:41
jaegeralso, if you have no schemas there's no reason to use gsettings, I think20:41
teK___an empty gschemas.compiled works!20:41
teK___so frinnst put an empty file /usr/share/glib-2.0/schemas/gschemas.compiled in your port :-)20:42
nrxtxalso long that does not result in further null pointer crashes :D20:51
teK___fucking freenode20:54
nrxtxdon't you own "teK_"?21:02
teK___this whole nick registration thing is b0rked21:03
teK___I don't care anymore21:03
nrxtxuh i should remove debug symbols from glib21:09
nrxtxhopefully not needed again21:10

Generated by 2.11.0 by Marius Gedminas - find it at!