IRC Logs for #crux-devel Wednesday, 2015-06-10

*** Feksclaus has quit IRC01:24
*** _mavrick61 has quit IRC02:35
*** _mavrick61 has joined #crux-devel02:36
*** deus_ex has joined #crux-devel05:43
*** frinnst has quit IRC09:52
*** frinnst has joined #crux-devel09:54
*** frinnst has quit IRC09:54
*** frinnst has joined #crux-devel09:54
*** Feksclaus has joined #crux-devel11:12
*** Feksclaus has quit IRC13:35
jaegerseems like the freetype update broke firefox builds13:44
frinnstoh?13:45
frinnstlets see if the beta is also broken13:46
jaegerfrinnst: http://pastebin.com/n5mbWxnD14:15
frinnstbeta built fine, opt/firefox not so much15:23
frinnstIm sure thunderbird is just as broken15:24
jaegerprobably15:24
frinnstteK_: https://www.mail-archive.com/qemu-devel@nongnu.org/msg302403.html15:36
frinnstcve not yet assigned i think15:37
frinnstsorry, obviously it has: CVE-2015-320915:37
frinnstac_add_options --enable-tree-freetype seems to solve it16:11
frinnstTricky finding any docs that describes what it really does though16:11
jaegerI would guess that means it will use an internal version instead of the system version16:12
frinnsti guess. but the flags are usually called something like --with-system-<library>16:14
jaegeryeah, I'm just guessing that "tree" means "firefox source tree"16:15
jaegerI could be wrong16:15
juefrinnst: are you running gcc 5.1?16:35
frinnstyes16:35
frinnstofcourse :D16:35
jueand probably the latest glibc and binutils? :)16:36
frinnstindeed16:36
frinnstgcc 5.1 was quite painless16:36
frinnstalmost everything built. i think ncurses failed16:37
frinnstbut that was more or less it16:37
jueI guess 5.1.1 isn't too far away16:37
frinnstyeah16:38
juelooking at TODO32 I see nothing really critical16:39
jueok, we have to decide about libwrap and dash, but that's all currently16:40
jueor are there other ideas not yet written down?16:41
jueah, I remember we had a discussion about runit16:42
jueis that something we should consider right now?16:43
frinnstI see no reason to consider runit16:45
jueTBH, I still like our super-simple init script and starting daemons isn't that important for me16:46
frinnstpersonally its very uninteresting16:46
frinnstyeah me too16:46
jue:)16:46
frinnstgperf to core?16:46
frinnstdbus-glib is deprecated but still breaks some stuff if removed16:47
frinnstdoubt much has changed there since I last looked at it back in february16:47
frinnstnot sure if that is included on the iso16:48
frinnstwe have a couple of "researching" tasks open in flyspray16:48
frinnstdosfstools on the iso, curl for pkgmk, manpages compressed as xz16:49
juegperf: I'd say so, it's a single binary and I'd prefer that over patching eudev all the time16:49
frinnstmanpages as xz seems pointless to me16:50
jueyeah, indeed16:50
jueand changing that would either mean we have to rebuild every port or use some nasty tickery16:51
frinnstyeah16:51
jue/usr/man is 49M on my box which is more or less nothing nowadays16:52
frinnstand just imagine what it will be like to open a xz manpage on a rpi1 :)16:52
jue:)16:53
jaegerI do think we should consider moving to /usr/share/man, it seems to be the default mostly now16:54
jaegerannoying amount of work, though16:54
frinnstyeah painful but probably easier in the long run16:55
jaegercan man support multiple paths? perhaps 3.2 can be a transitional release, supporting both locations16:55
juegood call, we should add that to TODO3216:55
frinnstyeah you can still open manpages in either location16:56
jueadded16:58
frinnstIm wondering if I should change firefox.. We currently use "official branding" but I'd bet we are not allowed to17:00
frinnstobviously nobody from mozilla cares17:00
frinnstbut still, since we include it on the iso it feels kinda "icky"17:01
jaegerhrmm, no idea there17:01
frinnstwe could 1) ignore it 2) ask mozilla 3) dont use official branding17:01
frinnstthe unofficial branding is just a blue icon17:02
frinnstanyways its not that big of a deal at the moment17:03
jaegerprobably the "proper" thing would be to ask mozilla and go from there17:03
frinnstyep. I was just afraid to poke the bear :)17:03
jaegerhehe17:03
*** frinnst has quit IRC17:09
*** frinnst has joined #crux-devel17:11
jueIIRC you both were for dropping tcp_wrappers?17:13
jaegerI am, but it's not a requirement for me17:15
jueI kinda like it, but the option to drop a unmaintained software is more compelling ;)17:18
jaegerfair enough. :)17:18
jaegerI just use iptables for stuff like that now17:18
*** Workster__ has quit IRC17:22
juejaeger: what do you think about using dash as /bin/sh?17:31
juebbl17:47
jaegerSo far I don't have any objections. I'm using it on my ISO build box without much trouble.17:50
frinnstsame here18:00
frinnstfor dropping tcp_wrappers, for using dash18:00
teK_frinnst19:07
teK_you built gcc 5.1, did you configure it to use the new (C++11?) ABI or not?19:08
frinnstisnt that the default?19:12
frinnstif not, no i did not19:12
teK_the default is incompatible with clang as it's 'gcc extension'19:12
frinnstthat would explain why everything worked :D19:12
*** Feksclaus has joined #crux-devel22:17
*** Workster has joined #crux-devel22:54
*** Workster has quit IRC22:54
*** Workster has joined #crux-devel22:54
frinnstThis is a minor bug release that fixes a regression in `ifconfig'.23:03
frinnstThe issue was reported by Juergen Daubert in23:03
frinnst\o/ :D23:03
jaegernice :)23:13

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