irclog2html for uclibc on 2002.10.15

00:05:28sandmanandersee: ping ?
00:08:07anderseesandman: pong
00:08:27sandmanhi erik
00:08:39anderseehey
00:08:43sandmanDid you get my mail regarding suid docu ?
00:09:24anderseeYup.  Havn't read it yet.  I'll try to get it reviewed, proof-read, and merged in this evening.
00:09:45sandmanProbably bad english ;)
00:11:13anderseeRegarless, bad english is far better than the nothing we had before. :)
00:11:26andersees/Regarless/Regardless/
00:11:43sandmanheh .. that's right .)
00:13:19sandmanI was very busy with Opie, but that is pretty stable now
00:13:55anderseecool
00:14:43anderseeI've been awfully busy myself, but a bit of the pressure is off me now so I'm trying to give busybox a bit more time.
00:15:43sandmanplans for 0.61 release any time soon ?
00:16:27sandman(or: are there any show stopper bugs in CVS currently ?)
00:20:09anderseesandman: I'd like to cut 0.60.5 tonight (yeah, I know I said it wouldn't happen) that should address the critical busybox in stable.
00:20:39anderseeThen for 0.61 we need to poke at bug1 and make him fix the archival stuff he broke.
00:20:53sandmanIs it still broken ?
00:21:01sandmanI haven't tested lately ...
00:21:35anderseeit would be nice to have 'make install' install a default busybox config file when support for that is enabled.
00:22:04sandmanHmm .. but it speeds things up, if this file is not present
00:22:20sandmanThis is why I don't install one in familiar
00:23:12anderseeperhaps so.
00:23:30sandmanMaybe we could install a busybox.conf.sample ?
00:23:33anderseeI wonder if we should change the defaults for ping, ping6, telnetd, etc
00:23:39anderseesandman: that works
00:25:03sandmanandersee: change to BB_SUID_MAYBE ?
00:25:17anderseesandman: that works
00:26:00anderseeping need root to do its thing.  telnetd needs root to listen on a low port.
00:26:29anderseeWe should probably add in thet tftpd applet, and that would need a low port as well.
00:26:36sandmanthen they should better have BB_SUID_ALWAYS ..
00:27:46BZFlagwhat does MAYBE do? leave suid for the default apps?
00:28:04BZFlagI'd be more happy with sane defaults, than having to have a config file.
00:28:14sandmanBZFlag: It just means don't drop any priviledges .. let the applet decide
00:28:49sandmanNEVER drops priv. before calling the applet
00:29:14sandmanALWAYS does not call the applet, if the process doesn't have sufficient priv.s
00:29:36BZFlagI'd prefer that applets register in a array someplace and then we drop suid if the applet name is not in the array.
00:30:27sandmanBZFlag: That's what the applet list in include/applets.h is for
00:30:35coobbrb
00:31:08sandmanIf you want to handle this user-defined, use /etc/busybox.conf
01:35:04sjhill-afkandersee: still around?
01:37:13sjhillis anyone on here that had problems with mips and dlopen?
02:15:34coobanyone know what it means when 'rc   :  0000000' flases by in 3 columns really quickly (too quick to see) after the kernel loads the ramdisk?
05:33:14anderseecoob: looks like part of an Oops....
14:40:11psypeteumm
14:40:45psypetei'm not sure why, but i just built an app of mine against uclibc, and it ended up being more than 3 times as large as the same program built for glibc
14:41:30psypetegcc uclibc toolchain 2.95.4 (20010319)
14:42:55sjhillpsypete: did ya' strip it?
14:43:57psypetesjhill: i just did, and it's the same size stripped as glibc binary is unstripped
14:44:18psypetestrange, eh?
14:44:48psypetedoes it automatically get compiled statically or something?
14:44:59sjhillcheck your options
14:45:10psypetegcc -o generic_uclibc generic.c
14:45:36sjhillyou might try using the latest 3.2 toolchain
14:47:12psypetewon't that just be bigger?
14:47:49psypeteit should be smaller, though, even with an older toolchain... uclibc apps are _not_ supposed to be bigger than glibc...
14:49:56sjhilli've had it happen very rarely
14:50:09sjhillunfortunately we don't have a lot of time to track that stuff down...knock yourself out
14:51:33psypetek
14:51:40psypetebecomes dissapointed
14:54:50psypeteok, this is weird
14:55:12psypetei also just tried to compile udhcp with plain gcc and the toolchain, and the toolchain was bigger
14:55:21psypetemaybe this isn't an app-specific problem
15:00:22psypetebusybox is the same size with glibc and uclibc
15:01:07psypeteare there any headers that *add* bloat to a program i should try to avoid or something?
17:11:01oyypsypete: what is the platform you are building for?
17:11:21oyyis that a MMU-full or MMU-less one?
17:37:59psypeteoyy: MMU-full. going for 486DX2 and higher
17:38:52psypetei'll still use uclibc as the library itself is so much smaller than glibc, but i'd like for the binary to at least be around the same size as a glibc app :\
18:31:25psypeteis away: pooh
21:03:47psypeteis back (gone 02:32:19)
21:42:15psypeteis away: sleep time nite
22:19:03anderseesjhill: thanks for getting that dlopen bug fixed
22:26:16sjhillandersee: my pleasure

Generated by irclog2html.pl by Jeff Waugh - find it at freshmeat.net! Modified by Tim Riker to work with infobot logs, split per channel and by date, etc.