00:01.28 | bug1 | waldi: i cant find it on my local machine, but it is in http://people.debian.org/~bug1/busybox/http://people.debian.org/~bug1/busybox/busybox-cvs_20031119-1.diff.gz i cant remember the status of it |
00:01.57 | bug1 | im not sure if it was the most upto date version i had, or what problems it had |
00:02.25 | andersee | waldi: I see you mentioned you plan to update busybox |
00:02.49 | andersee | waldi: be sure to include the mount fix I checked in after the -pre4 |
00:03.01 | andersee | s/-pre4/-pre4 release/ |
00:16.35 | *** join/#uclibc bug1_ (~bug1@m042-122.nv.iinet.net.au) |
00:16.54 | *** join/#uclibc bug1__ (~bug1@m042-122.nv.iinet.net.au) |
01:06.12 | *** join/#uclibc simmo (~simmo@adsl-18-27.swiftdsl.com.au) |
01:16.13 | *** join/#uclibc yashi (~yashi@dns1.atmark-techno.com) |
01:57.26 | ambassador | eek! |
01:57.37 | mjn3 | mouse? |
01:57.39 | ambassador | I just got back to see how my buildroot build went |
01:57.54 | ambassador | and it's stopped |
01:58.05 | ambassador | trying to get uClibc-local-030818.tgz |
01:58.12 | ambassador | and it can't get it! |
01:58.15 | ambassador | (horror) |
01:58.38 | ambassador | seems it doesn't keep a copy on sources/dl/ |
02:00.36 | mjn3 | ambassador: i'll look at it in a minute |
02:01.31 | ambassador | ok, thx |
02:05.37 | mjn3 | the locale data download is handled in the uClibc/Makefile at the moment. buildroot's uclibc.mk needs to be fixed |
02:06.08 | mjn3 | i just tried and it downloads fine for me |
02:06.38 | ambassador | I can't ping www.uclibc.org |
02:06.43 | ambassador | 63.223.66.155 |
02:06.57 | andersee | ambassador: there is a problem at the colocation facility |
02:07.06 | ambassador | oh, bummer |
02:07.23 | andersee | ambassador: I've exchanged a few emails with them trying to get things fixes |
02:09.02 | ambassador | :-( not a single copy on my system |
02:09.08 | ambassador | talk about weakest link |
02:11.37 | mjn3 | ambassador: what arch are you building for? |
02:12.23 | ambassador | mipsel-linux |
02:12.40 | mjn3 | i don't know if the pregenerated locale data will work for that |
02:12.53 | ambassador | I don't even know if I need it |
02:13.05 | ambassador | just something wanted iconv and I think that came along for the ride |
02:13.57 | mjn3 | yes, because currently iconv uses the codeset data from the locale support |
02:15.19 | ambassador | he, I faked it out with a 0 lenght file :-P |
02:15.20 | mjn3 | but the locale data is all stored in a struct, and the offsets are wrong on some archs due to struct padding |
02:15.48 | mjn3 | well, if you need locale support, the uclibc build will fail |
02:16.20 | andersee | ambassador: go to http://cvs.uclibc.org/ |
02:16.53 | andersee | ambassador: you can fetch things from there |
02:16.54 | mjn3 | in the next couple of weeks i'll be fixing some iconv and locale issues |
02:17.48 | mjn3 | andersee has already asked me for iconv support without requiring locales |
02:17.58 | mjn3 | :-) |
02:18.13 | ambassador | andersee, cool, that works, thank you |
02:18.40 | ambassador | mjn3, :-) sounds like exactly what I need :-) |
02:20.18 | ambassador | frightening experience |
02:20.42 | mjn3 | well, as i said, i'm not sure it will work for you |
02:20.53 | ambassador | it's been working |
02:20.53 | mjn3 | i know it doesn't work on arm |
02:21.02 | ambassador | I mean... I don't need locales |
02:21.03 | ambassador | I just need iconv |
02:21.11 | mjn3 | and iconv is working? |
02:21.13 | ambassador | and I don't even know why - just some app needed it to build |
02:21.18 | ambassador | maybe samba? |
02:21.41 | ambassador | anyway I enabled it, the app built, and worked to my satisfaction |
02:21.57 | mjn3 | ok |
02:23.13 | mjn3 | shame to add 250+K for that when you really only need about (maybe) 10K of that data for iconv |
02:23.55 | ambassador | yeah, but I don't know better :-P |
02:30.39 | mjn3 | andersee: busybox (from a couple of days ago)... tar xzf perl-5.8.2.tar.gz gives "tar: crc error" |
02:31.20 | mjn3 | andersee: never mind. let me look at it some more. could also be a bug in the new stdio code |
02:32.51 | ambassador | so when are you two going to co-author a uClibc book? |
02:34.03 | mjn3 | maybe after we finish with the actual library |
02:34.12 | mjn3 | we still need a new howto |
02:47.23 | ambassador | finish? :-) |
02:50.19 | mjn3 | well, at least get to 1.0 |
02:52.45 | mjn3 | andersee: hmm.. seems to be a stdio bug, so never mind |
03:03.10 | *** join/#uclibc randey (~randey@202.63.116.98) |
04:00.46 | andersee | mjn3: k |
04:01.46 | mjn3 | i'm trying a buildroot build with the new code now |
04:03.07 | andersee | mjn3: lots of abi changes in the new stdio code, right? |
04:03.49 | ambassador | when did new stdio code happen? |
04:03.53 | mjn3 | on the phone |
04:21.26 | andersee | ambassador: it is in progress... not checked in yet |
04:23.45 | ambassador | ah |
04:58.51 | *** join/#uclibc bug1_ (~bug1@m045-207.nv.iinet.net.au) |
05:27.06 | mjn3 | off the phone |
05:27.29 | mjn3 | ambassador: the new stdio core is something i've been working on for a while |
05:27.59 | ambassador | what does it do? :-) |
05:28.09 | mjn3 | andersee: well, the public api is pretty much the same. there are some internal changes though. the main app-visible thing is that the layout of the FILE struct is different |
05:28.36 | mjn3 | ambassador: performance enhancements primarily |
05:28.53 | mjn3 | especially for non-threaded apps using a thread-enabled uClibc |
05:29.15 | ambassador | oh, that sounds nice |
05:29.21 | ambassador | I shun threads |
05:33.39 | mjn3 | andersee: i'm amazed. the new code actually passed the python and perl self-tests |
05:33.49 | andersee | :-) |
05:33.57 | andersee | thats pretty cool |
05:33.59 | andersee | first try? |
05:34.22 | mjn3 | yep. but i know there are a couple of things i need to look at |
05:34.28 | andersee | mjn3: hmm |
05:34.56 | andersee | mjn3: the file struct is opaque |
05:34.56 | mjn3 | still, it's nice that it is passing those plus the glibc tests |
05:35.25 | mjn3 | well... not quite. it has to be visible for the getc/putc macros |
05:38.15 | mjn3 | those were breaking before when people changed wchar support. i fixed things so that doesn't happen anymore |
05:44.26 | andersee | k |
05:45.22 | mjn3 | i could make it completely opaque and do casting in the macros, but that has type safety implications. alternately, i could use inline functions i suppose |
05:47.00 | mjn3 | not seeing that tar problem anymore. things just seem to be working :-) |
05:47.25 | mjn3 | can't do any real speed tests at the moment because if got all the debugging and FILE validation stuff turned on |
05:47.25 | andersee | :-) |
06:08.31 | andersee | uclibc.org / busybox.net is back |
06:09.11 | andersee | the only evidence of anything happening in the log is this |
06:09.14 | andersee | Dec 13 05:46:06 torment kernel: e1000: eth0 NIC Link is Down |
06:11.01 | mjn3 | sigh |
16:23.30 | *** join/#uclibc TheMasterMind1 (~aman@h-69-3-1-37.MCLNVA23.dynamic.covad.net) |
18:47.09 | *** join/#uclibc waldi (~waldi@bblank.thinkmo.de) [NETSPLIT VICTIM] |
21:39.03 | *** join/#uclibc ShortWave (~root@ip68-100-100-183.dc.dc.cox.net) |
21:39.10 | ShortWave | Hey all |
21:40.59 | ShortWave | Got an issue |
21:41.08 | ShortWave | the root_fs I downloaded is missing libcrypt.so |
21:41.12 | ShortWave | Anyone know where I migth find it? |
21:43.26 | ambassador | $ ls build_mipsel_nofpu/root/lib/libcrypt* |
21:43.26 | ambassador | build_mipsel_nofpu/root/lib/libcrypt-0.9.23.so build_mipsel_nofpu/root/lib/libcrypt.so.0 |
21:43.38 | ambassador | so buildroot adds it |
21:43.43 | ambassador | have you tried buildroot? |
21:43.48 | ShortWave | What package provides? |
21:43.53 | ShortWave | buildroot fails on my setup for some reason |
21:44.12 | ambassador | build_mipsel_nofpu/uClibc/libcrypt/libcrypt-0.9.23.so |
21:44.34 | ShortWave | Hrm |
21:44.57 | ShortWave | DO you know what package provides that lib so I can download and compile it? |
21:44.58 | mjn3 | should be there.. in /lib |
21:45.11 | mjn3 | libcrypts.so is provided by uClibc |
21:45.43 | ShortWave | libcrypt? there's a libcrypto |
21:45.52 | ShortWave | can I flood without getting killed? |
21:46.06 | ambassador | flood private? |
21:46.10 | ShortWave | sure |
21:47.03 | ShortWave | libcrypt.so -> libcrypt.so.0, which isn't present |
21:47.03 | ambassador | looks like who build the root didn't copy over the .so |
21:47.12 | ambassador | (but it's not libcrypto) |
21:47.26 | ShortWave | hrm |
21:47.38 | ShortWave | ah, that was /usr/lib |
21:47.46 | ShortWave | in /lib I've got libcrypt-0.9.23 |
21:48.24 | ShortWave | hrm |
21:48.28 | ShortWave | Ok, here's the issue |
21:49.11 | ShortWave | Compiling X from source, the compile fails on undefined reference to 'crypt' |
21:50.55 | mjn3 | add -lcrypt to the link, since crypt is in /lib/libcrypt.so (or /usr/lib/libcrypt.a) |
21:51.04 | ShortWave | I was about to say... |