00:10.28 | JockeHome | psm: pong |
00:12.57 | psm | what do you think, should I try to hide malloc/calloc/realloc/free/memalign? |
00:15.21 | JockeHome | psm: nope, these need to be preemtive |
00:15.55 | psm | all of them ? |
00:15.57 | CIA-10 | 03vapier * r13272 10uClibc/test/misc/stdarg.c: simple stdarg test |
00:16.19 | CIA-10 | 03vapier * r13273 10uClibc/libc/inet/netlinkaccess.h: make sure linux/types.h doesnt screw us up |
00:16.26 | JockeHome | hmm, not sure. maybe not memalign |
00:16.33 | CIA-10 | 03vapier * r13274 10uClibc/libc/misc/sysvipc/ipc.h: fix sem funcs on 64bit arches |
00:16.52 | JockeHome | do the same as glibc |
00:17.07 | CIA-10 | 03vapier * r13275 10uClibc/libc/sysdeps/linux/common/ (fstat.c fstat64.c lstat.c lstat64.c stat.c stat64.c): dont include bits/kernel_stat.h as xstatconv.h already does and use sys/stat.h rather than bits/stat.h |
00:17.47 | JockeHome | there are a few packages that whant to use there own versions of malloc and friends, mostly debugging stuff |
00:20.43 | CIA-10 | 03vapier * r13276 10uClibc/ldso/include/dl-syscall.h: use regular headers rather than trying to get tricky |
00:28.24 | *** join/#uclibc andersee (n=andersee@codepoet.org) |
00:32.29 | SpanKY | make it an option JockeHome ? |
00:32.56 | SpanKY | yes the reason the mem related functions are exported is so debuggers can hook into em (ala dmalloc and such), but how often does that situation come up on uClibc |
00:33.34 | andersee | SpanKY: howdy |
00:34.19 | andersee | its good to have all my projects pretty well current |
00:34.32 | SpanKY | :) |
00:34.33 | andersee | and finally have a bit of time again |
00:35.08 | andersee | SpanKY: so whats been happening while I've been hiding? |
00:35.38 | SpanKY | some random guy has been getting nios/nios2 stuff working in uClibc/buildroot |
00:35.42 | SpanKY | and elf2flt integration |
00:36.09 | andersee | thats nice -- it'd be nice if all the other broken arches were made current as well |
00:36.21 | SpanKY | i bought a bfin board so i should be able to test that :) |
00:36.43 | andersee | :-) |
00:36.44 | andersee | cool |
00:36.46 | SpanKY | x86_64/uclibc is pretty far along (if you ignore threads) |
00:37.23 | SpanKY | psm has gotten relocations down a ton so LD_DEBUG=all isnt nearly as hard to parse |
00:37.27 | andersee | SpanKY: excellent. I just got a dual core opteron box... |
00:38.03 | andersee | SpanKY: cool re relocs -- LD_DEBUG=all used to print 100MB of junk.... |
00:38.22 | SpanKY | yeah, the # of relocs at first start is less than 60 iirc |
00:38.24 | SpanKY | maybe less |
00:38.30 | SpanKY | err much less heh |
00:40.20 | andersee | excellent |
00:41.01 | andersee | I've gotten a few reports in my inbox of problems with buildroot -- things failing to boot or init failing or some such |
00:41.25 | andersee | seen anything like that with the latest and greatest? |
00:41.45 | andersee | guess I should use this shiny new opteron box to do some builds and find out |
00:41.52 | andersee | :) |
00:42.03 | SpanKY | ive never used buildroot to generate a bootable image |
00:42.04 | *** join/#uclibc [g2] (n=Tom@nslu2-linux/g2) |
00:42.12 | [g2] | SpanKY hey |
00:42.14 | SpanKY | all the /dev and target stuff i have yet to touch and i probably wont |
00:42.25 | andersee | [g2]: evening |
00:42.31 | [g2] | andersee hey! |
00:42.51 | SpanKY | andersee: to be honest, ive gotten portage to the point where i use buildroot to verify things in Gentoo, not so much with actually using buildroot anymore :) |
00:43.09 | andersee | SpanKY: :-) |
00:43.13 | SpanKY | ive managed to use just Gentoo to bootstrap uClibc envs ... done it with sh4/ppc/arm so far |
00:43.37 | [g2] | SpanKY all the patchset for the kernel are in the nslu SF repo under kernel |
00:43.47 | andersee | SpanKY: I've been using my gentoo/uclibc box quite a bit this past month. |
00:43.50 | [g2] | we're running 2.6.15 and the latest Intel drivers |
00:44.11 | [g2] | and both LE and BE kernels runs and build |
00:44.35 | andersee | SpanKY: I probably need to update things, but havn't yet bothered to learn how to do anything gentoo specific beyond 'sudo emerge foo' |
00:45.13 | SpanKY | andersee: it does take a bit of work and knowning how emerge operates |
00:45.22 | SpanKY | which is what i work on fixing |
00:45.31 | SpanKY | [g2]: good news man |
00:45.40 | andersee | SpanKY: how would you estimate the latest uclibc is holding up then? |
00:45.42 | [g2] | excellent we're there |
00:45.57 | SpanKY | andersee: what do you mean ? |
00:46.08 | [g2] | jbowler builds and runs Thumb/non-thumb uclibc too |
00:46.55 | andersee | SpanKY: i.e. TODO items needed before cutting another release. I know I need to do some work on pthreads (the older stuff), to ensure mutexes are unlocked when a thread gets canceled. |
00:47.01 | [g2] | SpanKY do you know fx3|chilla ? He's running gentoo on the nslu2 |
00:47.13 | SpanKY | [g2]: nope |
00:47.26 | SpanKY | andersee: thats what i was worried you were going to ask :) |
00:47.29 | andersee | [g2]: cool |
00:47.40 | andersee | [g2]: I should order an nslu2 to play with |
00:47.49 | SpanKY | nah, order one of those boxes [g2] makes |
00:47.54 | [g2] | :) |
00:48.03 | andersee | [g2]: what boxes are those? url? |
00:48.05 | SpanKY | before we cut a release, i think we need to get the hidden_proto() stuff psm has been toying with in |
00:48.21 | [g2] | http://www.giantshoulderinc.com |
00:48.21 | SpanKY | andersee: i also committed the updated pthread forwarding stuff so that needs testing |
00:48.26 | [g2] | I'll be updating the web _very_ soon |
00:48.37 | SpanKY | that should solve the 'binary foo uses libbar.so which is linked against libpthread.so' issue |
00:49.02 | [g2] | the dev boxes are 533Mhz IXP425 64MB 3 MiniPCI, 2 serial (1 external/jTAG) USB 2.0, POE ..... |
00:49.06 | [g2] | $349 |
00:49.11 | [g2] | board, case and ps |
00:49.30 | [g2] | runs debian, gentoo, uclibc thumb BE LE ... |
00:49.35 | SpanKY | get the cases in yet ? :) |
00:49.45 | [g2] | yeah I did get a couple |
00:49.49 | andersee | [g2]: you badly need pics |
00:49.51 | SpanKY | my netwinders just cant hack it anymore when it comes to building stages |
00:50.02 | [g2] | http://www.giantshoulderinc.com/ab3/case.jpg |
00:50.07 | [g2] | http://www.giantshoulderinc.com/ab3/hw-4533 |
00:50.31 | [g2] | SpanKY I'm ready to ship you one |
00:50.34 | andersee | SpanKY: indeed. My netwinders (both of which I moded up to the max possible 256MB) are both boxed up now... |
00:50.50 | SpanKY | [g2]: want me to paypal you the cash or ... ? |
00:51.08 | [g2] | SpanKY can you send a check ? |
00:51.12 | SpanKY | sure |
00:51.13 | andersee | [g2]: http://www.giantshoulderinc.com/ab3/hw-4533 -> 404 |
00:51.15 | SpanKY | e-mail your addy |
00:51.24 | SpanKY | i'll send it out monday at the latest |
00:51.34 | SpanKY | [g2]: make sure you include your business cards btw :) |
00:52.00 | SpanKY | andersee: ignoring GLX issues, modular X seems pretty easy to build/use in uClibc btw |
00:52.01 | [g2] | http://www.giantshoulderinc.com/hw-4533 |
00:52.04 | [g2] | sorry |
00:52.16 | SpanKY | i got modular X running in uClibc-0.9.28/SuperH |
00:54.21 | andersee | SpanKY: very cool |
00:54.40 | SpanKY | andersee: how about i post a "TODO" list of what i think would make for a good 0.9.29 to the mailing list |
00:54.44 | SpanKY | and others can throw in what they have |
00:55.16 | andersee | SpanKY: please update uClibc/TODO |
00:55.32 | andersee | SpanKY: you can CC that to the list of course, but best to keep it checked in |
00:55.40 | SpanKY | rgr |
00:55.48 | SpanKY | oh and i threw together a very basic hppa port |
00:55.52 | [g2] | ring ring |
00:55.54 | SpanKY | static binaries and such work |
00:57.57 | andersee | SpanKY: cool. Do they still make hppa boxes? |
00:58.07 | SpanKY | i think so |
00:58.14 | SpanKY | they just made the last cpu last year though :/ |
00:58.20 | SpanKY | it's nasty, wish i could afford one |
00:59.17 | CIA-10 | 03psm * r13277 10uClibc/ (671 files in 70 dirs): hidden_def/hidden_proto: convert all users (I hope) termios split, add some missing headers, other jump relocs removed |
01:00.10 | andersee | [g2]: what are the 3 ribbon cable plugs on the top/right of the hw-4533 board for? |
01:00.15 | SpanKY | speak of the devil ha |
01:00.23 | CIA-10 | 03vapier * r13278 10uClibc/libc/sysdeps/linux/common/__syscall_fcntl.c: 64bit arches dont need no stinkin fcntl64 |
01:00.26 | andersee | [g2]: next to the serial cable |
01:01.34 | [g2] | andersee the 10/14/10 pin headers |
01:01.50 | andersee | yep |
01:02.02 | [g2] | COM2/JTAG/GPIOs |
01:02.08 | andersee | k |
01:02.31 | [g2] | it'll actually be COM2/JTAG/GPIO I2C on the next batch |
01:02.36 | psm | SpanKY: I am done, I can cleanup some stuff (unneeded prototypes and the like) |
01:02.36 | [g2] | 10/14/14 |
01:02.43 | psm | andersee: welcome back |
01:02.49 | SpanKY | psm: good, cause i got warnings/errors for you :) |
01:02.54 | SpanKY | want me to rafb.net/paste it ? |
01:02.57 | andersee | so I could hook up the GPIOs to some optical isolators and use then to turn on the house lights or some such |
01:03.01 | andersee | psm: thx |
01:03.30 | [g2] | andersee those are 2 USB 2.0 ports next to the db-9 |
01:03.35 | [g2] | high speed |
01:03.48 | psm | andersee: could you check my mails to ml about threads/cancellable syscalls? that and some other issues I asked there have to be decided before .29 |
01:03.52 | [g2] | and the board is all industrial temp rated |
01:03.52 | SpanKY | http://rafb.net/paste/results/xwUsb347.html |
01:03.58 | andersee | does the nslu2 have usb 2 hight speed? |
01:04.03 | andersee | psm: will do |
01:04.10 | [g2] | andersee yeah |
01:04.10 | andersee | psm: I badly need to catch up |
01:04.31 | [g2] | it's about 25% faster to disk than the slug running at 266 |
01:04.47 | [g2] | as it has the 533 MHz cpu |
01:04.53 | andersee | psm: so yes, I'll review your mails and try to answer |
01:05.08 | [g2] | I haven't tried to optimize the I/O |
01:05.54 | [g2] | bonnie++ measures 11MB disk I/O to external hd |
01:06.44 | CIA-10 | 03psm * r13279 10uClibc/libc/stdlib/stdlib.c: Correct atoi() |
01:08.56 | CIA-10 | 03psm * r13280 10uClibc/libc/stdlib/stdlib.c: Correct libc_hidden_proto |
01:10.36 | CIA-10 | 03psm * r13281 10uClibc/libc/string/x86_64/strpbrk.S: s/libc_hidden_proto/libc_hidden_def/ |
01:11.02 | psm | SpanKY: what compiler was it 3.4.5 didnt failed? |
01:11.21 | SpanKY | x86_64-gentoo-linux-uclibc-gcc (GCC) 3.4.5 (Gentoo 3.4.5, ssp-3.4.5-1.0, pie-8.7.9) |
01:12.53 | SpanKY | oh, x86_64/clone.S just needs features.h i think |
01:14.50 | psm | I have thought that syscall.h gets it, but errno is blocked |
01:15.22 | CIA-10 | 03psm * r13282 10uClibc/libc/sysdeps/linux/x86_64/clone.S: needs features.h |
01:15.34 | SpanKY | shouldnt it be HIDDEN_JUMPTARGET(exit) ? |
01:15.54 | SpanKY | otherwise it'll fail to link because it's looking for __GI__exit instead of __GI_exit |
01:16.09 | psm | it was originally _exit, that changed to _exit_internal, now back |
01:16.30 | psm | earlier exit changed to __exit |
01:16.32 | SpanKY | http://rafb.net/paste/results/r6Ztjg10.html |
01:18.05 | psm | could you look into libc/misc/statfs, I had to disabled the hidden usage in *vfs64.c files, due to this, if you enable the commented lines you should get the same |
01:20.34 | CIA-10 | 03psm * r13283 10uClibc/libc/sysdeps/linux/common/pread_write.c: s/__read/read/ and write in fake part |
01:20.53 | psm | the warning about mbstate is old |
01:27.00 | psm | if you set hidden_def/proto to empty, you'll get (almost) .28 similar behaviour (no duplicates x and __GI_x), if you dont like __GI_, we can change it to __ |
01:27.33 | SpanKY | psm: `scanelf -qRs __GI_exit libc` -> __GI_exit libc/stdlib/exit.os |
01:27.49 | SpanKY | $ scanelf -qRs __GI__exit libc |
01:27.49 | SpanKY | __GI__exit libc/stdio/popen.os |
01:27.49 | SpanKY | __GI__exit libc/sysdeps/linux/common/_exit.os |
01:27.49 | SpanKY | __GI__exit libc/sysdeps/linux/x86_64/clone.os |
01:28.32 | psm | there are 2 exits, exit() and _exit() |
01:28.54 | psm | I havent changed that behaviour (at least not on purpose) |
01:29.41 | SpanKY | well _exit() isnt defined anywhere |
01:29.53 | SpanKY | $ readelf -s libc/stdlib/exit.os | grep GI.*exit |
01:29.53 | SpanKY | <PROTECTED> |
01:29.53 | SpanKY | <PROTECTED> |
01:30.22 | SpanKY | $ readelf -s libc/sysdeps/linux/common/_exit.os | grep GI |
01:30.22 | SpanKY | <PROTECTED> |
01:30.38 | SpanKY | really need to update scanelf to print defined/undefined info heh |
01:32.32 | psm | common/_exit.c |
01:33.15 | SpanKY | hmm, maybe related to the INLINE_SYSCALL stuff |
01:35.03 | SpanKY | no, that doesnt seem to be it ... |
01:35.08 | SpanKY | my _exit.os isnt defining any symbols |
01:36.23 | psm | we should either use correctly INLINE_SYSCALL or dont at all |
01:36.33 | psm | maybe the loop again? |
01:36.42 | psm | while (1) ? |
01:36.44 | SpanKY | it doesnt matter, it still doesnt work |
01:37.30 | SpanKY | http://rafb.net/paste/results/xnRXbL49.html |
01:37.34 | SpanKY | that doesnt work |
01:38.44 | psm | without while as well? |
01:39.16 | SpanKY | correct |
01:42.31 | SpanKY | when i comment out libc_hidden_def(_exit), i get the symbol |
01:43.30 | psm | that should be equiv w/ strong_alias(_exit,__GI_exit) |
01:43.48 | SpanKY | you mean __GI__exit ;) |
01:43.52 | psm | sorry: strong_alias(_exit,__GI__exit) |
01:44.12 | SpanKY | and you mean hidden_strong_alias |
01:44.25 | SpanKY | using hidden_strong_alias(_exit,__GI__exit) works though :/ |
01:44.36 | SpanKY | $ readelf -s _exit.os | grep FUNC |
01:44.36 | SpanKY | <PROTECTED> |
01:44.37 | SpanKY | <PROTECTED> |
01:44.45 | SpanKY | -libc_hidden_def(_exit) |
01:44.45 | SpanKY | +hidden_strong_alias(_exit,__GI__exit) |
01:51.32 | CIA-10 | 03psm * r13284 10uClibc/libc/sysdeps/linux/ (mips/pipe.S mips/pipe.c sh/pipe.c): Add hidden versions of pipe() |
01:52.38 | SpanKY | hrm, looks like ftruncate.c is doing the same thing for me :/ |
01:54.43 | psm | are there more that you miss from libc.so? |
01:55.43 | SpanKY | yeah a bunch, but they're all the same issue |
01:55.50 | SpanKY | i'll try i386 to see if it's the same |
01:57.48 | CIA-10 | 03vapier * r13285 10uClibc/libc/sysdeps/linux/ (5 files in 5 dirs): need features.h for new HIDDEN_JUMPTARGET() |
02:00.25 | SpanKY | no, i386 is bad too :/ |
02:08.42 | CIA-10 | 03psm * r13286 10uClibc/libc/sysdeps/linux/common/setregid.c: missed to define internal setregid |
02:23.53 | CIA-10 | 03psm * r13287 10uClibc/libc/sysdeps/linux/common/seteuid.c: another missed hidden |
02:25.31 | khem | I get can't resolve symbol '__udivsi3' |
02:25.50 | khem | the symbo exist in /lib/libgcc_s.so on my rootfs |
02:26.18 | khem | somehow it is not finding this shared library it seems |
02:26.27 | khem | Has someone seen this before |
02:27.15 | psm | softfloat? |
02:27.40 | SpanKY | khem: is your stuff linked against libgcc_s.so.1 ? |
02:28.03 | khem | it is softfloat |
02:31.03 | khem | SpanKY: it seems it is linked with -lm thats what I see on commandline |
02:31.21 | SpanKY | readelf -d <booga> |
02:31.31 | SpanKY | psm: a ton of files are broken on my system :/ |
02:31.36 | psm | SpanKY: cache /etc/ld.so.cache is corrupt, could this be some commit from you? |
02:31.50 | SpanKY | maybe |
02:32.13 | SpanKY | cant test if i cant build :) |
02:32.32 | khem | SpanKY: it just shows libm.so.0 and libc.so.0 |
02:32.42 | khem | how come it got linked at first |
02:32.56 | khem | statically, it should have barfed |
02:33.12 | psm | for you only: set hidden_def to strong_alias(x,__GI_x) |
02:33.27 | SpanKY | khem: try LD_PRELOAD=libgcc_s.so <booga> |
02:34.09 | psm | I have 4 *truncate* defined |
02:34.18 | psm | also _exit |
02:34.50 | khem | SpanKY: that works |
02:35.12 | khem | how can I add it to DT_NEEDED in my binary |
02:35.18 | SpanKY | rebuild it ? ;) |
02:35.33 | SpanKY | psm: you have your .config somewhere ? |
02:38.44 | khem | SpanKY: It still does not link with libgcc |
02:39.14 | psm | I paste in a mom |
02:39.32 | psm | well, ld.so is faulty |
02:40.25 | psm | the one from 20060109 worked |
02:41.00 | SpanKY | http://www.uclibc.org/cgi-bin/viewcvs.cgi/trunk/uClibc/ldso/include/dl-syscall.h?rev=13276&r1=12489&r2=13276 |
02:41.06 | SpanKY | try reverting that |
02:41.56 | psm | SpanKY: my config http://rafb.net/paste/results/3mESSN46.html |
02:43.05 | psm | any easy way to revert w/ svn ? |
02:43.23 | SpanKY | svn up -r 13275 ldso/include/dl-syscall.h |
02:45.55 | SpanKY | hmm your config worked |
02:46.33 | psm | what is different? |
02:46.56 | SpanKY | lemme rebuild it one option at a time ;) |
02:50.00 | psm | it was that commit to dl-syscall.h |
02:50.09 | SpanKY | k, so revert it ;) |
02:54.03 | SpanKY | psm: ok, first enable ipv6/netlink |
02:54.14 | SpanKY | that'll cause like two failures |
02:54.52 | SpanKY | http://rafb.net/paste/results/0OfJwu71.html |
02:57.51 | CIA-10 | 03psm * r13288 10uClibc/libc/inet/ntop.c: Correct 2 issues if ipv6 is enabled found by SpanKY, thx |
02:59.18 | psm | others? |
03:01.19 | CIA-10 | 03psm * r13289 10uClibc/ldso/include/dl-syscall.h: Revert 13276, cache corrupted on x86 |
03:05.29 | SpanKY | disable these options: |
03:05.35 | SpanKY | UCLIBC_HAS_FOPEN_EXCLUSIVE_MODE UCLIBC_HAS_GLIBC_CUSTOM_STREAMS UCLIBC_HAS_PRINTF_M_SPEC |
03:06.06 | SpanKY | http://rafb.net/paste/results/ORN3wJ66.html |
03:10.38 | CIA-10 | 03psm * r13290 10uClibc/libc/stdio/ (_WRITE.c fclose.c): other failures w/ different options |
03:10.56 | psm | done |
03:11.01 | SpanKY | i got more, hold ;) |
03:16.23 | SpanKY | stime.c:(.text+0x31): undefined reference to `__settimeofday' |
03:16.53 | SpanKY | same config, but x86_64 target |
03:17.48 | CIA-10 | 03psm * r13291 10uClibc/libc/stdlib/arc4random.c: adapt to new hidden |
03:20.06 | CIA-10 | 03psm * r13292 10uClibc/libc/sysdeps/linux/common/stime.c: another from SpanKY |
03:26.17 | SpanKY | psm: ok, my problems stem from debugging it seems |
03:26.27 | SpanKY | enable DODEBUG |
03:26.49 | SpanKY | i was debugging x86_64/uClibc which is why i had it on |
03:27.55 | CIA-10 | 03psm * r13293 10uClibc/libc/misc/intl/intl.c: s/__strcmp/strcmp/ |
03:29.22 | SpanKY | yeah, with DODEBUG=n, everything is peachy |
03:29.31 | SpanKY | with DODEBUG=y, all hell breaks loose ;) |
03:29.48 | psm | do you have the brk redefine? I dont see where it comes from |
03:32.38 | SpanKY | with i386, everything works now if i turn off DODEBUG |
03:33.23 | psm | does malloc redefine brk if debug enabled? |
03:33.26 | SpanKY | x86_64 though hates lockf64() |
03:33.28 | SpanKY | <PROTECTED> |
03:33.28 | SpanKY | ../../../libc/misc/file/lockf64.c:53: warning: asm declaration conflicts with previous rename |
03:34.51 | psm | would it work if you change the line to fcntl64 ? |
03:36.32 | SpanKY | http://rafb.net/paste/results/iXvxXe97.html |
03:36.34 | SpanKY | that works for me |
03:37.21 | psm | readelf -s .os ? does it have non __GI_fcntl ? |
03:37.32 | SpanKY | before or after my patch |
03:37.37 | psm | after |
03:37.50 | SpanKY | http://rafb.net/paste/results/jQ0Jx561.html |
03:39.03 | psm | the same as in libc/misc/statfs, I had to disable the redirection for the *64(), you end up w/ jump reloc for fcntl64 |
03:39.20 | SpanKY | <PROTECTED> |
03:39.20 | SpanKY | +libc_hidden_proto(fcntl) |
03:39.20 | SpanKY | +#else |
03:39.20 | SpanKY | +libc_hidden_proto(fcntl64) |
03:39.20 | SpanKY | <PROTECTED> |
03:39.21 | SpanKY | ? |
03:39.50 | psm | not reversed ? |
03:40.14 | SpanKY | what i posted works for me on i386 and x86_64 |
03:40.35 | SpanKY | http://rafb.net/paste/results/x9NyCv95.html |
03:41.13 | psm | readelf -s? __GI_fcntlx or plain? |
03:41.44 | SpanKY | hrm true |
03:42.38 | SpanKY | doesnt seem to matter, i always end up with fcntl64 |
03:44.12 | psm | would be good to find some solution to this ... |
03:44.30 | SpanKY | indeed |
03:44.42 | SpanKY | i'll leave my local thing for now as it only breaks on 64bit host |
03:44.52 | SpanKY | fixing the DODEBUG is higher priority :) |
03:45.38 | CIA-10 | 03psm * r13294 10uClibc/libc/misc/file/lockf64.c: disable redirect for fcntl64 until a good way is found |
03:52.35 | psm | well, I found a solution, but that means reediting almost all files |
03:52.59 | psm | libc_hidden_proto should be moved before the function is defined |
03:53.32 | psm | what I do not understand, why DODEBUG influences it |
03:57.22 | psm | its too late, I cant think anymore ... |
03:58.11 | psm | maybe only the ones using syscalls? |
03:59.02 | psm | well, bye |
04:23.00 | CIA-10 | 03vapier * r13295 10uClibc/libc/sysdeps/linux/common/longjmp.c: enable _longjmp alias else less fails to build |
04:35.45 | CIA-10 | 03vapier * r13296 10uClibc/libc/sysdeps/linux/m68k/ (crt0.S crt1.S): update to new uClibc startup style |
04:38.55 | CIA-10 | 03vapier * r13297 10uClibc/libc/sysdeps/linux/m68k/bits/ (fcntl.h mman.h stat.h): sync with glibc |
04:41.23 | CIA-10 | 03vapier * r13298 10uClibc/libc/sysdeps/linux/m68k/bits/setjmp.h: sync with glibc |
04:42.40 | CIA-10 | 03vapier * r13299 10uClibc/libc/sysdeps/linux/m68k/crt0.c: old |
04:58.31 | SpanKY | Remove the booke stuff. There is no reason to punish people for |
04:58.31 | SpanKY | <PROTECTED> |
04:58.31 | SpanKY | <PROTECTED> |
05:57.09 | *** join/#uclibc ibot (i=ibot@pdpc/supporter/active/TimRiker/bot/apt) |
05:57.09 | *** topic/#uclibc is uClibc - for discussion of uClibc and Busybox | uClibc 0.9.28 released 17 August 2005 | busybox 1.1.0 released 11 Jan 2006 | for non uClibc/busybox such as buildroot and general setup issues try #elinux or #edev |
06:00.16 | *** join/#uclibc hgb (n=hgb@cm-80.111.151.063.chello.no) |
06:13.56 | CIA-10 | 03vapier * r13300 10uClibc/TODO: update TODO |
06:17.20 | CIA-10 | 03sjhill 07uClibc-nptl * r13301 10/libpthread/nptl/sysdeps/unix/sysv/linux/mips/ (fork-nptl.c fork.c): Rename file. |
06:35.30 | CIA-10 | 03vapier * r13302 10uClibc/libc/sysdeps/linux/m68k/bits/setjmp.h: shouldnt have trimmed the _ASM #if check |
06:35.37 | CIA-10 | 03vapier * r13303 10uClibc/libc/sysdeps/linux/m68k/__longjmp.S: need features.h |
06:40.00 | CIA-10 | 03vapier * r13304 10uClibc/libc/sysdeps/linux/m68k/bits/stat.h: forgot to cut out the other __USE_MISC section |
06:42.53 | CIA-10 | 03vapier * r13305 10uClibc/libc/sysdeps/linux/m68k/ (Makefile.arch __syscall_error.c clone.S): use __syscall_error trick |
06:50.35 | CIA-10 | 03vapier * r13306 10uClibc/test/misc/seek.c: use regular fseeko() |
06:54.33 | *** join/#uclibc landley (n=landley@adsl-71-145-190-167.dsl.austtx.sbcglobal.net) |
06:54.41 | landley | Hey Erik! |
06:55.32 | landley | Please put the busybox 1.1.0 stuff in my directory on busybox.net up in the downloads bit. |
06:57.03 | landley | (Or give the downloads directory a group access that I can put stuff in. I can come up with my own key to sign things with...) |
06:59.10 | CIA-10 | 03vapier * r13307 10uClibc/libc/sysdeps/linux/m68k/__syscall_error.c: make sure we set errno properly |
07:04.45 | andersee | landley: evening |
07:06.43 | landley | Hiya! |
07:06.49 | landley | Please put the busybox tarball up. |
07:07.06 | landley | It's not but-free, but it needs to ship. |
07:07.19 | landley | (Release early, release often. There will be a 1.1.1.) |
07:07.43 | landley | And if 1.1.0 isn't up on the website by the time 1.1.1 is ready, I'm going to send you a cake. |
07:07.58 | landley | (It's the strangest threat in the history of programming, I know...) |
07:08.33 | landley | That would be "bug-free" if I could spell. |
07:08.52 | landley | (Moving to pittsburgh. Brain fried. The boxes! the boxes!) |
07:09.00 | landley | I have waaaaay too many books. |
07:09.56 | andersee | moving again? |
07:10.01 | landley | Still pittsburgh. |
07:10.27 | landley | I get on a plane sunday morning, my first day of work is monday. |
07:10.49 | landley | That's part of the reason 1.1.0 had to ship, but I _meant_ to get it out weeks earlier. |
07:10.57 | andersee | work? |
07:10.59 | landley | (Deadlines are, unfortunately, highly motivational to me.) |
07:11.01 | andersee | whats that? |
07:11.01 | landley | Yup. Timesys. |
07:11.06 | landley | They hired me. |
07:11.08 | andersee | hey cool |
07:11.15 | landley | I'll be working about 15 feet from Garrett. |
07:11.22 | andersee | :-) |
07:11.33 | landley | This means FWL is likely to develop c++ support sooner rather than later. :) |
07:11.40 | landley | I don't suppose you know anything about cross-compiling? |
07:12.02 | landley | (I'm trying to replace the "run sed against the source code" method of configuring binutils and gcc with actual ./configure arguments. |
07:12.10 | landley | Unfortunately, there doesn't seem to be any way to specify what I want to. |
07:12.20 | landley | i explained what I was doing to Dan kegel, and his response was, in its' entirety: |
07:12.41 | landley | "... must ... not ... scream ..." |
07:12.47 | landley | That's a direct quote. |
07:13.26 | landley | (In theory, the cross-compiler mechanism can be used to build the Linux From Scratch /tools stage without requiring root access to create a /tools symlink. |
07:13.49 | landley | All I have to do is individually specify different build and runtime library paths. |
07:13.51 | andersee | sure |
07:13.59 | landley | And there is _no_way_ to do this from ./configure, that I can find. |
07:14.05 | andersee | I never do anything dev stuff as root |
07:14.17 | landley | I use User Mode Linux to simulate root access. |
07:14.29 | landley | But I'm trying to move to QEMU instead, which means I can't easily borrow the host system's root filesystem. |
07:14.32 | andersee | uml, qemu, fakeroot, etc |
07:14.35 | landley | Which I do during the /tools build stage. |
07:14.42 | CIA-10 | 03vapier * r13308 10uClibc/ldso/ldso/m68k/dl-startup.h: fixup so we can at least get to the ldso Standalone execution step |
07:14.48 | landley | qemu has no equivalent of UML's hostfs. |
07:14.57 | andersee | nfs? |
07:15.01 | landley | (It has samba mounts, but those are case insensitive and that make sthe build go "boing".) |
07:15.10 | landley | Configuring NFS on the host requires root access. |
07:15.21 | landley | (So far, anyway. Probably a way to do that without it...) |
07:15.40 | landley | But if I could get the darn cross-compiler to do what I want, I wouldn't need it. |
07:16.00 | andersee | you could always use sed.... |
07:16.03 | andersee | :) |
07:16.04 | landley | _And_ I could get the generated /tools to be for Arm or PPC or x86_64 and since I'm using qemu anyway... |
07:16.10 | landley | Yeah, that's what I've been doing. |
07:16.23 | landley | But using sed to specify a different target processor is a whole can of worms that I'm trying to avoid. |
07:16.29 | landley | (See Dan Kegel's head exploding, earlier.) |
07:16.50 | landley | I've been reading crosstool's source. |
07:16.58 | landley | And the Linux From Scratch cross-compiler book. |
07:16.59 | andersee | btw, busybox-1.1.0 is signed and in place |
07:17.03 | landley | And poking a bit at buildroot. |
07:17.08 | landley | Coolness, I'll crank out an announcement. |
07:17.20 | landley | Is there any way the download directory to grow a group access I could be added to? |
07:17.25 | andersee | congrats on the new job |
07:17.26 | landley | Or should I just bug you again next time? :) |
07:17.28 | landley | Thanks. |
07:17.46 | landley | There are a good half-dozen seperate cross compiler references, which all have one thing in common. |
07:17.49 | landley | Wnat to guess what that thing is? |
07:17.52 | andersee | landley: probably. bug me about it and we'll get it set up |
07:17.59 | landley | Cool. |
07:20.52 | landley | Thanks. |
07:21.27 | CIA-10 | 03vapier * r13309 10uClibc/ldso/ldso/m68k/dl-startup.h: add back in _dl_start_user |
07:21.27 | Newsome | landley: can you use this feature of qemu? - http://fabrice.bellard.free.fr/qemu/qemu-doc.html#SEC19 |
07:21.31 | andersee | and thanks for keeping on top of busybox |
07:23.14 | Newsome | looks similar to UML's hostfs |
07:23.37 | andersee | Newsome: yep, though looks to be rather more restrictive |
07:23.50 | Newsome | hmm, true |
07:27.16 | Newsome | could probably use FUSE + sshfs |
07:37.02 | SpanKY | heh, people are alive |
07:37.06 | SpanKY | which JockeHome was alive |
07:37.12 | SpanKY | i need him to fix m68k lazy relocs heh |
07:37.35 | CIA-10 | 03vapier * r13310 10uClibc/ldso/ldso/m68k/ (elfinterp.c resolve.S): m68k ldso works again (well, lazy relocation is broken atm) |
07:38.05 | Newsome | wow, looks like great service |
07:38.28 | Newsome | hrm, still broken I guess |
07:42.17 | CIA-10 | 03vapier * r13311 10uClibc/TODO: m68k is doing much better now |
07:43.05 | SpanKY | be nice if we could unify a lot of the arch elfinterp.c code |
07:43.24 | CIA-10 | 03landley * r13312 10busybox/docs/busybox.net/news.html: 1.1.0 is out. |
07:47.16 | andersee | SpanKY: indeed |
07:48.15 | SpanKY | make it a 0.9.30 goal :) |
07:50.35 | CIA-10 | 03vapier * r13313 10uClibc/ldso/ldso/m68k/elfinterp.c: grab some more fixes from x86_64 |
07:53.28 | SpanKY | andersee: btw, there's a neat "feature" atm where if you build with DODEBUG, libc.so fails with a lot of undefined references |
07:53.46 | SpanKY | psm is gonna fix it up when he gets back |
07:58.06 | CIA-10 | 03vapier * r13314 10uClibc/libc/signal/sigaction.c: dont hide __libc_sigaction |
07:59.04 | andersee | SpanKY: presumably that is what folks had complained about, re buildroot failing |
07:59.14 | andersee | hmm |
07:59.16 | andersee | maybe |
07:59.36 | SpanKY | andersee: well the bug was introduced a few hours ago so i dont think so :) |
07:59.43 | andersee | oh |
07:59.46 | andersee | maybe not |
07:59.47 | andersee | :) |
10:18.14 | CIA-10 | 03vapier * r13315 10uClibc/libc/sysdeps/linux/arm/crt1.S: fix loading of argc/argv for arm/nommu |
11:08.15 | *** join/#uclibc psm (n=mps@host-6.mikroweb.hu) |
13:23.25 | *** join/#uclibc psm (n=mps@host-6.mikroweb.hu) |
13:27.28 | *** join/#uclibc _zakalwe (i=zakalwe@you.sneakymonkey.net) |
13:32.02 | *** join/#uclibc tty56 (n=johannes@2001:6f8:1331:3:f005:17a9:285a:3429) |
13:34.56 | *** join/#uclibc Sgt-Donan (n=Loutre@feu30-1-82-242-59-35.fbx.proxad.net) |
14:25.21 | *** join/#uclibc ico2 (n=ico2@modem-872.chameleon.dialup.pol.co.uk) |
17:01.46 | *** join/#uclibc ico2 (n=ico2@modem-3324.elephant.dialup.pol.co.uk) |
17:01.52 | ico2 | hi |
17:02.06 | ico2 | how do i make busybox mdev put loop devices in /dev/loop/loop* |
17:02.10 | ico2 | wait |
17:02.13 | ico2 | /dev/loop/* |
17:02.29 | ico2 | currently it helpfully creates 255 /dev/loop? devices |
17:06.23 | *** join/#uclibc pinchartl (n=Unknown@32-81.245.81.adsl.skynet.be) |
17:06.24 | pinchartl | hi |
17:07.15 | pinchartl | I'm trying to create an intrd image which includes busybox. I've been able to mount the image at startup, and when the kernel starts linuxrc, I get the message "/bin/sh: Can't open auto". |
17:07.30 | pinchartl | could anyone help me ? |
17:09.18 | ico2 | pinchartl: never seen that error before, sorry |
17:10.22 | pinchartl | ok... I'll try to debug busybox to find out where it comes from |
17:10.51 | pinchartl | btw, does anyone know if there's a way to redirect the console to a network connection, like netconsole does with the kernel messags ? |
17:14.12 | ico2 | how do i make the busybox init handle ctrl-alt-del properly? it has ::ctrlaltdel:/sbin/reboot in its inittab |
17:39.18 | Newsome | ico2: busybox mdev doesn't do subdirectories |
17:41.31 | ico2 | Newsome: damn |
17:41.52 | Newsome | ico2: it just puts everything into /dev |
17:42.05 | ico2 | Newsome: Newsome will the next version be capable of doing this? |
17:44.28 | Newsome | ico2: I'm not really sure. It sounds like the sort of thing that would take up more bytes of code than it would be worth in busybox |
17:45.27 | solar | why would you need subdirs? |
17:45.38 | solar | or rather what needs to be in a subdir? |
17:46.12 | ico2 | solar: loop devices |
17:46.17 | ico2 | solar: i use 255 of them |
17:46.25 | ico2 | a bit messy in /dev |
17:49.38 | solar | thats not really a problem however is it? |
17:49.43 | solar | other than just how it looks |
17:50.18 | solar | when we hashed out the ideas for mdev we wanted it to be really simple and based on what the kernel was dishing out |
17:51.57 | Newsome | first I'd heard of anyone even using mdev yet :) |
17:52.19 | solar | it would indeed add a few extra bytes of you wanted to remap /dev/loop[0-255] to /dev/loop/[0-255] cuz when mdev first runs it processes /sys/block,decives and creates the device nods in /dev/ then later in it goes back and sets all the permissions etc. |
17:52.52 | solar | you can do quite a bit from the conf like shell actions. So if you wanted to say create /dev/loop/<symlinks> you could do that |
17:53.33 | solar | Newsome: likewise. |
17:53.44 | solar | I'm glad somebody is making use of it |
17:55.49 | Newsome | ico2: other than lots of devices in /dev, does it work? |
18:01.05 | JockeHome | SpanKY: http://rafb.net/paste/results/nBgjQi74.html |
18:12.39 | *** join/#uclibc ambroseL (n=bjb@ns1.linuxbutler.ca) |
18:17.43 | SpanKY | it must be my birthday ! |
18:17.44 | SpanKY | :) |
18:18.15 | Newsome | hmm, looking again at mdev, I don't think it would be that difficult to do path remappings from the conf file, but doing the sort of pattern matching that udev does would be quite painful, and it wouldn't automatically work with hotplug remove |
18:19.04 | Newsome | Also, I don't think the conf enables shell actions yet, just UID, GID, and permissions. |
18:25.13 | SpanKY | JockeHome: hmm, no worky, it just segfaults immediately now |
18:25.23 | SpanKY | JockeHome: non-lazy relocation works :) |
18:30.35 | *** join/#uclibc ico2_ (n=ico2@modem-2562.chameleon.dialup.pol.co.uk) |
19:18.54 | CIA-10 | 03sjhill 07uClibc-nptl * r13316 10/ldso/ldso/x86_64/ (. dl-debug.h elfinterp.c resolve.S): Merge from trunk. |
19:19.10 | CIA-10 | 03sjhill 07uClibc-nptl * r13317 10/Changelog.full: Merge from trunk. |
19:22.52 | CIA-10 | 03sjhill 07uClibc-nptl * r13318 10/libc/ (82 files in 4 dirs): Merge from trunk. |
19:24.16 | CIA-10 | 03sjhill 07uClibc-nptl * r13319 10/libc/sysdeps/linux/mips/ (bits/stat.h bits/syscalls.h brk.c pipe.S sigaction.c): Merge from trunk. |
19:32.14 | CIA-10 | 03sjhill 07uClibc-nptl * r13320 10/ (26 files in 16 dirs): (log message trimmed) |
19:32.14 | CIA-10 | Get rid of more glibc NPTL symbol and alias madness. Also disable usage of |
19:32.14 | CIA-10 | 'libc/sysdeps/linux/common/fork.c' when NPTL is being used. Finally, there |
19:45.46 | CIA-10 | 03psm * r13321 10uClibc/libc/ (405 files in 43 dirs): make DODEBUG=y happy, update sysdeps/common/* copyright |
19:45.57 | psm | SpanKY: could you please test ^^ ? |
19:58.40 | CIA-10 | 03psm * r13322 10uClibc/libc/string/_collate.c: DODEBUG=y and XLOCALE as well |
20:02.23 | CIA-10 | 03psm * r13323 10uClibc/libc/stdio/scanf.c: get rid of wcrtomb jump reloc |
20:05.11 | *** join/#uclibc zakalwe (i=zakalwe@you.sneakymonkey.net) |
20:07.27 | CIA-10 | 03psm * r13324 10uClibc/libc/string/_collate.c: get rid of strcmp jump reloc |
20:11.56 | CIA-10 | 03psm * r13325 10uClibc/libc/stdlib/strtod.c: same for iswspace* |
20:16.21 | CIA-10 | 03psm * r13326 10uClibc/libc/stdio/putwchar.c: same for __fputc_unlocked |
20:16.23 | SpanKY | almost psm |
20:16.58 | SpanKY | <PROTECTED> |
20:16.58 | SpanKY | libc/libc_so.a(getaddrinfo.os): In function `gaih_inet': |
20:16.58 | SpanKY | libc/inet/getaddrinfo.c:615: undefined reference to `__GI_in6addr_loopback' |
20:16.58 | SpanKY | make[1]: *** [lib/libc.so] Error 1 |
20:18.54 | CIA-10 | 03vapier * r13327 10uClibc/libc/sysdeps/linux/ (8 files in 8 dirs): fix broken commenting |
20:19.54 | CIA-10 | 03psm * r13328 10uClibc/libc/inet/in6_addr.c: I should also test w/ ipv6, missed that, thx SpanKY |
20:20.11 | JockeHome | SpanKY: http://rafb.net/paste/results/q1bl0794.html |
20:20.31 | JockeHome | I don't think this will work either, but one never knows |
20:20.57 | JockeHome | possibly there is something more that needs to be fixed |
20:27.06 | CIA-10 | 03psm * r13329 10uClibc/libc/misc/file/lockf64.c: get rid of fcntl64 jump reloc |
20:32.16 | *** join/#uclibc ico2__ (n=ico2@modem-1503.cougar.dialup.pol.co.uk) |
21:04.08 | CIA-10 | 03psm * r13330 10uClibc/libc/misc/statfs/ (fstatvfs64.c statvfs64.c): get rid of *stat*64 jump relocs |
21:04.29 | psm | SpanKY: could you check if the *64() changes work for you and are correct ? |
21:12.31 | *** join/#uclibc ulf_k (n=ulf_kypk@p54BDC0A8.dip0.t-ipconnect.de) |
21:19.59 | *** join/#uclibc khem_home (n=khem@adsl-69-236-91-54.dsl.pltn13.pacbell.net) |
21:21.13 | khem_home | hey I am getting this make problem make[1]: *** No rule to make target `|', needed by `../../extra/config/conf.o'. Stop. |
21:22.30 | khem_home | I figured out that this fails with make 3.79 and works ok with make 3.80 but my build machine has 3.79 and it is impossible to upgrade make on that box |
21:23.35 | khem_home | 0.9.28 worked fine but now the makefiles are changed completely |
21:24.17 | psm | I used | for dependencies |
21:24.21 | khem_home | the issue is due to ori'ng the dependencies. |
21:24.28 | psm | install 3.80 as gmake and use that |
21:24.30 | khem_home | psm, guten morgen |
21:24.42 | khem_home | psm, thats what I can not do |
21:24.47 | psm | abend hier |
21:25.40 | khem_home | I could do that but its a production machine so I can not upgrade to make |
21:25.49 | khem_home | psm, any workaround |
21:26.20 | psm | you could try to remove | from all the makefiles, but they are there on purpose |
21:26.23 | khem_home | I can build it locally on my fc4 |
21:27.16 | khem_home | psm, do you see any problem if I remove | |
21:28.06 | khem_home | what is | meant to do here |
21:28.10 | psm | you will loose the dependency chain (that is not bullet proof yet), so some parts will build before others |
21:29.36 | khem_home | hmm, that can cause issues. but I think we should still make it possible to use make 3.79 its widely used on distros still in use |
21:30.00 | psm | target: a | b ($^ matches a, but not b, but b has to be present, before running target) |
21:30.45 | khem_home | a: b.c dep |
21:30.51 | psm | 3.80 was done in 2002, we have already parts that need binutils-2.16.1 |
21:31.03 | khem_home | $CC a.c will do same |
21:32.15 | psm | if you remove | , then you also have to remove what is after it |
21:33.05 | psm | I am not intending to remove those, feel free to propose some solution that works w/ 3.79 |
21:34.00 | khem_home | psm, ok I will what I can do |
21:35.40 | psm | top level Makefile.in's(as of each lib) have to be changed |
21:35.54 | khem_home | psm, yes its a lot |
21:36.01 | khem_home | :( |
21:36.03 | solar | psm: all the redhat based distros seem to be at 3.79 still |
21:36.09 | psm | I know, I wrote it |
21:36.40 | psm | solar: what is the current toolchain state there? glibc/gcc/binutils ? |
21:39.42 | khem_home | psm, rhel3 has gcc 3.2.3 glibc 2.3.2 binutils 2.14.90 |
21:40.10 | psm | and fc4? |
21:40.16 | khem_home | psm, I build using cross toolchain to toolchain is not an issue for me |
21:41.04 | khem_home | fc4 has gcc 4.0.2 binutils 2.15.94 glibc 2.3.5 |
21:41.39 | psm | that is a bad combo for gcc/binutils |
21:42.44 | khem_home | psm, why is that |
21:43.03 | khem_home | I think they have right patches backported |
21:43.14 | khem_home | for binutils to supports gcc4 |
21:43.21 | psm | binutils is too old (and faulty) |
21:44.10 | khem_home | but it has make 3.80 so I had an escape route using host make and cross toolchain |
21:44.22 | khem_home | but unfortunately build system does not |
21:45.12 | ico2__ | Newsome: in response to your earlier q, yes |
21:45.19 | ico2__ | mdev did work |
21:47.11 | solar | psm: RHEL Taroon = binutils-2.14.90.0.4 glibc-2.3.2-95.33 gcc-3.2.3 ; Centos-4.2 make-3.80 binutils-2.15.92.0.2 gcc-3.4.4 glibc-2.3.4-2.13 ; Centos-3.6 = binutils-2.14.90.0.4-39 glibc-2.3.2-95.37 gcc-3.2.3-53 make-3.79.1-17.1 |
21:49.19 | SpanKY | sorry, but not having make-3.80 is pathetic |
21:49.29 | SpanKY | it was released over three years ago |
21:56.41 | khem_home | SpanKY, agreed but production systems dont get upgraded that easily |
21:57.23 | psm | khem_home: are you using buildroot? |
21:57.34 | khem_home | psm, no |
21:57.59 | khem_home | but I do have make as part of packages |
21:58.40 | psm | then build it first (buildroot has a similar solution iirc for sed-4) and use it |
21:58.42 | khem_home | so what I plan is to compile make 3.80 first and then use it instead of what is not host |
21:59.13 | khem_home | psm, yes I think thats what I gonna do |
22:01.02 | khem_home | psm, I think Q option in sed gave me some problem a while ago |
22:01.33 | khem_home | I wrote a small awk script as workaround |
22:01.52 | psm | I know redhat stayed long time w/ sed-3.0.2 iirc |
22:02.06 | khem_home | but I think compiling sed make makes sense |
22:02.21 | khem_home | psm, yes and suse too |
22:08.30 | solar | suse added pax-utils a few days ago :) |
22:11.04 | psm | with that name? |
22:11.10 | solar | yes |
22:11.32 | psm | fine then ;) |
22:12.14 | solar | time for me to call it a day. cya later |
22:17.29 | *** part/#uclibc khem_home (n=khem@adsl-69-236-91-54.dsl.pltn13.pacbell.net) |
22:18.46 | CIA-10 | 03psm * r13331 10uClibc/libc/signal/ (signal.c sysv_signal.c): get rid of duplicates, even bsd_signal is questionable |
22:20.38 | SpanKY | khem: 3 year lag time is no excuse :p |
22:42.26 | *** join/#uclibc ico2 (n=ico2@modem-250.chimpanzee.dialup.pol.co.uk) |
23:19.32 | *** join/#uclibc bug1 (n=me@202-180-70-42.jetbuster.co.nz) |
23:22.34 | CIA-10 | 03sjhill 07uClibc-nptl * r13332 10/test/nptl/ (tst-_res1.c tst-_res1mod1.c tst-_res1mod2.c): |
23:22.34 | CIA-10 | Remove invalid test. This test will not pass on uClibc because we are not |
23:22.34 | CIA-10 | concerned about backwards compatibility of the resolver code. |
23:33.40 | *** join/#uclibc ulf_k_ (n=ulf_kypk@p54BDADF6.dip0.t-ipconnect.de) |
23:35.03 | bug1 | ive moved to NZ ! |
23:35.26 | bug1 | now if only my computers where here :( |