irclog2html for #uclibc on 20030916

00:43.36*** join/#uclibc sjhill (~sjhill@pool-141-151-177-157.pitt.east.verizon.net)
00:53.12*** join/#uclibc andersee (~andersee@codepoet.org)
00:55.13*** join/#uclibc bug1 (~bug1@c211-28-122-16.sunsh3.vic.optusnet.com.au)
00:55.40bug1andersee: you forgot to add get_terminal_width_height to libbb/Makefile,i fixed it
00:56.32anderseebug1: oops
00:56.46anderseeI did do that.  But yes, I forgot to check that in.
00:56.48anderseesorry
00:57.21anderseeI also didn't mess with converting the relevant ioctl in stty.c
00:58.20anderseeIt seems to have a pile of hacks layered on that ioctl for solaris and such.  I decided to not mess with it till the rest was in
01:00.00bug1fair enough
01:00.38bug1can our testsutie handle dependencies ?
01:00.58anderseesuch as?
01:01.20bug1well, to test sed properly it will take lots of tests
01:01.44bug1and if i break something very basic, then heaps of them will fail
01:02.59anderseeso you would like to have the testsuite abort on the first failure?
01:03.00bug1and i will have to go through them all and see which is the real cause
01:03.28bug1not quite, in sed some commands require other commands to setup the test
01:03.51bug1i want to test the commands used to setup the test before the actual test
01:04.28anderseeTried adding "set -e"?
01:04.47anderseethat should cause it to about on the first failure
01:05.21bug1eg. i want to makes sure the test `echo foo | sed 's/foob/bar'` works beforee i try `echo foo | sed s/foo/bar/;s/bar/baz/`
01:06.10bug1i guess i can maniplate the test to run them in a specific order
01:09.28bug1looks like they run in alphabetic order at the moment, i should be able to jsut put a number infront to control the order i guess
01:56.23bug1Woohoo, i think ive recovered from all my "fixes" .GNU make's /configure works again
02:03.12*** part/#uclibc sjhill (~sjhill@pool-141-151-177-157.pitt.east.verizon.net)
02:03.49psypeteuh, guys
02:04.04psypeteoh nevermind
02:04.26psypetethough i found a bug... really i rm -rf'ed the build directory while compiling
02:04.34psypete:-X
02:11.49mjn3sigh...
02:13.16mjn3felix complains that i don't report dietlibc bugs.  so, i point out that his "fixed" fputc can write to stdin.  so he adds CANREAD and CANWRITE flags and checks for them in fgetc and fputc, but _not_ in fread or fwrite
02:14.54kergothsounds like you're fighting an uphill battle with these people
02:15.36mjn3the dietlibc philosophy seems to be "if it links and it looks like it runs, then it is good enough"
02:18.21mjn3i spent 20 minutes browsing their stdio code yesterday and spotted 5 or 6 definite bugs and 4 or 5 other questionable things i didn't feel like taking time to investigate
02:19.37mjn3on the plus side though, i made me check the C99 standard and realize that stdio output behavior when feof is true changed between c89 and c99
02:20.09mjn3and while i was looking, i found a site with all the C standard defect reports
02:20.24mjn3so it was 20 minutes well spent
02:30.14mjn3kergoth: the uphill battle i'm fighting is with ulrich drepper.  i swear, i don't know how to make things any clearer unless i pull out a box of crayons and start drawing cartoons to illustrate the problems
03:14.54mjn3sigh... dietlibc's ftell is only 2 lines long, but contains 2 bugs
03:15.18bug1hehe
03:17.24mjn3but it sure is small...
05:26.35bug1damnit, what did i jsut do to get_line_from_file.c
05:27.00psypetedraw from perl golf to keep it 2 lines long
05:27.07psypeteobfuscate and win!
05:29.14bug1well, that marks my sed work done for a while
05:29.40bug1there is now a config option to get it to follow GNU sed or posix, 3 differences
05:30.04psypetesweet
05:30.30psypeteis GNU basically posix + extras?
05:30.40psypeteor does one of them have stuff the other doesn't?
05:30.48psypeteand vice versa?
05:30.55bug1<PROTECTED>
05:31.32bug1biggest problem is 's/a/b/ g'
05:32.00bug1gnu treats it as 's/a/b/g' where as i believe it shoudl be 's/a/b/;g'
05:32.13bug1it does totally differnet things
05:32.44bug1ive seen configure scripts that do it the GNU way, so ./configure requires GNU compatable sed
05:34.07bug1the other difference is the standard doesnt allow '/regex/! <function>', it has to be '/regex/~<function>'
05:34.18bug1the other difference is the standard doesnt allow '/regex/! <function>', it has to be '/regex/!<function>'
05:35.01bug1the other difference is just that GNU sed outputs a blank line if 'N' is used at the end of the file
05:39.52psypete";g" does what?
05:40.14psypetethats a weird thing for GNU sed to do, about the N at the end of the file
05:41.03psypeteand btw, i'm with you that "/ g" shouldn't be "/g" just for syntax reasons, but i don't get the ";"
05:41.50bug1psypete: the ';' is a seperator between commands, g as a seperate command swaps pattenr space with hold space
05:42.31bug1g as a flag for the substute command make the substitue happen for the hwole line
05:43.21bug1g replaces pattern space with hold space
05:49.45bug1hold space is temperary storage, so you work on pattern space, store it, then use the g command bring it back and overrite your current data
05:50.27bug1so whenever the two usages conflict there is going to be major breakage
05:53.37psypeteyep
05:59.44bug1andersee: re your mail to busybox ml, ah, i knew there was a good answer to the "job control turned off" question somewhere, couldnt remember i saw it.
05:59.59anderseebug1: np
06:00.20anderseebug1: I _really_ should create a busybox FAQ, or wiki, or something
06:01.37bug1Gad has a webpage at http://yallara.cs.rmit.edu.au/~gabraham/choox.php about his project hes using busybxo for, its like tmsbrt
06:01.57bug1tomsrtbt
06:02.58bug1rmit.edu.au is jsut up the road from me
06:10.50anderseeit's great people are making these
06:19.09bug1i agree
06:20.39psypetea rescue disk based on 2.2?
06:20.47psypetewhat does he hope to rescue? slackware 7.1?
06:21.52bug1psypete: i imagine size had something to do with it
06:22.13psypetehmm
06:22.49bug1my 2.6 kernel wont even fit on a floppy anymore :(
06:23.02psypetepush it down with your foot
06:23.07psypeteworks for suitcases...
06:23.07bug1hehe
06:23.32psypetei've fit a whole lot into one 1.68MB floppy
06:24.03psypetebusybox w/2.4 kernel and uClibc, along with pppd, two httpd's, my web interface + daemon apps, etc.
06:24.36psypetethe trick with the kernel is to get the right balance of bzImage and modules, and then know how to compress the modules right
06:26.22bug1the bzImage from debians modular kernel was 1.0MB last time i looked
06:26.32bug1fro 2.6
06:26.53bug1its go up by almost 50%, used to be about 700kB
06:26.59psypetemine's 607kB
06:27.10psypeteso maybe i skimped on some extras ;)
06:27.54psypeteit's still got all the networking and filesystem still tho
06:28.51bug1i think its just some sections that really effect its size, like acpi, reiserfs
07:38.14*** join/#uclibc tahorg (~loic@perax.net1.nerim.net)
11:58.45*** join/#uclibc ofauchon (~olivier@mrs.mtice.net)
11:58.54ofauchonhi
12:01.48anderseeofauchon: hello
12:26.08*** join/#uclibc tahorg (~loic@perax.net1.nerim.net) [NETSPLIT VICTIM]
13:17.46*** join/#uclibc dsmith (~dsmith@mail.actron.com)
13:38.12*** join/#uclibc tahorg (~loic@perax.net1.nerim.net) [NETSPLIT VICTIM]
13:52.24*** join/#uclibc tahorg (~loic@perax.net1.nerim.net) [NETSPLIT VICTIM]
14:08.25*** join/#uclibc tahorg (~loic@perax.net1.nerim.net) [NETSPLIT VICTIM]
15:09.28*** join/#uclibc tahorg (~loic@perax.net1.nerim.net) [NETSPLIT VICTIM]
15:46.51*** join/#uclibc tahorg (~loic@perax.net1.nerim.net) [NETSPLIT VICTIM]
15:54.05*** join/#uclibc tahorg (~loic@perax.net1.nerim.net) [NETSPLIT VICTIM]
18:04.27*** join/#uclibc ofauchon (~olivier@AMarseille-206-1-16-137.w81-50.abo.wanadoo.fr)
18:19.28*** join/#uclibc kergoth (~kergoth@dsl081-228-056.chi1.dsl.speakeasy.net)
19:39.48ofauchonif i disable " Native shared library loader 'ldd' support ", can i still resolve dependencies with another ldd ?
19:40.19ofauchondoes the uclibc have its own ldd ?
19:42.21psypetepsypete is gaming.  !gameinfo for more info.
19:54.58mjn3ofauchon: yes, uclibc has an ldd.  also, the bug you were seeing the other day with ldd and glibc wasn't a bug in ldd.  it is just that libpthread for both glibc and uClibc have the same SONAME
20:00.16ofauchonmjn3: hi, ok for the tip. Can i disable "native ldd support " safely ? i need some more space
20:02.32mjn3ofauchon: got me.  try it and see.  i've apparently always got it enabled
20:02.50mjn3i somehow doubt it will save you much space though
20:03.24mjn3how much space do you need?
20:04.32ofauchoni'm working on a 1.44 boot disk . I can't strip the linux kernel anymore, so i must gain space on libs & applications
20:04.38ofauchoni need at least 50 K
20:10.31*** join/#uclibc ofauchon (~olivier@AMarseille-206-1-16-137.w81-50.abo.wanadoo.fr) [NETSPLIT VICTIM]
20:10.31*** join/#uclibc dsmith (~dsmith@mail.actron.com) [NETSPLIT VICTIM]
20:10.31*** join/#uclibc psypete (~psypete@adsl-211-141-3.mia.bellsouth.net) [NETSPLIT VICTIM]
20:10.32*** join/#uclibc mjn3 (~mjn3@68.226.136.114) [NETSPLIT VICTIM]
20:11.41mjn3ofauchon: disabling sys_errlist and the error message text and doing the same for signals will save you a few k.  using the old_vfprintf will as well
20:12.02ofauchongreat
20:12.19*** join/#uclibc solar (solar@linbsd.net) [NETSPLIT VICTIM]
20:12.47mjn3if you don't need gnu getopt (long option support), you can disable that as well and save a couple of k by using the susv3 getopt that will be built instead
20:13.47mjn3ofauchon: other things... thread support?  full libm?  full rpc?  or even no rpc support at all
20:14.32mjn3ofauchon: i forget what you're working on... do you need floating point support?
20:17.55ofauchoni'm working on a wireless authentification gateway . maybe a remplacement for some wifi access point
20:18.15ofauchoni removed rpc, i kept threads
21:27.27*** join/#uclibc andersee (~andersee@codepoet.org)
21:28.59mjn3andersee: hey
21:30.24mjn3andersee: some problems reported with tinylogin on arm.  storing getopt return code in a char results in an infinite loop since char is unsigned there
21:30.36anderseemjn3: morning
21:30.43mjn3andersee: sorry... morning.  :-)
21:30.54anderseemjn3: saw the report.
21:31.00anderseehttp://cvs.uclinux.org/cgi-bin/cvsweb/uClinux-dist/user/tinylogin/Makefile?rev=1.1.1.3&content-type=text/x-cvsweb-markup
21:31.17mjn3and someone wanting arch-optimized string funcs
21:31.18anderseeThis tells me that uClinux dist is no less than _5_ releases behind the current
21:31.26mjn3not surprising
21:31.35kergothhey mjn3, hows it goin?
21:31.45anderseeThe "signed char" vs "int" problem was fixed years ago
21:32.50mjn3kergoth: hey kergoth.  going ok is suppose.  wish i could find the bug i'm looking for though.  how about with you?
21:34.10kergothquite well, been ensuring that OE operates properly under ash rather than bash, and cygwin.. and consolodating autotools handling for all packages into a single place
21:34.13kergothbuild stuff. joy.
21:47.28ofauchondamned, seems busybox compilation complains about my uclibs customization. compile & compile again. so much fun
22:00.34kergothwhy do the openssh folks seem to want to add new autoconf tests with every release.
22:02.41*** join/#uclibc ChanServ (ChanServ@services.)
22:02.41*** mode/#uclibc [+o ChanServ] by zelazny.freenode.net
22:03.09*** join/#uclibc ChanServ (ChanServ@services.) [NETSPLIT VICTIM]
22:03.09*** mode/#uclibc [+o ChanServ] by zelazny.freenode.net
22:03.50mjn3kergoth: oooohhh... kick it again!
22:03.56kergothhehe
22:05.14anderseemjn3: perhaps drepped needs a kick as well
22:05.17anderseeerr
22:05.23anderseedrepper
22:05.27kergothhe probably needs more than a kick
22:06.09mjn3yeah...  well, i could always post to slashdot.  maybe they'd carry it
22:22.58mjn3ok... people should now be able to build locale support in a uClibc environment
23:02.20mjn3http://sources.redhat.com/ml/libc-alpha/2003-09/msg00148.html
23:38.55*** join/#uclibc DaveM (~david@h24-207-7-221.dlt.dccnet.com)

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