01:32.21 | *** join/#uclibc ambassador (~ambassado@h143.1.39.162.ip.alltel.net) |
01:34.28 | *** join/#uclibc ambassador (~ambassado@h143.1.39.162.ip.alltel.net) |
01:42.47 | *** join/#uclibc ambassador (~ambassado@h143.1.39.162.ip.alltel.net) |
02:33.52 | *** join/#uclibc ambassador (~ambassado@h143.1.39.162.ip.alltel.net) |
02:59.41 | *** join/#uclibc TheMasterMind1 (~aman@h-69-3-159-120.MCLNVA23.dynamic.covad.net) |
05:39.30 | *** join/#uclibc randey (~randey@202.63.124.234) |
05:45.50 | *** join/#uclibc randey (~randey@203.152.142.122) |
06:36.05 | *** join/#uclibc randey (~randey@202.63.124.234) |
08:43.35 | *** join/#uclibc TMM1 (~aman@h-68-166-71-14.MCLNVA23.dynamic.covad.net) |
08:56.18 | *** join/#uclibc tahorg (~loic@perax.net1.nerim.net) |
09:03.02 | *** join/#uclibc randey (~randey@202.63.124.234) |
09:48.17 | *** join/#uclibc randey (~randey@202.63.124.234) |
10:45.51 | *** join/#uclibc randey (~randey@202.63.124.234) |
11:31.40 | *** join/#uclibc kio (~chatzilla@204.Red-81-40-196.pooles.rima-tde.net) |
11:56.06 | *** join/#uclibc randey (~randey@202.63.124.234) |
13:04.42 | *** join/#uclibc randey (~randey@202.63.124.234) |
15:01.52 | *** join/#uclibc dsmith (~dsmith@mail.actron.com) |
16:38.07 | *** join/#uclibc ambassador_ (~ambassado@h143.1.39.162.ip.alltel.net) |
16:40.11 | *** join/#uclibc tahorg (~loic@perax.net1.nerim.net) |
17:03.05 | kergoth | morning mjn3 |
17:03.14 | mjn3 | kergoth: morning. how goes it? |
17:03.32 | kergoth | monday morning, cranky and half asleep |
17:03.33 | kergoth | but not bad |
17:03.34 | kergoth | you? |
17:03.41 | *** join/#uclibc DaveM_ (~david@h24-207-7-221.dlt.dccnet.com) |
17:04.09 | mjn3 | ok. got some errands to run today, but checking in for a bit |
17:04.20 | mjn3 | erik should be back sometime today from his travels |
17:04.58 | kergoth | ah cool |
17:10.24 | DaveM_ | mjn3: Hi Manuel. While generating the locales over the weekend, I see that a couple crashed (out of memory). I'm going to add some swap, but after is there a way to just rebuild the ones that crashed? |
17:22.43 | mjn3 | DaveM_: you mean for glibc? |
17:23.35 | DaveM_ | mjn3: yes |
17:23.55 | mjn3 | you could run localedef by hand |
17:24.17 | DaveM_ | I take it locale-gen is a script? I never even looked at that |
17:24.22 | mjn3 | yeah |
17:24.30 | DaveM_ | thanks |
17:24.34 | mjn3 | np |
17:25.10 | mjn3 | DaveM_: actually, i think if you rerun it then it will only process the ones that are missing. but i'm not 100% sure |
17:26.41 | DaveM_ | k, I'll let you know |
17:29.58 | *** join/#uclibc ShortWave (~bobby@ip68-100-100-183.dc.dc.cox.net) |
17:30.07 | *** part/#uclibc ShortWave (~bobby@ip68-100-100-183.dc.dc.cox.net) |
17:30.10 | *** join/#uclibc ShortWave (~bobby@ip68-100-100-183.dc.dc.cox.net) |
17:30.11 | ShortWave | hey all |
17:30.18 | ShortWave | Anyone here? |
17:30.25 | DaveM_ | hey |
17:31.22 | ShortWave | Hiya |
17:31.43 | ShortWave | How are you? |
17:31.56 | ShortWave | Me, I'm about to pull my hair out |
17:32.21 | DaveM_ | alright, you have a uClibc, busybox problem? |
17:32.27 | ShortWave | Kinda |
17:32.32 | ShortWave | It's more of a weird toolchain problem |
17:33.11 | ShortWave | I downloaded the i386 root_fs, and I booted that up under a user-mode linux instance (after I enlarged the filesystem somewhat) |
17:33.33 | ShortWave | My issue is that autoconf, specifically aclocal is segfaulting |
17:34.19 | ShortWave | I grabbed the source for autoconf, and tried compiling that, and it fails on "autom4ke" |
17:34.53 | ShortWave | rather, autom4te |
17:35.03 | ShortWave | that's actually a perl script, from the looks of it |
17:35.21 | DaveM_ | I was just about to say that |
17:35.21 | ShortWave | it seems that it's invoking something else, but the file is so huge, it's difficult to chase down what the actual problem is |
17:35.41 | ShortWave | heh, I ran file on it just to be sure |
17:36.14 | ShortWave | So what do you think? Perl crashing perhaps? |
17:36.47 | mjn3 | what are you trying to do when it crashes? |
17:37.04 | ShortWave | compile stuff |
17:37.09 | ShortWave | aclocal dies with segfault |
17:37.27 | DaveM_ | I remember some discussion on perl with uClibc about 3 weeks (I think). I don't remember what the discussion was, but you might want to read the archives |
17:37.35 | mjn3 | could you be a bit more specific? what are you trying to compile? |
17:37.45 | ShortWave | the aclocal distributed on the root_fs is segfaulting |
17:37.48 | ShortWave | [root@(none) root]# aclocal |
17:37.48 | ShortWave | Segmentation fault |
17:37.48 | ShortWave | [root@(none) root]# |
17:37.59 | ShortWave | It seems the binary on the root_fs itself is corrupt |
17:38.18 | ShortWave | The last thing I tried was compiling autoconf |
17:38.26 | ShortWave | a new version, just in case the local version was hosed |
17:38.52 | ShortWave | Which made me wonder if perl was somehow wonky |
17:39.05 | ShortWave | since aclocal is a perl script... |
17:39.11 | ShortWave | I wonder if recompiling perl would fix it |
17:40.22 | mjn3 | could be. i don't remember if perl is included in the downloadable root_fs or not. which version of perl, and did you build it yourself? |
17:40.35 | ShortWave | Did NOT build it myself |
17:40.50 | ShortWave | This is perl, v5.8.0 built for i386-linux |
17:40.55 | ShortWave | there you are |
17:42.53 | mjn3 | i just tried in an i386 rootfs i built myself. i built perl 5.8.2 (requires a couple of small patches) and aclocal ran fine |
17:43.00 | ShortWave | hrm |
17:43.27 | ShortWave | I tried buildroot, but that failed on me... |
17:43.31 | ShortWave | let me see why |
17:44.18 | ShortWave | oh yes |
17:44.24 | mjn3 | the perl patches in buildroot really need to be fixed. that's why i built perl inside the rootfs |
17:44.32 | ShortWave | a bunch of undefined references |
17:44.44 | ShortWave | mjn3: So you suggest rebuilding perl in the root_fs? |
17:45.02 | mjn3 | yes. but first, what undefined refs are you seeing with buildroot? |
17:45.31 | ShortWave | weird stuff |
17:45.41 | ShortWave | I did some digging and it seemed to be related to floating point stuff |
17:45.58 | ShortWave | give me a few minutes and I'll have a few errors for you :) |
17:46.01 | mjn3 | there was a bug that was fixed last week. |
17:46.33 | mjn3 | the i386 soft float patch was always being applied, when it was only supposed to be applied if soft float support was desired |
17:47.18 | ShortWave | ah |
17:47.23 | ShortWave | that might explain the problem |
17:47.35 | ShortWave | Which patches should I apply to perl 5.8? |
17:48.07 | mjn3 | if you want to build perl yourself, you will currently need to |
17:48.29 | mjn3 | apply at least the first 2 diffs in sources/perl.patch |
17:48.55 | ShortWave | ok |
17:49.00 | mjn3 | specificly adding -lpthread to the link helps work around a uClibc shared loader bug |
17:49.45 | ShortWave | hrm, ok |
17:50.14 | mjn3 | only an issue if you build perl with thread support |
17:50.35 | ShortWave | that perl patch is located where? |
17:50.57 | mjn3 | buildroot/sources/perl.patch |
17:51.03 | ShortWave | ah |
17:51.12 | ShortWave | will buildroot handle it? |
17:51.48 | mjn3 | not if you build perl yourself inside the rootfs |
17:51.53 | ShortWave | hrm |
17:51.59 | ShortWave | so I have to decide now... |
17:52.05 | ShortWave | use buildroot...or compile perl myself |
17:52.22 | ShortWave | I'm gonna try the buildroot method first, since I really would prefer a clean build |
17:53.00 | ShortWave | What must I do to work around the undefined references? |
17:53.58 | ShortWave | actually, better idea |
17:54.04 | ShortWave | I'll do them both at the same time! :) |
17:54.27 | mjn3 | you don't have to do anything. the i386 soft float patch was always being applied (as i said above). this resulted in a broken toolchain if you weren't using soft-float |
17:54.36 | mjn3 | the problem was fixed last week |
17:54.46 | ShortWave | OK |
17:54.59 | ShortWave | well I'll see what happens with this buildroot then |
17:55.12 | ShortWave | Tho it occurs me I might be using an older buildroot package... |
17:55.19 | mjn3 | did you cvs up to this mornings version? |
17:55.34 | ShortWave | nope |
17:55.40 | ShortWave | I probably should |
17:55.43 | ShortWave | I was using a tarfile |
17:56.38 | *** join/#uclibc loid (zebra@200-102-064-202.paemt7014.e.brasiltelecom.net.br) |
17:56.40 | loid | somebody know how make an only one floppy linux (diskless) with X and rdesktop? |
17:57.28 | ShortWave | what's the -p level I need for this perl.patch to build it? |
17:58.26 | mjn3 | loid: that would be challenging. you might be able to do it with tinyx. i've never looked at rdesktop though |
17:59.13 | loid | but |
17:59.20 | loid | how? |
17:59.28 | loid | i know an site that teach this? |
17:59.55 | mjn3 | loid: no, i don't know of a site that would show you how |
18:00.13 | mjn3 | ShortWave: depends on if you are in the perl-5.8.0 dir or in the parent dir |
18:01.44 | mjn3 | ShortWave: if you are in the perl-5.8.0 dir, then -p 1 i would think |
18:04.10 | ShortWave | mjn3: yeah it was 1 |
18:06.12 | mjn3 | loid: try www.angelfire.com/linux/floorzat/2diskXwin.htm |
18:06.23 | ShortWave | mjn3: do you suggest using uml or chroot to do the compile? |
18:07.55 | mjn3 | either should work. if normally chroot. but if you chroot and run some tests as root there, expect a few failures. they pass if you aren't root |
18:08.05 | ShortWave | Ok |
18:08.07 | mjn3 | ur.. i normally chroot |
18:08.23 | *** join/#uclibc rfs (~ralphs@mail.blueairnetworks.com) |
18:09.03 | ShortWave | Okee |
18:09.19 | ShortWave | I'm building perl in chroot whilst I'm running buildroot seperately, so we'll see what happens |
18:09.26 | ShortWave | I gotta run out for a bit, but I shall return |
18:11.15 | ShortWave | one more question, how much space does buildroot require? |
18:23.47 | mjn3 | ShortWave: depends on what you're building. for a full-ish build, figure about 1.5G perhaps. that's rough |
18:28.06 | *** part/#uclibc loid (zebra@200-102-064-202.paemt7014.e.brasiltelecom.net.br) |
18:59.43 | *** join/#uclibc ade|desk (~adrian@modem-3715.gorilla.dialup.pol.co.uk) |
19:10.37 | *** join/#uclibc jc (~jcw@adsl-80-64-175.asm.bellsouth.net) |
19:19.23 | ShortWave | well then |
19:19.33 | ShortWave | recompiling perl appears to have solved the segfault problem |
19:27.21 | *** join/#uclibc ade|desk (~adrian@modem-2159.gazelle.dialup.pol.co.uk) |
19:35.32 | ShortWave | hrm |
20:17.57 | mjn3 | ShortWave: good |
20:20.53 | *** join/#uclibc Qui_Gon (~fox@lns-vlq-14-62-147-221-82.adsl.proxad.net) |
20:33.16 | *** join/#uclibc andersee (~andersee@codepoet.org) |
20:33.25 | mjn3 | andersee: welcome back |
20:33.30 | kergoth | andersee: hey, wb |
20:34.38 | andersee | mjn3: thx |
20:35.29 | andersee | kergoth: thx |
20:54.09 | rfs | hola andersee |
20:56.51 | andersee | rfs: howdy |
20:57.26 | rfs | andersee: I've found some odd behaviour between uClibc and madplay... |
20:59.06 | rfs | basically, the output is noisier - like the math is being done with fewer bits |
20:59.48 | andersee | rfs: arch? soft float? |
21:00.05 | rfs | andersee: oh, sorry :) armv4l (netwinder), nwfpe. |
21:00.29 | rfs | its uclibc-0.9.21 |
21:00.56 | rfs | i would have tried 0.9.23 but the installation options appear to have changed :{ |
21:01.27 | rfs | but I didn't see anything obvious |
21:01.49 | andersee | rfs: ok |
21:02.11 | rfs | afaik madplay is all integer arithmetic, so i am quite puzzled... |
21:02.17 | andersee | rfs: indeed |
21:02.19 | andersee | rfs: it is |
21:03.00 | andersee | it is totally integer |
21:03.31 | andersee | So I wouldn't expect any additional noise unless we have real live math bugs somewhere |
21:03.33 | rfs | andersee: madplay now has some cpu-specific assembly, but I turned that off, because it was causing distortion (with both uClibg and glibc-2.3.2) |
21:04.17 | andersee | rfs: perhaps you just need to recalibrate your ears then. ;-) |
21:04.23 | rfs | hehe |
21:04.42 | rfs | i'm comparing against old madplay-0.13beta binary from 2 years ago... :) |
21:05.26 | andersee | rfs: you may want to try comparing straces -- uClibc w/noise present vs your old binary w/o |
21:05.34 | rfs | andersee: incidentally, I have now managed a kernel+initrd with much of busybox enabled, and loads of kernel modules, still under the magic 1MB size limit :) |
21:05.37 | andersee | see if the same stuff is happening |
21:05.46 | andersee | rfs: :-) |
21:06.09 | rfs | andersee: yes, good idea, maybe ltrace too |
21:07.36 | andersee | I'll perhaps be able to do stuff this evening |
21:08.03 | andersee | I'm tending my 1.5 year old daughter the rest of the day |
21:10.05 | rfs | andersee: "tending"... sounds almost like farming term :) |
21:10.32 | rfs | oh, one more bit that may be significant: i'm aiming for a static binary of madplay. |
21:10.53 | andersee | rfs: ok |
21:11.12 | andersee | rfs: you may then want to tune uClibc a bit for small size |
21:11.12 | rfs | using the mad-0.14.2b.tar.gz which seems to be current stable version. |
21:11.39 | andersee | rfs: i.e. disable pthreads and everything else you can |
21:12.09 | rfs | andersee: i'll upload my .config at some point and perhaps you can review it... |
21:14.08 | rfs | andersee: btw, the httpd in busybox is very cool :) I have an almost endless list of possibilities for it ;) |
21:14.36 | andersee | rfs: heh |
21:17.28 | rfs | andersee: so i can theoretically boot ext3/reiserfs with the ramdisk & modules. |
21:17.54 | rfs | however i've yet to work out how to handle the 2nd boot (with a different kernel). |