| 01:17.16 | *** join/#uclibc wrobbie (n=rob@cm74.kappa84.maxonline.com.sg) |
| 02:22.42 | mcgrof | I'm puzzled. I got busybox-1.5.0, using kexec-1.101. What I do is mount /dev/hda1 on /mnt and then do the kexec -l --command-line="foo" --initrd= etc.. and it starts booting fine but it then cannot find the root dev specified (/dev/hda1) |
| 02:22.45 | mcgrof | anyone see this before? |
| 02:31.08 | *** join/#uclibc Skapare (n=phil@sedna.ipal.net) |
| 06:05.10 | *** join/#uclibc dalias (i=dalias@brightrain.aerifal.cx) |
| 06:05.14 | dalias | solar, you around? |
| 06:10.22 | *** join/#uclibc blindvt__ (n=bf@M996P028.adsl.highway.telekom.at) |
| 07:25.01 | *** join/#uclibc hw_ (n=hw@87.139.57.5) |
| 08:27.45 | *** join/#uclibc reppel (n=reppel@213-140-11-128.fastres.net) |
| 08:31.14 | reppel | hi, is mklibs actually working for you in buildroot? |
| 09:19.06 | blindvt` | reppel, it'd help if you could describe what is failing resp. what you expect it to do. I'm not familiar with it, sorry |
| 09:37.32 | reppel | blindvt`: i selected "run mklibs" from the buildroot menu but it doesn't run it |
| 10:16.41 | CIA-9 | 03vda * r18318 10busybox/ipsvd/ (tcpudp.c udp_io.c): (log message trimmed) |
| 10:16.41 | CIA-9 | ipsvd: use IP:PORT syntax for environment vars. Pros: |
| 10:16.41 | CIA-9 | 1. it's saner (matches internal libc sockaddr abstraction). |
| 11:02.57 | CIA-9 | 03vda * r18319 10busybox/ (include/usage.h ipsvd/tcpudp.c ipsvd/udp_io.c): Add/improve comment, shorten some usage texts. No code changes. |
| 12:22.48 | *** join/#uclibc wrobbie (n=rob@cm74.kappa84.maxonline.com.sg) |
| 12:44.11 | *** join/#uclibc vrm (n=vrm@210.23.98-84.rev.gaoland.net) |
| 12:52.34 | vrm | hi |
| 13:06.53 | *** join/#uclibc svolpe_gerrath (n=Gerrath_@unaffiliated/gerrath) |
| 13:35.59 | *** join/#uclibc prpplague (n=dave@mail.americanmicrosystems.com) |
| 13:45.25 | *** join/#uclibc kasc (n=kasc@dslb-084-060-115-069.pools.arcor-ip.net) |
| 13:45.43 | kasc | hi there |
| 13:52.11 | kasc | i've got a problem with building busybox from within a busybox environment. the Makefiles use -prune option with find, but the busybox find implementation doesnt seem to understand -prune |
| 13:52.17 | kasc | any ideas? ^^ |
| 13:56.03 | blindvt` | kasc, USE_DESKTOP( ACTS(prune)) |
| 13:56.52 | blindvt` | kasc, do away with this silly CONFIG_DESKTOP thing for find and instead add an option for the args of find that are currently hidden with DESKTOP |
| 13:58.35 | CIA-9 | 03aldot * r18320 10busybox/coreutils/stat.c: (log message trimmed) |
| 13:58.35 | CIA-9 | - remove superfluous bss user (flags) and manually unswitch some areas: |
| 13:58.35 | CIA-9 | <PROTECTED> |
| 13:59.50 | CIA-9 | 03aldot * r18321 10busybox/coreutils/stty.c: (log message trimmed) |
| 13:59.50 | CIA-9 | - remove some bss users. |
| 13:59.50 | CIA-9 | <PROTECTED> |
| 14:01.26 | CIA-9 | 03aldot * r18322 10busybox/coreutils/dd.c: (log message trimmed) |
| 14:01.26 | CIA-9 | - remove bss users. Shrinkage while at it. See XXX for further, pre-existing bugs |
| 14:01.26 | CIA-9 | <PROTECTED> |
| 14:06.40 | kasc | prune seems to be quite well embedded within all those desktop stuff. is it safe to just move the code outside the #ifs or will that run me into more trouble? |
| 14:07.25 | blindvt` | kasc, introduce a CONFIG_FIND_PRUNE and use that instead of DESKTOP |
| 14:07.52 | blindvt` | kasc, the same for all other args of find that are hidden below DESKTOP |
| 14:08.12 | blindvt` | kasc, then, after cursory testing, send the patch to the mailing-list, please |
| 14:08.23 | blindvt` | TIA |
| 14:08.59 | kasc | oh, ok |
| 14:09.40 | blindvt` | s/CONFIG_FIND_PRUNE/FEATURE_FIND_PRUNE/;# i ment to type |
| 14:27.54 | *** join/#uclibc blindvt__ (n=bf@M990P015.adsl.highway.telekom.at) |
| 14:39.08 | *** join/#uclibc KhemHome (n=kraj@adsl-71-146-1-192.dsl.pltn13.sbcglobal.net) |
| 15:10.35 | kasc | blindvt`: patch sent ;) |
| 15:43.31 | *** join/#uclibc sjhill (n=sjhill@real.realitydiluted.com) |
| 15:55.25 | blindvt` | kasc, what subject? Don't see it yet.. |
| 16:00.44 | kasc | [PATCH] un-desktopize find |
| 16:08.12 | kasc | looks like the busybox implementation of patch is kinda broken... |
| 16:24.06 | kasc | patch relies on the original file being there, which under usual conditions is not. |
| 16:25.21 | *** join/#uclibc ambroseL (n=bjb@router.fidus.ca) |
| 16:39.04 | blindvt` | kasc, i don't think your mail got through. Please send a copy to me directly |
| 16:46.15 | kasc | do i, by chance, have to be subscribed to the list to send to it? |
| 17:01.35 | blindvt` | kasc, indeed :) |
| 17:02.34 | blindvt` | kasc, on the bright side, that subscription is free of charge and done easily |
| 17:02.48 | blindvt` | ;) |
| 17:03.42 | blindvt` | it's a relatively low-volume list |
| 17:05.53 | kasc | i was hoping i was getting some simple fix without to subscribing to yet another list, but since it looks like i have to also fix the busybox patch, it might really be better to actually subscribe |
| 17:15.12 | kasc | ok, resent the patch |
| 17:19.45 | blindvt` | kasc, that one made it to the list, good. You forgot the attachment, which is suboptimal |
| 17:20.08 | kasc | uh well ^^ |
| 17:20.58 | kasc | resent, with patch this time |
| 17:21.23 | kasc | anyways, bbl in some hours or so |
| 17:40.05 | blindvt` | kasc, that patch must be against an older version, isn't it? Can you please diff against trunk? TIA for your patience |
| 17:49.49 | CIA-9 | 03vda * r18323 10busybox/ipsvd/tcpudp.c: reword comment |
| 17:57.56 | CIA-9 | 03aldot * r18324 10busybox/TODO: - note cosmetic glitch in tail |
| 17:58.33 | *** join/#uclibc KhemHome (i=KhemHome@nat/montavista/x-3262d46c4d653ab8) |
| 17:58.40 | CIA-9 | 03aldot * r18325 10busybox/TODO: - remove debugging output. sorry.. |
| 18:00.29 | blindvt` | KhemHome, hi |
| 18:00.42 | KhemHome | hi there blindvt` |
| 18:06.20 | blindvt` | KhemHome, quick question inline: uclibc.org/~aldot/bb.tail-shrinkage.00.diff |
| 18:06.20 | blindvt` | ideas? |
| 18:06.20 | blindvt` | lately going over the accumulated "bloat" in coreutils.. |
| 18:08.18 | blindvt` | up until recently, i've largely avoided big parts of those, i admit, just fooled with a bit of coreutils here and there and cowardly avoided the risk of causing breakage with random shrinkage ;) |
| 18:09.44 | blindvt` | they are relatively concise, still offer opportunities upon closer inspection, fwiw |
| 18:10.30 | KhemHome | I guess you are targetting code size ? |
| 18:11.15 | blindvt` | i guess that putting effort into coreutils is most benefical, overall |
| 18:11.34 | KhemHome | probably yes |
| 18:12.26 | KhemHome | bool is 1 byte long .. |
| 18:13.18 | blindvt` | is mixing ops on fd's and filps hazardous? IIRC it's not, but didn't look |
| 18:14.12 | blindvt` | KhemHome, bool is often char, depending on arch and toolchain. For some bool is benefical, compared to int |
| 18:14.57 | KhemHome | yeah |
| 18:15.28 | blindvt` | VRP often doesn't grok stuff, so i prefer to rely on it where it makes sense |
| 18:15.42 | blindvt` | to not rely, even |
| 18:16.27 | KhemHome | right |
| 18:16.34 | KhemHome | what is bb_common_bufsiz1 type? |
| 18:16.48 | blindvt` | char*, ISTR |
| 18:17.31 | blindvt` | we flag globals with attribute data. quite a nice trick |
| 18:18.02 | blindvt` | helps on nommu, i'm told, so i'm moving towards it when i touch stuff |
| 18:20.11 | KhemHome | I think the patch looks ok to me |
| 18:20.16 | blindvt` | i'm still unclear how i would go about emulating an ia32 without mmu.. or any random arch qemu supportsfor that matter |
| 18:20.21 | KhemHome | you save the intructions to access bss I think |
| 18:20.37 | KhemHome | as you get rid of static in there |
| 18:21.18 | blindvt` | KhemHome, yes, it's ok, but i ment to do away with the fileno handling and go to FILE* throughout. Guess i can't avoid to pass the filename in then.. hmz |
| 18:21.31 | KhemHome | blindvt`: is there a way to compare the performance ? |
| 18:22.18 | blindvt` | KhemHome, well, i once wrote an applet speed patch (see --speed site:busybox.net) that counted the cycles spent in an applet |
| 18:23.47 | blindvt` | http://busybox.net/lists/busybox/2005-September/016309.html |
| 18:24.21 | KhemHome | blindvt`: fstat will need a file descriptor so you can not avoid fd |
| 18:24.46 | blindvt` | there was even a yearly reminder ;) http://www.busybox.net/lists/busybox/2006-September/024612.html |
| 18:25.23 | blindvt` | KhemHome, yea, but i could pass in the filename and stat that instead |
| 18:26.55 | KhemHome | blindvt`: as it seems from patch lseek also uses fd |
| 18:28.05 | blindvt` | KhemHome, i didn't fixup tail_read yet but wanted to hear if someone would have a better idea than to additionally pass in the filename to use there along the FILE* |
| 18:28.47 | blindvt` | KhemHome, what do you think about that applet --speed patch? Sounds usefull for debugging/devel ? |
| 18:29.17 | blindvt` | i didn't get much feedback but questions on how rdtsc would work :-( |
| 18:29.58 | blindvt` | which is certainly not the kind of feedback i'd wanted to have for a patch, but oh well |
| 18:30.20 | KhemHome | passing an additional argument prolly is same in terms of code size |
| 18:30.31 | KhemHome | blindvt`: the applet idea is nice |
| 18:30.43 | KhemHome | blindvt`: it could be a config option |
| 18:31.15 | blindvt` | KhemHome, the patch had it as a config-option already, IIRC :) |
| 18:31.35 | KhemHome | yes I see |
| 18:31.36 | KhemHome | that |
| 18:32.11 | KhemHome | I think this is useful |
| 19:32.06 | *** join/#uclibc quitte (n=quitte@stgt-d9bea933.pool.mediaWays.net) |
| 20:29.19 | CIA-9 | 03aldot * r18326 10busybox/coreutils/tail.c: (log message trimmed) |
| 20:29.19 | CIA-9 | - minor shrinkage |
| 20:29.19 | CIA-9 | <PROTECTED> |
| 20:30.50 | blindvt` | that was the wrong stats. was 1362 afterwards *shrug* |
| 20:52.06 | CIA-9 | 03aldot * r18327 10busybox/ (33 files in 15 dirs): - sed -e "s/char[[:space:]]*\*[[:space:]]*argv\[\]/char **argv/g" |
| 21:05.05 | kasc | re |
| 21:26.31 | kasc | blindvt`: i'm not sure about what to do with the "and" "or" and "not" stuff within #if ENABLE_DESKTOP in find.c near line 345 |
| 21:30.16 | kasc | should the long argument versions stay in DESKTOP? |
| 22:10.46 | *** join/#uclibc blindvt_ (n=bf@M911P001.adsl.highway.telekom.at) |
| 23:10.59 | *** join/#uclibc common (n=common@port-212-202-20-216.dynamic.qsc.de) |
| 23:41.36 | CIA-9 | 03vda * r18328 10busybox/ (findutils/Config.in findutils/find.c include/usage.h): (log message trimmed) |
| 23:41.36 | CIA-9 | find: un-DESKTOPize (Kai Schwenzfeier <niteblade@gmx.net>) |
| 23:41.36 | CIA-9 | find: -group, -depth (Natanael Copa <natanael.copa@gmail.com>) |