01:17.21 | *** join/#uclibc Xires (n=Xires@68-184-15-187.dhcp.stls.mo.charter.com) |
01:23.31 | *** join/#uclibc sjhill (n=sjhill@eth13.com-link.com) |
01:56.58 | *** join/#uclibc Xires012 (n=Xires@68-184-15-187.dhcp.stls.mo.charter.com) |
02:34.47 | *** join/#uclibc Xires012 (n=Xires@68-184-15-187.dhcp.stls.mo.charter.com) |
02:54.11 | *** join/#uclibc El-Lotso (n=EnergonC@202.146.86.86) |
02:54.23 | *** part/#uclibc El-Lotso (n=EnergonC@202.146.86.86) |
06:43.04 | *** join/#uclibc blindvt_ (n=bf@M872P028.adsl.highway.telekom.at) |
06:59.52 | *** join/#uclibc vapier (i=UserBah@wh0rd.org) |
07:50.37 | *** join/#uclibc vodz (i=1000@dzo.simtreas.ru) |
07:50.56 | vodz | hi all |
07:52.19 | vodz | blindvt_: why into arping you do not use cfg |= opt & 127 ? |
07:52.56 | vodz | vs if(opt & 1) cfg |= 1; if(opt & 2) cfg |= 2... |
07:55.00 | blindvt` | vodz, that would be better, yes. |
07:56.53 | blindvt` | vodz, btw. in arping, i didn't use bb_common_buf1, because if someone would set the BUFSIZ too low, it would brake. |
07:57.38 | vodz | blindvt`: i do not know with my 20 y experience system with BUFSIZ less 256 |
07:57.40 | blindvt` | well, more stuff would brake for e.g -DBUFSIZ=0, i guess |
07:59.28 | vodz | blindvt`: and is not problem too. See coreutil/realpath.c |
08:00.50 | blindvt` | vodz, hm. That would of course work |
08:04.28 | blindvt` | vodz, did you see the mail on the list about getting the symbols of libbb.h and scanning the applet source to emit dependencies to these functions, so they are built transparently? |
08:05.50 | vodz | blindvt`: and RESERVE_CONFIG_BUFFER in while(1) is bad idea, RELEASE.... must moved to finish() or atexit, and RESERVE to top of main() |
08:07.42 | blindvt` | vodz, this thread: http://busybox.net/lists/busybox/2006-March/019914.html |
08:09.30 | *** join/#uclibc psaksa (n=pate@ip212-226-134-137.adsl.kpnqwest.fi) |
08:09.53 | vodz | blindvt`: I last time look the list very selectively. At me interest was gone to be engaged in the project full time. Only the fan now. Something last time is a lot of bazaar, instead of real a work. |
08:15.54 | vodz | blindvt`: why not. your URL is "more busyboxes" a work. |
08:23.46 | blindvt` | vodz, i don't 100% agree about the RESERVE_CONFIG_BUFFER in while(1). It trades binary size vs. runtime. I.e. if you choosed to malloc your buffers, then you are willing to be slower as you want a smaller binary |
08:24.16 | blindvt` | vodz, i certainly see why it's not ideal, runtime-wise |
08:27.16 | vodz | blindvt`: for me, RESERVE_CONFIG_BUFFER must moved to applets_struct to busybox_main and all RELEASE to atexit() in run_applet_by_name() |
08:27.48 | vodz | its clear and beautifully |
08:29.35 | vodz | and more use as first need buffer - common_bufsiz1[max(4097, BUFSIZ+1)] |
08:31.04 | vodz | it make clear, small and fast |
08:33.35 | vodz | My unique consent with Rob - that may to be the name of the buffer not beautiful. ;-) |
08:33.49 | blindvt` | so you would do {count,{size1,size2,size3,...} in (or around) the applet_struct and then use applet_buf[0..count] in the applet-source. Did i understand this right? |
08:37.22 | vodz | blindvt`: Somehow so. The truth it will demand very serious attitude to support of change as to watch it is necessary in a complex. I do not see it now. |
09:02.32 | blindvt` | vodz, yes, i don't see us changing this soon |
09:19.51 | vodz | wtf "Woot" ? Is a interjection? |
09:34.05 | blindvt` | vodz, yes. http://en.wikipedia.org/wiki/Woot |
09:39.39 | vodz | aha, "it originated via music and culture." |
09:40.17 | vodz | i think, use culture-depend modern word in international mailist is bad idea |
09:46.31 | CIA-10 | 03aldot * r14734 10busybox/networking/arping.c: (log message trimmed) |
09:46.31 | CIA-10 | - set flags in a more reasonable manner. Thanks, vodz. |
09:46.31 | CIA-10 | <PROTECTED> |
10:32.12 | *** join/#uclibc FireEgl (i=Atlantic@Atlantica.Tcldrop.US) |
11:25.49 | CIA-10 | 03aldot * r14735 10busybox/util-linux/ipcs.c: (log message trimmed) |
11:25.50 | CIA-10 | - do away with useless fwd-decls, redundant bb_printf calls, silence warning |
11:25.50 | CIA-10 | <PROTECTED> |
11:45.43 | *** join/#uclibc Jenna (n=cherryRe@209.8.233.208) |
11:46.34 | *** part/#uclibc Jenna (n=cherryRe@209.8.233.208) |
11:52.06 | CIA-10 | 03aldot * r14736 10busybox/networking/arping.c: - use bb_opt_complementally; Saves another 14 bytes. (thanks again, vodz) |
11:59.08 | blindvt` | is there a convention in busybox about treating stdin/out buffers on exit? AFAICS there isn't.. |
12:03.02 | blindvt` | i lean towards dropping fflush() before calls to exit() as most likely, the closing I/O will drain the buffers for me |
12:04.50 | blindvt` | (and no, i'm not talking about using bb_fflush_stdout_and_exit() instead) |
12:05.21 | Xires012 | any reason not to use ffsync? |
12:10.08 | blindvt` | Xires012, ffsync? Do you mean fasync? |
12:12.08 | Xires012 | no, ffsync..it's a macro that does an fsync on the provided fd(if provided) & then does similar to an fflush |
12:15.18 | blindvt` | Xires012, i can't find ffsync in uClibc. |
12:15.43 | Xires012 | oh, crap..I forgot it wasn't there |
12:16.18 | Xires012 | sorry, I'll fish for the patches when I get some time, K? |
12:18.42 | blindvt` | Xires012, ah, FFSYNC is there |
12:20.25 | Xires012 | blindvt_; ah, K..make sure it's the same thing though |
12:20.40 | Xires012 | I had to make mine after someone told me about it |
12:21.21 | blindvt` | but it's a BSD'ism and from the looks boils down to O_SYNC. I'm pretty sure that i don't need that on stdio for the most part :) |
12:21.52 | blindvt` | s/stdio/stdout/; |
12:29.04 | blindvt` | wonder how we could a callgraph analysis to generally warn about these printf bloat.. |
12:29.14 | CIA-10 | 03aldot * r14737 10busybox/networking/arping.c: (log message trimmed) |
12:29.14 | CIA-10 | - coalesce printf calls. |
12:29.14 | CIA-10 | <PROTECTED> |
12:29.52 | blindvt` | can't type. could do a, even |
13:04.42 | *** join/#uclibc prpplague (n=billybob@72.22.159.157) |
13:14.52 | *** join/#uclibc sjhill (n=sjhill@eth13.com-link.com) |
13:56.20 | *** join/#uclibc prpplague (n=billybob@72.22.159.157) |
13:57.51 | *** join/#uclibc prpplague (n=billybob@72.22.159.157) |
14:00.26 | mjn3 | blindvt_, BUFSIZ is defined by stdio.h and ansi/iso requires it be at least 256 |
14:12.13 | *** join/#uclibc El-Lotso (n=EnergonC@60.50.203.198) |
14:15.45 | blindvt` | mjn3, i see. So we should better make sure that in libbb.h, #ifndef BUFSIZ #define BUFSIZ 4096 ends up >= 256, which doesn't seem to be the case currently |
14:16.17 | *** join/#uclibc Han (n=han@unaffiliated/Han) |
14:17.55 | Han | Hi. Is there a livecd with uclibc installed so I can try building an application from the livecd? |
14:18.19 | YazzY | Han: yes |
14:18.36 | Han | And perhaps it would be nice if the FAQ would mention the -nostdinc gcc-option. |
14:19.43 | YazzY | Han: http://www.vyatta.com/ has one |
14:19.49 | YazzY | they use busybox etc |
14:20.39 | YazzY | or maybe http://www.xorp.org/livecd.html |
14:22.46 | Han | I wonder if they include gcc etc... |
14:23.46 | *** join/#uclibc prpplague (n=billybob@72.22.159.157) |
14:43.33 | *** join/#uclibc blindvt_ (n=bf@M913P019.adsl.highway.telekom.at) |
15:00.30 | *** join/#uclibc El-Lotso (n=EnergonC@60.50.203.198) |
15:46.20 | CIA-10 | 03aldot * r14738 10busybox/ (coreutils/printf.c modutils/modprobe.c): - make append_option and multiconvert static. |
16:39.33 | CIA-10 | 03aldot * r14739 10busybox/ (18 files in 4 dirs): |
16:39.33 | CIA-10 | - move buffer allocation schemes to libbb.h |
16:39.33 | CIA-10 | - include the correct headers: applets need busybox.h while lib* need libbb.h |
16:43.00 | *** join/#uclibc andersee (n=andersee@codepoet.org) |
16:43.41 | blindvt` | andersee, hi. please add a 1.1.1 version to the versions of busybox in the bug-tracker |
16:48.21 | andersee | blindvt_: how about just 1.1.x? |
16:48.31 | andersee | and 1.2.x? |
16:49.30 | *** part/#uclibc Han (n=han@unaffiliated/Han) |
17:10.45 | prpplague | or X.X.X |
17:12.30 | andersee | prpplague: nah, apparently ICANN scrapped .xxx |
17:12.46 | andersee | prpplague: slashdot says so, so it must be true |
17:12.59 | prpplague | hehe |
17:15.36 | andersee | and a one legged man in a butt kicking contest |
17:42.11 | *** join/#uclibc thraxisp (n=thraxisp@ottgate.precidia.com) |
18:07.37 | *** join/#uclibc tux1800 (n=eric@c66.110.147-33.clta.globetrotter.net) |
18:56.27 | *** join/#uclibc khem (n=khem@63.81.120.158) |
19:38.27 | *** join/#uclibc tux1800 (n=eric@c66.110.147-33.clta.globetrotter.net) |
20:53.51 | blindvt_ | andersee, thanks |
21:00.03 | *** join/#uclibc zakalwe (i=zakalwe@you.sneakymonkey.net) |
21:00.03 | *** join/#uclibc Masta-G (n=mastag@5355BE2D.cable.casema.nl) |
21:04.58 | Masta-G | hey I tried building a toolchain for sh4 little endian gcc-3.4.4 but libstdc++-v3 fails with "macro "index" requires 2 arguments, but only 1 given" |
21:05.11 | Masta-G | it does the same thing for 3.4.5 and 3.4.6 |
21:07.23 | Masta-G | could someone please have a quick look at http://pastebin.com/638580 |
21:07.31 | Masta-G | I think its because i using fc5 |
21:07.42 | Masta-G | *i'm |
21:16.56 | CIA-10 | 03aldot * r14740 10busybox/Makefile: - make sure that we see all lib*/.c in IMA-mode even when not building the .so |
21:40.54 | khem | Masta-G: have you looked into 304-index_macro.patch |
21:41.21 | khem | Masta-G: this is available for GCC 4.1.0 so you may not get it for 3.4.x directly in buildroot |
21:42.23 | Masta-G | i see |
21:42.31 | Masta-G | thanks, i will give it a try |
22:44.02 | *** join/#uclibc blindvt_ (n=bf@M785P024.adsl.highway.telekom.at) |
23:04.29 | *** part/#uclibc YazzY (n=yazzy@unaffiliated/yazzy) |
23:05.07 | *** join/#uclibc thraxisp (n=thraxisp@ottawa-hs-209-217-123-76.d-ip.magma.ca) |
23:09.32 | *** join/#uclibc FireEgl (i=Atlantic@Atlantica.Tcldrop.US) |
23:38.34 | *** join/#uclibc tux1800 (n=eric@modemcable211.23-80-70.mc.videotron.ca) |