irclog2html for #uclibc on 20051220

03:01.38*** join/#uclibc fishhead (i=gerber@unaffiliated/fishhead)
05:19.46*** join/#uclibc Newsome (n=sorenson@216-190-206-130.customer.csolutions.net)
05:32.50*** join/#uclibc Jenna (n=cherryRe@209.8.233.248)
05:33.09*** part/#uclibc Jenna (n=cherryRe@209.8.233.248)
06:03.50*** join/#uclibc Jenna_zzz (n=cherryRe@209.8.233.226)
06:04.04*** part/#uclibc Jenna_zzz (n=cherryRe@209.8.233.226)
07:47.44*** join/#uclibc blindvt_ (n=bf@M774P027.adsl.highway.telekom.at)
07:58.47*** join/#uclibc vrm (n=vrm@94.55.101-84.rev.gaoland.net)
08:15.50*** join/#uclibc Kaloz (i=kaloz@arrakis.dune.hu)
09:30.20*** join/#uclibc ashes (n=ashes@modemcable085.56-130-66.mc.videotron.ca)
09:45.34*** join/#uclibc psm (n=mps@host-6.mikroweb.hu)
11:02.56CIA-1203vodz * r12954 10busybox/networking/httpd.c: remove debug feature for production
11:07.36blindvt`vapier, ping.. is it ok for you if i update libpcap to 0.9.4 and tcpdump to 3.9.4 (both were at .3)?
11:10.54*** join/#uclibc psaksa (n=pate@ip212-226-134-137.adsl.kpnqwest.fi)
11:12.04psmblindvt: are you using gcc-4.0.2 w/ uClibc? do you use the buildroot patches for it (and enabled c++)?
11:13.58blindvt`psm, no, i'm using 4.2 (just disabling c++ for the target as we speek). Why do you ask?
11:14.53blindvt`psm, a dynamic ldd segfaults for me, btw. Didn't delve into it yet, though. I guess this isn't expected, right?
11:15.37psmw/o c++ it is not relevant, the libstdc++-pic is faulty for 4.0.2
11:17.17psmldd segfaults on apps/libs that miss DT_NEEDED (so unresolved symbols)
11:17.41psmis this your case ?
11:19.35blindvt`psm, not sure.. ldd $(which ldd) triggers it, but iirc i asked for a segfault
11:21.29psmthat does not segv for me
11:28.26*** join/#uclibc T`2 (n=total@tor/session/x-1e2ee471c94128db)
12:03.58*** join/#uclibc prpplague (n=billybob@72.22.141.111)
13:18.03ZtaHm, how do I from my C (uclibc) program execute rdate and pipe its output into an unnamed pipe?
13:19.05ZtaI did (approximately) this using systen, but it only works on my pc with glibc: system("rdate -p 10.0.0.1 >&10");
13:19.26ZtaSo instead, I tried using execv -- perhaps it worked better.
13:19.50ZtaBut execv simply exits my entire program when finished
13:23.09Ztaany known problems on this?
13:23.23ZtaShould I do it differently?
13:24.52psaksafork before exec
13:29.54Ztashouldn't system work?
13:29.58Ztasystem is easier to use.
13:30.34Ztasystem("/bin/sh rdate -p 192.168.20.2 >&10");  and  system("rdate -p 192.168.20.2 >&10");  gives me the error:  sh: Syntax error: Bad fd number
13:30.47ZtaI did open my pipe and 10 is the fd.
13:31.54psaksawhat's the default for close-on-exec flag?
13:32.07psaksasomeone remembers or do I have to hit the man pages? :-)
13:32.47Ztasince noone else is alive, and I don't know what you're talking about... =)
13:33.23psaksaeach fd has close-on-exec flag..  if that flag is set, your fd is closed when system() calls exec
13:34.24ZtaAh...
13:34.41ZtaStill, the approach using system() works on PC.
13:34.54ZtaBut perhaps glibc and uclibc have different defaults.
13:35.02ZtaHow do I check it and set it?
13:35.23psaksacan you add one function call yo your program and print it's output?
13:35.35Ztasure
13:35.43psaksaclose-on-exec is handled with call to fcntl()
13:37.02ZtaSo I should use system() right now?
13:39.37psaksasec.. I'll try google since my man page search fails
13:40.10ZtaI'll try and set it to F_DUPFD
13:43.43psaksawell anyway, if you print return value of fcntl(fd, F_GETFD); that should give answer if it's closed
13:44.07psaksacan't seem to find the default value from docs
13:45.26Ztaint res = pipe( fds );
13:45.34Ztain = fds[0];
13:45.34Ztaout = fds[1];
13:45.34Ztaint outCOE = fcntl(fd, F_GETFD)
13:45.34Ztacout << "in=" << in << ", out=" << out << ", out close-on-exec=" << outCOE << endl;
13:45.37ZtaoutCOE = fcntl(out,F_DUPFD);
13:45.40Ztacout << "in=" << in << ", out=" << out << ", out close-on-exec=" << outCOE << endl;
13:45.49Ztasomething like this.  Ignore the tabs and don't comment the c++.
13:47.56Ztain=9, out=10, out close-on-exec=0
13:47.57Ztain=9, out=10, out close-on-exec=-1
13:48.13psaksafcntl(out, F_GETFD)  ?
13:48.37ZtaSo first it's 0.  Then I set it to -1 = F_DUPFD)  This gives same error:  sh: Syntax error: Bad fd number
13:48.41Ztapsaksa, you're right.
13:50.43Ztaint outCOE = fcntl(out, F_GETFD);
13:50.59ZtaoutCOE = fcntl(out,F_SETFD);
13:51.25Ztaprintint outCOE in betwee gives first 0 and then -1.
13:51.58Ztafuck I suck as C.
13:52.02Ztaat
13:52.04Ztawell..
13:52.12Ztasame same =)
13:52.13psaksaok well that 0 answers the question anyway...  it's not closed by that
13:54.18Ztaok.
13:54.23ZtaSo it should be allright?
13:54.37ZtaI can't see any aparrent close-on-exit flags in man open(2)
13:56.40Ztagrrr... for a person who's used to java, this low-level shit slowly by surely kills me.
13:56.45Ztabut
13:56.49ZtaI can't spell anymore.
13:57.07ZtaThat's probably becasue of lame abbrevations like fcntl
13:57.24Ztaman..
13:57.24psaksayeah that close-on-exec flag should be ok if it's 0....
13:57.37ZtaI'll try using execv and fork
13:57.40psaksathis is the point where I'd start using fork+exec calls :-)
13:57.55Zta=)
13:58.18Ztahttp://beej.us/guide/ipc/pipes.html
13:58.20Ztathis seem useful
14:03.20*** join/#uclibc Newsome (n=sorenson@216-190-206-130.customer.csolutions.net)
14:05.42Zta/usr/sbin/rdate: 1: Syntax error: word unexpected (expecting ")")
14:05.46Ztatime to go home.
14:06.11ZtaBut thanks for the help anyway =)  fork will make it happen.
14:16.29*** part/#uclibc Newsome (n=sorenson@216-190-206-130.customer.csolutions.net)
14:43.46*** join/#uclibc sjhill (n=sjhill@eth13.com-link.com)
15:09.43CIA-1203aldot * r12955 10buildroot/package/libsysfs/libsysfs.mk: - add dirclean target to libsysfs
15:15.55CIA-1203aldot * r12956 10buildroot/package/vtun/Config.in: - vtun depends on lzo and openssl just like openvpn
15:16.58vapierblindvt_: vtun already depended on lzo/ssl
15:23.23CIA-1203landley * r12957 10busybox/docs/style-guide.txt: Actually, static buffers look like _this_...
15:25.31CIA-1203landley * r12958 10busybox/loginutils/Config.in: (log message trimmed)
15:25.31CIA-12Shadow password support went beyond the dependency event horizon, just make
15:25.31CIA-12it an independent menu. And make internal shadow password support a subset
15:48.23*** join/#uclibc blindvt_ (n=bf@M779P022.adsl.highway.telekom.at)
16:04.31blindvt`vapier, as these two were not selected in the config, their build dirs were not cleaned.
16:05.10blindvt`vapier, did you saw my earlier question about updating libpcap and tcpdump? ok with you?
16:39.27*** join/#uclibc thraxisp (n=thraxisp@ottgate.precidia.com)
16:52.54*** join/#uclibc samrobb_ (n=samrobb_@65-117-135-105.dia.cust.qwest.net)
17:18.04CIA-1203landley * r12959 10busybox/util-linux/ (mount.c nfsmount.c): (log message trimmed)
17:18.04CIA-12Some minor cleanups/bugfixes split off from the big remount work:
17:18.04CIA-12<PROTECTED>
17:25.52CIA-1203landley * r12960 10busybox/util-linux/switch_root.c:
17:25.52CIA-12Fix chroot, leave console alone if -c not specified, and yank debug code.
17:25.52CIA-12(I still haven't set up a test environment to confirm this works...)
20:09.04*** part/#uclibc fishhead (i=gerber@unaffiliated/fishhead)
20:51.45*** join/#uclibc prpplague (n=billybob@72.22.141.111)
20:54.00*** join/#uclibc ulf_k (n=ulf_kypk@p54BDB07E.dip0.t-ipconnect.de)
22:51.31*** join/#uclibc psm (n=mps@host-6.mikroweb.hu)
23:48.15*** join/#uclibc blindvt_ (n=bf@M860P014.adsl.highway.telekom.at)
23:49.36*** join/#uclibc ulf_k_ (n=ulf_kypk@p54BD89DD.dip0.t-ipconnect.de)
23:53.38CIA-1203vapier * r12961 10buildroot/toolchain/binutils/ (8 files in 2 dirs): binutils 2.16.91.0.5 has been released

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.