irclog2html for uclibc on 2002.09.16

01:23:42bug1anyone understand the purpose of the two dup(fd) statements in the busybox openvt applet posted to the n-l
01:27:44mjn3bug1: you still there?
01:32:21mjn3bug1: oh well, anyway... he closes fd 0 (stdin) and then open()s the new vt.  so the fd returned for it is 0 since that is the lowest available.  then he closes 1 (stdout) and 2 (stderr) and dup()s twice so that fds 1 and 2 now point to the new vt as well
01:33:36bug1yep
01:34:17bug1cool
01:35:25bug1that makes sense, i couldnt work it out by looking at it though
01:35:44bug1ill commit the applet
01:36:41bug1i better have a look at that cp butg as well that 2 people have mentioned
01:37:02mjn3if you're adding stuff, i posted a mesg applet to the list in may... in response to someone else's rather bloated version.  i never did add it.  it would need to be integrated.  it is message 2002-May/011409.html in the mailing list archive
01:37:54bug1ok, ill look into it
01:40:31mjn3thanks.  glad i could help with the openvt thing.  later
01:40:42bug1k
01:45:52bug1hi Erik, im going to commit the openvt applet soon (unless you have reservations)
01:46:25anderseeNo worries.
01:46:34anderseeI see you got awk, thanks
01:46:42bug1Ill also look into that cp bug that 2 people have mentioned
01:47:00bug1i think awk will be usefull to a few people adds a lot of flexibility
01:47:19bug1not sure what cp is usppoed to do though
01:59:10bug1andersee: another thing im unsure about is vlad's dhcpcd applet
01:59:41bug1i think he put a fair bit of effort into it, havent tried it yet htough
02:02:05anderseeMe either
02:09:49bug1openvt going in now
03:44:48anderseebug1: I'm adding tainting support to busybox lsmod/insmod...
03:47:08bug1ok, that allows it to load lgpl'ed (etc) modules i asusme
03:48:18bug1such as plex86
03:48:42andersee$ ./busybox lsmod | head -n 1
03:48:42anderseeModule                  Size  Used by    Not tainted
03:48:44andersee$ ./busybox lsmod | head -n 1
03:48:44anderseeModule                  Size  Used by    Tainted: PF
03:49:06anderseeI just loaded the vmware kernel module to taint my kernel.  lsmod works now.
03:49:16bug1cool
03:49:32anderseeNow to fixup tainting in insmod
03:49:33bug1do you understand how cp is supposed to work wit hspecial files ?
03:49:43anderseeWhat do you mean?
03:50:08andersee'cp /dev/fd0 /tmp/foo' should copy the contents of the floppy
03:50:15bug1its been reported by 2 people the busybox cp behaves differently to gnu cp with pipes, devices etc
03:50:18andersee'cp -a /dev/fd0 /tmp/foo' should copy the device special file itself
03:50:31anderseeHmm
03:50:42anderseeI guess its worth looking at.
03:51:19anderseeI know I tested cp behavior pretty carefully, but people have changed things since then....
03:51:57bug1ill do some tests anyway
03:52:39anderseecool, thx
03:55:36bug1cp /dev/null . copies the node itself not the contents.... dont you need to use dd to read the contents
03:56:28anderseeNope.
03:56:56andersee$ cp /dev/null . && ls -l ./null
03:56:57andersee-rw-r--r--    1 andersen andersen        0 Sep 15 23:03 ./null
03:57:11anderseeIt creates a zero length file.  Not a device node.
03:57:29anderseeBusybox cp should not copy device nodes by default....
03:58:30bug1hmmm, the current bb behaviour seems to make more sense to me
03:59:33bug1whats the point it creating an empty file
04:00:09anderseeTry testing something else, like /dev/fd0
04:00:28bug1ahh cp /dev/null -a creates the device
04:04:15anderseeright
04:11:30anderseehehe
04:11:35anderseeIt works
04:11:36andersee$ sudo ./busybox insmod vmmon
04:11:37anderseeUsing /lib/modules/2.4.20-pre7-erik/misc/vmmon.o
04:11:37anderseeWarning: loading vmmon will taint the kernel: no license
04:11:37andersee  See http://www.tux.org/lkml/#export-tainted for information about tainted modules
04:12:37bug1cool
04:17:23bug1just sent a one line patch to the mailing list re the cp bug... ill be back later
06:52:37anderseebug1: you still around?
06:54:01bug1yes
06:54:26bug1(reading up on my zaurus i just got)
06:54:43andersee:-)
06:54:48anderseewants to get one too
06:55:21anderseeSee any other bugs I should fix before releasing 0.60.4?
06:55:28anderseehttp://bugs.busybox.net/db/pa/lbusybox.html
06:55:36anderseeAny idea on the dpkg bug?
06:55:53bug1the segfault bug ?
06:56:14bug1im suspect it may be caused by dmalloc
06:56:48bug1the fix just got rid of a possible free(NULL) if i remember correctly
06:57:17bug1dmalloc defaults to deliberatly not handleing free(NULL)
06:58:14anderseeHmm.  Do you know if he was using dmalloc?
06:58:28anderseeI didn't see that mentioned
06:58:31bug1he didnt say
06:59:12anderseeSo perhaps we should write it off as user error?
06:59:24anderseeI suppose we could memset(file_entry, 0, sizeof(foo))
06:59:28anderseeJust in case....
07:00:42bug1yea i guesss it couldnt hurt
07:00:59bug1i actually havnet done anything with dpkg for ages
07:01:41anderseePerhaps its best to ignore it...
07:02:01bug1yea, dont worry about it
07:02:59anderseeI'm looking over the debian bug reports now.
07:03:00anderseehttp://bugs.debian.org/cgi-bin/pkgreport.cgi?src=busybox&repeatmerged=yes
07:03:22anderseeLooks like #158706 is easy to fix.  :)
07:05:05bug1i wouldnt worry about the bb ash bug reports too much, ash is still improving
07:05:59anderseeI'm actually considering backporting ash from busybox unstable...  I think the unstable on is actually more solid at the moment.
07:06:56bug1you mean backporting to unstable ?
07:07:50bug1ill do that uname patch if you like
07:08:18anderseeI mean taking busybox/shell/ash.c and busybox/libbb/arith.c and backporting then to busybox.stable/ash.c and busybox.stable/libbb/arith.c
07:08:31anderseebug1: Already finished the uname patch.
07:08:33andersee:)
07:08:40bug1k
07:09:23bug1isnt that the way ash should be going anyway ?
07:09:58bug1is stable going to stay the old layout ?
07:10:36anderseeWell, ideally, I'd like this to be the end of the line for the 0.60.x...  So yeah the one from unstable is where new and interesting things happen.
07:11:01bug1that mv bug (127522) sounds pretty nasty
07:11:26anderseeLooks it.  Probably some error checking gone awry
07:14:20bug1mv works for me
07:14:33anderseeHmm.
07:14:34bug1must have been something specific to what he was doing
07:14:40anderseePerhaps so.
07:14:51anderseeShall I close it (unreproducable)
07:14:53andersee?
07:16:16bug1yea
07:16:23bug1or ask for more information
07:16:35bug1but its probably too late for that
07:17:10bug1#138060: losetup missing that wouldnt take long to put in
07:18:02bug1but if bb mount supports -o loop than losetup isnt needed is it ?
07:22:26bug1#148396 looks like free might be converting kB incorrectly
07:23:40bug1hmm
07:25:54anderseebug1: I added losetup to the .deb so we are good there.
07:26:16anderseebug1: depends on what one is using losetup for.
07:26:55anderseebug1: I'll just enable it (cuz its tiny) to close the bug and let aph worry about the result....
07:28:56bug1i think you close #148396
07:29:27bug1free and bb free display different amount of free space becasue the buffer increases everytime you run it :)
07:31:41bug1there is the preserve hardlinks build option in unstable to handle 122860
07:41:25bug1busybox.org bug #1224 could be done, the watch applet
07:44:24anderseebug1: hehe
07:50:00bug1ive got the watch applet ready to commit
07:50:44bug1except it doesnt compile.. hmm
08:04:26bug1can close #1224 watch is in there
08:07:27anderseewill do, thx.
08:13:13anderseeI'm looking at that 'cp' bug just reported
08:17:21bug1my approach was to just trieat it like a file unless the -a option was used
08:17:43bug1regular file rather than a special  file
08:19:28anderseeI just checked in a fix.  I did the simple/dumb thing that looked simplest...
08:20:04anderseehttp://busybox.net/cgi-bin/cvsweb/busybox.stable/libbb/copy_file.c.diff?r1=1.16&r2=1.17
08:26:07bug1i guess it will do, cp and mv might need a bit of attention in future though, it seems overly complex
08:26:53anderseeindeed
08:27:07anderseeI was tempted to pull things into a common function
08:27:38anderseeBut decided to make things simple/obvious.  An easy target to optimize later...
08:29:17bug1fair enough
08:29:50bug1back in half hour or so
08:30:16anderseeThe only thing I can see remaining is ash...  I'll put it to the list whether to backport ash from unstable for the 0.60.4 release...
13:03:27mjn3andersee: morning
13:04:29anderseemorning
13:04:46anderseeI got busybox in reasonable shape last night...
13:05:11mjn3i'm just reading the posted messages
13:05:19anderseeAlso (finally!) got the gcc 2.95 toolchain to build with the STLport C++ lib.
13:05:25mjn3cool
13:05:33anderseeI'll be releasing that and the gcc-3.2 toolchains today
13:06:04mjn3alright.  guess i'll do some rebuilding then... ;-)
13:06:22andersee:-)
13:11:48anderseemjn3: cool.  The gcc-2.95/STLport toolchain works perfectly.  I just need to fixup the default c++ include dir and I'm (finally) done.
13:12:03mjn3anyone know of a small implementation of diff?  the only one i've been able to find is a rather limited version from the minix src
13:45:41mjn3andersee: ugh... i just tried the new busybox watch applet.  to clean the screen it outputs 25 newlines.  :-(
13:48:12anderseeick
13:48:19anderseehasn
13:48:25anderseehasn't read it yet

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