Random fork failures

2011-07-11 Thread yoni levi
Hi, We have a problem with fork for a very long time. From time to time, fork just hangs (with 2 process doing busy loop). The problem occurs when we do allot of spawns (make system). It is very easy to recreate the problem, but unfortunately, there is too much code involve to send. I did a li

Re: Random fork failures

2011-07-11 Thread yoni levi
That's a clean installation. what can cause this kind of crashes? I tried again with cygwin1.dll from latest snapshot (http://cygwin.org/snapshots/cygwin1-20110711.dll.bz2) I am not sure if this will be helpful, but these are the md5sum for the relevant files: ~$ md5sum /usr/bin/install /us

Re: Random fork failures

2011-07-11 Thread yoni levi
Actually, I have not the faintest idea. Can you build your own Cygwin DLL without -O2, see if it still occurs, and run that scenario through GDB? I am on it. -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http:/

Re: Random fork failures

2011-07-12 Thread yoni levi
On 12/7/2011 10:23, Corinna Vinschen wrote: > I am also attaching strace.txt with the output of: > ~$ strace install a /etc/b And it worked, didn't it? Your strace looks normal, no exception occured. Right. But rxvt crash. -- Problem reports: http://cygwin.com/problems.html FAQ:

Fwd: Re: Random fork failures

2011-07-12 Thread yoni levi
On 12/7/2011 10:23, Corinna Vinschen wrote: > I am also attaching strace.txt with the output of: > ~$ strace install a /etc/b And it worked, didn't it? Your strace looks normal, no exception occured. Right. But rxvt crash. bash.exe is still running though. -- Problem reports:

Re: Random fork failures

2011-07-12 Thread yoni levi
On 12/7/2011 16:18, Fergus wrote: (People recommend mintty for 1.7 and probably with very good reasons, but I still quite like rxvt.) my rxvt does not hang, all rxvt windows just disappear. with mintty it works perfectly. -- Problem reports: http://cygwin.com/problems.html FAQ:

Re: Random fork failures

2011-07-13 Thread yoni levi
On 13/7/2011 10:32, Corinna Vinschen wrote: The access violation occurs inside the Cygwin DLL, but the address is useless without the DLL. Since you built the DLL yourself, you would have to look where this crash occurs. You should have a rxvt.exe.stackdump file with function addresses, kind of

Re: Random fork failures

2011-07-13 Thread yoni levi
On 13/7/2011 10:32, Corinna Vinschen wrote: Please call addr2line for all addresses starting with 61. In the above example that would be: $ addr2line -e /bin/cygwin1.dll 6100749B 61004EFC 61004F84 61006499 $ addr2line -e /bin/cygwin1.dll 61116FEF 610C30EB 610C7605 0040524B 0040107B 610

Re: Random fork failures

2011-07-13 Thread yoni levi
On 13/7/2011 10:32, Corinna Vinschen wrote: I still have to point out, though, that I can not reproduce this problem. Neither with a debug DLL, nor with an optimized DLL I managed to catch this inside gdb (run "gdb rxvt" from mintty). but the backtrace seems useless to me. Any ideas how to impr

Re: Random fork failures

2011-07-13 Thread yoni levi
On 13/7/2011 12:55, Corinna Vinschen wrote: Building Cygwin in debug mode and access to the source tree helps a lot. The only problem is that this crash only reproduce when compiling with -O2 -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/f

tar 1.25 fail to handle symliks correctly

2011-07-14 Thread yoni levi
Hi, I know tar version 1.25 sporadically fail to untar symlinks under cygwin. http://www.cygwin.com/ml/cygwin/2011-04/msg00335.html I switch to an older tar version (1.23) and everything is working perfectly. Since this is a critical bug (tar is practically unusable) - I think you should conside