This is the mail archive of the
mailing list for the Cygwin project.
Re: 1.5.1 issue?
Charles Wilson wrote:
Christopher Faylor wrote:
It is not necessarily a non-problem. If you can reproduce this
consistently, then it is a problem, either in make or cygwin.
It is not consistent. (is it ever? :-)
Make -j10 will probably be more useful, since running a threaded make is
usually what triggers the issue for me. I've also observed that the
likelihood of creating what I call an "orphaned" process (when the
parent returns but the child doesn't receive the proper SIG) happens to
increase significantly as the average load increases. While this is
most prevalent in Make, I have seen it happen with shell scripts that
call multiple, concurrent sub-shells. Another observation is that this
behavior seems to happen more frequently on Win9X boxes then on Win2k.
What this means, I have no idea, but Chuck is right, it isn't consistent
at all. I'll try to grab an strace the next time I find a situation
which does this.
Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Problem reports: http://cygwin.com/problems.html