Re: [PATCH][RFA] Reap children on stack trace generation.

[Casey Marshall <csm at gnu dot org>]

>>>>> "Tom" == Tom Tromey <tromey@redhat.com> writes:

>>>>> "David" == David Daney <ddaney@avtrex.com> writes:
David> natPosixProcess.cc has a related problem.  First it forks, but
David> if the exec fails, the child is not reaped.

Tom> If the exec fails, do we really need the kill()?

Tom> Otherwise this is looking good.  But also see
Tom> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11801 The thing is,
Tom> even with this patch we need a fix for the case where the user
Tom> just expects the process to be GCd and reaped without an explicit
Tom> waitFor call.

I remember this thread:
<http://gcc.gnu.org/ml/java-patches/2004-q2/msg00140.html>

I thought it was generally agreed that the most reliable
implementation for this sort of thing would be a seperate thread that
reaped child processes. That is, a single thread forks, execs, then
waits for processes.

In which case, wouldn't Classpath's VM reference implementation of
this be a better choice than fixing this version?

-- 
Casey Marshall || csm@gnu.org