Re: scoped_lock

From: bill_kempf (williamkempf_at_[hidden])
Date: 2002-06-19 13:57:39

--- In Boost-Users_at_y..., Tobias McNulty <tmcnulty_at_d...> wrote:
> At 7:12 PM +0000 6/18/02, bill_kempf wrote:
> >--- In Boost-Users_at_y..., "tmcnulty1982" <tmcnulty_at_d...> wrote:
> >>  boost::recursive_mutex::scoped_lock is crashing with some
> >>  unknown exception in its destructor, no matter whether the lock
> >>  is locked or not.  Has anyone else experienced this problem and
> >>  possibly have a solution?
> >
> >Nope.  Several questions for you.  What compiler/platform?  Have 
you
> >run the regression tests in $BOOST_ROOT/libs/thread/test?  Does the
> >regression test produce the same exceptions?
> >
> >Bill Kempf
> >
> >
> 
> Win2k with Metrowerks CodeWarrior 7.  I ran the regression test, 
> which reported "no errors detected".

I don't have access to Metrowerks, so it's going to be difficult for 
me to determine if this is a compiler specific issue.  However, if 
the regression tests pass, it's not likely.  Can you reduce the code 
to a short example that reproduces the issue?

> There is no requirement that the mutex be used in a boost thread is 
> there?  I'm using CodeWarrior to build this library, but my final 
> application is run by Sun's JRE 1.4.0 and also contains some legacy 
> Visual C++ code.  The mutex is used on a Java thread (which maps to 
a 
> native Windows thread).

No, there's no requirement that the mutex be used in a boost thread.  
This trouble report surprises me, and I hope that together we can 
discover what the issue is.  We'll have to find a reproducable 
example to do so, however.

Bill Kempf