Re: [boost] Re: ?python: Throwing an error within a class init causes a segment fault

From: David Abrahams (abrahams_at_[hidden])
Date: 2001-05-18 18:28:24

I have been in touch with the GCC developers at Cygnus/Redhat about using an
exception-handling testsuite I wrote to validate GCC 3.0, which they have in
hand. They were receptive, but it definitely sounded like it wasn't being
treated as a high-priority item. If you care about this, probably the best
thing you can do is to make noise in some forum that the GCC developers
listen to.

-Dave

----- Original Message -----
From: "Thomas Matelich" <sosedada_at_[hidden]>
To: <boost_at_[hidden]>
Sent: Friday, May 18, 2001 4:32 PM
Subject: Re: [boost] Re: ?python: Throwing an error within a class init
causes a segment fault

> "Norman Shelley (rrdn60)" wrote:
>
> > >> We had problems like this with gcc before. In general, the  gcc
exception handling
> > >> seems to be unreliable. Have you tried your code on a different
platform?
> > >> Ralf
> > Nope. This code only has to run on HPUX systems. There is no desire to
> > try other platforms.
> > Norman
>
> Ooh, I'm sorry.  In my experience, on HP-UX with gcc, an exception thrown
and caught within a
> shared library causes a crash.  An exception escaping a shared library
actually has a slightly
> better shot at working.  This is the main reason I began investigating HP
aCC for our
> projects.  Unfortunately, at this time aCC is also unsatisfactory for
other reasons.  An aCC
> support contract has been purchased for me at work and I will be
investigating more
> workarounds when I receive it.  Right now, my fingers are crossed for gcc
3.0 (supposedly due
> out June 15).
>
> --
> Thomas O Matelich
> Senior Software Designer
> Zetec, Inc.
> sosedada_at_[hidden]
> tmatelich_at_[hidden]
>
>
>
>
> To unsubscribe, send email to: <mailto:boost-unsubscribe_at_[hidden]>
>
>
> Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/
>
>