[gccsdk] [GCCSDK commit] joty - r2459 - in trunk/gcc: gcc/libgcc unixlib/Docs unixlib/source/clib unixlib/source/clib/bits unixlib/source/clib/unixlib unixlib/source/gcccompat unixlib/source/netlib unixlib/source/pthread unixlib/source/scl unixlib/source/signal unixlib/source/stdlib unixlib/source/sys unixlib/source/unix

Alex Waugh alex at alexwaugh.com
Wed Dec 20 22:49:56 GMT 2006


John Tytgat <John.Tytgat at aaug.net> wrote:

> In message <d4885c984e.ajw498 at caramel.cp15.org>
>           Alex Waugh <alex at alexwaugh.com> wrote:
> > I though I
> > fixed that a while ago by making sure a command window got opened when
> > not in a taskwindow.
> 
> Probably this change you're referring:
> 
> --8<--
> Sat Mar 26 23:32:09 2005  Alex Waugh  <alex at alexwaugh.com>
> 
> 	* source/signal/post.c (post_signal): Open a command window if the
> 	  backtrace would otherwise get lost.
> --8<--
> 
> but that's a different case.  The routine we have here is __unixlib_fatal
> which is the global panic routine where the absolute minimum is done for
> a possible graceful return to calling environment.  Basically setup a stack,
> restore the RISC OS environment handlers, output an error string via
> OS_Write0 (which does not get redirected unless you use RISC OS style
> of redirection) and exit via _exit().  I wished we could have this routine in
> read-only memory because if this gets stuffed, a lockup is almost garanteed.

Yes, that was the code I was thinking of, and I can see now that it
won't get called in this case. It would be simple to add it to
__unixlib_fatal as well (but without the stderr check), although
Wimp_ReportError would be equally as good.

Alex

-- 
Alex Waugh                                           alex at alexwaugh.com

PHP, Roots, Subversion, WebJames and more from http://www.alexwaugh.com/



More information about the gcc mailing list