[PING][RFC] breakpoint.c: Fix nasty problem with msvcrt DLL on Windows

["Pierre Muller" <pierre.muller@ics-cnrs.unistra.fr>]

  We still didn't take any decision on this patch.
The problem is reported in
http://sourceware.org/bugzilla/show_bug.cgi?id=11886

  I cannot debug anything on one of my Windows PC
because it loads and unloads msvcrt.dll at startup;
which leads to the problem.


Pierre Muller
Pascal language support maintainer for GDB




> -----Message d'origine-----
> De : gdb-patches-owner@sourceware.org [mailto:gdb-patches-
> owner@sourceware.org] De la part de Pierre Muller
> Envoyé : Friday, August 13, 2010 3:50 PM
> À : 'Pedro Alves'; gdb-patches@sourceware.org
> Objet : RE: [RFC] breakpoint.c: Fix nasty problem with msvcrt DLL on
> Windows
> 
> > > > -11     longjmp master keep n   0x77c06d74  inf 1
> > > >
> > > > (gdb) n
> > > > Warning:
> > > > Cannot insert breakpoint -11.
> > > > Error accessing memory address 0x77c06d74: Input/Output error.
> 
> > Your analysis is correct.  But then why did gdb print
> > "Cannot insert breakpoint -11.", with -11 being the
> > number of one of the internal longjmp master breakpoints?
> > Are we copying the breakpoint number when creating the
> > longjmp momentary breakpoints from the master, meaning that's
> > a red herring (though one that I'd like fixed), or is there
> > something else really broken?
> 
> 
>   I was unable to reproduce that issue,
> and I suspect that this is just an error on me side:
> I probably just copied parts of two different gdb sessions.
> 
>   When trying to reproduce it, I always got internal (negative) numbers
> that were different (more negative, meaning created later)
> for the "longjmp" breakpoints as compared to their "longjmp master"
> equivalent.
> 
>   Sorry to have generated that confusion,
> 
> Pierre
> 
>   PS: could you test and confirm the bug report 11886?