Re: [Denemo-devel] The symbol Denemo

[Richard Shann]

On Wed, 2013-07-31 at 15:37 +0200, Éloi Rivard wrote:
> Do you mean it does not happens in 1.0.6 but happens in some more recent
> code of yours ?
yes, 1.0.6 is ok, the latest git is bad
> Can git bisect help there ?
yes, I was hoping you might recall using the name Denemo for something
(in some way, which I don't understand) recently...
> Are you sure you check after the structure is initialized ?
It is initialized statically by the compiler:
main.c first line of code:

struct DenemoRoot Denemo;

(I don't think the C language defines what values this holds, but the
problem is that gdb responds to an enquiry to print the pointer Denemo
with 0, while printing out the value at the start of main() naturally
shows that this is not the value; ultimately, this must be a problem
that gdb is having with the optimizer, but I just wondered if you were
aware of any changes that could have triggered the problem.

Richard


> 
> Le mercredi 31 juillet 2013 à 14:28 +0100, Richard Shann a écrit :
> > Eloi,
> > 
> > I have noticed when running Denemo under gdb that the symbol Denemo
> > (which is the first thing defined in main.c) is in some way eclipsed -
> > gdb ascribes a value 0 to it, and denies the existence of a struct named
> > DenemoRoot.
> > 
> > This is a recent occurrence (it doesn't happen running the 1.0.6
> > release) so I wonder if you know what is triggering this behavior. As it
> > is I can't easily debug, since global variables are all brought together
> > in that structure.
> > 
> > Richard
> > 
> > 
> > 
> > _______________________________________________
> > Denemo-devel mailing list
> > address@hidden
> > https://lists.gnu.org/mailman/listinfo/denemo-devel
> 
>