Re: Bison GLR buffer-overflow and invalid pointer fixes

[Frank Heckenbach]

Paul Eggert wrote:

> Paul Hilfinger <address@hidden> writes:
> 
> > Presumably, we should use sigsetjmp/siglongjmp and decide now whether
> > we want the signal context to be restored.  If you contemplate using
> > some other mechanism, then presumably your preference is that the
> > signal mask NOT be reset by failure (which would be my preference).
> 
> Yes, that sounds right to me as well.
> 
> Older systems won't have sigsetjmp/siglongjmp, but glr.c already
> assumes a relatively modern system,

Does it? I remember doing some changes to lower the requirement, but
I'm not sure exactly what's required at the moment.

> so I suppose we needn't worry
> about that.

If not available, one could of course always #define them to
setjmp/longjmp, I suppose.

Frank

-- 
Frank Heckenbach, address@hidden
http://fjf.gnu.de/
GnuPG and PGP keys: http://fjf.gnu.de/plan (7977168E)