[Bug middle-end/125767] can_div_away_from_zero_p incorrectly depends on can_div_trunc_p succeeding

[Chris.Bazley at arm dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=125767

--- Comment #6 from Christopher Bazley <Chris.Bazley at arm dot com> ---
(In reply to Richard Sandiford from comment #5)
> (In reply to Christopher Bazley from comment #4)
> > it would be possible to work around the issue
> I wasn't suggesting that.  I agree the case should be added.  I was just
> saying that...
> 
> > but I reported it as a bug
> > because the function does not seem to behave according to its documented
> > contract.
> ...this function, like several others in poly-int.h, is "best effort". 
> Returning false does not mean that no solution is possible.
> 
> So the function is behaving according to its contract.  It's just not
> handling all cases in the best possible way.
> 
> This is similar to how, for example, the routines that ask "can these two
> memory accesses alias?" act conservatively.  Returning true (in that case)
> does not mean that no technique exists for telling them apart.  New
> techniques can be added as we need them.

Thanks for your explanation.