[Bug middle-end/78501] [7 regression] SEGV in vrp_val_max

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=78501

--- Comment #6 from Richard Biener <rguenth at gcc dot gnu.org> ---
(In reply to Jakub Jelinek from comment #4)
> Even for Fortran and other non-C FEs I think ptrdiff_type_node isn't what
> you expect it to be.  Wouldn't it be better to use signed_type_for (type) as
> the type instead (where type is the type of strlen's lhs, i.e. usually
> size_t)?

As ptrdiff_type_node is appearantly used (otherwise we'd not have the LTO
variant) it should be a middle-end defined type (it's defined by the target
via PTRDIFF_TYPE).