[Bug tree-optimization/81297] [8 Regression] ICE in get_single_symbol

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=81297

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |ASSIGNED
           Assignee|unassigned at gcc dot gnu.org      |rguenth at gcc dot gnu.org

--- Comment #4 from Richard Biener <rguenth at gcc dot gnu.org> ---
Hmpf.  So the assert was to make sure we do not introduce OVF stuff in symbolic
ranges (it should better have been in set_value_range).  But yes, generic
folding is susceptible to introduce it, esp. with -fwrapv :/

The easiest "fix" is to remove the assert.  But OTOH the best fix would be to
stop setting TREE_OVERFLOW alltogether... or at least for where overflow
is (implementation) defined.

I will eventually look into this after I return from vacation.