[Bug target/83158] [8 regression] gcc.target/i386/pr78057.c fail

[glisse at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=83158

Marc Glisse <glisse at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2017-11-25
     Ever confirmed|0                           |1

--- Comment #1 from Marc Glisse <glisse at gcc dot gnu.org> ---
It is related to PR 83073 but not quite a dup, it looks like another case where
we have more information than we can store in a range, and choose a subset that
hurts in this testcase. The opposite choice would hurt a different testcase,
the issue is finding a reasonable heuristic (say give priority to excluding 0
from the range for instance). The (longer term) plans to have 2+-component
ranges would also help.