[Bug tree-optimization/77479] [7 Regression] Compile time hog w/ -O2 (-Os)

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=77479

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|memory-hog                  |
             Status|NEW                         |ASSIGNED
   Last reconfirmed|2016-09-05 00:00:00         |
           Assignee|unassigned at gcc dot gnu.org      |rguenth at gcc dot gnu.org
   Target Milestone|---                         |7.0

--- Comment #2 from Richard Biener <rguenth at gcc dot gnu.org> ---
Confirmed and mine.  Bah, still VRP and (OVF) ... (as I suspected)

Visiting PHI node: lm_5 = PHI <lm_12(14), lm_21(9)>
    Argument #0 (14 -> 10 executable)
        lm_12: [-INF, 0]
    Argument #1 (9 -> 10 executable)
        lm_21: [-INF(OVF), 0]
Meeting
  [-INF, 0]
and
  [-INF(OVF), 0]
to
  VARYING
Intersecting
  VARYING
and
  [-INF, 0]
to
  [-INF, 0]
Found new range for lm_5: [-INF, 0]

so the main issue triggering the recent endless iterations is not honoring the
PHI node visiting bail-out via VARYING but using previous pass range-info to
create sth else from it.