[Bug tree-optimization/77824] unreachable code in SLSR GIMPLE pass

[wschmidt at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=77824

Bill Schmidt <wschmidt at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2016-10-05
     Ever confirmed|0                           |1

--- Comment #5 from Bill Schmidt <wschmidt at gcc dot gnu.org> ---
OK, no problem.

Regarding the "minor nit," c->kind tells us what kind of candidate expression
the statement is equivalent to (a CAND_MULT vs. a CAND_ADD, etc.), but that
isn't sufficient to tell us whether the statement is a copy or something else
that isn't profitable to replace.  So if we have a copy statement S1: Y = X,
where S0 is the statement that defines X, S0 and S1 will both have the same
"kind" and represent the same expression, but S1 is a copy while S0 may not be.
 We might want to replace S1, but we don't want to replace S0.

Confirmed, btw.