[Bug c++/35182] [4.2/4.3 Regression] ICE in coalesce_abnormal_edges

[rguenth at gcc dot gnu dot org]

------- Comment #5 from rguenth at gcc dot gnu dot org  2008-02-14 13:46 -------
That is, somehow we split an edge:

+<bb 6>:
+  goto <bb 7>;
+
 <bb 2>:
-  # iter_1(ab) = PHI <iter_3(D)(0), iter_2(ab)(5), iter_7(ab)(4)>
+  # iter_6 = PHI <iter_7(ab)(4), iter_2(ab)(5)>
+
+<bb 7>:
+  # iter_1(ab) = PHI <iter_6(2), iter_3(D)(6)>
   D.2093_4 = operator new (4);

 <bb 3>:

but mess up somehow.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=35182