[Bug middle-end/28545] [4.1 Regression] Wrong code for hoisted multiplication

[rguenth at gcc dot gnu dot org]

------- Comment #7 from rguenth at gcc dot gnu dot org  2006-10-14 09:26 -------
It's a VRP bug :/  Before VRP we have

  D.6958_313 = cr_285 * cr_285;
  D.6959_314 = (size_t) D.6958_313;
  D.6960_315 = cr_285 + 1;
  memset (number_287, D.6960_315, D.6959_314);

and after we get introduced a spurious, uninitialized cr_2738

  D.6958_313 = cr_285 * cr_2738;
  D.6959_314 = (size_t) D.6958_313;
  D.6960_315 = cr_285 + 1;
  memset (number_287, D.6960_315, D.6959_314);

this get's generated by

cr_2738 = ASSERT_EXPR <cr_285, x_1 < cr_285>;


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=28545