[Bug fortran/34566] [4.3 Regression] Matmul of logical values

[tkoenig at gcc dot gnu dot org]

------- Comment #3 from tkoenig at gcc dot gnu dot org  2007-12-23 21:48 -------
Here's a patch.

$ svn diff
Index: matmull.m4
===================================================================
--- matmull.m4  (revision 131146)
+++ matmull.m4  (working copy)
@@ -154,7 +154,7 @@ sinclude(`matmul_asm_'rtype_code`.m4')dn
     {
       astride = a->dim[1].stride * a_kind;
       count = a->dim[1].ubound + 1 - a->dim[1].lbound;
-      xstride = a->dim[0].stride;
+      xstride = a->dim[0].stride * a_kind;
       xcount = a->dim[0].ubound + 1 - a->dim[0].lbound;
     }
   if (GFC_DESCRIPTOR_RANK (b) == 1)
@@ -169,7 +169,7 @@ sinclude(`matmul_asm_'rtype_code`.m4')dn
     {
       bstride = b->dim[0].stride * b_kind;
       assert(count == b->dim[0].ubound + 1 - b->dim[0].lbound);
-      ystride = b->dim[1].stride;
+      ystride = b->dim[1].stride * b_kind;
       ycount = b->dim[1].ubound + 1 - b->dim[1].lbound;
     }


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34566