[Bug fortran/32027] [4.3 Regression] The "DO" variable isn't consistent after loop

[FX Coudert]

Attached is the patch I committed to mainline after regtesting on  
i686-linux, under the "obvious" rule. I just didn't took care of the  
final value of the loop counter when I initially wrote this patch. I  
do apologize for the inconvenience.

FX

-------------- next part --------------
A non-text attachment was scrubbed...
Name: loop-regress.diff
Type: application/octet-stream
Size: 8752 bytes
Desc: not available
URL: <http://gcc.gnu.org/pipermail/fortran/attachments/20070521/4afdb61b/attachment.obj>