On 1/16/17 3:09 PM, Aaron Sawdey wrote:
Here is an updated version of this patch.
Tulio noted that glibc's strncmp test was failing. This turned out to
be the use of signed HOST_WIDE_INT for handling strncmp length. The
glibc test calls strncmp with length 2^64-1, presumably to provoke
exactly this type of bug. Fixing the issue required changing
select_block_compare_mode() and expand_block_compare() as well.
If glibc's testsuite exposed a bug, then we should also add a similar
bug to our testsuite. I scanned the patch and I'm not sure I see
that exact test scenario. Is it there and I'm just not seeing it?
Peter
