On Sun, 11 Jan 2015 17:49:22 -0800 Jim Meyering <j...@meyering.net> wrote:
> On Sun, Jan 11, 2015 at 4:31 PM, Norihiro Tanaka <nori...@kcn.ne.jp> wrote: > ... > > How about the attachments instead for the second patch? > > Thank you for the suggestion. > > However, I do not see a problem with Yuliy's fix, so have pushed it, > along with the other two commits. > > Comparing your change to Yuliy's, I have a slight preference > for his, since it adds work only to the rarely-used code path on > which this bug was introduced, and keeps the handling of > "out of bounds TP" closer to the code that makes TP too large. > > If you can provide justification for this proposed change, > would you please do so in the commit log of a rebased patch? I understood. However, if fill d == 0 before reach in memchr(), even if fill ep <= tp, bm_delta2_search() can be called, and it is not buggy. So It is difficult for me to understand that we must exit the loop if tp <= ep at the point, although I understand that his fix is correct.
