On Mon, May 2, 2011 at 12:45 PM, William A. Rowe Jr. <wr...@rowe-clan.net> wrote: > On 5/2/2011 8:11 AM, Jeff Trawick wrote: >> Anticipated timeline for release: >> >> * Jeff to review/test wrowe's fnmatch rewrite today (Monday) >> * someone else do the same??? > > Stefan was interested in it. I'm building upon a list of patterns in > test/testfnmatch.c designed to tickle bugs, not looking at the code > again. I'll commit Bert's root dir and my url path 'case-ignore' > patch tonight, and spend just 20 min looking at the apr_file_stat's > behavior w.r.t. new win7 softlinks with and without APR_FSTAT_LINK, > just to make sure we are on the right page. > > I like the single announce idea. > > FYI there are two remaining warnings in apr_fnmatch.c, one is the > assignment within conditional that is trivial to fix (it was paren > delimited so the compiler shouldn't be complaining anyways), but > the other is a fixed loop with no useful break expression, e.g. > for (matchptr = pattern, matchlen = 0; 1; ++matchlen) termination > conditions all had specific additional processing. If there was > a worthwhile alternative to avoid the warning, I didn't see it. > If there are any other warnings at -Wall I don't use the same > compilers as you had; please share.
You fixed one of those in the meantime, and I'm not seeing anything here with my particular flavor of gcc or Sun Studio; clang is happy too. I'm planning to roll 1.4.3 in an hour or so; if the code is ready to ship, great; if having a tarball to decide on brings others out of the woodwork, great :)