[FFmpeg-cvslog] r25165 - in trunk: Changelog configure doc/filters.texi libavfilter/Makefile libavfilter/allfilters.c libavfilter/avfilter.h libavfilter/vf_frei0r.c
Fri Sep 24 11:41:08 CEST 2010
Stefano Sabatini <stefano.sabatini-lala at poste.it> writes:
>> >> > @@ -1050,6 +1052,7 @@ HAVE_LIST="
>> >> > poll_h
>> >> > setrlimit
>> >> > strerror_r
>> >> > + strtok_r
>> >> > struct_addrinfo
>> >> > struct_ipv6_mreq
>> >> > struct_sockaddr_in6
>> >> I told you not to check for strtok_r. It exists on all POSIX systems,
>> >> and frei0r doesn't work anywhere else (dlopen). Please remove this
>> >> pointless check, and also please respect the review comments in future.
>> > POSIX => dlopen
>> > but not:
>> > dlopen => POSIX
>> > so if dlopen is present you can't infer that strtok_r exists.
>> > And note that we already have a check for strerror_r, which is
>> > supported by POSIX but is not supported in MinGW, following the same
>> > logic it makes sense to keep the check for strtok_r.
>> We check for strerror_r because it causes an actual build failure if
>> we do not. That is not the case for strtok_r. Gratuitous checks only
>> make configure run slower and bloat the build system for no gain.
>> For just about any function we use, no matter how standard, there
>> surely exists some system, somewhere that doesn't have it.
> Check attached.
Just to be sure, does frei0r work on Windows?
>> >> You have changed this substantially from what was reviewed, and not in
>> >> the way I requested. What is the point of reviews if you are going to
>> >> ignore them?
>> > Basically I rewrote it to be more strict with the frei0r specs.
>> Without bothering to respond to my comments, and without posting it
>> for review.
> I believed I addressed to your comments with that change,
You did not.
> and the patch was posted since more than a week (09-13).
Maybe I missed it. You could have reminded me. I've seen you on IRC
several times since then.
mans at mansr.com
More information about the ffmpeg-cvslog