On 04.01.2016 20:10, Derek Buitenhuis wrote:
On 1/4/2016 6:05 PM, Ganesh Ajjanagadde wrote:
Personally, I think it should be ok to use rename here for now,
especially since even projects like Python had trouble with this
aspect: https://bugs.python.org/issue8828. Someone with greater
Windows expertise can then examine the validity of
https://stackoverflow.com/questions/167414/is-an-atomic-file-rename-with-overwrite-possible-on-windows#,
where there seems to be a lack of clarity, and perhaps make a better
effort at an atomic rename operation for Windows.
Windows API can lock files from read/write by other procs, which may "emulated"
it... kinda. Probably a bad idea.
We don't really need to discuss this now though, the overall point is
that it uses rename now through ff_rename, and the new call should also
use rename on all the same systems.
rename is used unconditionally in ff_rename as far as I can tell, so it
would probably be OK to remove the unistd.h check in file.c for rename?
--
Hendrik
_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
