[FFmpeg-user] ebuR128 display of live source

[Carl Zwanzig]

On 3/19/2023 3:52 AM, Phil Rhodes via ffmpeg-user wrote:
> I could be wrong here as I'm trying to remember something from eight or
> ten years ago, but I seem to recall having a similar delay problem
> parsing standard output from a windows commandline application of my own
> devising. The reason I mention it is that the problem ultimately wasn't
> the application itself but the standard I/O handling in Windows, which
> was buffering some stuff up internally under at least some
> circumstances.

Although... IIRC the fflush() call works as expected even on windoze. One 
could argue that most writes to stderr should be followed by that, at least 
for things that might be considered close to real-time.

(Without spending a lot of time or untangling things, it looks like the log 
message from f_ebur128.c is written by av_log in libavutil/log.c, and that 
doesn't contain any flushes. Could also be that I missed something.)

Later,

z!