[Ffmpeg-devel] libavcodec pcm.c is this correct?

[phil]

line 411 of pcm.c contains a #define for a max buffer size.  This limits the size of the passed buffer that is used. This was introduced in cvs version 1.15 to account for a buffer overflow. This #define is divided by 2, I assume to convert bytes to 16-bit samples, but all the code in this routine seems to work with bytes not samples and applies its own additional scaling to the used buffer size.  Should the divide by 2 be removed?

Phil