avcodec/encode: only allow undersized audio frames if they are the last

Otherwise the user might get a silence padded frame in the beginning or in the
middle of the encoding.

Some other bug uncovered this:

./ffmpeg -loglevel verbose -y -f data -i /dev/zero \
-filter_complex "nullsrc=s=60x60:d=10[v0];sine=d=10[a]" \
-map '[v0]' -c:v:0 rawvideo \
-map '[a]'  -c:a:0 mp2 \
-f mpegts out.ts

Signed-off-by: Marton Balint <cus@passwd.hu>

0 files changed, 0 insertions, 0 deletions