You can do all reading, seeking and writing yourself and leave only
decoding, encoding to library.

For reading:

first, get the input format with av_probe_input_format

then call init_put_byte to redirect seek and read to your functions

then call av_open_input_stream (instead of av_open_file)

and that's it - now you can open unicode file yourself


For writing:

av_alloc_put_byte (and redirect seek and write to your functions) instead of
url_fopen

Above approach works fine for me.


I used this as an example:

http://ffmpeg.arrozcru.org/forum/viewtopic.php?f=8&t=736

2 notes :

long lSize = 2048; -- I set it to 32768 instead (after inspecting libavcodec
sources)

pAVInputFormat->flags |= AVFMT_NOFILE;  -- wasn't working if I set this, so
I didn't set it at all

hope it helps



On Sun, Mar 8, 2009 at 10:38 AM, LRN <[email protected]> wrote:

> av_open_input_file() does many things, but when it comes down to opening
> files, the file protocol does that with open().
> Unfortunately, open doesn't know anything about unicode, and the best
> thing it can open is a multi-byte character filename. But not all
> multibyte conversions are possible.
> _______________________________________________
> libav-user mailing list
> [email protected]
> https://lists.mplayerhq.hu/mailman/listinfo/libav-user
>
_______________________________________________
libav-user mailing list
[email protected]
https://lists.mplayerhq.hu/mailman/listinfo/libav-user

Reply via email to