[Ffmpeg-devel] [PATCH] support for vorbis language in native ogg demuxer

Rich Felker dalias
Sat Jun 24 06:44:24 CEST 2006


On Thu, Jun 22, 2006 at 08:25:30PM +0100, M?ns Rullg?rd wrote:
> Rich Felker <dalias at aerifal.cx> writes:
> 
> > On Thu, Jun 22, 2006 at 06:54:08PM +0100, M?ns Rullg?rd wrote:
> >> Rich Felker <dalias at aerifal.cx> writes:
> >> 
> >> > On Wed, Jun 21, 2006 at 12:49:28AM +0200, Nico Sabbi wrote:
> >> >> $Subject says everything, but there's to notice that (as for everything
> >> >> that has to deal with stuff done by xiph) there are hideous shortcomings:
> >> >> - the language is identified in a non-standardized way, so it's
> >> >> longer than 3 characters and in a mixed-case format (e.g. English,
> >> >> Japanese, etc.)
> >> >> - I didn't strtolower() them because afaik this depends on the charset used
> >> >
> >> > Charset is ASCII [A-Za-z]. Just or each byte with 0x20, IMO. :)
> >> 
> >> Vorbis comments are utf8.  Furthermore, there is no official field for
> >> language.  Thus we must assume that a language field, if present,
> >> might contain anything.
> >
> > However, anything but established language names will be meaningless,
> > so does it matter if we mangle invalid data in the field?
> 
> Randomly flipping bits in utf8 data is not a good idea.  With bad
> luck, the result won't be valid utf8, and even if it is valid, it
> could be anything.

GIGO

Rich





More information about the ffmpeg-devel mailing list