[vlc-devel] [PATCH] kate decoder: track and use basic formatting

[ogg.k.ogg.k at googlemail.com]

> What if only p_sys->ki.original_canvas_width is known. That might
> result in some very weird w:h ratio's if you blindly take p_vout-
>  >fmt_in.i_height; I think that should be handled with a bit more care.

Yes, but this would be quite unusual, though possible. Any other way
to handle this would probably be not what was intended either. I'd be
inclined to leave this as is, as it's what the stream intends, technically,
Unless you have another suggestion.

> + char *copy = (char*)malloc(ev->len0+strlen(prefix)+strlen(suffix));
> + strcpy(copy, prefix);
> + memcpy(copy+strlen(prefix), ev->text, len0)
>
> is len0 the length of the string or of the data (ergo does it include
> the \0) ?
> I wanna make sure we malloc properly here.

It is the length of the string plus any necessary zero space (always +1
at the moment, but it's there so if UTF-16 is ever supported it might be
+2, or even +4).

> And do we have kate samples anywhere ? I haven't been able to test
> Kate on OSX yes.

Oh, if you're running OSX, I'd be interested to know if it builds, and
a list of warnings, in case something might show up!

For samples, not in the wild yet that I know of. ffmpeg2theora can make
them. For DVD ones, it becomes complex as the patches to GStreamer
and Thoggen are not applied (though you're welcome to try them out,
I use them myself, the diffs are in the libkate source distribution).
If you want to test sample streams, I can send raw streams (eg, without
video as they're, er, large) which you can remux with a random video
of a same size. Mail me privately if you want to do that.

Thanks