[ogg-dev] [PATCH] skeleton.c

[Conrad Parker]

On 01/02/2008, Conrad Parker <conrad at metadecks.org> wrote:
> On 31/01/2008, ogg.k.ogg.k at googlemail.com <ogg.k.ogg.k at googlemail.com> wrote:
> > This fixes an off by one bug in the user of snprintf, and returns
> > negative if writing the
> > header returns negative (otherwise we'd just get a short write, losing
> > the error).
> > This patch isn't tested though, but I compiled it :)
>
> ok, I'll test it out :-)

I've checked the patch. The check for fwrite returning negative is ok,
but i think the off-by-one check for snprintf is not necessary -- the
intention is not to write a trailing NUL as this function is used to
incrementally add lines of text to a buffer.

Adapted in changeset 3396: http://trac.annodex.net/changeset/3396

> > Also, I've not fixed this (yet ?) but I believe the filling of the ogg
> > packets is bogus as it
> > will fill it with values in whatever endianness the current host has.
> > It should be using
> > oggpack_write and friends instead, I think. I may be missing something though.
>
> eww, you're right, it's broken on big-endian hosts. I'll follow up
> with a patch unless someone else does first ...

I've written up a patch for this, applied to liboggz/src/tools/skeleton.c:

http://trac.annodex.net/changeset/3395

Thank you very much for noticing the problems :-)

Conrad.