Re: "Patch" libdvdread.

Tue Aug 20 2002, Olaf Beck wrote:
> Hello Hakan,
> 
> Thanks for a quick answer,
> 
> --- H}kan Hjort <[email protected]> wrote:
> [snip]
> > > The case is that dvdread doesn't export the whole
> > ifo structure some
> > > small parts are missing. It's offset tables ...
> > 
> > Do you really want them?  I gues they only matter if
> > you want to make
> > minimal non-intrusive edits.  The are just the
> > offsets to were the
> > info is stored.  When you layout the suff in the IFO
> > file you createt
> > just make up new offsets that will work for the data
> > you generate.
> 
> You hit the nail - I want to be able to do minimal
> non-intrusive edits. This is about doing it the right
> way (tm) and preserve as much as possible of a
> structure. In the case that I make very small changes.
> 
Well ok, I guess I don't mind that much... just a tiny fraction
larger memmory foot print for ap using libdvdread.

> > > and the time map in the VTS IFO that I main issues
> > 
> > I've actually just implemented this, see current cvs
> >
> (http://cvs.berlios.de/cgi-bin/viewcvs.cgi/ogle/libdvdread/)
> 
> Ahh, nice and how do I do a anon cvs checkout of the
> latest and greatest :) dvdread. 
>  
Sorry should have pasted that link instead, it's basicaly a European
Sourceforge.  See http://developer.berlios.de/cvs/?group_id=180

> > > although I have also bumped into some incorrect
> > things. 
> > > Anyhow I have attached a patch for ifo_types.h
> > that will "export" all
> > > I need for ifoedit.
> > >
> > The audio extensions stuff wasn't exactly a clean
> > addition.
> > Where is that information derived from?   Having a
> > bunch of pointers
> > in the ifo_handle_t instead of at the end of the
> > vmgi_mat_t and 
> > vtsi_mat_t was to share the common ones (and it
> > looked a bit cleaner).
> 
> Nope it for sure wasn't but I didn't really see where
> to put it without trashing your structres. Which I
> ended up doing after all later on. I got the info from
> mpucoder and ifoedit, and reading DVD Demystified. I'm
> not totally happy with it since there is a lot of
> questions marks at least in my view but it's a start.
>  
I'll take a look at where to add it.
 
[snip]
> 
> I kind of counted that there would be suggestions,
> comments etc. hence my header changes was a one eveing
> quick hack from notes I made during the dev of
> ifoedits "ifo type" headers. My question is are you
> okay with my additions if I make them work? I
> understand that you hold your reservations for the
> audio multi channel extensions so I can skip that one
> for the moment? If so I will download the CVS version
> and start implementing. 
> 

Yeah this should be fine, so what's left?  Two or three 
additional pointers that will hold a reference to a 
memmory area that we now free in libdvdread when we are
done with it.  A bunch of tweeks for the audio / video /
subtitles attributes that where wrong.  A quick fix for
just preserving the TXTDT stuff, and at some later time a
real TXTDT datatype / parser.
Shouldn't be any big problems here I think.

[snip]
> > 
> > We have a few (two or three) DVDs that have TXTDT if
> > you need more examples.
> 
> Thanks for the offer can you possible send me the
> VIDEO_TS.IFO files along with a note about what disk
> you got each from. Ahh, while I'm on it you don't have
> any karaoke DVD's ??
> 
No karaoke DVDs, though I friend of mine tried to force
one on me (ording it delivered to me from Amazon...).  
It (un)fortunately turned out to be just a normal DVD 
with some 'hilighting text' overlayed on the video (not
even spu/subtitles, but in the video) and normal 2.0
audio.

-- 
Håkan Hjort