Re: [Gnash-dev] [PATCH] Initial implementation of a cairo renderer

[Timothy Lee]

strk wrote:
> On Tue, May 02, 2006 at 11:04:47PM +0200, Bastiaan Jacques wrote:
>   
> > On Tuesday 02 May 2006 09:45, Timothy Lee wrote:
> >     
> > >     m_texture_id = (unsigned int)cairo_pattern_create_for_surface(m_image);
> > >       
> > This is a cast from a pointer to 32-bit integer, which is 64-bit unsafe. I've 
> > added an m_pattern member to replace m_texture_id.
> >
> > I've added a configure option --enable-renderer=[renderer], where renderer is 
> > either opengl or cairo (the default is still opengl).
> >
> > This work increases the importance of a cleanup of gnash.cpp (which is already 
> > underway in the form of a frontend library) - when the cairo backend is 
> > enabled, currently two windows are created. Fixing that is nontrivial due to 
> > the goop in gnash.cpp.
> >     
>
> I tried hoping to run gnash on a non-opengl X system, but with no luck.
> Is cairo supposed to work w/out opengl ?
> Currently, gnash fails when it failed before, at SDL_SetVideoMode().
> Should the cairo rendered still use SDL at all ?
>   
gnash.cpp needs to be cleaned up:  cairo renderer doesn't need SDL's 
OpenGL support.  There is an abundance of OpenGL calls there.

Regards,
Timothy