Re: [Gnash] gnash testing

[Jens Petersen]

Rob Savoye wrote:
> Jens Petersen wrote:
>
> > Hmmm, pkgconfig doesn't seem an unreasonable dependency to me, 
> > considering that libxml2, gtk2, cairo, pango, glib2, and gtkglext all 
> > support it.
>
>   Right. But those are new dependencies for the GTK version of the 
> player. pkgconfig isn't supported on all platforms though, for instance 
> it would make cross-compiling with MingW or VC++ impossible, or other 
> embedded platforms like eCOS or VxWorks. Gnash isn't just for the 
> desktop...

Ok, fair point. :)

> > Anyway the main problem with current configure file seems to be for 
> > the gtk family of libs.  This is what configure gives on my FC5 x86_64 
> > box:
[snip]
> > Configured paths for x86_64-redhat-linux-gnu are:
[examples of misconfig snipped]
> > The simple pkgconfig hack patch I posted in my original mail at least 
> > makes it build on x86_64 for me on both FC4 and FC5 (and on i386 too 
> > of course).

>   All this has added is the path to the config files, which I've never 
> needed to do to build Gnash on any platforms.

Sorry, not sure what you mean here?  (I was going to comment in the 
previous mail that x86_64 Linux seems an important arch for gnash on 
the desktop.:)

> They should probably be 
> included though, so I'll add them to the macros. I don't mind a hack 
> that'll use pkgconfig if present, and then drop into my own tests if 
> not. I'll look into it when I add these additional include paths.

Anyway, I agree using pkgconfig if available and otherwise falling 
back to something like the current configure tests should make most 
people happy. :)

Thanks, Jens