On Tue, 05 Jun 2012 10:41:45 +0200, Diego Biurrun <[email protected]> wrote:
> On Mon, Jun 04, 2012 at 12:04:50AM +0200, Anton Khirnov wrote:
> > On Sun, 03 Jun 2012 23:56:29 +0200, Diego Biurrun <[email protected]> wrote:
> > > On Sun, Jun 03, 2012 at 11:43:23PM +0200, Anton Khirnov wrote:
> > > > On Sun, 03 Jun 2012 17:53:22 +0200, Diego Biurrun <[email protected]> 
> > > > wrote:
> > > > > They are not really header files, so treating them as templates is 
> > > > > more
> > > > > consistent.  This also fixes generation of Doxygen documentation, 
> > > > > which
> > > > > got confused by #including files within function declarations.
> > > > 
> > > > side note: I'd really appreciate it if you could integrate doxy
> > > > generation into the build system
> > > > so only doxy for the installed headers is generated.
> > > 
> > > I don't think that is a good idea.  While much of our doxy should not
> > > really be doxy, restricting doxy generation to installed headers is
> > > not the right solution.
> > 
> > Care to elaborate?
> > 
> > I think it would be an improvement. Our users are only interested in the
> > public API, having file and symbol lists filled with internal stuff just
> > makes it harder for them to find what they need.
> > Libav devs AFAIK don't read doxy for internal stuff either (or does
> > anyone?).
> 
> Doxygen is not a tool intended for external API documentation;

Oh? Why is that? Seems to me we use it for exactly this purpose.

> it does
> not make a difference where the API is used.  If you want to separate
> internal and external API more clearly, group them appropriately in
> Doxygen.  There is no reason not to use Doxygen for internal APIs.

Sorry, what? Nobody uses generated doxy for internal APIs and they
clutter the output. IMO that is a very good reason not to use doxygen
for internal APIs.

If you want to go through all our files and setup the groups properly,
go ahead. But I think it's a lot of pointless work, so generating the
doxy only for installed headers is a much better solution.

-- 
Anton Khirnov
_______________________________________________
libav-devel mailing list
[email protected]
https://lists.libav.org/mailman/listinfo/libav-devel

Reply via email to