Re: [sigc] Forward declaration of signal.h

["Timothy M. Shead" <tshead k-3d com>]

martin-ml hippogriff de wrote:
> > Is there any plans for forward declaring headers for signal and/or slots
> > akin to <iosfwd>?
>
>
> Sorry, I don't understand what you mean by "forward declearing headers"?

See "Exceptional C++" Item 26, and "More Exceptional C++" Item 37.

> > signal.h and the files it includes is hugh and not something you'd
> > willingly add to a header file.

> Add to what header file?

He's saying that if you are defining an API that uses signals, you have
to pull <signal.h> into your API header, where it gets pulled into every
compilation unit that uses that API.  With a special libsigc header
file, one that provides forward declarations for slots/signals/etc,
compile times for programs that use libsigc++ could be improved.

The hypothetical <signalfwd.h> would forward-declare the contents of
<signal.h>, in the same way that <iosfwd> forward-declares <iostream>

Cheers,
Tim