Re: texinfo documentation

[Marko Rauhamaa <marko@pacujo.net>]

Kevin Ryde <user42@zip.com.au>:

> Marko Rauhamaa <marko@pacujo.net> writes:
> >
> > I have now changed all the rectangle-inputting procedures to input
> > GdkRectangle's instead of (x,y,w,h).
> 
> That function really wants to be a plain multiple-values thingie, not
> have its x,y,w,h transmuted into a rectangle (if I read the code
> rightly).
> 
> Propose a new routine in the latter style if you have a burning need
> for it, but the function called gdk-window-get-geometry really should
> follow the general conventions.

I found the origin (no pun intended) of my audaciousness:

   gdk-window-get-origin

was written prior to my time, and it returns (x . y) (i.e., GdkPoint)
instead of (x y) as suggested by

   gint gdk_window_get_origin (GdkWindow *window, gint *x, gint *y);

That precedent led to a whole slew of other analogous GdkPoint routines,
and it was only natural to extend the principle to GdkRectangle as well.

While I could back out on my changes, this one would require a higher
authority. Now, gdk-window-get-origin is not included in my SuSE 7.3
release (guile-gtk ver. 0.19). Was it ever released out to the public?
It is a vintage function according to the ChangeLog (2001-02-04).

So it seems to me we are stuck with the deviation. The good side is that
the deviation is more natural for scheme programmers. The bad side is
that we can't simply refer the application developers to the C
documentation. Each guile-gtk function signature needs to be mentioned.


Marko

-- 
Marko Rauhamaa      mailto:marko@pacujo.net     http://pacujo.net/marko/