Re: Last few pieces of GTK+ documentation needed

[Torsten Landschoff <t landschoff gmx net>]

On Thu, Oct 21, 1999 at 09:35:30PM -0700, David Benson wrote:
>=20
> Yes, I'm seeing =3D20 at the end of every line of your post. :)
> But the list itself has just low traffic, I think.

Hmm, I also got that on my post. That does not happen to my posts to=20
the Debian lists though...

> First off, widgets are pretty hard.  So be aware you're
> tackling a tough issue.  Be prepared to spend a _lot_ of time.
> It took me forever to write signals and widgets seems a lot harder.

Hehe, I wanted to write the signal documentation originally. But after=20
I understood how everything worked I learned that the docs are=20
already in place ;)

The code path of gtkwidget.c seems a lot easier to follow than the one=20
of gtksignal.c

> It'll be a great learning experience though.

Sure. It is very interesting to read how gtk was written.=20

> I see some references to "deprecated" in the .h files.
> Note that usually the private functions are marked "Non-public" in
> the .h file.

Yep.=20

> I'm sure people on this, and the dev list, will
> be happy to answer questions.  (I want to thank
> Tim Janik for being extremely helpful in
> proofing the signal docs.)
>=20
> The real trick is that you have to get the whole
> thing into your head, chew on it, and spit it out
> as cleanly and cogently as possible.
> I want to emphasize getting beyond a description
> of each function and primarily instead writing a nice summary.
> If you have that, the function descriptions will
> follow naturally.  Otherwise, they won't be that helpful,
> and they'll prevent others from trying.

Right, you have to find the right approach to reading source. Originally=20
I started reading a function and on the first call I followed into the=20
function called. Now I knew it is better to read the whole beast to get=20
an overall picture and read it again to figure the exact workings out.

cu
    Torsten

PGP signature