To: jerome.marant at free.fr
Cc: lablgtk at kaba.or.jp
Subject: Re: Using the direct bindings
In-Reply-To: <1046770247.3e647247dd6ac at imp.free.fr>
References: <1046767789.3e6468ad4355a at imp.free.fr>
	<20030304182023J.garrigue@kurims.kyoto-u.ac.jp>
	<1046770247.3e647247dd6ac@imp.free.fr>
Mime-Version: 1.0
Content-Type: Text/Plain; charset=iso-8859-1
Content-Transfer-Encoding: quoted-printable
Message-Id: <20030304184657U.garrigue at kurims.kyoto-u.ac.jp>
Date: Tue, 04 Mar 2003 18:46:57 +0900
From: Jacques Garrigue <garrigue at kurims.kyoto-u.ac.jp>
Lines: 23

From: J=E9r=F4me Marant <jerome.marant at free.fr>

> AFAIK, you don't provide .mli files for most gtk*.ml files because
> they are currently ment to be used by the upper layer only.

Not at all. This is just because they are basically a big bunch of
externals, and the .mli would be remarkably similar to the .ml.
There is almost nothing to abstract in them.

I do not document them actively because it is much simpler to use the
object layer, but if you want to use tham, and even request some
improvements to make them more practical, this should be possible.

> So, ocamlbrowser is capable of production mli files from ml files,
> is that was you meant?

Not exactly: it can display the contents of a .cmi.
It's true that you can then cut and past this contents to a file if
you really prefer it in that form.
For a library the size of lablgtk, I believe you really need to use
something like ocamlbrowser to navigate the types.

Jacques=