[Xlibs] Janitorial works
Jamey Sharp
jamey@minilop.net
Sun, 16 Nov 2003 11:56:57 -0800
--C7Ke/meiCZutM6I/
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable
These sound like fine plans to me, but what do I know? Still, note that
part of my work with XCL and XCB involves replacing a big chunk of the
protocol implementation functions with automatically generated code.
(You can see how this works by looking at the old XCL code in CVS at
freedesktop.org:/cvs/xcb/xcl/src/proto.m4 and xcbwrap.m4.) So I'd
generally advise against touching too much in the core protocol or
extension protocol implementations.
By the way, while I haven't studied it, I wouldn't be surprised if the
I18N/iconv glue could be generated in a similar manner. I bet other
parts of Xlib could be glued to other libraries using the same trick
too. It has a lot of advantages; Bart and I tried to describe them in
our paper on XCL (see http://xcb.freedesktop.org), so I won't try again
here.
As for my own curiousity about janitorial matters: There is surely dead
code in Xlib. I have convinced myself that _Xbcopy is an example, and
gcc whines because there are unused static functions in the Thai input
method. I'm trying to figure out if I should just strip out things like
that as I see them, or leave them for other people to clean up. My only
hesitation is that I'm already making enough big changes that I have to
test, so it maybe seems unwise to change more. Advice would be
appreciated.
On 11/16 11:15AM, Luciano Montanaro wrote:
> I have been browsing through the X11 sources lately, thinking what kind
> of stuff I could do without damage.
>=20
> Two things came to me:
> - the removal of the _Xconst macro - since the K&R prototypes are going,
> this should go too, or is there any different reason for it?
> - doxygen documentation for the interface functions=20
> (see http://www.doxygen.org). the documentation should already be=20
> available for the man pages. It could be processed with some script,
> and inserted near the function definition.
--=20
Jamey Sharp <jamey@minilop.net> - http://minilop.net/
--C7Ke/meiCZutM6I/
Content-Type: application/pgp-signature
Content-Disposition: inline
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3 (GNU/Linux)
iD8DBQE/t9aJNgAXSpyH6VcRAmEMAJ9QNMngA/JDso5wqrHY23Hbjxo/aQCcDxGU
q0uBwWwobtDx/9QA/AbrwAU=
=X3VE
-----END PGP SIGNATURE-----
--C7Ke/meiCZutM6I/--