PolyText/PolyPoint fixes, mi cleanups, ChangeGC rework

[Keith Packard]

On Sat,  8 May 2010 16:39:15 -0700, Jamey Sharp <jamey at minilop.net> wrote:

> Wow, the rabbit hole just keeps getting deeper. Here are five new
> cleanup patches, two new bug fixes, and an improved version of my
> previous ChangeGC rework patch series.

There is an infinite amount of cleanup work that can be done on almost
any code base...

> So that leaves ProcChangeGC and CreateGC.

Even simpler.

> Not at all. :-) But I think it makes the most sense to let ProcChangeGC
> and CreateGC share the guts of processing the wire-level value-list, so
> I've kept an exposed ChangeGCXIDs in this reworked patch series. It's
> just used only by those two functions.

Yeah, sounds good.

> Can I drop the _X_EXPORT from its prototype, and maybe move it to
> dix/dispatch.h or something? If I did, would you feel better about
> leaving it in?

Yeah, don't stick _X_EXPORT on it.

> Much better. I didn't much like the names I'd picked. :-)

I'm not usually good with names, but this one seemed pretty simple.

> By the way, I think all the ChangeGC callers pass effectively constant
> value-masks. I was considering splitting ChangeGC into a separate setter
> function for each field, plus a FinishChangeGC function to call

Icky. That's a lot of API churn for no particularly good reason. If we
wanted to clean things up, we could use the Xlib form where you fill in
fields in a structure and pass that into the function, but that would
require almost as much change.

Something to consider for later, I think

-- 
keith.packard at intel.com
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
URL: <http://lists.x.org/archives/xorg-devel/attachments/20100508/3ed9345c/attachment.pgp>