XC-MISC / X-Resource / core protocol extension proposal

[Adam Jackson]

On Mon, 2009-02-09 at 15:13 -0500, Keith Packard wrote:
> On Mon, 2009-02-09 at 10:17 -0500, Adam Jackson wrote:
> > I can't think of anything obvious this would break.  There's a minor
> > concern about what to do about the server's XID space, but (assuming
> > xrestop isn't lying to me) the server is only holding 67 or so, so
> > probably one block of whatever size is more than sufficient.  Also, yes,
> > this would be a fairly invasive change to the server, and possibly to
> > Xlib, but when has that ever stopped anyone.
> 
> There are a bunch of random debug-ish apps that depend on a stable mask
> to pull a client ID out of a resource ID.

They'll still work, as long as the initial XID block is big enough for
the client.  Empirically, most apps seem to need way less than 64k XIDs
at a time.  This is actually an argument in favor of keeping the same
mask size for 11.0 and 11.1 clients though.

Presumably if you're debugging a sufficiently greedy app, you're willing
to upgrade your debugger.

> One (easier?) alternative would be to take back the top three bits of
> the integer to give us both more clients and more resources?

You're still kind of losing 11.0 compat if you do, CLX definitely
relies on those bits being its to abuse.

- ajax
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 197 bytes
Desc: This is a digitally signed message part
Url : http://lists.x.org/archives/xorg-devel/attachments/20090209/68297e56/attachment.pgp