SetPointerMapping protocol braindeadedness

[Keith Packard]

On Mon, 2006-07-10 at 15:17 -0700, Alan Coopersmith wrote:

> Are there any apps other than test suites that depend on SetPointerMapping
> not allowing multiple buttons with the same mapping?

We have had a policy for years that accurate reporting of errors is just
as much a part of the spec as accurately executing valid requests. This
is a stark contrast with many other protocols where syntax and semantics
are arbitrarily extended to fill in former error states. I'd suggest
that applications wanting to use the new semantics should have to ask
for it explicitly, most obviously by using a separate XFixes request.
The other alternative is an XFixes request which 'opens' up the core
requests, but the former seems easier in this case.

> It would seem much more likely they'd depend on GetPointerMapping not
> returning a mapping with buttons that are mapped to the same value,
> so that's where you'ld want to add something to XFixes (perhaps making
> the non-XFixes GetPointerMapping return 0 for those buttons).

Yeah, suitable definition for how existing core requests note the effect
of the changed semantics will also be necessary. This seems as sensible
as anything.

-- 
keith.packard at intel.com
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part
URL: <http://lists.x.org/archives/xorg/attachments/20060710/45c2a186/attachment.pgp>