[Xcb] [PATCH] Don't set full sequence for ge events

[Jamey Sharp]

On Fri, Feb 22, 2013 at 05:15:37PM +0100, Daniel Martin wrote:
> On Fri, Feb 22, 2013 at 07:54:25AM -0500, Peter Harris wrote:
> > That's not a show-stopper. Given that the first thing every user does
> > is memmove the full_sequence out, or the equivalent, I'd say it's a
> > good thing. But we have to be careful to bump an ABI version in the
> > autofoo.
> 
> Okay, if we would do this change, would it be possible to break even
> more? I'm thinking of this bug:
>     https://bugs.freedesktop.org/show_bug.cgi?id=6105
> where Jamey gave additional comments:
>     http://lists.freedesktop.org/archives/xcb/2013-February/008172.html
> Then we could simplify a few things.
> 
> If that would realy be possible I will sum up what I would like to do,
> i.e. change
>     xcb_poll_for_event(xcb_connection_t *c)
> to
>     xcb_poll_for_event(xcb_connection_t *c, unsigned long *sequence)
> (or directly with uint64_t).

I wish we could, but that's way too big of a change. Also, don't use
long for new XCB API; uint64_t is right.

I'd suggest fixing both problems the way I suggested before: introduce
new API that returns sane events, and hack up the old API to memmove the
end of events and insert full_sequence.

Jamey
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 198 bytes
Desc: Digital signature
URL: <http://lists.freedesktop.org/archives/xcb/attachments/20130222/90cb3e13/attachment.pgp>