[Xcb] client message event strangeness

[Vincent Torri]

>> Another question about events. If an extension has an event (like
>> damage, for example), is that 7th bit also set for the first_event
>> member of the reply structure ?
> Hi Vincent,
>
> I am not quite sure if I understood you right, so maybe this is way
> of:
>
> the 7th (counting from 0) bit is only set, when the event was sent
> using SendEvent. As I understand it, "first_event" describes the
> offset of all the events in the extension. Extensions start to count
> their own events from 0, which would collide with the xproto event 0.
> Therefore an offset is needed (e.g. 35).
>
> "first_event" is remembered on server side whenever an extension is
> added. Since the initialization of extensions is not ordered (or some
> may be missing), the value of first_event may change with each X
> server.

I asked that question because there is that code in ecore_xcb :

(ev->response_type - rep->first_event) != XCB_RANDR_SCREEN_CHANGE_NOTIFY

I gues that it should instead be:

((ev->response_type & ~0x80) - rep->first_event) != 
XCB_RANDR_SCREEN_CHANGE_NOTIFY

Vincent