[RFC PATCH] damageproto: Explicitly document Damage lifetimes

[Adam Jackson]

On Thu, 2012-03-29 at 09:54 -0700, Keith Packard wrote:
> On Thu, 29 Mar 2012 12:02:28 -0400, Adam Jackson <ajax at redhat.com> wrote:
> 
> > One possibility would be to have a 1.2 version of the protocol that
> > simply doesn't generate BadDamage in any event; or, that tracks client
> > timestamps and suppresses BadDamage if the drawable has been destroyed
> > since the last Damage request; or, that generates DamageDestroyed events
> > as well.
> 
> XIDs get reused, so you might end up talking about an unexpected damage
> object, which means we need the client to be able to know the damage
> lifetime. Either a damage destroyed event, a pixmap destroyed event or
> having the damage live until explicitly destroyed would all work.

Changing the lifetime would mean needing to change the extension major
number, since otherwise currently legal 1.x clients would leak
themselves to death, and that's not nice.  And then you'd basically have
to support both at runtime anyway, and boo to that.

DamageDestroyed would be 1.2 not 2.0, and gives you "PixmapDestroyed"
for nearly free.  Completely free if we add DamageReportLevelNone.

- ajax
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 198 bytes
Desc: This is a digitally signed message part
URL: <http://lists.x.org/archives/xorg-devel/attachments/20120329/11dff240/attachment.pgp>