[cairo-bugs] [Bug 43397] EXTEND_NONE is used instead of EXTEND_PAD when src sample area is entirely within the surface

[bugzilla-daemon at freedesktop.org]

https://bugs.freedesktop.org/show_bug.cgi?id=43397

--- Comment #25 from Andrea Canciani <ranma42 at gmail.com> 2012-01-13 00:15:41 PST ---
(In reply to comment #24)
> It seems wasteful to have another call to distinguish the "any" case from
> "none", as Cairo must have already done the work to figure this out. Unless
> cairo caches this result it will have to run the calculation again.
> 
> Would prefer that a new symbol be added that indicates "any" repeat mode is ok,
> since it is apparent that Cairo figured this out just so it could send "none".
> 
> So rather than a new call I would vastly prefer just an extra enumeration
> value.

That "new enumeration value" would not be within Cairo scope.
It would be a new RENDER protocol value *and* a new value for video drivers.

A much simpler change would be to leave the RENDER protocol intact and only add
it internally to Xorg and the drivers (the computation is quite cheap, you just
have to transform a rectangle and check if it's inside another one).
This would still require changes to all the drivers that make use of it
(obviously) and it would probably require something like an EXA version bump or
something like new caps values (to avoid breaking older driver which do not
support the new enum value), too.

I believe that if the developers of the ati driver think that this change can
provide major performance boost, it might be an interesting suggestion, yet
this would  would better be discussed in xorg-devel.
Are they following this discussion or should they be pinged about it?

-- 
Configure bugmail: https://bugs.freedesktop.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the QA Contact for the bug.