[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 #32 from Bill Spitzak <spitzak at gmail.com> 2012-01-22 19:52:26 PST ---
I don't know if you are being deliberately misleading, but comment #6
explicitly mentions the need for an "EXTEND_ANY" to fix this. I will quote it:

"I think problem (1) is something that should be fixed inside Cairo (and might
be already, I haven't looked at recent changes), either by introducing an
internal EXTEND_ANY value that the backends transform to their fastest extend
mode on demand. Or by having code in the backends that can check if extend
matters for a certain information and then can pick the per-backend/operation
fastest extend mode."

The first statement is exactly what I am proposing. The second statement is a
repeat of the exact same problem everybody is complaining about: the complete
useless test in Cairo when the driver has to repeat the test!

I do not believe this bug should be closed unless one of these happens:

1. Cairo stops testing for clipping and sends the clip unchanged to the backend
(unlikely)

2. A new enumeration called EXTEND_ANY is added to the api between Cairo and
the backend so that the backend can distinguish between EXTEND_NONE and "extend
does not matter"

3. The backend can inform Cairo which mode is fastest and it chooses that
instead of EXTEND_NONE when running the clipping test. I don't like this
because the fastest mode may depend a lot on the current graphics state.

Any solution that requires the backend to run a clipping test that Cairo
already ran is unacceptable.

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