751668 - Avoid incorrectly using EXTEND_NONE

Status: RESOLVED FIXED

(Core :: Graphics, defect)

Description

[Jeff Muizelaar [:jrmuizel]]

Attachment: Avoid incorrectly using EXTEND_NONE

It seems like we sometimes end up using EXTEND_NONE unintentionally, this is likely caused by different rounding between cairo and pixman. Pixman already does a similar reduction to EXTEND_NONE so there's not point in duplicating it here.

Comment 1

[:ehsan akhgari]

Comment on attachment 620807 [details] [diff] [review]
Avoid incorrectly using EXTEND_NONE

Review of attachment 620807 [details] [diff] [review]:
-----------------------------------------------------------------

Sorry, I know nothing about this...

Comment 2

[Joe Drew (not getting mail)]

Comment on attachment 620807 [details] [diff] [review]
Avoid incorrectly using EXTEND_NONE

This is totally safe, considering it's just an optimization, but it sure doesn't make me happy.

Don't forget to add it as a cairo patch.

Comment 3

[Jeff Muizelaar [:jrmuizel]]

https://hg.mozilla.org/integration/mozilla-inbound/rev/f561a638fe1e

Comment 4

[Jeff Muizelaar [:jrmuizel]]

Comment on attachment 620807 [details] [diff] [review]
Avoid incorrectly using EXTEND_NONE

[Approval Request Comment]
Regression caused by (bug #): none
User impact if declined: black edges on images
Testing completed (on m-c, etc.): locally and on inbound
Risk to taking this patch (and alternatives if risky): Possible performance regression, but very unlikely. If there is a performance regression, it is likely to only effect mobile.
String changes made by this patch: none

Comment 5

[Ed Morley [:emorley]]

https://hg.mozilla.org/mozilla-central/rev/f561a638fe1e

Comment 6

[Brad Lassey [:blassey] (use needinfo?)]

(In reply to Jeff Muizelaar [:jrmuizel] from comment #4)
> Comment on attachment 620807 [details] [diff] [review]
> Avoid incorrectly using EXTEND_NONE
> 
> [Approval Request Comment]
> Regression caused by (bug #): none
> User impact if declined: black edges on images
> Testing completed (on m-c, etc.): locally and on inbound
> Risk to taking this patch (and alternatives if risky): Possible performance
> regression, but very unlikely. If there is a performance regression, it is
> likely to only effect mobile.
> String changes made by this patch: none

let's let this bake on trunk for a couple nightlies

Comment 7

[JP Rosevear [:jpr]]

Comment on attachment 620807 [details] [diff] [review]
Avoid incorrectly using EXTEND_NONE

Baked on on central, giving approval.

Comment 8

[JP Rosevear [:jpr]]

Backfilling beta+ so its on our radar.

Comment 9

[Jeff Muizelaar [:jrmuizel]]

https://hg.mozilla.org/releases/mozilla-aurora/rev/e73465a6bfcb

Comment 10

[JP Rosevear [:jpr]]

Aurora is actually 14.