Closed
Bug 1204824
Opened 9 years ago
Closed 9 years ago
Transform perspective is dependent on devicePixelRatio
Categories
(Core :: Layout, defect)
Core
Layout
Tracking
()
RESOLVED
FIXED
mozilla43
People
(Reporter: cers, Assigned: tnikkel)
References
Details
(Keywords: regression)
Attachments
(2 files)
36.83 KB,
text/html
|
Details | |
846 bytes,
patch
|
bas.schouten
:
review+
Sylvestre
:
approval-mozilla-beta+
|
Details | Diff | Splinter Review |
In Firefox (but not in Chrome or Safari) the transform perspective is dependent on devicePixelRatio.
Steps to reproduce:
1) load attached test case on non-retina screen
2) compare rendering to reference included in test case
3) move tab/window to retina screen
4) repeat 2)
Result:
In Firefox on a retina screen, elements to not line up. They do on non-retina screen.
Expected result:
Rendering should be the same on both screens.
Note: for elements to line up on retina in Firefox (but only Firefox), the perspective needs to be doubled (in this case to 500px) which matches the devicePixelRatio of 2 on a retina screen.
![]() |
||
Comment 1•9 years ago
|
||
Timothy, is this something you'd know about, or someone else?
Flags: needinfo?(tnikkel)
Assignee | ||
Comment 2•9 years ago
|
||
This was caused by bug 1181240 exposing a bug in Matrix4x4. Specifically Matrix4x4::PreScale should be scaling all values in the first three rows, not just the first three columns. That code came from bug 825928 (the layers refactoring). Patch coming up.
Blocks: 1181240, layers-refactoring
Flags: needinfo?(tnikkel)
Assignee | ||
Comment 3•9 years ago
|
||
Assignee: nobody → tnikkel
Attachment #8661957 -
Flags: review?(bas)
Updated•9 years ago
|
Attachment #8661957 -
Flags: review?(bas) → review+
Comment 5•9 years ago
|
||
Status: NEW → RESOLVED
Closed: 9 years ago
Resolution: --- → FIXED
Target Milestone: --- → mozilla43
[Tracking Requested - why for this release]:
A regression in 42, effecting matrix computation on devices with different than 1 pixel ratio. This is across the board on CSS computations and would have far reaching effects.
tracking-firefox42:
--- → ?
Comment on attachment 8661957 [details] [diff] [review]
Patch
Approval Request Comment
[Feature/regressing bug #]: 1181240
[User impact if declined]: Wrong CSS transformations.
[Describe test coverage new/current, TreeHerder]: none
[Risks and why]: low, there was a bug in the matrix computation.
[String/UUID change made/needed]:
The Aurora request is assuming 42 is still in Aurora
Attachment #8661957 -
Flags: approval-mozilla-aurora?
Comment 10•9 years ago
|
||
Comment on attachment 8661957 [details] [diff] [review]
Patch
This missed aurora 42. Moving the flag for beta approval instead.
Attachment #8661957 -
Flags: approval-mozilla-aurora? → approval-mozilla-beta?
Updated•9 years ago
|
Keywords: regression
Comment 12•9 years ago
|
||
Comment on attachment 8661957 [details] [diff] [review]
Patch
Regression, taking it. Should be in 42 beta 2.
Attachment #8661957 -
Flags: approval-mozilla-beta? → approval-mozilla-beta+
Comment 13•9 years ago
|
||
You need to log in
before you can comment on or make changes to this bug.
Description
•