Join GitHub today
GitHub is home to over 50 million developers working together to host and review code, manage projects, and build software together.
Sign upWR update #22973
WR update #22973
Conversation
highfive
commented
Mar 5, 2019
|
Heads up! This PR modifies the following files:
|
highfive
commented
Mar 5, 2019
|
If we remove the use of the icon_loading feature (and replace it with code from https://github.com/tomaka/winit/pull/799/files) then can we skip the winit issue? |
|
|
| @@ -2012,6 +2012,7 @@ impl Fragment { | |||
| offset: LayoutVector2D::new(shadow.horizontal.px(), shadow.vertical.px()), | |||
| color: self.style.resolve_color(shadow.color).to_layout(), | |||
| blur_radius: shadow.blur.px(), | |||
| should_inflate: false, | |||
This comment has been minimized.
This comment has been minimized.
emilio
Mar 5, 2019
Member
This should be true (Gecko inflates shadow bounds already to account for the blurs, Servo does not).
| @@ -233,7 +243,9 @@ impl WebRenderDisplayItemConverter for DisplayItem { | |||
| stacking_context.transform_style, | |||
| stacking_context.mix_blend_mode, | |||
| &stacking_context.filters, | |||
| &[], | |||
This comment has been minimized.
This comment has been minimized.
| RasterSpace::Screen, | ||
| /* cache_tiles = */ false, |
This comment has been minimized.
This comment has been minimized.
emilio
Mar 5, 2019
Member
You should probably set this to true only for the root stacking context. @gw3583 can probably confirm that.
This comment has been minimized.
This comment has been minimized.
gw3583
Mar 5, 2019
Contributor
I would probably just leave it as false unconditionally, for now. That will preserve existing behavior. Enabling picture caching could be done as a separate step / follow up.
9cf87d3
to
a40570c
|
@jdm:
Yes, this would have been easier. But at that point I've updated servo and skia to the new winit/glutin versions, so I'd rather go with it. We can remove the dependency on the the icon_loading feature later. |
|
The |
1a7a15f
to
721e68a
|
LGTM |
| let face = font.create_font_face(); | ||
| let files = face.get_files(); | ||
| if files.len() >= 1 { | ||
| if let Some(path) = files[0].get_font_file_path() { |
This comment has been minimized.
This comment has been minimized.
emilio
Mar 8, 2019
Member
This could be written more clearly like:
if self.bytes.is_some() {
return None;
}
let font = ...
let face = ...
let files = ...
let path = files.iter().next()?.get_font_file_path()?;
Some(NativeFontHandle { .. })
| @@ -209,12 +211,20 @@ impl WebRenderDisplayItemConverter for DisplayItem { | |||
| stacking_context.perspective.is_some() | |||
| ); | |||
This comment has been minimized.
This comment has been minimized.
emilio
Mar 8, 2019
Member
I'd remove the assertion since it's now a release assertion effectively right below.
|
winit 0.19/glutin 0.20 are now out. |
1e05b4a
to
9adaa15
|
@bors-servo r=emilio |
|
|
WR update ~Need rust-windowing/winit#803 @emilio can I ask you to look at the *"WR udpate: layout"* commit? There are a few changes I'm not sure about (should_inflate, the new filters_data and cache_tiles). Fix: #22993 <!-- Reviewable:start --> --- This change is [<img src="https://reviewable.io/review_button.svg" height="34" align="absmiddle" alt="Reviewable"/>](https://reviewable.io/reviews/servo/servo/22973) <!-- Reviewable:end -->
|
|
|
I'll go ahead and land this. I had to mark 3 tests as FAIL, but the only one that's actually worrying is the mix-blend test on background behavior. Hope to fix that soon. See servo/webrender#3594 |
|
@bors-servo r=emilio |
|
|
WR update ~Need rust-windowing/winit#803 @emilio can I ask you to look at the *"WR udpate: layout"* commit? There are a few changes I'm not sure about (should_inflate, the new filters_data and cache_tiles). Fix: #22993 <!-- Reviewable:start --> --- This change is [<img src="https://reviewable.io/review_button.svg" height="34" align="absmiddle" alt="Reviewable"/>](https://reviewable.io/reviews/servo/servo/22973) <!-- Reviewable:end -->
|
|
|
@bors-servo retry |
WR update ~Need rust-windowing/winit#803 @emilio can I ask you to look at the *"WR udpate: layout"* commit? There are a few changes I'm not sure about (should_inflate, the new filters_data and cache_tiles). Fix: #22993 <!-- Reviewable:start --> --- This change is [<img src="https://reviewable.io/review_button.svg" height="34" align="absmiddle" alt="Reviewable"/>](https://reviewable.io/reviews/servo/servo/22973) <!-- Reviewable:end -->
|
|
https://phabricator.services.mozilla.com/D20608 Caused problems in Servo servo/servo#22973 and MotionMark - https://bugzilla.mozilla.org/show_bug.cgi?id=1519718 Differential Revision: https://phabricator.services.mozilla.com/D23318 UltraBlame original commit: 84bc38dc1fdeb810790f1b7987df59377c027ae9
https://phabricator.services.mozilla.com/D20608 Caused problems in Servo servo/servo#22973 and MotionMark - https://bugzilla.mozilla.org/show_bug.cgi?id=1519718 Differential Revision: https://phabricator.services.mozilla.com/D23318 UltraBlame original commit: 84bc38dc1fdeb810790f1b7987df59377c027ae9
https://phabricator.services.mozilla.com/D20608 Caused problems in Servo servo/servo#22973 and MotionMark - https://bugzilla.mozilla.org/show_bug.cgi?id=1519718 Differential Revision: https://phabricator.services.mozilla.com/D23318 UltraBlame original commit: 84bc38dc1fdeb810790f1b7987df59377c027ae9
paulrouget commentedMar 5, 2019
•
edited
Need rust-windowing/winit#803@emilio can I ask you to look at the "WR udpate: layout" commit? There are a few changes I'm not sure about (should_inflate, the new filters_data and cache_tiles).
Fix: #22993
This change is