layout: Rewrite whitespace stripping.
This patch makes Servo unconditionally strip whitespace before text run
scanning (assuming that the `white-space` property allows it). Whitespace
stripping during reflow is now only used for handling whitespace at the ends of
lines; reflow now never attempts to handle ignorable whitespace.
Many CSS tests pass now. There are some new failures, however.
The following reference tests now fail due to a pre-existing bug whereby
whitespace is used to calculate the position of inline hypothetical boxes for
elements with `display: inline; position: absolute`:
* `absolute-replaced-height-036.htm`
* `vertical-align-sub-001.htm`
* `vertical-align-super-001.htm`
The following reference tests fail due to a pre-existing bug whereby we don't
handle `font-size: 0` properly in inline reflow:
* `font-size-zero-1.htm`
* `font-size-zero-2.htm`
The following reference test fails due to the fact that it relied on our
incorrect insertion of whitespace to make room for the black background:
* `inline-formatting-context-007.htm`
r? @mbrubeck
<!-- Reviewable:start -->
[<img src="https://reviewable.io/review_button.png" height=40 alt="Review on Reviewable"/>](https://reviewable.io/reviews/servo/servo/7024)
<!-- Reviewable:end -->
This patch makes Servo unconditionally strip whitespace before text run
scanning (assuming that the `white-space` property allows it). Whitespace
stripping during reflow is now only used for handling whitespace at the ends of
lines; reflow now never attempts to handle ignorable whitespace.
Many CSS tests pass now. There are some new failures, however.
The following reference tests now fail due to a pre-existing bug whereby
whitespace is used to calculate the position of inline hypothetical boxes for
elements with `display: inline; position: absolute`:
* `absolute-replaced-height-036.htm`
* `vertical-align-sub-001.htm`
* `vertical-align-super-001.htm`
The following reference tests fail due to a pre-existing bug whereby we don't
handle `font-size: 0` properly in inline reflow:
* `font-size-zero-1.htm`
* `font-size-zero-2.htm`
The following reference test fails due to the fact that it relied on our
incorrect insertion of whitespace to make room for the black background:
* `inline-formatting-context-007.htm`
Snap rectangles to nearest pixels consistently.
Snapping the top-left and bottom-right corners separately can cause a rectangle to change size or even become empty when offset by a subpixel amount. Instead, this patch snaps the top-left corner, then snaps the size to a whole pixel amount, so any rectangle of a given original size will always have the same snapped size.
Fixes#7152. r? @pcwalton or @glennw
<!-- Reviewable:start -->
[<img src="https://reviewable.io/review_button.png" height=40 alt="Review on Reviewable"/>](https://reviewable.io/reviews/servo/servo/7161)
<!-- Reviewable:end -->
Snapping the top-left and bottom-right corners separately can cause a
rectangle to change size when offset by a subpixel amount. Instead, this
patch snaps the top-left corner, then rounds the size to a whole pixel amount,
so any rectangle of a given size will always have the same snapped size.
Fixes#7152.
Known issues:
* Display list optimization can sometimes optimize out elements that
should be shown. This affects the Enyo demo.
* The `overflow: scroll` container doesn't clip the inner layer properly
when borders, border radius, etc. are present.
* `overflow-x: scroll` and `overflow-y: scroll` don't work individually;
elements are scrolled all at once.
* Scrolling only works on absolutely-positioned elements.
new styles are set.
Tying transitions to the DOM node avoids quadratic complexity when
updating them.
Finishing transitions instantly when styles are updated makes our
behavior more correct.
Now GLRasteizationContexts require having an active GLContext. This will
allow preserving GLContexts and possibly framebuffers between
rasterization sessions, improving GL Rasterization performance.
Linux Before:
+ Painting Per Tile 4.5559 4.3392 1.6920 18.5548 74
Painting 170.1554 151.8353 0.0008 350.1093 28
Linux After:
+ Painting Per Tile 3.8726 3.1299 1.5848 12.6732 62
Painting 13.5480 10.8947 0.0029 39.1198 23
To actually make the multiprocess communication work, we'll need to
reroute the task creation to the pipeline or the compositor. But this
works as a first step.
profile: Make the time and memory profilers run over IPC.
Uses a couple of extra threads to work around the lack of cross-process
boxed trait objects.
r? @nnethercote
<!-- Reviewable:start -->
[<img src="https://reviewable.io/review_button.png" height=40 alt="Review on Reviewable"/>](https://reviewable.io/reviews/servo/servo/6629)
<!-- Reviewable:end -->
We currently store LayerBuffers, because previously NativeSurfaces did
not record their own size. Now we can store NativeSurfaces directly,
which saves a bit of space in the surface cache and allows us to create
LayerBuffers only in the PaintTask.
This also means that instead of sending cached LayerBuffers, the
compositor can just send cached NativeSurfaces to the PaintTask.
