Because this is a bottom-up traversal it can generates flows and throw them away. To prevent that, this cascades an internal `-servo-under-display-none` property and then checks that during flow construction. Fixes#1536.
r? @pcwalton
Source-Repo: https://github.com/servo/servo
Source-Revision: 92ae502765d8b6028db57c2d2ff5cda886ab9439
Partial implementation of the issue #10692 (the easy part).
Source-Repo: https://github.com/servo/servo
Source-Revision: d926b5d3762a101d5280266f806f2b979f73b695
The old code tried to do the speculation as a single bottom-up pass
after intrinsic inline-size calculation, which was unable to handle
cases like this:
<div>
<div style="float: left">Foo</div>
</div>
<div>
<div style="overflow: hidden">Bar</div>
</div>
No single bottom-up pass could possibly handle this case, because the
inline-size of the float flowing out of the "Foo" block could never make
it down to the "Bar" block, where it is needed for speculation.
On the pages I tried, this regresses layout performance by 1%-2%.
I first noticed this breaking some pages, like the Google SERPs, several
months ago.
r? @mbrubeck
Source-Repo: https://github.com/servo/servo
Source-Revision: 1554331f06900e69f246ed9986a08aae91a0a71e
This reduces CPU usage when mousing over simple pages (example.com). More complex pages (Wikipedia) still reflow a lot due to other bugs.
Additionally, this change causes Servo to stop painting the results of hit test queries. This is also a win for CPU usage.
This significantly improves #9999, though there's more that can be done. I'll leave it open in case @paulrouget thinks this PR isn't enough.
r? @mbrubeck
Source-Repo: https://github.com/servo/servo
Source-Revision: 9813d11f862a61a7cffb4cf2d7de0fa7d269c62a
Right now, the only reason that overflow calculation works is that we
rely on script inducing extra reflows that are sent for display. This
was preventing #10021 from landing.
This change regresses layout performance by about 1% in my tests.
Fixes#7797 properly.
r? @mbrubeck
cc @glennw
Source-Repo: https://github.com/servo/servo
Source-Revision: 8120fb5c2d170d5d9838696a7a5b871a43f57bb2
Right now, there's a huge amount of complexity in T{Node,Element,Document} and friends because of the lifetime parameter.
Before I started generalizing this code for use by Gecko, these wrappers were plain structs. They had (and still have) a phantom lifetime associated with them to prevent references to DOM nodes from leaking past the end of restyle, when they might be invalidated by a GC.
When I generalized them, I decided to put the lifetime on the trait as well, since there are some situations where the lifetime is, in fact, necessary. Specifically, they are necessary for the compiler to understand that all the things borrowed from all the nodes and elements and so on have the same lifetime (the lifetime of the restyle), rather than the lifetime of whichever particular element or node pointer the value was borrowed from. This come up in situations where we do |let el = node.as_element()| or |let n = el.as_node()| and then borrow something from the result. The compiler thinks the borrow lifetime is that of |el| or |n|, when it's actually longer.
In practice though, I think the style and layout algorithms we use don't run into this issue much, and we can hack around it where it comes up. So I think we should remove the lifetimes from the traits, which will let us aggregate the embedding-provided traits together onto a single meta-trait and significantly simplify the code.
Source-Repo: https://github.com/servo/servo
Source-Revision: aea8d8959dcb157a8cc381f1403246ce8ca1ca00
Flat display lists were a 2x regression on the spheres demo. This patch series fixes that.
See the individual commits for more details.
r? @mrobinson
Source-Repo: https://github.com/servo/servo
Source-Revision: 55fc48e4c46917a0f036d0054fac296bb5719434
Instead of producing a tree of stacking contexts, display list
generation now produces a flat list of display items and a tree of
stacking contexts. This will eventually allow display list construction
to produce and modify WebRender vertex buffers directly, removing the
overhead of display list conversion. This change also moves
layerization of the display list to the paint thread, since it isn't
currently useful for WebRender.
To accomplish this, display list generation now takes three passes of
the flow tree:
1. Calculation of absolute positions.
2. Collection of a tree of stacking contexts.
3. Creation of a list of display items.
After collection of display items, they are sorted based upon the index
of their parent stacking contexts and their position in CSS 2.1
Appendeix E stacking order.
This is a big change, but it actually simplifies display list generation.
Source-Repo: https://github.com/servo/servo
Source-Revision: 62814f7cb486bc267a796b7ce58c51d59240fad0
This allows, among other things, having different implementations for parsing pseudo{elements, classes} in both `ports/geckolib` and in servo.
Source-Repo: https://github.com/servo/servo
Source-Revision: c11844cbf28054784c8d65781cff20045d8ee48b
This should be the last major hoisting necessary to use the style system standalone. \o/
Source-Repo: https://github.com/servo/servo
Source-Revision: ddc3b7942eea8328e9eb22d864d34fce572d5535
This is a step towards removing the dependency of stylo on layout/.
This PR depends on #9004.
Source-Repo: https://github.com/servo/servo
Source-Revision: 0f5c614609fd8867a9e7c27b8a398ea7d877c714
This gets us to where we need to be in order to write a Gecko implementation of the layout wrapper and have things Just Work.
Note that this is somewhat more than we need for just running the style system. But there wasn't a clean place to cut, and I thought it was a good idea to just do a complete job on it now, which may save us effort and mismatches in the long run.
Source-Repo: https://github.com/servo/servo
Source-Revision: 1f732cfbb404bfd96c3ed3f71c4b9900b7ffbba2
BUBBLE_ISIZES and REPAINT can become "stuck" on in the default Servo
configuration once they are activated. This is solved by removing these
damage bits after they no longer apply. There isn't a good way to test
this, other than noting that it doesn't break any existing CSS tests.
This will become more important in the future as the REPAINT bit is used
to implement display list patching.
Source-Repo: https://github.com/servo/servo
Source-Revision: 8b6bfb615a8a04177c321c4940f79f090b0a7971
There's still more refactoring to do, but this is the core stuff that's most likely to bitrot.
Source-Repo: https://github.com/servo/servo
Source-Revision: a5babb89a02b9b84a8cd62554a5ceef9efb0d481
Also include absolutely positioned elements in the overflow rect calculation.
Fixes#7797.
Source-Repo: https://github.com/servo/servo
Source-Revision: ca56ebbb09f3c258d10e7a7fa276d42fe258d893
This isn't doing anything right now, and we're not even setting it properly
in dirty_impl the |dirty_subtree(self)| was causing us to hit the skip case
for step 3.
Source-Repo: https://github.com/servo/servo
Source-Revision: 50ec2353845bf2a3971d5b01db37d2c3741d3912
(#[cfg(debug)] is false in every normal servo configuration, and the
code in question doesn't compile.)
Source-Repo: https://github.com/servo/servo
Source-Revision: d4f5948294545860a88d6a266acaa4cf9b9ea1c6
This is a direct extract from my abandoned PR for a lint (#7546), along with some rather clumsy modifications (only on `components/script/dom/mod.rs` and `components/style/lib.rs`), because I had to sort some of the files again to make peace with tidy, which hasn't been educated about sorting yet!
Source-Repo: https://github.com/servo/servo
Source-Revision: a7208869f2903e36f9b2f540b55b50283d7df466
This necessitated changing overflow to be calculated by the parent flow
if relatively positioned children are present. That is because the
overflow regions cannot be calculated without knowing relative offsets,
which themselves cannot be calculated without knowing the parent size
(because of percentages). To accomplish this without sacrificing
parallelism in the non-relative case, this patch splits overflow into
"early" and "late" computation. Late overflow computation cannot be
parallelized across children, while early overflow computation can.
Makes the "Apple Music" text show up over the full-bleed promotional
background on apple.com.
r? @SimonSapin -- would appreciate a look over the iframe test case that was changed.
Source-Repo: https://github.com/servo/servo
Source-Revision: dcaf66397a06246b9b4fdca3a10af1508a11f1e8
`LOCAL_CONTEXT_KEY` is currently a `Cell<*mut LocalLayoutContext>`. The use
of the raw pointer means that the `LocalLayoutContext` is not dropped when
the thread dies; this leaks FreeType instances and probably other
things. There are also some unsafe getter functions in `LayoutContext`
(`font_context`, `applicable_declarations_cache` and
`style_sharing_candidate_cache`) that @eddyb says involve undefined
behaviour.
This changeset changes `LOCAL_CONTEXT_KEY` to
`RefCell<Option<Rc<LocalLayoutContext>>>`. This fixes the leak and also
results in safe getters.
(Fixes #6282.)
Source-Repo: https://github.com/servo/servo
Source-Revision: 0dec64caf01c98d10e72b73e35b994127c23e81f
This improves Servo's performance on large pages.
Please double-check the logic when it comes to nested layers—I'm sure I've messed up some of the geometry calculations :)
r? @glennw
Source-Repo: https://github.com/servo/servo
Source-Revision: 0880e54f987bac7c34c934ef6ee36f46475b06e3
Makes qz.com visible.
In order to work around a compiler bug involving Sized, this patch moves
`store_overflow` to be a virtual method.
r? @glennw
Source-Repo: https://github.com/servo/servo
Source-Revision: 5a13cae064c4418923fe35bb7e07c71f5bfd851e
* Fix queries involving stacking contexts
* The code was double accumulating stacking context origins.
* Handle queries of inline elements.
* The node addresses being compared were incorrect (CharacterData vs. Span)
* Handle ScriptQuery reflows correctly.
* The layout task was skipping the compute absolute positions traversal, so failed before window.onload.
Source-Repo: https://github.com/servo/servo
Source-Revision: 5f6a740190e1e5912d84162c92c6b79365df165a
