`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.)
This reverts commit 945adabd48.
The CSS Working Group resolved to drop this value from the spec:
http://log.csswg.org/irc.w3.org/css/2015-05-20/#e555680
The group was unable to come up with even a theoretical use case.
Gecko only implemented this value for completeness.
Other browsers vendors have clearly expressed they have no interest
in implementing this.
I've done a bit of job to get this done. Right now readback is still used, but we have a `LayerId` -> `CanvasRenderer` map on the paint task, that we can use to get rid of that.
I'd want review, to see if this is a good approach (I know it's not the initial `CanvasId` -> renderer approach, but it's pretty similar, since a canvas involves a `PaintLayer`).
I had to do a bit of refactoring to avoid cyclic dependencies between canvas and gfx. I'd want you to review them too.
It's mergeable and doesn't break any tests :P
Some of my main concerns:
* Does the canvas render really need to be behind an `Arc<Mutex<T>>`?
* I can't clone a `NativeSurface` right now (that's why the `SendNativeSurface()` msg is unimplemented in the WebGL task). It should be easy to add that to rust-layers, supposing the caller is responsible to mark it as non-leaking, any reason to not do it?
cc @jdm @pcwalton
<!-- Reviewable:start -->
[<img src="https://reviewable.io/review_button.png" height=40 alt="Review on Reviewable"/>](https://reviewable.io/reviews/servo/servo/6083)
<!-- Reviewable:end -->
image fragments.
This also changes the way the placeholder is handled in the image cache
task to decode it up front instead of each time an image fails to load,
both because it was more convenient to implement that way and because
it saves CPU cycles to do so.
This matches the behavior of Gecko and WebKit. It improves the look of
our cached copy of Wikipedia.
absolutely-positioned elements.
This also implements a little bit of the infrastructure needed to
support for fragmentation via support for multiple positioned fragments
in one flow.
Improves Google.
computing the intrinsic widths of the associated fragment.
Fixes sites that use spacer gifs for table layout, such as the comments
page on Hacker News.
Improves the Google SERPs.
We mark `html/rendering/replaced-elements/images/space.html` as failing.
This test tested whether `<img hspace>` and inline margins do the same
thing. Since this was trivially the case before (since we implemented
neither) and now is not, this test now fails.
blocks.
* Stop double-counting border and padding for inline-block fragments.
(Test case: `inline_block_border_intrinsic_size_a.html`.)
* Take clearance into account when determining intrinsic widths of
blocks containing floats.
Improves the Amazon headers.
* 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.
Known issues:
* Collapsed borders do not correctly affect the border-box of the table
itself.
* The content widths of all cells in a column and the content height of
all cells in a row is the same in this patch, but not in Gecko and
WebKit.
* Corners are not painted well. The spec does not say what to do here.
* Column spans are not handled well. The spec does not say what to do
here either.
wrapping on newlines rather than searching for a newline character.
Since the newline character might have been stripped out during
whitespace stripping, this was incorrect.
Fixes the "jumpiness" seen on the Google home page, Wikipedia, and many
other places.
* Simpler image cache API for clients to use.
* Significantly fewer threads.
* One thread for image cache task (multiplexes commands, decoder threads and async resource requests).
* 4 threads for decoder worker tasks.
* Removed ReflowEvent hacks in script and layout tasks.
* Image elements pass a Trusted<T> to image cache, which is used to dirty nodes via script task. Previous use of Untrusted addresses was unsafe.
* Image requests such as background-image on layout / paint threads trigger repaint only rather than full reflow.
* Add reflow batching for when multiple images load quickly.
* Reduces the number of paints loading wikipedia from ~95 to ~35.
* Reasonably simple to add proper prefetch support in a follow up PR.
* Async loaded images always construct Image fragments now, instead of generic.
* Image fragments support the image not being present.
* Simpler implementation of synchronous image loading for reftests.
* Removed image holder.
* image.onload support.
* image NaturalWidth and NaturalHeight support.
* Updated WPT expectations.
text layout, and unify the inline layout paths for pre- and
normally-formatted text.
This fixes a lot of "jumpiness" and removes the `new_line_pos` stuff.
Closes#2260.
