Parallel display list construction hasn't been shown to give any
performance gains. It is also incompatible with the current flat display
list implementation. Once flat display lists have landed, we can explore
possible benefits of parallel construction once again.
This patch does a number of things, unfortunately all at once:
* Hoists a large subset of the layout wrapper functionality into the style system.
* Merges TElementAttributes into the newly-created TElement.
* Reorganizes LayoutData by style vs layout, and removes LayoutDataShared.
* Simplifies the API for borrowing style/layout data.
There's still more to do to make the style system usable standalone, but
this is a good start.
There wasn't a good way to split this up, unfortunately.
With this change, the only remaining usage of the Servo-specific structures is
in layout_task, where the root node is received from the script task. \o/
There is no good reason to have the two types.
This also means that the result of LayoutTask::profiler_metadata no longer
borrows the LayoutTask, which I'll need later.
Use a dedicated type for the opaque type parameter to WorkQueue in layout.
Currently, we use UnsafeFlow and UnsafeLayoutNode, both of which are aliases
for (usize, usize) and thus interconvertible. This change should make it
clearer that the WorkQueue is not limited to one particular type.
<!-- Reviewable:start -->
[<img src="https://reviewable.io/review_button.png" height=40 alt="Review on Reviewable"/>](https://reviewable.io/reviews/servo/servo/6461)
<!-- Reviewable:end -->
Currently, we use UnsafeFlow and UnsafeLayoutNode, both of which are aliases
for (usize, usize) and thus interconvertible. This change should make it
clearer that the WorkQueue is not limited to one particular type.
those instead of working on nodes one-by-one.
This reduces the overhead of the work-stealing traversal function
significantly. It's especially important on ARM, where memory barriers are
expensive.
A rebuild after touching components/profile/mem.rs now takes 48 seconds (and
only rebuilds `profile` and `servo`) which is much lower than it used to be.
In comparison, a rebuild after touching components/profile_traits/mem.rs takes
294 seconds and rebuilds many more crates.
This change also removes some unnecessary crate dependencies in `net` and
`net_traits`.