Commit graph

322 commits

Author SHA1 Message Date
Nicholas Nethercote
48a0725c01 Measure LocalLayoutContexts in LayoutTask and LayoutWorkers.
The FreeType instances in layout workers are reasonably large and worth
measuring. The one in LayoutTask is smaller but it's easy to measure at the
same time. Sample output:
```
|    8.33 MiB -- pages
|       8.33 MiB -- url(file:///home/njn/moz/servo/../servo-static-suite/wikipedia/Guardians%20of%20the%20Galaxy%20(film)%20-%20Wikipedia,%20the%20free%20encyclopedia.html)
|          1.32 MiB -- layout-worker-0-local-context
|          1.31 MiB -- layout-worker-1-local-context
|          1.24 MiB -- layout-worker-3-local-context
|          1.17 MiB -- layout-worker-4-local-context
|          1.08 MiB -- layout-worker-2-local-context
|          1.06 MiB -- layout-worker-5-local-context
|          0.78 MiB -- paint-task
|             0.78 MiB -- buffer-map
|          0.38 MiB -- layout-task
|             0.30 MiB -- display-list
|             0.07 MiB -- local-context
```
This required adding a mechanism to WorkQueue to measure worker TLSes.
2015-06-10 21:06:59 -07:00
Corey Farwell
4f47b41fa7 Remove fnv & smallvec crate reexports from util
The util component specified fnv and smallvec as dependencies and publicly
reexported both of them. Several other components utilized these reexports,
presumably because fnv and smallvec used to live in the tree so reexporting
made the transition easier.

These indirect dependencies through the util component are unnecessary.

This commit removes the fnv & smallvec crate reexports in the util component.
It exchange, it adds fnv & smallvec as dependencies to non-util components
wherever needed. Finally, it removes the fnv dependency from util as it is not
utilized anywhere in the util component.
2015-06-10 07:14:55 -07:00
bors-servo
0dec64caf0 Auto merge of #6289 - nnethercote:unleak-LOCAL_CONTEXT_KEY, r=pcwalton
`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.)

<!-- Reviewable:start -->
[<img src="https://reviewable.io/review_button.png" height=40 alt="Review on Reviewable"/>](https://reviewable.io/reviews/servo/servo/6289)
<!-- Reviewable:end -->
2015-06-09 15:52:45 -06:00
Manish Goregaokar
af364a412e Audit and reduce unstable usage in layout
Reasons behind existing unstable features:

alloc:

 - `make_unique()`
 - direct calls into `heap::allocate()`, etc
 - `boxed::into_raw()` (naming)

collections:

 - `slice_chars()` (needs to prove its worth)

core:

 - lots and lots of pointer manip
 - `raw` stuff

std_misc:

 - Handle stuff
 - hasher stuff

str_char:

 - CharRange
2015-06-10 00:54:14 +05:30
Eduard Burtescu
00e8e5f77d Use the correct log crate and setup env_logger in main. 2015-06-06 23:20:40 +03:00
Nicholas Nethercote
9b4d39d6d1 Make LOCAL_CONTEXT_KEY safe and non-leaky.
`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.)
2015-06-04 20:57:34 -07:00
Glenn Watson
f47ba6fd33 First step towards 3d transforms.
* Add parser support for 3d transforms.
 * Change ComputedMatrix to a representation that suits interpolation.
 * Switch stacking contexts to use 4x4 matrices.

The transforms themselves are still converted to 2d and handled by azure for now, but this is a small standalone part that can be landed now to make it easier to review.
2015-05-29 09:47:38 +10:00
ecoal95
3350522306 Layerize canvas
Note that this keeps using readback right now, `NativeSurface` painting
will be implemented soon.

Also see https://github.com/servo/servo/issues/6142
2015-05-20 19:10:50 +02:00
bors-servo
c51e9f0455 Auto merge of #6131 - glennw:jquery-exit-fix, r=jdm
This fixes a hang found while testing the jQuery test suite.

<!-- Reviewable:start -->
[<img src="https://reviewable.io/review_button.png" height=40 alt="Review on Reviewable"/>](https://reviewable.io/reviews/servo/servo/6131)
<!-- Reviewable:end -->
2015-05-19 21:44:45 -05:00
Patrick Walton
6a197719b3 compositing: Implement display ports and avoid creating display lists
for items outside it.

This improves Servo's performance on large pages.
2015-05-19 16:53:51 -07:00
Patrick Walton
72f031e2a1 layout: Support inline incremental reflow, and stop reconstructing all
flows when mousing over the document.

This exposes more "jumpiness" on sites like Hacker News, but the bug
that causes it was pre-existing.
2015-05-19 11:08:27 -07:00
Glenn Watson
35a570ab66 Fix several hangs / panics during pipeline cleanup of in progress loads.
This fixes a hang found while testing the jQuery test suite.
2015-05-19 16:14:25 +10:00
bors-servo
5e61ebaa05 Auto merge of #6031 - glennw:reftest-race-conditions, r=larsberg,jdm
The basic idea is it's safe to output an image for reftest by testing:
 - That the compositor doesn't have any animations active.
 - That the compositor is not waiting on any outstanding paint messages to arrive.
 - That the script tasks are "idle" and therefore won't cause reflow.
    - This currently means page loaded, onload fired, reftest-wait not active, first reflow triggered.
    - It could easily be expanded to handle pending timers etc.
 - That the "epoch" that the layout tasks have last laid out after script went idle, is reflected by the compositor in all visible layers for that pipeline.

<!-- Reviewable:start -->
[<img src="https://reviewable.io/review_button.png" height=40 alt="Review on Reviewable"/>](https://reviewable.io/reviews/servo/servo/6031)
<!-- Reviewable:end -->
2015-05-13 18:37:54 -05:00
Glenn Watson
eec3fad93d Fixes a number of race conditions and reliability issues with reftests and compositor.
The basic idea is it's safe to output an image for reftest by testing:
 - That the compositor doesn't have any animations active.
 - That the compositor is not waiting on any outstanding paint messages to arrive.
 - That the script tasks are "idle" and therefore won't cause reflow.
    - This currently means page loaded, onload fired, reftest-wait not active, first reflow triggered.
    - It could easily be expanded to handle pending timers etc.
 - That the "epoch" that the layout tasks have last laid out after script went idle, is reflected by the compositor in all visible layers for that pipeline.
2015-05-14 09:35:53 +10:00
Patrick Walton
711993eb46 gfx: Print out stacking context info when dumping display lists. 2015-05-13 12:00:54 -07:00
bors-servo
2baa69595e Auto merge of #5804 - jdm:docloader, r=Ms2ger
This implements a simple load-tracking system and tracks stylesheet loads as an example of how it fits together. This is a simplified and rebased version of #3714; I do not believe that the main thrust of hsivonen's comments (related to tracking navigation in browsing contexts) affect this part of the work. 

r? @Ms2ger

<!-- Reviewable:start -->
[<img src="https://reviewable.io/review_button.png" height=40 alt="Review on Reviewable"/>](https://reviewable.io/reviews/servo/servo/5804)
<!-- Reviewable:end -->
2015-05-11 14:35:33 -05:00
Josh Matthews
a1ae53a230 Delay stylesheet load in test to increase confidence. 2015-05-11 15:35:10 -04:00
Josh Matthews
f3cdba6b8b Make link elements fire a load event. 2015-05-11 13:41:52 -04:00
Josh Matthews
32a89c9455 Make stylesheets block page load. 2015-05-11 13:41:52 -04:00
Patrick Walton
be8a5a2939 layout: Don't rerun selector matching unless we need to when resizing
the window.
2015-05-07 15:48:14 -07:00
James Gilbertson
138596e861 Constrain the viewport & zoom when reflowing 2015-05-06 00:34:33 -06:00
Simon Sapin
8b522f2e7d Rename Au methods with f32/f64 instead of frac32/frac/subpx 2015-05-05 18:23:29 +02:00
Simon Sapin
ef8edd4e87 Upgrade to rustc 551a74dddd84cf01440ee84148ebd18bc68bd7c8. 2015-05-05 10:07:34 -04:00
bors-servo
5f6a740190 Auto merge of #5913 - glennw:bounding-rect, r=pcwalton
* 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.

<!-- Reviewable:start -->
[<img src="https://reviewable.io/review_button.png" height=40 alt="Review on Reviewable"/>](https://reviewable.io/reviews/servo/servo/5913)
<!-- Reviewable:end -->
2015-05-01 13:15:23 -05:00
Glenn Watson
a5a5214783 Various fixes to getClientBoundingRect()
* 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.
2015-05-01 12:40:37 +10:00
Nicholas Nethercote
092507d23c Add a profile_traits crate to reduce compile times.
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`.
2015-04-30 17:58:47 -07:00
Prabhjyot Singh Sodhi
b980278d90 Uniformise the various Msg types [#5882] 2015-04-30 01:22:16 +05:30
Ms2ger
6a55ae06d7 Remove some as_slice calls. 2015-04-24 17:44:47 +02:00
bors-servo
ac0645c236 Auto merge of #5767 - glennw:image-cache, r=larsbergstrom,jdm
* 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.

<!-- Reviewable:start -->
[<img src="https://reviewable.io/review_button.png" height=40 alt="Review on Reviewable"/>](https://reviewable.io/reviews/servo/servo/5767)
<!-- Reviewable:end -->
2015-04-22 19:16:46 -05:00
Glenn Watson
d8aef7208e Refactored image cache task - details below.
* 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.
2015-04-23 09:40:24 +10:00
Adenilson Cavalcanti
2c7edeb2bf Renaming the color function to just *transparent*,
as suggested on review.
2015-04-22 14:51:39 -07:00
Joseph Seaton
d843526cb8 Animate many more CSS properties 2015-04-16 17:42:57 +01:00
bors-servo
4fac8b6810 Auto merge of #5593 - shinglyu:ismap, r=jdm
This implements issue 4873
2015-04-14 00:10:08 -05:00
Shing Lyu
7a65b95ae5 4873 - Support the image map processing for <img ismap/> inside an <a/> 2015-04-14 10:09:23 +08:00
bors-servo
22e570a06e Auto merge of #5553 - luniv:css-rule-iterators, r=SimonSapin 2015-04-10 19:17:49 -05:00
James Gilbertson
85dcc2deaf Use iterators for iterating through CSS rules 2015-04-07 22:28:06 -06:00
Thomas Jespersen
0c1eeb2fc7 Only send unused buffers messages if there are actually any buffers
Some debugging reveals that the send_back_unused_buffers() quite often
sends empty vectors back to the paint task. This still incurs an
communication overhead though. Instead check that the there actually are
buffers to send back.
2015-04-07 17:52:42 +02:00
Gilles Leblanc
ba36a108c1 Split out shared networking code into net_traits crate
Fixes #4476
2015-04-03 13:38:10 -04:00
aditya
a3eaacccf6 Move iframe and url fields from Reflow to LayoutTask and LayoutTaskFactory
* Fix #5466
2015-04-02 16:09:48 -04:00
Ms2ger
37b5e78a2c Avoid a cast in LayoutImageResponder. 2015-04-02 14:56:05 +02:00
Ms2ger
232dadc81d Use isize for FragmentIndex. 2015-04-02 14:53:10 +02:00
Patrick Walton
750bbed2cb gfx: Perform more aggressive caching in
`FontContext::get_layout_font_group_for_style()`.

There are several optimizations here:

* We make font families atoms, to allow for quicker comparisons.

* We precalculate an FNV hash of the relevant fields of the font style
  structure.

* When obtaining a platform font group, we first check pointer equality
  for the font style. If there's no match, we go to the FNV hash. Only
  if both caches miss do we construct and cache a font group. Note that
  individual fonts are *also* cached; thus there are two layers of
  caching here.

15% improvement in total layout thread time for Facebook Timeline.
2015-04-01 08:58:16 -07:00
bors-servo
8976bbfe7d auto merge of #5441 : servo/servo/heap-unclassified, r=jdm 2015-03-31 15:03:56 -06:00
Patrick Walton
66dd8c8a6c layout: Implement CSS transitions per CSS-TRANSITIONS § 2.
Transition events are not yet supported, and the only animatable
properties are `top`, `right`, `bottom`, and `left`. However, all other
features of transitions are supported. There are no automated tests at
present because I'm not sure how best to test it, but three manual tests
are included.
2015-03-31 08:46:11 -07:00
Nicholas Nethercote
66902d2c90 Use usize rather than u64 in MemoryReport.
I should have used usize in the first place.
2015-03-31 15:37:15 +02:00
Patrick Walton
16db7bab28 layout: Stop going to the DOM to determine the root flow's background
color.

This allows us to do layout off the main thread, which is necessary for
transitions.
2015-03-30 10:00:32 -07:00
bors-servo
39556cc832 auto merge of #5422 : bdero/servo/bdero/box-syntax, r=jdm
Closes #5417
2015-03-29 17:10:01 -06:00
Brandon DeRosier
ccc55d658f Use box syntax instead of Box::new()
Closes #5417
2015-03-29 13:41:14 -04:00
Ms2ger
9904216d76 Use u32 for generation numbers. 2015-03-28 22:46:16 +01:00
Fabrice Desré
0580ad9a82 Suppress unused import warnings 2015-03-25 22:57:27 -07:00