…s for frame treese in script tasks.
This underreports by a significant amount, since only Document, Window and CharacterData (ie. text) nodes are fully represented. That being said, every HTML element in the tree is measured, but only counted as a Node. It's easy to improve this, it just requires adding the appropriate HeapSizeOf derives and increasing the granularity of `measure_memory_for_eventtarget`. google.com shows a dom-tree value of 0.24 MB for me at the moment.
r? @nnethercote
Source-Repo: https://github.com/servo/servo
Source-Revision: 84e25befdd97cf74fb00707dbe150d59d980e977
For now, this just gives some purpose to the abandoned receiver and later selects over the two receivers (for #6767). (oh wait, forgot to check the local build - there are still a few errors)...
Source-Repo: https://github.com/servo/servo
Source-Revision: effb17b3368769baaf5247908b26b7f0dd370ea3
This change makes Servo use serialized messages over IPC channels for resource loading. The goal is to make it easier to make Servo multiprocess in the future. This patch does not make Servo multiprocess now; there are many other channels that need to be changed to IPC before that can happen. It does introduce a dependency on https://github.com/serde-rs/serde and https://github.com/pcwalton/ipc-channel for the first time.
At the moment, `ipc-channel` uses JSON for serialization. This is because serde does not yet have official support for bincode. When serde gains support for bincode, I'll switch to that. For now, however, the JSON encoding and decoding will constitute a significant performance regression in resource loading.
To avoid having to send boxed `AsyncResponseTarget` trait objects across process boundaries, this series of commits changes `AsyncResponseTarget` to wrap a sender only. It is then the client's responsibility to spawn a thread to proxy calls from that sender to the consumer of the resource data. This only had to be done in a few places. In the future, we may want to collapse those threads into one per process to reduce overhead. (It is impossible to continue to use `AsyncResponseTarget` as a boxed trait object across processes, regardless of how much work is done on `ipc-channel`. Vtables are fundamentally incompatible with IPC across mutually untrusting processes.)
In general, I was pretty pleased with how this turned out. The main changes are adding serialization functionality to various objects that `serde` does not know how to serialize natively—the most complicated being Hyper objects—and reworking `AsyncResponseTarget`. The overall structure of the code is unchanged, and other than `AsyncResponseTarget` no functionality was lost in moving to serialization and IPC.
r? @jdm
Source-Repo: https://github.com/servo/servo
Source-Revision: 2eb122f394651232abd683fc576a5c4288bf277f
Credits for Mike Blumenkrantz (@zmike), I just rebased against trunk and
fixed the url serialization.
Fixes#6178.
Source-Repo: https://github.com/servo/servo
Source-Revision: 82e476fd07564f2822a5af1f397751027ddf7258
This prevents them from piling up endlessly if the JS
`requestAnimationFrame()` callback takes a long time to complete.
r? @jdm
Source-Repo: https://github.com/servo/servo
Source-Revision: 12c46ddd6131533454e353fa8fdd8cb4b517ffd0
This is used for two memory reporting improvements.
- It's used to distinguish "explicit" memory reports from others. This
mirrors the same categorization that is used in Firefox, and gives a single
tree that's the best place to look. It replaces the "pages" tree which
was always intended to be a temporary stand-in for "explicit".
- It's used to computed "heap-unclassified" values for both the jemalloc
and system heaps, both of which are placed into the "explicit" tree.
Example output:
```
| 114.99 MiB -- explicit
| 52.34 MiB -- jemalloc-heap-unclassified
| 46.14 MiB -- system-heap-unclassified
| 14.95 MiB -- url(file:///home/njn/moz/servo2/../servo-static-suite/wikipe
dia/Guardians%20of%20the%20Galaxy%20(film)%20-%20Wikipedia,%20the%20free%20encyc
lopedia.html)
| 7.32 MiB -- js
| 3.07 MiB -- malloc-heap
| 3.00 MiB -- gc-heap
| 2.49 MiB -- used
| 0.34 MiB -- decommitted
| 0.09 MiB -- unused
| 0.09 MiB -- admin
| 1.25 MiB -- non-heap
| 1.36 MiB -- layout-worker-3-local-context
| 1.34 MiB -- layout-worker-0-local-context
| 1.24 MiB -- layout-worker-1-local-context
| 1.24 MiB -- layout-worker-4-local-context
| 1.16 MiB -- layout-worker-2-local-context
| 0.89 MiB -- layout-worker-5-local-context
| 0.38 MiB -- layout-task
| 0.31 MiB -- display-list
| 0.07 MiB -- local-context
| 1.56 MiB -- compositor-task
| 0.78 MiB -- surface-map
| 0.78 MiB -- layer-tree
```
The heap-unclassified values dominate the "explicit" tree because reporter
coverage is still quite poor.
Source-Repo: https://github.com/servo/servo
Source-Revision: bff5e325a89ab6621a049ac55c1da66e901c776c
This necessitated getting rid of the boxed trait object that was being
be passed between the script task and the image cache task.
r? @jdm
Source-Repo: https://github.com/servo/servo
Source-Revision: e13ebf712de444132a6cc90f394c121d8d751c4c
Uses a couple of extra threads to work around the lack of cross-process
boxed trait objects.
r? @nnethercote
Source-Repo: https://github.com/servo/servo
Source-Revision: f778e0eecf7cd8a2b870d18c3c305ff10d6b1894
- Use SmallVec<[T; N]>
- Make find_iframe a free function
- Make ProgressEvent use enums for bubbles and cancelable
- Change README, as `rust-snapshot-hash` is just a text file
Source-Repo: https://github.com/servo/servo
Source-Revision: aafc3dfa963b466303d5f241d69036f211aaad00
SpiderMonkey provides an extremely fine-grained breakdown of memory
usage, but for Servo we aggregate the measurements into a small number
of coarse buckets, which seems appropriate for the current level of
detail provided by Servo's memory profiler. Sample output:
```
| 17.41 MiB -- url(file:///home/njn/moz/servo/../servo-static-suite/wikipedia/Guardians%20of%20the%20Galaxy%20(film)%20-%20Wikipedia,%20the%20free%20encyclopedia.html)
| 7.32 MiB -- js
| 3.07 MiB -- malloc-heap
| 3.00 MiB -- gc-heap
| 2.48 MiB -- used
| 0.34 MiB -- decommitted
| 0.09 MiB -- unused
| 0.09 MiB -- admin
| 1.25 MiB -- non-heap
```
Most of the changes are plumbing to get the script task communicating
with the memory profiler task.
Source-Repo: https://github.com/servo/servo
Source-Revision: 2b0bdbe1c195f2f6dd7671981999d622c505fbc5
This will make it easier to adapt to IPC.
The trickiest part here was to make script tasks spawn new layout tasks
directly instead of having the pipeline do it for them. The latter
approach will not work in multiprocess mode, because layout and script
must run in the same address space and the pipeline cannot inject tasks
into another process.
r? @larsbergstrom
Source-Repo: https://github.com/servo/servo
Source-Revision: e06eaa0064f49bc215e3851f0a3686e1191b356a
fixes#6524. I had to make an additional change not mentioned in the ticket - adding the `#[feature]` to enable deriving custom traits but I assume that's expected at this time.
Source-Repo: https://github.com/servo/servo
Source-Revision: bbb39082e0f640400546d2084a450a8675820a82
> Here it is.
>
> ~~There's two major things that are unfinished here:~~
> - ~~Dealing with the unroot_must_root lint. I'm not sure about the value of this lint with the new rooting API.~~ Done.
> - ~~Updating the Cargo.locks to point to the new SM and SM binding.~~ Done.
>
> I also included my fixes for the rust update, but these will disappear in a rebase. A rust update is necessary to support calling `Drop` on `Heap<T>` correctly when `Heap<T>` is inside a `Rc<T>`. Otherwise `&self` points to the wrong location.
>
> Incremental GC is disabled here. I'm not sure how to deal with the incremental barriers so that's left for later.
>
> Generational GC works. SM doesn't work without it.
>
> The biggest change here is to the rooting API. `Root` was made movable, and `Temporary` and `JSRef` was removed. Movable `Root`s means there's no need for `Temporary`, and `JSRef`s aren't needed generally since it can be assumed that being able to obtain a reference to a dom object means it's already rooted. References have their lifetime bound to the Roots that provided them. DOM objects that haven't passed through `reflect_dom_object` don't need to be rooted, and DOM objects that have passed through `reflect_dom_object` can't be obtained without being rooted through `native_from_reflector_jsmanaged` or `JS::<T>::root()`.
>
> Support for `Heap<T>` ended up messier than I expected. It's split into two commits, but only because it's a bit difficult to fold them together. Supporting `Heap<T>` properly requires that that `Heap::<T>::set()` be called on something that won't move. I removed the Copy and Clone trait from `Heap<T>` so `Cell` can't hold `Heap<T>` - only `UnsafeCell` can hold it.
>
> `CallbackObject` is a bit tricky - I moved all callbacks into `Rc<T>` in order to make sure that the pointer inside to a `*mut JSObject` doesn't move. This is necessary for supporting `Heap<T>`.
>
> `RootedCollectionSet` is very general purpose now. Anything with `JSTraceable` can be rooted by `RootedCollectionSet`/`RootedTraceable`. Right now, `RootedTraceable` is only used to hold down dom objects before they're fully attached to their reflector. I had to make a custom mechanism to dispatch the trace call - couldn't figure out how to get trait objects working for this case.
>
> This has been tested with the following zeal settings:
>
> GC after every allocation
> JS_GC_ZEAL=2,1
>
> GC after every 100 allocations (important for catching use-after-free bugs)
> JS_GC_ZEAL=2,100
>
> Verify pre barriers
> JS_GC_ZEAL=4,1
>
> Verify post barriers
> JS_GC_ZEAL=11,1
Source-Repo: https://github.com/servo/servo
Source-Revision: e7808c526c348fea5e3b48af70b7f1a066652097
Sorry for not doing it yesterday, I couldn't.
cc @metajack @SimonSapin
Source-Repo: https://github.com/servo/servo
Source-Revision: 4ebb95ccd8e034007eacb447a054919ef4af2bf7
This moves webdriver_traits into msg to avoid a circular dependency.
Source-Repo: https://github.com/servo/servo
Source-Revision: d08995e1a94fa093b7fc1e5d918d9dca79f260d6
It would be nice if HTML parsing didn't have to hog the event loop, so I didn't have to do this polling in `content_changed()`, but maybe the way we do it is unavoidable.
r? @jdm
Source-Repo: https://github.com/servo/servo
Source-Revision: e604b663f4e664b9a94d61569082b2b62de2151d
...r parsing. Hook up document loading to async networking events.
Relies on https://github.com/servo/html5ever/pull/107, so we'll likely need to backport it rather than wait for the next rustc upgrade.
Source-Repo: https://github.com/servo/servo
Source-Revision: dd319c1a998bbd3eeb84fdc4ca8a41ee7877ca37
This fixes a hang found while testing the jQuery test suite.
Source-Repo: https://github.com/servo/servo
Source-Revision: c51e9f04559f04f1e820b792261e1653c6869ee5
This allows most of the jquery test suite to run without exhausting thread resources.
Source-Repo: https://github.com/servo/servo
Source-Revision: 7e022b25a8ab87897180dcf1b6aff8d9b57adcb4
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.
Source-Repo: https://github.com/servo/servo
Source-Revision: 5e61ebaa05e5babb7b2fdd1347b6cdd23df38e62
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
Source-Repo: https://github.com/servo/servo
Source-Revision: 2baa69595e2d57213c34b7e168b60885948fa85b
Also adds example support for getting the name and text properties of the elements.
Source-Repo: https://github.com/servo/servo
Source-Revision: 8df824998f108d8ef0472bcf6c006b3270e2800e
This adds support for compositing to a PNG without actually quiting
the browser.
Cargo bits need to be updated after the upstream changes to rust-png land.
Source-Repo: https://github.com/servo/servo
Source-Revision: 63ba1cb69b80d70c1d893ba80edd802cd326316d
Following the (short) in #5930 it was decided that we should remove this
FIXME.
Fixes#5930
Source-Repo: https://github.com/servo/servo
Source-Revision: 2e4d2eb6a9fd135ee19475664a608789880f37b6
