Instead of running paint requests during microtasks, we run them after updating
layout for the current frame. After draining all the paint requests, we then
drain whatever commits are available.
It's still possible that requestPaint callbacks can dirty style and layout
information, so we need to clean that information after servicing the requests.
Ideally we'd block these callbacks from dirtying style or layout information.
R=ojan@chromium.org
Review URL: https://codereview.chromium.org/1027563002
This CL adds just enough custom painting to Sky to make
sky/examples/painting/circle.sky draw a circle. Over time, we should be able to
elaborate this system into something interesting and to make it actually work
in a reasonable way.
R=ojan@chromium.org
Review URL: https://codereview.chromium.org/1017593005
This CL implements custom elements. The design is as follows:
1) Authors subclass Element and call registerElement.
2) When we create C++ elements for custom elements, we call the author's
constructor synchronously.
3) The attach/detach/attributeChanged callbacks are called either:
a) when exiting the current custom element callback scoped (e.g., before
returning from appendChild), or
b) when draining the microtask queue.
The implementation in this CL is a bit fragile because we don't detect name
registration conflicts and we let you create custom elements with the same name
as built-in elements. Also, not every part of the engine is prepared to execute
script synchronously below createElement. We'll need to iron out these issues
over time, but this CL is a start.
R=ojan@chromium.org
Review URL: https://codereview.chromium.org/943013002
This CL is progress towards deleting the concept of an HTMLElement entirely. We
won't actually get all the way there in this CL series, but we're getting
closer. This CL also will let us make custom elements just be Elements instead
of HTMLElements.
R=eseidel@chromium.org
Review URL: https://codereview.chromium.org/942933003
We now only preserve the whitespace inside a <t> element inside
the parser. This removes the known n^2 from reattaching whitespace
which should make parsing and appending nodes faster. I also
removed the dead WhitespaceMode code from the parser, and made
the dom-seralizer.sky auto indent the markup so the test output
would be readable.
R=abarth@chromium.org
Review URL: https://codereview.chromium.org/867963006
Instead lets just iterate the list of sheets and ask each one questions. This
shouldn't be much slower since most widgets only have one sheet anyway.
I also moved the media query matching to the sheet collection. We weren't
correctly matching them since we only guarded the feature set, not the
addition to the list of active sheets.
R=abarth@chromium.org, ojan@chromium.org
Review URL: https://codereview.chromium.org/858423002
Instead just use textContent() and teach HTMLStyleElement how to atomize
its children. I also added the optimization to Node::textContent so that
Text just returns its value instead of allocating a new buffer which
avoids malloc and copy when getting the textContent of text.
R=abarth@chromium.org
Review URL: https://codereview.chromium.org/864613002
This removes the attributes sky doesn't intend to support. It
removes references to the attrs, but leaves behind a lot of
plumbing that will be cleaned up in the future. This at least
removes the API surface.
R=ojan@chromium.org
Review URL: https://codereview.chromium.org/855803002
This patch is largely just a rename since updateLayoutIgnorePendingStylesheets()
didn't really do anything except call updateLayout() now as nothing was reading
the ignorePendingStyleSheets state, and no callers used the synchronous post
layout task option which was meant for plugins (which Sky doesn't have).
Placeholder styles were related and just add confusion since things pretend to
be display: none while <import>'s are loading. We should expose a real API for
avoiding FOUC instead of pumping frames with display: none elements when
imports are loading.
I had to skip the layout/continuations.sky test since it always crashes now
with an ASSERT failure about a bad cast (filed as bug 446739). The bug already
existed, this patch just makes that one test hit it.
BUG=446739
R=ojan@chromium.org
Review URL: https://codereview.chromium.org/834693007
This now means that using a class/id/attribute selector in one scope will
not cause style recalcs in other scopes in the page when something with
that feature changes.
It also removes an iteration over all the scopes that used to collect
features.
I also removed all the extra cases around adding/removing classes since
the general case can handle them.
In addition I removed the check for classStringHasClassName which looked
to see if the string was all whitespace. This check dated way back to
fixing an assert in code we don't even have anymore. Assertion fix aside
all the extra check optimized for was if you wrote class=" " with no
names, which will now cause an extra malloc of the SpaceSplitString::Data.
This seems super rare, it makes more sense to not scan the class string
every time the list of classes changes which is far more common than
setting a string of only whitespace.
R=ojan@chromium.org
Review URL: https://codereview.chromium.org/837883002
Sky doesn't really have many built in elements, we no longer support
<foo is="my-element">. This does mean you can't extend <import>,
<script> or <style>, but we'll figure that out later.
R=abarth@chromium.org
Review URL: https://codereview.chromium.org/831993005
In Sky we can style share if our TreeScopes have the same styles, our :host
styles are the same, and we'd inherit the same styles. This allows a lot of
simplification to the style sharing logic since we don't need to deal with
descendant selectors or tree boundary crossing rules:
- We can remove the logic that was checking that we were distributed
to the same insertion points since there's no ::content selectors.
- We can check the actual inherited values instead of looking at the
parentOrSHadowHostNode(). We used to look at the node in Blink because we
were checking that you'd get the same descendant selectors applied. In Sky
we instead want to make sure you'd inherit the same values. This also
means we don't need the element().parentOrShadowHostElement() != parent case
in the SharedStyleFinder which was trying to deal with descendant selectors
again.
I also removed the checks that were redundant with the checks inside
supportsStyleSharing() which we always check before adding sharing
candidates.
Finally by refactoring the code to make the TreeScope style check work it
exposed that the Document::styleSheets() and TreeScope::styleSheets() APIs
are now dead. A future patch will delete the now dead StyleSheetList class
as well.
This change makes the city-list application share between all the items in
the list, and all of the headers of the same type now share as well.
R=ojan@chromium.org
Review URL: https://codereview.chromium.org/796713002
Particularly, this deletes RenderLayerCompositor. After this,
there's just CompositingState left to remove.
This is all dead code, so there should be no change in behavior.
R=abarth@chromium.org
Review URL: https://codereview.chromium.org/758843004
This is another step towards making attach non-virtual. InsertionPoints are
core to the engine, so it seems fine to bake them directly into recalcStyle.
I did this already to willRecalcStyle.
R=ojan@chromium.org
Review URL: https://codereview.chromium.org/765523002
This hook existed because some elements wanted to run script inside
::insertedInto, but that wasn't a safe time. Now nothing can run script
inside there so we don't need this extra hook.
HTMLStyleElement's usage of this hook is not clear, but it seems like
it might have related to a pre-lazy attach bug.
R=abarth@chromium.org
Review URL: https://codereview.chromium.org/758623002
This caused us to lose our gn check certification. :(
Turns out gn check was just ignoring all the header
paths it didn't understand and so gn check passing
for sky wasn't meaning much. I tried to straighten
out some of the mess in this CL, but its going to take
several more rounds of massaging before gn check
passes again. On the bright side (almost) all of
our headers are absolute now. Turns out my script
(attached to the bug) didn't notice ../ includes
but I'll fix that in the next patch.
R=abarth@chromium.org
BUG=435361
Review URL: https://codereview.chromium.org/746023002
