Before this change, we were going through the chain of base classes for
each IDL interface object and having them set the prototype to their
prototype.
Instead of doing that, reorder things so that we set the right prototype
immediately in Foo::initialize(), and then don't bother in all the base
class overrides.
This knocks off a ~1% profile item on Speedometer 3.
Many elements never end up needing this string, so instead of eagerly
generating it in the Element constructor, let's defer it until someone
actually requests it.
Knocks off a ~1% profile item on Speedometer3's jQuery test.
Instead of using the more generic define_native_accessor() here,
we poke directly at indexed property storage for the parameter map.
We can also construct the NativeFunction objects directly, without
giving them names like "get 0" etc, since these are not observable
by userspace.
Finally, by using default property attributes (not observable anyway),
we get simple indexed storage instead of generic (hash map) storage.
Instead, let JS::NativeFunction store the AK::Function directly, and
take care of conservatively marking its captured data.
This avoids an extra GC allocation for every JS::NativeFunction.
This will cause an exception to be thrown if user attempts to read from
the response stream of a failed request.
This is unfortunately not testable in CI. It requires a network response
(i.e. not a file:// URL). We also cannot import relevant WPT tests; they
exercise this condition with a python-generated response.
The HTTPS server used by WPT will close TLS connections without sending
a "close notify" alert. For responses that did not have a Content-Length
header, curl treats this as an error.
The basic idea is that style sheets can block script execution under
some circumstances. With this commit, we now handle the simplest cases
where a parser-inserted link element gets to download its style sheet
before script execution continues.
This improves performance on Speedometer 3 where JavaScript APIs that
depend on layout results (like Element.scrollIntoView()) would get
called too early (before the relevant CSS was downloaded), and so we'd
perform premature layout work. This work then had to be redone after
downloading the CSS anyway, wasting time.
Note that our Text/input/link-re-enable-crash.html test had to be
tweaked after these changes, since it relied on the old, incorrect,
behavior where scripts would run before downloading CSS.
No code now relies on using URL's valid state.
A URL can still be _technically_ invalid through use of the URL
constructor or by directly changing URL fields.
However, all URLs should be constructed through the URL parser,
and we should ideally be getting rid of the default constructor
at some stage.
Also, any code which is manually setting URL fields need to be
aware that this is full of pitfalls since there are many different
forms of canonicalization which is bypassed by not going through
the URL parser.
Creating a URL should almost always go through the URLParser to
handle all of the small edge cases involved. This reduces the
need for URL valid state.
Before this change, we were waiting for Document to lazily evaluate
sheet media and media rules. This often meant that we'd get two
full-document style invalidations: one from adding a new style sheet,
and one from the media queries changing state.
By evaluating the rules eagerly on insertion, we coalesce the two
invalidations into one. This reduces the number of full-document
invalidations on Speedometer 3 from 51 to 34.
We now cache the containing block (box) once at the start of layout,
which allows Layout::Node::containing_block() to return instantly
instead of doing tree traversal.
Removes a 0.7% profile item on Speedometer 3.
We already had a really nice hash that had a single issue, this commit
fixes that and makes it *the* hash for the hash table, so we avoid
double-hashing and making a long chain.
This is an easy 10% perf gain.
These are slightly unfortunate as we're crossing the library boundary,
but there's precedent with Object::is_dom_node(), and these are just
knocking down a few more items that were showing up in profiles.
The remaining AOs can stay where they are. This patch just sorts them in
spec order to match the other AO files. Section 8.1 is last, however, as
these are all templates and need the declarations of other AOs above.
The main streams AO file has gotten very large, and is a bit difficult
to navigate. In an effort to improve DX, this migrates TransformStream
AOs to their own file.
The main streams AO file has gotten very large, and is a bit difficult
to navigate. In an effort to improve DX, this migrates WritableStream
AOs to their own file.
