If a property is custom or contains a `var()` reference, it cannot be
parsed into a proper StyleValue immediately, so we store it as an
UnresolvedStyleValue until the property is compute. Then, at compute
time, we resolve them by expanding out any `var()` references, and
parsing the result.
The implementation here is very naive, and involves copying the
UnresolvedStyleValue's tree of StyleComponentValueRules while copying
the contents of any `var()`s it finds along the way. This is quite an
expensive operation to do every time that the style is computed.
This represents a property value that hasn't been converted to a
"proper" StyleValue yet. That is, it's either a custom property's value,
or a value that includes `var()` references, (or both!) since neither of
those can be fully resolved at parse time.
This allows document.implementation to keep the underlying document
alive for as long as we need it (for example, if someone holds on to a
DOMImplementation JS wrapper after the document is GC'd.)
The "completely finish loading" algorithm (from the HTML spec) is
responsible for sending a "load" event to nested browsing context
containers (iframes).
This patch removes the old mechanism for sending "load" events, which we
had mistakenly kept around, causing two events to be sent instead of
one. :^)
DOMImplementation.createDocument() should set the content type of the
newly created document, not replace the content type of the
DOMImplementation's own host document.
Now that we can serialize CSS tokens, we can just hold a string and then
re-parse it when the Supports is evaluated. This feels a little weird,
but it only happens once so it's not going to slow it down much, and it
keep the API cleaner.
This is `<general-enclosed>` in CSS grammar. It represents a section of
a `@media` or `@supports` rule that exists in some future standard that
we don't understand yet, but don't want to make the entire rule invalid.
There's not much that it can usefully do, but we store a string
representation of its contents so that it can be serialized out.
This changes allows for nested browser contexts to be embedded in the
serialized JSON of their container element (like `iframe`) and enables
their inspection in the DOM Inspector.
An image document is the synthetic DOM::Document we create to wrap an
image when you open the URL of an image directly in a web view.
The path that creates these documents will now also call out to the
separate ImageDecoder process for the actual decoding work.
This saves user code from having to parse the numbers, as we already did
that while Tokenizing. :^)
As a bonus, we now round extremely large integers to the closest
available value, like the spec tells us to.
All CSS loading is now done by the relevant classes:
- CSSImportRule, which loads its linked stylesheet
- HTMLStyleElement, which "loads" its contents
- HTMLLinkElement, which loads its linked stylesheet
The changes in commit d462a6720a meant
that `CSSLoader::load_next_import_if_needed()` was no longer being
called, so no `@import` rules were loading.
The spec does not seem to mention when that loading should take place,
but WebKit does this at parse time, as each rule is added to the style
sheet. If it works for them, it's probably not that bad. :^)
So, that's what we now do here. The `CSSImportRule` creates a fetch
request when it is constructed, so each one is responsible for its own
contents.
