Same as with `BlocksRingBuffer`: Instead of a potentially-null pointer to a
`ProfileBufferEntryWriter`, we are now providing a
`Maybe<ProfileBufferEntryWriter>`, which is safer.
Differential Revision: https://phabricator.services.mozilla.com/D71287
Instead of a potentially-null pointer to a `ProfileBufferEntryWriter`, we are now providing a `Maybe<ProfileBufferEntryWriter>`, which is safer.
Differential Revision: https://phabricator.services.mozilla.com/D71286
This way, something like:
*:where(.foo, .bar)
Will end up twice on the selector map, just as if you would've written
.foo, .bar.
But we're a bit careful to not be wasteful, so:
.foo:where(div, span)
Will still end up using the .foo bucket.
It needs a bit of borrow-checker gymnastics to avoid cloning the entry
in the common path. It's a bit gross but not too terrible I think.
Differential Revision: https://phabricator.services.mozilla.com/D71457
There are a bunch of missing tests, and there are some tests that don't
match the current spec text, that I need to write or fix before I enable the
feature everywhere.
But there are fairly complex invalidation tests, that we pass flawlessly :)
Differential Revision: https://phabricator.services.mozilla.com/D71424
See the comment about why this is valuable. For a selector like:
.foo:is(.bar) > .baz
Before this patch we'd generate an Dependency for .bar like this:
Dependency {
selector: .bar,
offset: 0,
parent: Some(Dependency {
selector: .foo:is(.bar) > .baz,
offset: 1, // Pointing to the `>` combinator.
parent: None,
}),
}
After this patch we'd generate just:
Dependency {
selector: .foo:is(.bar) > .baz,
offset: 1, // Pointing to the `>` combinator.
parent: None,
}
This is not only less memory but also less work. The reason for that is that,
before this patch, when .bar changes, we'd look the dependency, and see there's
a parent, and then scan that, so we'd match `.bar` two times, one for the
initial dependency, and one for .foo:is(.bar).
Instead, with this we'd only check `.foo:is(.bar)` once.
Differential Revision: https://phabricator.services.mozilla.com/D71423
That way we can look at the parent dependency as described in the previous
patch. An alternative would be to add a:
parent_dependency: Option<&'a Dependency>
on construction to `Invalidation`, but this way seems slightly better to avoid
growing the struct. It's not even one more indirection because the selector is
contained directly in the Dependency struct.
Differential Revision: https://phabricator.services.mozilla.com/D71422
The tricky part of :is() and :where() is that they can have combinators inside,
so something like this is valid:
foo:is(#bar > .baz) ~ taz
The current invalidation logic is based on the assumption that you can
represent a combinator as a (selector, offset) tuple, which are stored in the
Dependency struct. This assumption breaks with :is() and :where(), so we need
to make them be able to represent a combinator in an "inner" selector.
For this purpose, we add a `parent` dependency. With it, when invalidating
inside the `:is()` we can represent combinators inside as a stack.
The basic idea is that, for the example above, when an id of "bar" is added or
removed, we'd find a dependency like:
Dependency {
selector: #bar > .baz,
offset: 1, // pointing to the `>` combinator
parent: Some(Dependency {
selector: foo:is(#bar > .baz) > taz,
offset: 1, // Pointing to the `~` combinator.
parent: None,
})
}
That way, we'd start matching at the element that changed, towards the right,
and if we find an element that matches .baz, instead of invalidating that
element, we'd look at the parent dependency, then double-check that the whole
left-hand-side of the selector (foo:is(#bar > .baz)) actually changed, and then
keep invalidating to the right using the parent dependency as usual.
This patch only builds the data structure and keeps the code compiling, the
actual invalidation work will come in a following patch.
Differential Revision: https://phabricator.services.mozilla.com/D71421
addA11yLoadEvent gets the contentWindow and wait for the document from that window to fire the load event.
Enabling the DocumentChannel for parent process load (or here in non-e10s mode) we have one extra event loop before the load starts.
So the window passed to addAllyLoadEvent would have been of the about:blank page.
The current code was based on an observable behaviour which was that the load was occuring synchronously.
DocumentChannel broke that assumption.
Differential Revision: https://phabricator.services.mozilla.com/D70799
When starting a load via the ParentProcessDocumentChannel, the event http-on-modify-request will be fired before the DocumentLoadListener has a chance to set the notificationCallback attribute.
When using a DocumentChannel, this test will not trigger the expected codepath as the DOMWindowCreated event will be fired once the channel is fully up and running; which in effect is also a fix of the original bug 1339722
Instead we use the document-on-modify-request event when the DocumentChannel is enabled.
Differential Revision: https://phabricator.services.mozilla.com/D70011
Some tests rely on this event to start action. The DocumentChannel had no equivalent. We make the ParentProcessDocumentChannel listen to this event and if it matches the nsIChannel currently in use in the DocumentLoadListener than we emit a similar document-on-modify-request event on the DocumentChannel.
Differential Revision: https://phabricator.services.mozilla.com/D70010
Add ParentProcessDocumentChannel object. This object is a DocumentChannel that will start a channel load from the parent process via a DocumentChannel.
The aim of this task is two-fold.
1- Be consistent on how we handle redirects before continuing the load on the final channel.
2- Prepare to initiate a process switch when needed without having to go via an intermediary content process, saving a process switch. This task will be done in a follow-up task.
The behaviour of the ParentProcessDocumentChannel is similar in logic to the DocumentChannelChild/DocumentChannelParent pair. The ParentProcessDocumentChannel sets up a DocumentLoadListener, have it handle the redirects and upon completion continue the load on the final channel.
Differential Revision: https://phabricator.services.mozilla.com/D70009
This prepares the code so that it doesn't always have to deal with dealing with another process as is the case when using DocumentChannelChild/DocumentChannelParent
Depends on D72271
Differential Revision: https://phabricator.services.mozilla.com/D72272
Move DocumentChannelChild methods to the base class as they will be needed with the ParentProcessDocumentChannel
Depends on D70008
Differential Revision: https://phabricator.services.mozilla.com/D72271
We must wait for the iframe created by extension1 to be fully loaded before we can attempt to send a message to it.
So we send a message when it is ready, and suspend the execution until this message is received.
Differential Revision: https://phabricator.services.mozilla.com/D70008
The code assumed that nsJARChannel::RetargetDeliveryTo would have been called synchronously from nsJARChannel::OnStartRequest, which would be true if we weren't using a DocumentChannel.
The DocumentLoadListener queue the calls to OnStartRequest until the final redirect.
nsJARChannel::RetargetDelivery mRequest member is be set to forward the call.
So we need to only reset it once OnStopRequest has been received.
Differential Revision: https://phabricator.services.mozilla.com/D70007
The DocumentLoadListener is setting up a ParentChannelListener to go in between the normal listener->channel chain.
ParentChannelListener not implementing nsIThreadRetargetableStreamListener would prevent a nsHtml5StreamParser settings things up so that OnDataAvailable could be sent to a html parser thread off the main thread; improving performance.
Differential Revision: https://phabricator.services.mozilla.com/D70006
Following the changes to DocumentChannel the test was failing.
With DC, a load may take a few event loops to start. This current test was only waiting for the load to start to the URL about:preferences#privacy-logins and would immediately tear down the window.
However, this URL redirects to about:preferences#privacy ; destroying the window midway could cause XML parsing error.
So now we wait for the page to fully load, and make sure we've been through both addresses.
Differential Revision: https://phabricator.services.mozilla.com/D70003
This also removes the need to call SerializeRedirectData to set mRedirectChannelId to the proper value and register the channel.
Differential Revision: https://phabricator.services.mozilla.com/D69998
It's inherited only by `PlaceholderTransaction` and used only for QI.
Therefore, we can get rid of it.
Additionally, this makes storing `PlaceholderTransaction` and
`CompositionTransaction` in `PlaceholderTransaction` faster and safer with
`WeakPtr`.
Finally, this makes `PlaceholderTransaction` always have non-null name
because it caused a lot of useless warnings in
`EditAggregationTransaction::GetName()` and
`PlaceholderTransaction::GetTxnName()`.
Differential Revision: https://phabricator.services.mozilla.com/D71910
In a lot of places in libeditor, we do nothing if given transaction is not
our edit transaction classes' instance. Therefore, it's better to have
casting virtual method in `nsITransaction` for performance because QI cost
may not be cheap.
Differential Revision: https://phabricator.services.mozilla.com/D71908
* Prefer autocomplete=new-password fields when the field we filled was autocomplete=new-password
* Only fill a password confirmation field when the next input is/was a password type, not-disabled, and empty
* Limit matching to password fields no more than 3 fields away from the filled password field
Differential Revision: https://phabricator.services.mozilla.com/D71130