When a new document is loaded in a WindowContext, various pieces of state need
to be updated in the parent process. This is currently done in an ad-hoc manner
in nsGlobalWindowOuter::SetNewDocument. This change moves the updating logic
into a common method on WindowGlobalChild.
Differential Revision: https://phabricator.services.mozilla.com/D74325
This should make the flow of how data gets into the initial WindowContext state
more clear, and allows the setting of initial synced WindowContext fields.
Differential Revision: https://phabricator.services.mozilla.com/D74324
The entire CookieJarSettingsArgs is currently being synced into every content
process, when only two fields of that structure are actually needed.
Those two fields are extracted from the CookieJarSettingsArgs and synchronized
separately to avoid leaking information such as principals into every content
process.
Differential Revision: https://phabricator.services.mozilla.com/D74258
This requires --build-peers-said-large-imports-were-ok since
third_party/rust/mp4parse/src/lib.rs is 113KB. This code is just moving from
media/mp4parse-rust to third_party/rust, so it's not really adding to net code
size.
Differential Revision: https://phabricator.services.mozilla.com/D74488
Direct Manipulation uses a different input model from processing messages that Windows sends.
Windows asks us if we want to start a direct manipulation session by sending us the DM_POINTERHITTEST message, and we call SetContact if we do. After that Windows won't send us messages until the user ends the gesture. Instead Direct Manipulation will update a transform (that's invisible to the user). We pull that transform and turn it into pan and pinch gestures.
So DealWithPopups is not called and popups don't get rolled up. Instead I call it in the function where we send all events that come from dmanip.
Differential Revision: https://phabricator.services.mozilla.com/D74215
The old code didn't handle content prevent defaulting the input.
The pinch gesture code doesn't seem to fully work properly, it will allow a little pinch zooming before halting it if content is prevent defaulting it. Not sure what is up yet.
Differential Revision: https://phabricator.services.mozilla.com/D73220
We can't just get pinch events, we need to handle both.
This state machine code is basically copied from Chrome's implementation.
Differential Revision: https://phabricator.services.mozilla.com/D71307
I had to mess with the refcounting of Display (and hence destructors) because we create a NewRunnableMethod inside Display that holds a pointer to |this|. There are versions of NewRunnableMethod that don't take a ref but I'm not sure of the lifetime of Display, so easier to just take a ref since several of the subclasses are already refcounted.
Differential Revision: https://phabricator.services.mozilla.com/D71303
The APZ was keeping a raw pointer to the controller thread. This was a dangerous exercise.
This was okay on desktop, as the controller thread was the main thread and would have outlived everything else. On Android however it's the UI thread and it could get deleted before we received a last input event.
So we use a strong pointer instead to prevent the thread from being deleted and as such, we now needs to explicitly clear it on shutdown.
This requires the various methods in APZThreadUtils to be made thread-safe so that the controller thread can be shutdown mid-air.
Differential Revision: https://phabricator.services.mozilla.com/D73830
RemoteContentController::PostDelayedTask must be called on the controller thread; and all the 3 implementations are doing is dispatching the task on the current MessageLoop.
The naming and that the method was pure virtual made it a bit confusing as it gave the impression we would dispatch the task on the controller's internal thread, which wasn't the case.
So we make a generic implementation and assert to the documented use.
Differential Revision: https://phabricator.services.mozilla.com/D73828
The test is a permafail without any of those changes on all my windows machine so I've been unable to diagnostic what could happen.
The test was also to be nominated for disabling to be too high failure rates.
https://bugzilla.mozilla.org/show_bug.cgi?id=1411477#c99
According to :kats, the test was never rewritten to use the new APZ. Disabling on the platform where there's high intermittent failure (about 70%)
Differential Revision: https://phabricator.services.mozilla.com/D74391
MessagePool brings no benefit over the traditional nsIThread.
Additonally, replace some incorrect use of RefPtr for xpcom objects.
Differential Revision: https://phabricator.services.mozilla.com/D73827
CompositorBridgeParent::ScheduleTask was always called from the compositor thread ; so make it explicit that we are dispatching the task to the compositor thread.
We inline the method instead.
Differential Revision: https://phabricator.services.mozilla.com/D73824
already_AddRefed destructor assert that it's nullptr.
DelayedDispatch check that the value passed isn't 0 and return an error code, leaving the already_AddRefed untouched.
Differential Revision: https://phabricator.services.mozilla.com/D73821
When we animate text, we rasterize the glyphs in an arbitrary local
space once, and scale them during the animation. Some glyphs may be
pressed against the edge of the texture, resulting in artifacts due to
how the sampling works in the shader. This patch fixes the sampling
issues by padding glyph textures with an extra transparent pixel border.
This only applies to glyphs that are rasterized in local space.
This patch does not add the extra padding for Mac because it is already
padding its glyphs for Mac-specific reasons, and does not appear to be
as suspectible to the problem.
Differential Revision: https://phabricator.services.mozilla.com/D74457
Depends on D62624
With the previous implementation, an uninitialized document could be returned as a root node.
Here we try to be more explicit and wait for a correct root node. However in some cases a document can remain uninitialized and will never transition to any other state.
If the document is uninitialized but is not currently loading, we should consider it as a valid root node.
Differential Revision: https://phabricator.services.mozilla.com/D62625
Depends on D62623
`new-root` is no longer a mutation, but an event emitted by the `walker` actor.
Tests watching for mutations should be updated accordingly. They also need to call watchRootNode explicitly.
Differential Revision: https://phabricator.services.mozilla.com/D62624
The fix for bug 1621390 extended the tile backdrop computation to recognize
images. This patch extends that to consider gradients as potential
backdrops as well.
Differential Revision: https://phabricator.services.mozilla.com/D70458
Starting with Firefox 76 we're seeing this (seemingly benign) error being reported much more commonly, causing unnecessary load on our Sentry instance. Adding .noReport=true keeps this from being reported to Sentry.
Differential Revision: https://phabricator.services.mozilla.com/D74471
These pragmas can be used to influence stack trace filenames, and to affect
how and where files show up in developer tools. In some circumstances, it can
be nice to disable allof that functionality in order to ensure that you get
the stack trace and debug information as SpiderMonkey sees it.
Differential Revision: https://phabricator.services.mozilla.com/D72103
`Query.start` shouldn't `await provider.tryMethod("isActive")`. Instead it
should call all providers at once and then await all the promises. Also make
some other simplifications and changes to `Query.start` in preparation for later
patches.
Factor out `notifyResults` into `Query._notifyResultsFromProvider` because the
next patch will use it.
Differential Revision: https://phabricator.services.mozilla.com/D74191
The hashes of certain strings is used indirectly in generating IDL C++ files. Before, we were using the `hash()` function to generate these hashes, which in Python 3 is non-deterministic over subsequent `python3` processes, causing bugs like bug 1635755. Instead, use a specific, deterministic hash to avoid spurious diff failures.
The `md5` hash function isn't completely cryptographically secure but the security of the hashes isn't important for us in this case since we're just using them as a per-string identifier. We could use a more robust hash function but there may be performance implications from doing so.
Differential Revision: https://phabricator.services.mozilla.com/D74446
Assert that trace list visits children before parents. Make this explicit
instead of relying on existing publishDeferredFunctions checks so that we
check for non-lazy functions as well for more certainty.
Differential Revision: https://phabricator.services.mozilla.com/D74440
