Previously, this was passed around using the nsHashPropertyBag base
class on HTTP channels. In bug 1907269, this base was added to
nsExtProtocolChannel to support reading this property for external
protocols.
This unfortunately ended up not working when a HTTP channel redirects to
an external protocol, as the hash property propagation logic in
nsHttpChannel happens after the check that the target channel is a HTTP
channel, so the property was lost when redirecting to an external
protocol.
Rather than refactoring HTTP redirect logic, this patch simplifies by
moving the property onto the LoadInfo where it will be trivially
preserved across redirects, and available in the places which need it.
Because the nsHashPropertyBag base is no longer required for
nsExtProtocolChannel, it is also removed.
Differential Revision: https://phabricator.services.mozilla.com/D220744
It would seem sometimes in Firefox 77 we regressed the ability to set the TRR
mode for a browsing context when opening a new tab.
I confirmed that this bug didn't happen in Fx 77 when setting the
`browser.tabs.documentchannel.parent-initiated` pref to `false`.
The nsIWebNavigation::LOAD_FLAGS_DISABLE_TRR is correctly passed into
CanonicalBrowsingContext, but it doesn't end up getting used.
This sets the appropriate DefaultLoadFlags for BrowsingContext
when the LOAD_FLAGS_DISABLE_TRR or LOAD_TRR_ONLY_MODE flags are present
in nsDocShellLoadState::LoadFlags()
Differential Revision: https://phabricator.services.mozilla.com/D220550
Your patch looks good to me. Not sure if there's a non-racy way of
testing the internal full-zoom value here...
But in general, assuming consistent zoom lacking other sources of
information seems like the right thing to do.
I agree that eventually we want to probably move the site-specific zoom
impl into Gecko (if only because Android is also going to want that, see
bug 1547181, where I added a comment).
Co-authored-by: Nika Layzell <nika@thelayzells.com>
Differential Revision: https://phabricator.services.mozilla.com/D220277
This patch switches the main logic from a boolean pref to an integer with different feature modes.
See nsIBounceTrackingProtection for the new modes supported. Dry-run mode has been merged into the mode
pref.
A new state is introduced MODE_ENABLED_STANDBY in which BTP still collects user activation signals, but
does not classify or purge bounce trackers. This new mode should be used when it's expected that the
feature may be toggled on/off after rollout. Collecting user activation even in the "disabled" state
ensures that when the feature gets re-enabled the user activation map is already warmed up and sites
important to the user are not classified + purged.
The BounceTrackingProtection singleton listens for changes on the mode pref and inits / tears down accordingly.
However it never shuts down fully. To fully disable the feature a restart is required.
Differential Revision: https://phabricator.services.mozilla.com/D217848
Previously these checks were largely diagnostic tools for finding bugs
in other code as it evolves. This unifies the checks a bit more and
makes them stronger for BrowsingContexts created over IPC, providing a
place for more coherency checks to be added in the future.
Differential Revision: https://phabricator.services.mozilla.com/D218860
This patch switches the main logic from a boolean pref to an integer with different feature modes.
See nsIBounceTrackingProtection for the new modes supported. Dry-run mode has been merged into the mode
pref.
A new state is introduced MODE_ENABLED_STANDBY in which BTP still collects user activation signals, but
does not classify or purge bounce trackers. This new mode should be used when it's expected that the
feature may be toggled on/off after rollout. Collecting user activation even in the "disabled" state
ensures that when the feature gets re-enabled the user activation map is already warmed up and sites
important to the user are not classified + purged.
The BounceTrackingProtection singleton listens for changes on the mode pref and inits / tears down accordingly.
However it never shuts down fully. To fully disable the feature a restart is required.
Differential Revision: https://phabricator.services.mozilla.com/D217848
There are many other uses of OtherPid which could be switched over to
OtherChildID, but these were a couple of obvious low-hanging fruit use-cases
which will work better when using OtherChildID.
Differential Revision: https://phabricator.services.mozilla.com/D217120
This patch removes test_iframe.html. We remove it because the newly introduced test covers the tests done in that test. The reason for removing it in the first place is now that screen properties are inherited/spoofed xorigin, we get a 4px difference. The reasosn for 4px difference is the test runner runs tests in an iframe with a 2px border on each side.
Differential Revision: https://phabricator.services.mozilla.com/D215509
Previously this load would be allowed due to the "web-controlled" check
failing to reject the load. This patch instead bases it on the
docshell.newWindowTarget flag, which is set for the first load in a new
DocShell due to a call to window.open or target=_blank.
This required making the external protocol channel be a property bag so
that the flag can be set on the channel. In the future we may want to
switch this flag to being set in a more generic way which is less
specific to the channel.
Differential Revision: https://phabricator.services.mozilla.com/D217484
Previously this load would be allowed due to the "web-controlled" check
failing to reject the load. This patch instead bases it on the
docshell.newWindowTarget flag, which is set for the first load in a new
DocShell due to a call to window.open or target=_blank.
This required making the external protocol channel be a property bag so
that the flag can be set on the channel. In the future we may want to
switch this flag to being set in a more generic way which is less
specific to the channel.
Differential Revision: https://phabricator.services.mozilla.com/D217484
Previously it was possible to bypass specific BCG selection based on
cross-origin isolated status if the site was allowed to load file URIs
using enterprise policies, which could lead to a crash.
This patch changes the behaviour such that BCG selection now happens
correctly. The site will still not be cross-origin isolated due to being
loaded into a file content process.
Differential Revision: https://phabricator.services.mozilla.com/D217007
The code path that was used for loading a document
with fission disabled did not take user activation into account.
This made almost all WPTs for Text Fragments fail,
since they depend on this flag to decide if
a text directive is allowed to be scrolled to.
This patch makes sure that the user activation is
also available for code paths that run through
`CanonicalBrowsingContext::FixupAndLoadURIString()`
to load a document.
This is done by adding a flag `hasValidUserGestureActivation`
to the `nsIOpenWindowInfo`, which carries it through
the JS stack for loading a document.
Differential Revision: https://phabricator.services.mozilla.com/D213401
This commit is a prerequisite for part 5 of this patch set.
This flag is set in the docshell, based on `user gesture activation`.
This is not necessarily sufficient,
as the text fragments spec bases the value of
this flag also on User Involvement, which is not
implemented in Gecko yet.
The flag is then plumbed through the parent process
(`DocumentLoadListener::CreateDocumentLoadInfo()`)
to the document, where it can be consumed.
This includes client-side redirects as described in [0].
There is no changed behavior in this commit.
[0] https://wicg.github.io/scroll-to-text-fragment/#restricting-the-text-fragment
Differential Revision: https://phabricator.services.mozilla.com/D212817
This commit is a prerequisite for part 5 of this patch set
and introduces the `text directive user activation` flag
as well as the necessary boilerplate to transport the value
through process boundaries in `nsDocShellLoadState` and `LoadInfo`.
There is no changed behavior in this commit.
Differential Revision: https://phabricator.services.mozilla.com/D212816
This is a fairly significant patch, however it would be difficult to break it
down into smaller patches:
1) The various mechanisms used to manage ContentParent lifecycles have been
merged together into a common "KeepAlive" system. A process will
begin shutdown when its keepalive count reaches 0. (though it will
still wait for all BrowserParents to also be dead before sending the
actual shutdown message as before).
This replaces a number of bespoke systems for tracking BrowserParent
instances in different lifecycle states, remote workers, ongoing
process switches, and preallocated processes.
2) KeepAlives are now managed automatically by a UniquePtr variant
(Unique[Threadsafe]ContentParentKeepAlive). This makes the hand-off
over KeepAlive lifecycles explicit, even for workers.
3) All KeepAlives are now keyed by a BrowserId, which will be 0 for keepalives
not associated with a specific tab. This allows the new process
selection logic to count all tabs other than the one being navigated
when deciding which process to use.
4) The process switching logic now tracks it's KeepAlive with a BrowserId,
meaning that ongoing process switches are considered when performing
process selection, even if the BrowserParent hasn't been created yet.
Differential Revision: https://phabricator.services.mozilla.com/D213338
