To ensure correct usage of TLSs in the profiler, they are now better encapsulated so that:
- init() is called once and its result is cached. (TLSREGISTEREDThread::Init() doesn't need proof of the PSLock, because it's using thread-safe function-static initializers.)
- get() and set() always init() as needed, or in some particular cases strongly assert that init() was successful.
Also, a null-check was missing in profiler_init_threadmanager().
Depends on D87588
Differential Revision: https://phabricator.services.mozilla.com/D87589
Assertions are also clarified with messages, to better distinguish the same tests in different locations.
Assertions should now cover all cases:
- NEW: After registering a thread in the profiler with `CorePS::AppendRegisteredThread`, the TLS should be set to that thread.
- NEW: If `profiler_register_thread` is called again, the TLS should still be set to that thread.
- When `profiler_unregister_thread` is first called, the TLS should still be set to that thread (that's the assertion currently trigering this bug 1657174),
- NEW: When `profiler_unregister_thread` is first called and after we remove the thread with `CorePS::RemoveRegisteredThread`, the TLS should now be null.
- If `profiler_unregister_thread` is called again (or with a never-registered thread), the TLS should be null.
This is a further exploratory patch for bug 1657174, this will not prevent crashes, but hopefully it should give a bit more information, at least a smaller range in which the possible presumed registration/TLS race happens.
Differential Revision: https://phabricator.services.mozilla.com/D87588
This change consists of;
1) Use switch statement for the given PrintPreview Navigation type.
2) Simply iterate over the children of nsPageSequenceFrame for
PRINTPREVIEW_GOTO_PAGENUM
3) Use GetCurrentSheetFrameAndPageNumber for PRINTPREVIEW_PREV_PAGE and
PRINTPREVIEW_NEXT_PAGE so that it should now match
printPreviewCurrentPageNumber
(that means the edge case where the current scroll position is in
the gap between pages has been fixed by this change)
4) Scroll to the position where the target frame is positioned at the center of
the print preview scroll port in the cases of PRINTPREVIEW_PREV_PAGE,
PRINTPREVIEW_NEXT_PAGE and PRINTPREVIEW_GOTO_PAGENUM
4) is a bit debatable but it can be now easily modified by changing
ComputeScrollPositionFrameAtCenter later if it turns out the current way is not
reasonable.
Differential Revision: https://phabricator.services.mozilla.com/D87548
So that we can keep using the logic in the old print preview UI.
For the new print preview UI, the original PrintPreviewScrollToPage will be
modified gradually in subsequent changes.
Differential Revision: https://phabricator.services.mozilla.com/D87543
This data structure is the same between ScriptStencil and BaseScript forms so
it belongs in SharedStencil.h. Avoid pulling in CompileOptions.h by passing
lineno/column explicitly.
Differential Revision: https://phabricator.services.mozilla.com/D87070
The fact that this control is a slider is communicated through its role.
Having the word "slider" in the label is thus redundant and results in unnecessary verbosity for screen reader users.
Differential Revision: https://phabricator.services.mozilla.com/D87401
1. Label it.
2. Expose the position and duration as its value text.
The raw numeric value exposed previously wasn't meaningful to users.
3. Hide the progressBar for a11y.
While the scrubber and progressBar are visually combined, they are exposed as entirely separate controls for a11y.
As well as duplicating information, progress bars are automatically reported by screen readers even if they aren't focused, which intrudes on the audio being played.
4. Remove tabindex="-1" on the progressBar.
A progress element isn't focusable by default.
tabindex="-1" makes it programmatically focusable (but not tabbable).
This isn't interactive, so there's no reason for this to be programmatically focusable.
Moreover, this is now aria-hidden and aria-hidden elements should never be focusable.
Differential Revision: https://phabricator.services.mozilla.com/D87400
The PostUpdate task must always be called as the unelevated user, even if we didn't use the service, in order to ensure that we register the WDBA. Additionally, the PostUpdate task should always be run synchronously so that the elevated and unelevated PostUpdate tasks are guaranteed to run in order. This is important since the elevated PostUpdate unregisters the task and the unelevated PostUpdate re-registers it.
Differential Revision: https://phabricator.services.mozilla.com/D87509
GeckoView uses these flags for the same purpose: to determine
whether or not to allow certain effects, such as pull to refresh
or dynamic toolbar movement, that should be limited to cases
where an event manipulates the root scrollable.
In a subsequent patch, we will enhance the API between APZ and
GeckoView such that in cases where we don't know for sure whether
an event will manipulate the root scrollable until it is
dispatched to content, we tell GV the answer asynchronously.
This is easier if there is only one flag to report.
Differential Revision: https://phabricator.services.mozilla.com/D79929
2020-08-19 Kevin Jacobs <kjacobs@mozilla.com>
* tests/libpkix/certs/PayPalEE.cert:
Bug 1659792 - Update libpkix tests with unexpired PayPal cert. r=jcj
The in-tree `PayPalEE.cert `expired today. This patch replaces it
with a current copy that expires on 12 Jan 2022.
CI breakage before patch: https://treeherder.mozilla.org/#/jobs?repo
=nss&revision=2890f342de631bf6774ac747515a8b5736e20d3f CI with the
fix applied: https://treeherder.mozilla.org/#/jobs?repo=nss-
try&revision=bd28f21d8acbcb15502bd4fc606fc9c0ed09c810
[52c965eaffa1] [NSS_3_56_BETA1]
2020-08-18 Kevin Jacobs <kjacobs@mozilla.com>
* tests/interop/interop.sh:
Bug 1659814 - Pull updated tls-interop for dependency fix. r=jcj
[70376af425ae]
* automation/release/nspr-version.txt:
Bug 1656519 - NSS 3.56 should depend on NSPR 4.28. r=kaie
[2890f342de63]
Differential Revision: https://phabricator.services.mozilla.com/D87648
Keep L10n repacks from trying to include defaultagent_localized.ini when it's
not needed. Thunderbird L10n repacks fail because this file is not present.
Differential Revision: https://phabricator.services.mozilla.com/D87622
This affects HTMLMediaElement, MediaRecorder and RTCPeerConnection, which are
the implementors of NotifyEnabledStateChanged.
This use case is going to become more common as users will be able to globally
mute cameras before requesting one. While muted, the camera is off and no frames
will flow. The old logic for showing disabled video tracks as black relied on a
frame appearing that could be turned black.
With this patch in this case we will create a frame if none has been seen yet,
and it will have a hardcoded size.
Depends on D87127
Differential Revision: https://phabricator.services.mozilla.com/D87128
This allows newly created window listeners, in MediaManager instances where the
cameras have already been muted, to get the right state.
Differential Revision: https://phabricator.services.mozilla.com/D87127
Previously it listened for them on main thread, as notified by the
MediaStreamTrack it was transmitting directly. With this patch it
listens for them through the TrackListener in the graph, and will
get notified when any track prior to the MediaStreamTrack's main
track in the graph changes enabled state as well.
Depends on D86926.
Differential Revision: https://phabricator.services.mozilla.com/D86931
Before this patch, re-enabling a camera track would first render an old frame
(from when it got disabled), then when the camera has started anew, render the
new frames.
With the patch, the black frame from disabling remains until the camera has
started and is sending new frames.
Differential Revision: https://phabricator.services.mozilla.com/D86926
This enables a use case where MediaManager disables a MediaStreamTrack's input
track, and all listeners of NotifyEnabledStateChanged on the MediaStreamTrack's
main track will get notified.
There can be 1 input track for N MediaStreamTracks, so this simplifies how many
tracks will need to be tracked by MediaManager and explicitly disabled in the
graph.
Depends on D86072
Differential Revision: https://phabricator.services.mozilla.com/D86925
Add UA (user agent) muting, a spec-supported feature that somewhat mirrors
track enabling/disabling, except only the browser controls it. The effect
on track sinks is additive: must be unmuted and enabled for there to be
output.
Fire mute/unmute events on JS, and observably set track.muted independent
of track.enabled (reusing existing infrastructure already in use by
RTCPeerConnection tracks).
Low-level: add mDeviceMuted and SetMutedFor() modeled after mDeviceEnabled
and SetEnabledFor() as parallel device state for both camera and microphone for
symmetry and maintenance.
High-level: Only expose messages to mute/unmute camera at the moment, since
that is what is immediately required for Android in bug 1564451.
Differential Revision: https://phabricator.services.mozilla.com/D84222
Bogdan Tara
Gerald Squelart
Agi Sferro
John Lin
Hiroyuki Ikezoe
James Teh
Ted Campbell
Michael Kaply
Emma Malysz
Chun-Min Chang
Kirk Steuber
James Willcox
Botond Ballo
Kevin Jacobs
Nika Layzell
Rob Lemley
Jan-Ivar Bruaroey
Andreas Pehrson