The dirty bit fiddling is nontrivial, but it's pretty much what we do for
invalidation and allows to keep this incrementally easily.
The only caller that we cared about for the GetFlattenedTreeParent check in
DestroyFramesForAndRestyle was the old ShadowRoot invalidation functions that
went away.
MozReview-Commit-ID: GmgPPJ6d2qX
This was useful because nsRuleData in the old style system may not hold
all the data, but the only subclass ServoSpecifiedValues is always able
to hold any data, and thus passes NS_STYLE_INHERIT_MASK to mSIDS. Given
this, this method and mSIDs seems to be useless and can be removed.
MozReview-Commit-ID: 4vWcV4DRS2i
--HG--
extra : rebase_source : 4f364d431821e29838082e391c6dde08af7f5343
We can get button indexes in HID usage reports that do not actually
correspond to a button we store, meaning we can overstep bounds of the
button array. Check validity before accessing array.
MozReview-Commit-ID: AAQJLEgy2Ua
Update head.js so that calls to the gUM helper will check the loopback and fake
device prefs and update the frequency output by test devices accordingly. This
should hopefully avoid issues where tests could change prefs, and the output
tone would no longer be correct for the new configuration.
No longer have audioDevice and videoDevice as globals, this way tests
explicitly fetch direct from prefs.
MozReview-Commit-ID: 9jhVScaBfTX
--HG--
extra : rebase_source : 64b17e44ef166983361f94cf7d287faed3c2cdc5
Update the DeviceEnumerationType enum to scoped style enum for safety and
consistency. Add extra logging to aid in debugging test issues.
MozReview-Commit-ID: cm5bdlIcEG
--HG--
extra : rebase_source : bfcae3d49f809edfd0ed6ab3e516b8be0c0d517e
Update head.js so that logic for fake audio and fake video is more granular.
Now code handles fake audio and fake video separately, and will set the fake
pref is either is needed. This allows for a single loopback device to be set
and fake used for the other. Previously both devices needed to be loopback or
fakes would be used.
MozReview-Commit-ID: K4blmPrSVeh
--HG--
extra : rebase_source : 369ea7f788d29a02db68e517d324977f251c9a98
With changes giving loopback devices higher precedence in testing, various tests
in dom/media/tests/mochitest have been updated to accommodate this. Many tests
just worked, but some require tweaks, or to explicitly request fake devices.
Also update webaudio's test_mediaStreamAudioSourceNodeGC test to explicitly use
fake devices. This test does not have the loopback tone exposed to it in JS, so
is unabel to adjust the loopback tone to suit its needs.
Various tests are updated to set fake device prefs instead of requesting via
gUM to move away from non-standard behaviour per bug 1436424.
MozReview-Commit-ID: 5GAVZFzF2hq
--HG--
extra : rebase_source : 27f39e3573eda321025ce0739e1d5f4101fc5d12
NullPrincipal::Create() (will null OA) may cause an OriginAttributes bypass.
We change Create() so OriginAttributes is no longer optional, and rename
Create() with no arguments to make it more explicit about what the caller is doing.
MozReview-Commit-ID: 7DQGlgh1tgJ
This changeset adds a gUM_support.js to dom/media/test/. This file provides
functions to setup prefs for loopback or fake device selection before gUM calls
are made. This is useful for configuring tests and providing an explicit point
of reference for settings, rather than the implicit ones provided by the
harness.
Updates tests so that the new helper functions are called before gUM. This
will result in loopback prefs being set if loopback device names are detected,
if not then fake devices will be used. This also removes the use of the fake
constraint in gUM calls.
Update touched tests to use some more modern JS. No behavioural changes were
made (except in minor cases, but functionality should be the same). These
changes are largely as follows:
- var -> let
- async is used in places where I felt it improved readability
- semicolons added to various event handler assignments
MozReview-Commit-ID: 1HuE8thBA6w
--HG--
extra : rebase_source : b866056b2821436cf34ea683421c200b4bb4e55f
Change the media manager so that if fake and loopback devices are requested,
loopback is preferred. With this change loopback and fake devices can also be
mixed. Since the fake flag is coarse, and does not specify fake audio or video,
we would previously just make everything fake. As loopback sets flags for video
and audio separately, we can now request a single loopback device, while also
setting the fake flag to get a mix. E.g. if we request a loopback audio device,
and set the fake flag, we should get loopback audio and fake video.
This change also attempts to somewhat consolidate where these settings take
place. Previously, EnumerateRawDevices did much of the loopback setup. However,
other steps around fingerprint resistance or fake devices were done in earlier
functions (EnumerateDevices and GetUserMedia). This changeset moves the loopback
setup so that it's more consolidated with the other setup code in these
functions.
MozReview-Commit-ID: FF0bR0Nyws9
--HG--
extra : rebase_source : 374a6fd0842a430e27c695bcf956e2e072a77fc3
Prior to this patch various bools are used to track if we requst fake or
loopback devices during enumeration. However, since the normal, fake, and
loopback cases are mutually exclusive, a enum can be used. This provides allows
for having descriptive enum values in code and makes clear that the settings are
mutually exclusive.
MozReview-Commit-ID: FF0bR0Nyws9
--HG--
extra : rebase_source : 498513bdc6673fa299080097364a6d0dff00a073