Currently we run the risk of missing MediaManager::OnNavigation() at shutdown in
some cases. When it happens, cleanup of window listeners is delayed to
MediaManager::Shutdown().
Triggering window listener shutdown through the destructor doesn't work when the
destructor is invoked by WindowTable::Clear(), as clearing out the
window listener's SourceListeners will trigger a call back into
WindowTable::Remove() to remove the now empty window listener.
This is not allowed by the hash table implementation.
This patch fixes this by changing strategy from cleanup-in-dtor to explicit-cleanup-before-dtor and making a non-clean state in the dtor an error.
MozReview-Commit-ID: C9FryVlv28d
--HG--
extra : rebase_source : 55ccf96b15eb3261d5176364e6952f51415bc668
This delays setting sInShutdown (renamed to sHasShutdown for better semantics)
until all that remains of shutdown is to shut down the media thread.
From JS this is no change as JS cannot run until Shutdown() is done.
The difference is that we let members and still-active listeners get destroyed
and cause any members that are media-thread only (in this case devices), to shut
down before the thread goes away.
MozReview-Commit-ID: 5NTWwh8Umn8
--HG--
extra : rebase_source : f6b508abc95434f027a6544fdac53949138764da
These were originally exposed directly as static methods on nsGlobalWindow, but
as they are clearly associated with either the inner or outer window, it makes
more sense for them to be called as such.
MozReview-Commit-ID: LFq8EfnhDlo
This is a large patch which tries to switch many of the external consumers of
nsGlobalWindow to instead use the new Inner or Outer variants.
MozReview-Commit-ID: 99648Lm46T5
The NS_LITERAL_STRING macro creates a temporary nsLiteralString to encapsulate the char16_t string literal and its length, but AssignLiteral() can determine the char16_t string literal's length at compile-time without nsLiteralString.
MozReview-Commit-ID: 6vgQiU8zN3o
--HG--
extra : rebase_source : 1b536b92ef43f610db057ace6f108620e8d8b4d5
extra : source : 336e21386d5eeb16f1c9893c29377f23b67cc4b0
XPCOM's string API doesn't have the notion of a "null string". But it does have
the notion of a "void string" (or "voided string"), and that's what these
functions are returning. So the names should reflect that.
--HG--
extra : rebase_source : 4e3f982e0873877174a08a25413595ff66f7d20e
Replace it with NS_INTERFACE_MAP_BEGIN_CYCLE_COLLECTION, because it
has been the same for a while.
MozReview-Commit-ID: 5agRGFyUry1
--HG--
extra : rebase_source : 5388c56b2f6905c6ef969150f0c5b77bf247624d
As well as the obvious #ifdefs, this allows DOMHwMediaStream to be
removed, and also the "phone-state-changed" observer.
--HG--
extra : rebase_source : 373280183e228bd4b9bd9d866959409f2444c77e
Right before `MediaManager::EnumerateRawDevices` posts its enumeration
task to the background thread, add the option to send out a
"getUserMedia:ask-device-permission" notification, whose purpose is to
give frontend UI a chance to ask for device permissions (e.g. permission
to enumerate cameras on Android).
The subject parameter of the notification is an `nsIRunnable` task,
which is passed back to `MediaManager`, after receiving permission,
through the `getUserMedia:got-device-permission` notification. The
notification listener then posts the task and enumeration proceeds as
before.
MozReview-Commit-ID: IhRXTdIplvN
Andreas Pehrson
Tom Ritter
Chung-Sheng Fu
Nika Layzell
Masatoshi Kimura
Jan-Ivar Bruaroey
Wes Kocher
Andreas Farre
Munro Mengjue Chiang
Chris Peterson
Nicholas Nethercote
Andrew McCreight
Matthew Gregan
Eric Rahm
Sebastian Hengst
Carsten "Tomcat" Book
Paul Adenot
Alex Chronopoulos
Bill McCloskey
Jim Chen