The distinction between moz_malloc/moz_free and malloc/free is not
interesting. We are inconsistent in our use of one or the other, and
I wouldn't be surprised if we are mixing them anyways.
Currently when sending a beacon, HttpBaseChannel::ShouldIntercept tries
to get access to the nsINetworkInterceptController interface through the
channel's notification callbacks, but in this case the notification
callback is the nsPingListener object.
This patch extends nsPingListener to make it aware of
nsINetworkInterceptController, and have it route the request for
nsINetworkInterceptController correctly to the docshell without the need
to mess with the notification callbacks.
This will be tested in bug 1147699.
SharedWorker's interface requestor needs to know how to forward requests
for nsINetworkInterceptController to the docshell.
This will be tested in bug 1148935.
Currently when sending a CSP report, HttpBaseChannel::ShouldIntercept tries to
get access to the nsINetworkInterceptController interface through the channel's
notification callbacks, but in this case the notification callback is the
CSPReportRedirectSink object (thanks to nsCORSListenerProxy::Init).
This patch extends CSPReportRedirectSink to make it aware of
nsINetworkInterceptController, and have it route the request for
nsINetworkInterceptController correctly to the docshell without the need to
mess with the notification callbacks.
This will be tested in bug 1147699.
Currently when sending a beacon, HttpBaseChannel::ShouldIntercept tries
to get access to the nsINetworkInterceptController interface through the
channel's notification callbacks, but in this case the notification
callback is the nsCORSListenerProxy object (thanks to
nsCORSListenerProxy::Init).
nsCORSListenerProxy already knows how to forward calls to
nsIInterfaceRequestor::GetInterface to mOuterNotificationCallbacks, and
ShouldIntercept calls GetInterfce. But mOuterNotificationCallbacks is
set by default to the callbacks of the channel at the time
nsCORSListenerProxy is called, and the callbacks on this channel is
intentionally null, so ShouldIntercept bails out and the beacon never
gets intercepted.
This patch extends nsCORSListenerProxy to make it aware of
nsINetworkInterceptController, and have it route the request for
nsINetworkInterceptController correctly to the docshell without the need
to mess with the notification callbacks.
This will be tested in bug 1147699.
This patch renames the confusing IsRunning method since IsRunning() is *not*
the same as (PlayState() == AnimationPlayState::Running). It also removes
the old definition to make better re-use of PlayState() and IsInPlay().
This patch adds a method for testing if an animation is "in play" which is
a term defined in the Web Animations spec. This is in preparation for removing
some slightly redundant code in IsRunning and aligning better with the spec.
In preparation for introducing IsInPlay (where "in play" is a term in the Web
Animations spec), this patch aligns the existing IsCurrent with the definition
in the spec that says an animation effect is only current if it is attached
to an animation (player in our current naming) that is not finished. In order
to ensure that we need to pass the animation/player into the method.
This actually changes the behavior of IsCurrent since now we will return false
for animations that are finished. As far as I can tell, all the call sites that
are requesting current animations should only be concerned with animations that
are actually running. If not, they need to be adjusted to look for animations
that are either current or in effect.
