A grid item spanning multiple grid rows/columns which include a flexible
track should not participate in the auto min-sizing calculation.
Resolves#2688
This is relevant if scripting is terminated due to execution timeout but there are setTimeout/Interval callbacks pending
or if scripting is disabled altogether.
QIing to CC interfaces is a hot path for a few widely-used classes.
Presumably there are many of these objects being created and destroyed.
By making these classes check first for the CC interfaces directly, rather
than going up the inheritance chain, the resulting overhead is reduced.
This callback is only used in very limited ways, so just require that the
caller passes in the canonical supports pointer plus the participant.
This probably won't affect performance much.
All DOMProxyHandler::ClearExternalRefsForWrapperRelease now does is clear
a backpointer to the expando, and that's handled during reflector
finalization already, making it redundant.
The nsISupports objects added to the purple buffer are already canonical,
so we can avoid some overhead by not QIing them to
`nsCycleCollectionISupports`.
This is a stub implementation because of the considerable abuse risk and
bad web code risk causing DoS, e.g. when using bad timeout values and
callbacks on the AbortSignal (e.g. recursive).
Considering the abort-ability of these signals otherwise there is never
a guarantee to websites that these signals will be triggered anyway, so
it should be considered an optional signal - meaning we can safely just
ignore the timeout, which allows us to avoid a ton of complexity and
potential sec/privacy issues.
Primary use in the wild seems to be detailed profiling/ad metrics anyway;
we don't really want to encourage that!
baseDomain is used as a key for cookies in a hashmap. For cookie
operations to work properly, the baseDomain generation must be reliable.
This is, however, not the case, because the result can change by updates
to the public suffix list (PSL).
Since the stored baseDomain is not reliable, the value must be recomputed
when the database is read. This causes a minor performance hit but is
needed if we want to keep abreast of PSL changes without clobbering the
cookies database.
Resolves#2670
This should avoid the intermittent crashes that are still being reported
on this issue that end up crashing because of a proxied object that can't
be unwrapped after being passed around.
