When we have a backwards fill and we sample at *exactly* the start of the
animation on the next refresh driver tick, when we get to
RestyleManager::ComputeStyleChangeFor (or more specifically
ElementRestyler::CaptureChange) we notice that the style hasn't changed (since
the first frame of the animation produces the same value as the backwards fill)
and end up with an empty change list. As a result we never schedule a view
manager flush and rebuild the layer. Hence, the animation never gets sent to the
compositor thread. On the next tick we're already throttling the main thread.
This patch fixes this by applying the same approach as is used for transitions,
that is, explicitly marking which animations are running on the compositor
thread so we know if we need to trigger a layer transaction or not. This should
not only be more robust than the previous code but also facilitate aligning
animations and transitions code (bug 880596).
Animations with a delay are not put on the compositor thread until the end of
the delay phase. However, there is currently nothing that explicitly triggers
this transaction. It may occur due to flushes that arise from UI events but it
is not guaranteed.
This patch detects the end of a delay phase and turns off throttling for that
sample. It re-uses the mLastNotification member which is not ideal but
a subsequent patch in this queue removes this and replaces it with the approach
used for transitions.
This changes the behavior of the CanPerformOnCompositorThread methods of
both ElementAnimations and ElementTransitions to check that the
respective animations or transitions are actually running. This is ok
because:
- The main caller is nsLayoutUtils::HasAnimationsForCompositor, and all
of its callers pretty clearly want the more restricted behavior (they're
concerned with layer activity)
- The only other callers of these functions are
nsAnimationManager::FlushAnimations and
nsTransitionManager::FlushTransitions (determining when to do
throttling), nsAnimationManager::GetAnimationsForCompositor (whose
only caller,
nsDisplayListBuilder::AddAnimationsAndTransitionsToLayer, also checks
IsRunningAt). I think these also all want or are fine with having
the IsRunningAt check.
As to the actual changes:
- In the animation manager, I think it's a mistake that
ElementAnimation::IsRunningAt didn't already check
mIterationDuration, since we throw out animations with a bad
iteration-duration in ElementAnimations::EnsureStyleRuleFor. So this
makes that change as well.
- In the transition manager, IsRunningAt already checks
!IsRemovedSentinel().
I've confirmed in gdb on a device that this fixes the repeated
nsIFrame::SchedulePaint calls that were the symptom of this bug.
I believe this patch also makes it so that a short animation of a
property that can't be animated on the compositor doesn't prevent the
entire duration of the animation of a property that can from being
throttled (having the main thread style updates suppressed).
Patch co-authored by Emmanuele Bassi <ebassi@gmail.com>
This defines a CSSVariableDeclarations class that holds a set of
variable declarations. This is at the specified value stage, so values
can either be 'initial', 'inherit' or a token stream (which is what you
normally have). The variables are stored in a hash table. Although
it's a bit of a hack, we store 'initial' and 'inherit' using special
string values that can't be valid token streams (we use "!" and ";").
Declaration objects now can have two CSSVariableDeclarations objects
on them, to store normal and !important variable declarations. So that
we keep preserving the order of declarations on the object, we inflate
mOrder to store uint32_ts, where values from eCSSProperty_COUNT onwards
represent custom properties. mVariableOrder stores the names of the
variables corresponding to those entries in mOrder.
We also add a new nsCSSProperty value, eCSSPropertyExtra_variable, which
is used to represent any custom property name.
nsCSSProps::LookupProperty can return this value.
The changes to nsCSSParser are straightforward. Custom properties
are parsed and checked for syntactic validity (e.g. "var(a,)" being
invalid) and stored on the Declaration. We use nsCSSScanner's
recording ability to grab the unparsed CSS string corresponding to
the variable's value.
This also changes the functionality a little bit to track independent
per-property mutation counts and independent "content active" status.
--HG--
extra : rebase_source : e69b8e7a95d36720bd38d74f0789ede603e58a09
This patch does the following:
* Move nsIFrame::IntrinsicSize to mozilla::IntrinsicSize so that it can
be forward-declared.
* Move a number of templated inline nsLayoutUtils methods to nsIFrame.
* Use mozilla::layout::FrameChildListID instead of the
nsIFrame::ChildListID typedef in nsLayoutUtils.h.
* Move nsReflowFrameRunnable to its only user, nsProgressMeterFrame.cpp.
* Make a number of functions requiring nsIFrame.h out-of-line.
* Remove the nsIFrame.h #include from nsLayoutUtils.h and add it to the
places which require it implicitly.
The fixes to the miniflush code
(nsTransitionManager::UpdateThrottledStyle and UpdateAllThrottledStyles)
fix the case where we constructed totally incorrect style contexts for
outer table frames (which have special style contexts inheriting from
the table frame) during the miniflush, leading to inconsistent style
data and other bad things, when we should have been touching the style
on the table frame instead.
The fixes to the other OMTA codepaths lead to layer tests being
performed on the same frame that the styles will be applied to, and
probably fix real bugs (which would occur when animating opacity or
transform on a table).
The fixes to the miniflush code
(nsTransitionManager::UpdateThrottledStyle and UpdateAllThrottledStyles)
fix the case where we constructed totally incorrect style contexts for
outer table frames (which have special style contexts inheriting from
the table frame) during the miniflush, leading to inconsistent style
data and other bad things, when we should have been touching the style
on the table frame instead.
The fixes to the other OMTA codepaths lead to layer tests being
performed on the same frame that the styles will be applied to, and
probably fix real bugs (which would occur when animating opacity or
transform on a table).
Note that this patch has a little bit of a belt-and-braces aspect to it.
In each file, either one of the changes should be sufficient, but one of
them prevents us from doing unneeded work and the other one ensures that
we never apply style resulting from transitions and animations even if
somehow we do that work.
Also note that the tests don't actually test anything usefully, since
the reftest harness doesn't currently make the pres context non-dynamic.
(Thus they're marked as failing.) I'm not sure what I should do about
that, though I'm considering just deleting the tests entirely.
Except for the changes in:
layout/generic/nsIFrame.h (part)
layout/style/nsComputedDOMStyle.h (all)
layout/style/nsRuleNode.cpp (part)
layout/style/nsStyleContext.cpp (part)
layout/style/nsStyleContext.h (part)
(see patch 3b in the bug), this patch was written with the sed script:
s/\<GetStyle\(Font\|Color\|List\|Text\|Visibility\|Quotes\|UserInterface\|TableBorder\|SVG\|Background\|Position\|TextReset\|Display\|Content\|UIReset\|Table\|Margin\|Padding\|Border\|Outline\|XUL\|SVGReset\|Column\)\>/Style\1/g
NOTE: The tests in test_animations.html fail without the patch; the
tests in test_shorthand_property_getters.html are only tangentially
related and pass both with and without the patch.
This implements my proposal in
http://lists.w3.org/Archives/Public/www-style/2011Apr/0381.html and
http://lists.w3.org/Archives/Public/www-style/2011Apr/0387.html . I
think it was a serious mistake to implement what the spec says, and I'm
fixing that mistake so that we have a chance to change the spec.
In other words, when an @keyframes rule has two key selectors at the
same time, the later one no longer overrides the entirety of the earlier
one. The overriding is done for each property that's in the later rule.
(And the -moz-animation-timing-function is taken only from the keyframe
actually used for the given property; if there's no declaration there
then the computed value of the property is used.)
The test for @keyframes cascade fails without the patch; the test for
@keyframes cascade2 tests behavior that works both before and after the
patch.
This inverts the relationship between segments and properties in the
animation data structures: now each property has a set of segments,
since the segments differ between properties.
Furthermore, we now handle inability to interpolate between values by
dropping the entire property rather than dropping a single segment.
Brian Birtles
Masayuki Nakano
L. David Baron
Nicholas Cameron
Cameron McCormack
Robert O'Callahan
Phil Ringnalda
Wes Kocher
Markus Stange
Ehsan Akhgari
Catalin Iacob
David Zbarsky
Olli Pettay
Daniel Holbert
Jeff Walden
Isaac Aggrey
Mark Finkle
Aryeh Gregor
Ed Morley
Matt Brubeck
Gervase Markham
Zack Weinberg
Nicholas Nethercote
Boris Zbarsky
Michael Wu
Craig Topper