reftest is disabled on windows due to localized try readback errors that can't be reproduced.
MozReview-Commit-ID: 379PZsRE5d6
--HG--
extra : rebase_source : b990572c0f33860998eb5485824e417387d3e204
XPT has some functions and macros for freeing memory. However, they (a) are
only used on error paths, and (b) don't actually free memory -- they just
optionally log the "freeing" -- because piecewise freeing doesn't make sense
with arena allocation.
This patch removes all that unnecessary machinery.
--HG--
extra : rebase_source : 40fb8dfd3851a89aaf0501ae290b3a6ca8ac90bc
RegisterBuffer() is the only place that creates an XPTState, and it also
destroys it. So the XPTState can be allocated on the stack, which voids the
need for the creation of an XPTArena.
--HG--
extra : rebase_source : b25f0e798d72b8742efc96793a927f8a060101cf
XPT supports annotations but xpt.py doesn't generate them except for a single
empty annotation (to indicate there are no real annotations). So we can remove
almost all support for them. This also allows XPTString to be removed.
--HG--
extra : rebase_source : 889e635e9167f31c38c015bafe943f2c577f1daf
Currently XPT can both encode and decode, but encoding has been handled by
Python code since bug 643817, so the encoding support can be removed. This
results in many simplifications. Some notable changes:
- All the XPTHashTable code (including XPTDatapool::offset_map) is no longer
necessary.
- PrimitiveTest.cpp and SimpleTypeLib.cpp both don't make much sense without
encoding support, so I removed them.
- A lot of the version code was already unused, e.g. XPT_VERSION_*,
XPT_TYPELIB_VERSIONS_STRUCT, XPT_TYPELIB_VERSIONS.
XPT_MAJOR_INCOMPATIBLE_VERSION is the only thing actually used in version
checks.
- The patch also removes some code that was dead even before encoding removal,
such as XPT_ParseVersionString().
--HG--
extra : rebase_source : 11cfe0b01efde4e2ff0c74b02b408baebedd3dd8
Currently, CommonAnimationManager::GetAnimationCollection returns early when
the referenced list of animation collections is empty. So, for example, if
we try to get the collection of CSS animations on an element on a page with no
CSS animations, we will quickly return null without possibly expensive property
lookup. However, if there is just one CSS animation on the page, we will do the
property lookup for every element in the page where this method is called.
In this bug, we would like to remove the linked list of animation collections
since this is now the only place where it is used. So, in place if this
optimization, we introduce quite a different one based on the changes from bug
1226091 which makes MayHaveAnimations() apply to animations on the element
itself as well as pseudo elements. Using this, we can return early for any
element that has never had any kind of animation on it. The page may have
dozens of other animations but we will still return early. However, if the
element has ever had any kind of animation on it, we will not return early. It
is expected that this optimization is at least as good as the one it replaces.
In this bug we will trim off unnecessary functionality from the animation
managers and make AnimationCollection into an independent data type
so in this patch we separate it into its own file.
It is also generally easier to navigate the source code and eliminate
cyclic dependencies between header files when there is a rough
correspondance between class names and file names (e.g. rather than having
#include "AnimationCommon.h" // For mozilla::AnimationCollection).
This patch also makes a few simplifications to include dependencies since
they're a bit of a mess (making it hard to move code around). The changes to
IncrementalClearCOMRuleArray.cpp are due to the changes to the unified build
introduced by adding AnimationCollection.cpp exposing a missing include from
that file.
This was added in bug 780692 to work around assertions that arose due to the
inconsistent state introduced by mini-flushes. However, that workaround
no longer seems necessary. In particular, the crashtest for bug 813372 no
longer reports failed assertions when we remove this method and nor do any
other tests.
I'm not sure exactly what changed about how we do mini-flushes but I suspect
it was bug 960465 or one of the related follow-ups.
After patch 7, the order of animations in MutationObserver has been reversed.
When we want to use the animations ordering by something, we need to sort it first.
MozReview-Commit-ID: 7xOYNsuSbvA
This patch removes a loop for the new temporary animation collection in
CheckAnimationRule. The old collection is passed to CSSAnimationBuilder,
and CSSAnimationBuilder removes each animation which matches to new animation
name in it.
:birtles took care of storing animations in AnimationCollection in reverse order.
Thanks so much!
MozReview-Commit-ID: KmlnjFptKdv
