Adds another restyle generation which represents the dirty state of raw
style changes, so that getComputedStyle() wont be confused by optimizations
made by style engines.
MozReview-Commit-ID: 7RYeNCzFygO
Before this refactoring, getComputedStyle could have side effects, and left the
style data in the element, so we could never arrive there without data.
There are a few crashtests that caught this, but this was already broken if you
called animate() on an element deep in a display: none subtree.
MozReview-Commit-ID: 1AvOvhAyOP3
This patch just moves the calculation of the pseudo type earlier in the method
so we can re-use it in the next patch in this series to get the style context
from the Servo backend for the case where we don't need to resolve style.
This patch also renames the local variable from 'type' to 'pseudoType' since
that seems less ambiguous.
MozReview-Commit-ID: 6pi2F1vZYHJ
This change does the following:
* Introduce a new smart pointer called CounterStylePtr which either
holds an AnonymousCounterStyle strongly, or a named CounterStyle
managed by CounterStyleManager weakly, and use it to replace all
RefPtr<CounterStyle> around the codebase.
* Rename CounterStyleManager::mCacheTable to mStyles to reflect the
fact that it is used to manage all styles, not just for caching.
* Add a retired styles list which collect all named CounterStyle
evicted from mStyles, and post a PostRefreshObserver to destroy
objects in that list after next flush.
* Remove helper functions for counter style in nsStyleList and expose
mCounterStyle directly, to make code simpler with the new pointer.
Reason for adding a new smart pointer type rather than making their
AddRef/Release behave like BuiltinCounterStyle is that, it is possible
that after a flush, some stale style structs may still be alive. They
can contain pointer to destroyed CounterStyle objects. Although the
actual content may never be accessed anymore, RefPtr may still access
the object for refcounting during destruction.
MozReview-Commit-ID: xxegwSDhNb
In our current setup, in which links with an href attribute always match either
:link or :visited, no matter whether that attribute's value is a valid URI,
changes to the attribute always put the element into either the "match nothing"
state or the "match :link" state, via calls to Link::ResetLinkState.
The only thing FlushPendingLinkUpdates is needed for is (lazily, in case it
turns out to not be needed because the element got removed from the DOM anyway)
registering a history observer to switch the link state to :visited as needed.
This means that selector matching consumers that would never expose :visited
state to start with don't need to worry about calling FlushPendingLinkUpdates.
Specifically:
* This patch uses a flag added in a prior patch to let us use the author's
chosen prefix (-webkit or -moz) when serializing. (We treat the -moz version
as a special case, because that makes it more straightforward to unsupport
-moz if/when we can.)
* This patch makes us share the linear-gradient() side-or-corner serialization
codepath when serializing points for -webkit-linear-gradient. (The
alternative is the -moz-linear-gradient codepath, which defaults to
serializing with percent values 0%/100% for sides & corners -- and raw
percentages are invalid in -webkit-linear-gradient(), so we can't share that
codepath.) Notably, we have to skip the "to " token that the
linear-gradient() codepath would normally print out -- that was a late
addition to the spec and so it only exists in the unprefixed modern syntax.
(Instead, -webkit-linear-gradient syntax is implicitly "from" the given
point).
MozReview-Commit-ID: 9Oqo8nG1XDU
This patch doesn't change our behavior -- we'll still produce the same serialization that we would've before.
MOTIVATION: A later patch will make us share this codepath to serialize into
-webkit-linear-gradient() syntax. That syntax uses the same representation for
points as unprefixed modern linear-gradient() (with box-side-or-corner keywords
"top", "right", etc.), but it does *not* use the word "to". So we'd like to
allow "to"-and-its-subsequent-space-character to be optional. Hence, this patch
groups the space together with "to", rather than as a prefix on the next token,
so that we can skip right to printing the point (without a leading space)
when we add support for -webkit-linear-gradient() serialization.
MozReview-Commit-ID: 5fCzx4NmpcC
In this patch, we implement nsIDOMWindowUtils::GetUnanimatedComputedStyle
which returns computed value of given CSS property without animation rule. This
method is used from the DevTools animation inspector to fill in keyframe values
when the property value is null (indicating that the underlying/base value is
being used).
In order to implement this, we extend nsComputedDOMStyle constructor to fetch
the computed style minus animation style (i.e. the base style). This is somewhat
complicated by the fact that for discrete animation.
StyleAnimationValue::ExtractComputedValue may return ‘unset’, ‘initial’ or
‘inherit’. For example, if the author uses the 'unset' 'initial' or 'inherit'
keyword for a discrete property (e.g. 'align-content’), ExtractComputedValue
returns the keywords as-is. Furthermore, if the user does not set any specific
keyword, ExtractComputedValue returns ‘unset’. We use this new
nsComputedDOMStyle mechanism to resolve these keywords into a valid keyword for
computed style in the same way as other properties (e.g. ‘opacity’).
MozReview-Commit-ID: HffJ9SCDf2k
