The <fill-rule> in the polygon() syntax is not handled because it doesn't
matter to shape-outside at all.
The reftests are numbered from 018 to avoid conflict with the w3c upstream
ones according to this list.
https://test.csswg.org/harness/results/css-shapes-1_dev/grouped/
Reftest 018 to 025 are under various writing-modes, 026 to 029 are testing
empty float area, 030 to 031 are testing polygon containing horizontal
lines.
MozReview-Commit-ID: FPUbMdZsvu6
Remove WritingMode parameter from LineRight() and LineLeft() in both
FloatInfo and ShapeInfo.
Bug 1316549 Part 3 added the parameter to compute the border radii under
writing-modes correctly. However, bug 1326407 Part 6 later made us cache
border radii, so the WritingMode parameter is no longer needed.
MozReview-Commit-ID: DFhm5d51OXJ
The radii has been cached in the BoxShapeInfo in the previous part. Hence
the rename.
This class will be used to implement inset() in the next part, so the rect
stored isn't necessarily the rect of the <shape-box>. It could be the inset
rectangle. Therefore I rename mShapeBoxRect to mRect to avoid any confusion.
MozReview-Commit-ID: J0hpQDsbMyN
The radii can be computed when creating BoxShapeInfo. No need to compute
them every time in the LineLeft() and LineRight().
MozReview-Commit-ID: GIDSLgickCT
Make the name and the order of arguments be consistent with the
ConvertPhysicalToLogical() method added in the previous part.
MozReview-Commit-ID: 1YARDzI3cyr
The life cycle of nsFloatManager managed by mNew is same as
nsAutoFloatManager, which lives only in nsBlockFrame::Reflow(). Therefore,
other nsFloatManager pointers are all non-owning ref to the
nsAutoFloatManager::mNew.
MozReview-Commit-ID: B34BOcsjE2X
The difference between CircleShapeInfo's constructor and EllipseShapeInfo's
is the computation of the radii. Therefore, this patch creates a factory
function to distinguish that, so shape-outside: circle() could be
implemented by using EllipseShapeInfo.
MozReview-Commit-ID: 9ZBQu8zCSrM
Instead of manually adding (aLineLeft, aBlockStart) when creating a shape,
add Translate() to let the shapes implement their only way to move their
origin. FloatInfo could then move the shapes after they're created.
MozReview-Commit-ID: ApZBHnkng74
circle() allows the user to define an empty flow area, so IsEmpty() needs to
be overridden.
The flow area defined by a shape needs to be clipped to the margin-box per
https://drafts.csswg.org/css-shapes/#relation-to-box-model-and-float-behavior
In the reftests, both clip-path and shape-outside uses the same value so
that it's easier to debug visually.
Add LogicalPoint::LineRelative() because we need to convert a point's I() to
the line-axis in nsFloatManager. LineRelative() differs from I() in all
'rtl' direction per
https://drafts.csswg.org/css-writing-modes-3/#logical-to-physical
MozReview-Commit-ID: FxQaFPrEQ73
Create ShapeInfo as a base class for implementing all the shapes. In this
design, we only need to create the correct subclass in FloatInfo's
constructor whenever shape-outside is used rather than manually branching on
StyleShapeSourceType and StyleBasicShape in all the methods like
LineRight(), LineLeft(), etc.
The concrete subclass of ShapeInfo could focus on implementing how its shape
influence the flow area by overriding the needed methods in ShapeInfo.
Move ComputeEllipseLineInterceptDiff() and XInterceptAtY() under the scope
of ShapeInfo so that they could be used by BoxShapeInfo and all the other
ShpapeInfo subclasses yet to come.
MozReview-Commit-ID: ETVc5FdGNha
Use move constructor for two reasons. 1) The copy constructor is needed only
when appending FloatInfo to mFloats, so using move constructor will likely
be more efficient if some of the member variables support move constructor.
2) Part 6 will added a UniquePtr member to FloatInfo, so using move
constructor becomes necessary.
Also change the return value of AddFloat() to void to simplify the code,
since all the other callers do not check the return value, and
BlockReflowInput::FloatAndPlaceFloat() only asserts in debug mode. I assume
it's safe to omit the OOM check.
MozReview-Commit-ID: GVbbsdBjr7b
circle() allows the user to define an empty flow area, so IsEmpty() needs to
be overridden.
The flow area defined by a shape needs to be clipped to the margin-box per
https://drafts.csswg.org/css-shapes/#relation-to-box-model-and-float-behavior
In the reftests, both clip-path and shape-outside uses the same value so
that it's easier to debug visually.
Add LogicalPoint::LineRelative() because we need to convert a point's I() to
the line-axis in nsFloatManager. LineRelative() differs from I() in all
'rtl' direction per
https://drafts.csswg.org/css-writing-modes-3/#logical-to-physical
MozReview-Commit-ID: FxQaFPrEQ73
Create ShapeInfo as a base class for implementing all the shapes. In this
design, we only need to create the correct subclass in FloatInfo's
constructor whenever shape-outside is used rather than manually branching on
StyleShapeSourceType and StyleBasicShape in all the methods like
LineRight(), LineLeft(), etc.
The concrete subclass of ShapeInfo could focus on implementing how its shape
influence the flow area by overriding the needed methods in ShapeInfo.
Move ComputeEllipseLineInterceptDiff() and XInterceptAtY() under the scope
of ShapeInfo so that they could be used by BoxShapeInfo and all the other
ShpapeInfo subclasses yet to come.
MozReview-Commit-ID: ETVc5FdGNha
Use move constructor for two reasons. 1) The copy constructor is needed only
when appending FloatInfo to mFloats, so using move constructor will likely
be more efficient if some of the member variables support move constructor.
2) Part 6 will added a UniquePtr member to FloatInfo, so using move
constructor becomes necessary.
Also change the return value of AddFloat() to void to simplify the code,
since all the other callers do not check the return value, and
BlockReflowInput::FloatAndPlaceFloat() only asserts in debug mode. I assume
it's safe to omit the OOM check.
MozReview-Commit-ID: GVbbsdBjr7b
circle() allows the user to define an empty flow area, so IsEmpty() needs to
be overridden.
The flow area defined by a shape needs to be clipped to the margin-box per
https://drafts.csswg.org/css-shapes/#relation-to-box-model-and-float-behavior
In the reftests, both clip-path and shape-outside uses the same value so
that it's easier to debug visually.
Add LogicalPoint::LineRelative() because we need to convert a point's I() to
the line-axis in nsFloatManager. LineRelative() differs from I() in all
'rtl' direction per
https://drafts.csswg.org/css-writing-modes-3/#logical-to-physical
MozReview-Commit-ID: FxQaFPrEQ73
Create ShapeInfo as a base class for implementing all the shapes. In this
design, we only need to create the correct subclass in FloatInfo's
constructor whenever shape-outside is used rather than manually branching on
StyleShapeSourceType and StyleBasicShape in all the methods like
LineRight(), LineLeft(), etc.
The concrete subclass of ShapeInfo could focus on implementing how its shape
influence the flow area by overriding the needed methods in ShapeInfo.
Move ComputeEllipseLineInterceptDiff() and XInterceptAtY() under the scope
nsFloatManager so that they could be used by BoxShapeInfo and all the other
ShpapeInfo subclasses yet to come.
MozReview-Commit-ID: ETVc5FdGNha
Use move constructor for two reasons. 1) The copy constructor is needed only
when appending FloatInfo to mFloats, so using move constructor will likely
be more efficient if some of the member variables support move constructor.
2) Part 6 will added a UniquePtr member to FloatInfo, so using move
constructor becomes necessary.
Also change the return value of AddFloat() to void to simplify the code,
since all the other callers do not check the return value, and
BlockReflowInput::FloatAndPlaceFloat() only asserts in debug mode. I assume
it's safe to omit the OOM check.
MozReview-Commit-ID: GVbbsdBjr7b
The tests cases are designed based on the integer solution to the ellipse
equation (x/a)^2 + (y/b)^2 = 1, where x=36, y=32, a=60, b=40.
MozReview-Commit-ID: De2fXcb6ypP
In GetFlowArea(), "Shrink our band's height if needed." computation was
moved to the end because we need to pass the unmodified |blockEnd| to
compute LineRight() and LineLeft().
Revamp OutsetBorderRadii() to allow negative margin to reduce the radius,
but not below zero. Also implement the cubic formula required by the spec.
https://drafts.csswg.org/css-shapes/#valdef-shape-box-margin-box
OutsetBorderRadii() is now tailored only for margin-box with border-radius,
so it might no longer be suitable for other scenarios.
MozReview-Commit-ID: HKxW7rp6sIA
Per spec, float positioning and stacking is not affected by defining a float
area with a shape.
https://drafts.csswg.org/css-shapes/#relation-to-box-model-and-float-behavior
So all the call sites of GetFloatAvailableSpace() related to adding a
float are replaced by GetFloatAvailableSpaceForPlacingFloat().
<shape-box> with border-radius will be implemented in next part.
MozReview-Commit-ID: 1RXEeXDhdWo
Also exchange the order of the second argument |aBSize| and the third
argument |aBandInfo| for GetFlowArea() so that aBCoord and aBSize are
grouped together. And it'll be prettier to add ShapeInfo behind BandInfo
in later part.
MozReview-Commit-ID: cusdomjUyu
The operator new is infallable. Also change the return value to void
accordingly. The only caller doesn't even check the return value.
MozReview-Commit-ID: 3whf7s1d35q
After using enum class, a switch-case warning in CombineBreakType is caught.
This is one of such kind safty checks that we would like to gain.
Fix it by adding default case for switch-case in CombineBreakType.
MozReview-Commit-ID: BdS3LPN6qzX
