The core of this change is in gfxContext.*:
- change gfxContext::CurrentMatrix() and gfxContext::SetMatrix() to
return and take a Matrix respectively, instead of converting to
and from a gfxMatrix (which uses doubles). These functions therefore
will now match the native representation of the transform in gfxContext.
- add two new functions CurrentMatrixDouble() and SetMatrixDouble() that
do what the old CurrentMatrix() and SetMatrix() used to do, i.e.
convert between the float matrix and the double matrix.
The rest of the change is just updating the call sites to avoid round-
tripping between floats and doubles where possible. Call sites that are
hard to fix are migrated to the new XXXDouble functions which preserves
the existing behaviour.
MozReview-Commit-ID: 5sbBpLUus3U
This patch was generated automatically by the "modeline.py" script, available
here: https://github.com/amccreight/moz-source-tools/blob/master/modeline.py
For every file that is modified in this patch, the changes are as follows:
(1) The patch changes the file to use the exact C++ mode lines from the
Mozilla coding style guide, available here:
https://developer.mozilla.org/en-US/docs/Mozilla/Developer_guide/Coding_Style#Mode_Line
(2) The patch deletes any blank lines between the mode line & the MPL
boilerplate comment.
(3) If the file previously had the mode lines and MPL boilerplate in a
single contiguous C++ comment, then the patch splits them into
separate C++ comments, to match the boilerplate in the coding style.
MozReview-Commit-ID: EuRsDue63tK
There are two callers of nsFilterInstance::PaintFilteredFrame:
1. nsSVGUtils::PaintFrameWithEffects at [1]
This function is used while painting a filtered element on a path which
display item painting is not supported yet, such as drawing elements inside a
indirect painted SVG object, such as a mask, a clipPath or a gradient object.
Let's say we have a masked element, which refers to an SVG mask, and there is
a filtered element inside that SVG mask.
Using nsFilterInstance::PaintFilteredFrame to paint that filtered frame in
the mask, we have to pass a gfxContext and a transform matrix to it. The
transform of the gfxContext 'target' that we pass in consists of a transform
from the referenced frame, of the masked frame, to the masked frame. We also
pass in a transform matrix 'aTransform', this matrix contains a transform
from the the masked frame to the filtered frame in *device units*, which
means it contains css-to-dev-px scaling factor.
2. nsSVGIntegrationUtils::PaintFilter at [2]
This function is used by normal display item painting.
The same, we pass a gfxContext 'context' and a transform matrix 'tm' into
nsFilterInstance::PaintFilteredFrame. The transform matrix of 'context'
consists of a transform from the referenced frame, of the filtered frame,
to this filtered frame, but the scale factor was taken out . The transform
matrix 'tm' we pass in contains scale transform from the referenced frame to
the filtered frame in *device unit*.
Inside nsFilterInstance::PaintFilteredFrame, we treat the transform matrix of
'aCtx' and 'aTransform' as parameters we pass in in #2 caller. So it can be
failed in #1. For example, if the filtered frame inside a masked frame has a
translation transform applied, since that translation was put in 'aTransfrom',
but we only use the scale factor of 'aTransform' in
nsFilterInstance::PaintFilteredFrame, translation factor disappears.
In this patch, I unified the definition of parameters of
nsFilterInstance::PaintFilteredFrame:
1. nsFilterInstance::PaintFilteredFrame(aCtx): the transform matrix of aCtx
should be a transform from the referenced frame to the filtered frame in
*css units*. Originally, the aCtx we passed in #1 is in device units, which
should be fixed; the aCtx we passed in #2 does not even include css scaling
factor, need be fixed too.
2. nsFilterInstance::PaintFilteredFrame(aTransform): this transform matrix
should contain only scaling factor in device units. And I removed it in the
end since I found we can get this value easily right inside the callee.
[1]
https://hg.mozilla.org/mozilla-central/file/ef585ac7c476/layout/svg/nsSVGUtils.cpp#l857
[2]
https://hg.mozilla.org/mozilla-central/file/ef585ac7c476/layout/svg/nsSVGIntegrationUtils.cpp#l1114
MozReview-Commit-ID: gRV128NyQv
One thing to note here is that the Scale function on gfxRect has a
different implementation than that in gfx::Rect which is replacing it.
The former just scales the width/height directly whereas the latter
scales the XMost/YMost and recomputes the width/height.
MozReview-Commit-ID: 5FImdIaNfC3
Most of this patch is updating a few places that use gfxMatrix to use
the equivalent-but-differently-named functions on MatrixDouble:
- Translate/Rotate/Scale get turned into PreTranslate/PreRotate/PreScale
- Transform(Point) gets turned into TransformPoint(Point)
- gfxMatrix::TransformBounds(gfxRect) gets turned into
gfxRect::TransformBoundsBy(gfxMatrix).
- gfxMatrix::Transform(gfxRect) gets turned into
gfxRect::TransformBy(gfxMatrix).
The last two functions are added in this patch as convenience wrappers
to gfxRect instead of Matrix.h because we don't want Matrix.h to "know"
about gfxRect (to avoid adding gecko dependencies on Moz2D). Once we
turn gfxRect into a typedef for RectDouble these will be eliminated
anyway.
MozReview-Commit-ID: BnOjHzmOSKn
As I've said before, as module owner I prefer that MOZ_ASSERT_IF not be
used in the module because I consider it to be unreadable. However, a
few uses have crept in, and this patch removes them.
I consider it to be unreadable because the name looks like a name that
uses smalltalk-ish naming conventions, i.e., with a part of the name
corresponding to each parameter, in order. However, the parameters are
in the order opposite the name.
This was written primarily with the vim commands:
:%s/MOZ_ASSERT_IF(\([^,]*\),/MOZ_ASSERT(!\1 ||/
:wn
followed by manual cleanup for indentation and removal of !!.
MozReview-Commit-ID: G6rLbOn7k8d
The reason of this change is the same as for Part 2, except that this commit fixes
nsSVGPaintServerFrame::GetPaintServerPattern rather than PaintSVG.
Commit-ID: 691YrKZ0Lm9
MozReview-Commit-ID: KSnFhCndFUk
The reason of this change is the same as for Part 2, except this commit fixes
nsSVGMaskFrame::GetMaskForMaskedFrame rather than PaintSVG.
MozReview-Commit-ID: DS0eG6eKDgs
The DrawResult return was not in fact anything to do with the success or
failure of that method, but was actually passing out a very specific piece of information
about the success or failure of any imagelib drawing that may not have occurred
under the various PaintSVG calls.
The signature of PaintSVG is changed from
DrawResult PaintSVG(...., uint32 flags);
to
void PaintSVG(...., imgDrawingParams& aPackage);
imgDrawingParams wraps DrawResult and imgIContainer::FLAG_* as a pack, pass through
PaintSVG to imagelib draw calls under beneath.
MozReview-Commit-ID: IOq2evUAOQF
The reason of this change is the same as for Part 2, except that this commit fixes
nsSVGPaintServerFrame::GetPaintServerPattern rather than PaintSVG.
Commit-ID: 691YrKZ0Lm9
MozReview-Commit-ID: KSnFhCndFUk
The reason of this change is the same as for Part 2, except this commit fixes
nsSVGMaskFrame::GetMaskForMaskedFrame rather than PaintSVG.
MozReview-Commit-ID: DS0eG6eKDgs
The DrawResult return was not in fact anything to do with the success or
failure of that method, but was actually passing out a very specific piece of information
about the success or failure of any imagelib drawing that may not have occurred
under the various PaintSVG calls.
The signature of PaintSVG is changed from
DrawResult PaintSVG(...., uint32 flags);
to
void PaintSVG(...., imgDrawingParams& aPackage);
imgDrawingParams wraps DrawResult and imgIContainer::FLAG_* as a pack, pass through
PaintSVG to imagelib draw calls under beneath.
MozReview-Commit-ID: IOq2evUAOQF
GetPreEffectsVisualOverflowRect is a function that be used in both reflowing
and painting. PreEffectsBBoxProperty of continuations is not set correctly yet
in reflow process, so we should skip this assertion check during reflow.
PS:
I will change MOZ_ASSERT, in GetPreEffectsVisualOverflowRect, back to NS_ASSERTION
if we still encounter this assertion in the future.
MozReview-Commit-ID: 6Gvg3XOd2mD
I did many change in many files in this patch. But the goal is pretty simple: To
pass the return value of nsSVGPatternFrame::PaintPattern back to the caller
(nsDisplaySVGGeometry). My suggestion is to review this patch right from
nsSVGPatternFrame.cpp.
I made two mistakes in bug 1258510
1. We should not return directly at [1]. RemoveStateBits at l418 will be skip.
2. nsSVGPatternFrame::PaintPattern should return both SourceSurface and draw
result, so that we can update UpdateDrawResult in display item.
All the other changes are to
1. make sure the return value of nsSVGPatternFrame::PaintPattern goes back to
nsDisplaySVGGeometry::Paint correctly.
2. Since the return value of nsSVGPatternFrame::PaintPattern change, we need
modify all existed callers.
I also filed bug 1346124 for handle the returning value of PaintMarkers.
[1] https://hg.mozilla.org/mozilla-central/file/c0700bedb4f7/layout/svg/nsSVGPatternFrame.cpp#l415
MozReview-Commit-ID: Iq9RPQ6Omz0
I did many change in many files in this patch. But the goal is pretty simple: To
pass the return value of nsSVGPatternFrame::PaintPattern back to the caller
(nsDisplaySVGGeometry). My suggestion is to review this patch right from
nsSVGPatternFrame.cpp.
I made two mistakes in bug 1258510
1. We should not return directly at [1]. RemoveStateBits at l418 will be skip.
2. nsSVGPatternFrame::PaintPattern should return both SourceSurface and draw
result, so that we can update UpdateDrawResult in display item.
All the other changes are to
1. make sure the return value of nsSVGPatternFrame::PaintPattern goes back to
nsDisplaySVGGeometry::Paint correctly.
2. Since the return value of nsSVGPatternFrame::PaintPattern change, we need
modify all existed callers.
I also filed bug 1346124 for handle the returning value of PaintMarkers.
[1] https://hg.mozilla.org/mozilla-central/file/c0700bedb4f7/layout/svg/nsSVGPatternFrame.cpp#l415
MozReview-Commit-ID: Iq9RPQ6Omz0
After fighting with this assertion several months, I decided to remove it for
two reasons:
This assertion allows PreEffectBBoxProperty not being cached only under specific
condition. But the condition is wider then we expect.
1.
PreEffectsBBoxProperty is cached by nsIFrame::FinishAndStoreOverflow(this
function calls ComputeEffectsRect which cache this property actually) and
it is called from nsXXXFrame::Reflow on demand. Yes, *on demand*, not always.
And this is the fist reason that why I think we should just remove
this assertion.
For example, nsBlockFrame::Reflow calls FinishAndStoreOverflow to store this
property. But like BRFrame, it does not call FinishAndStoreOverflow at all.
In anohter word, if you apply any SVG effect to a BRFrame, you will always hit
this assertion. Here is an example:
<br style="filter: saturate(0%);"/>
So, if we still want to keep this assertion, we may need to create a list which
list all frame types that cache PreEffectsBBoxProperty, and do this check only if
the type of aFrame is listed. This is error prone since we may introduce a
new frame type at any time and forget to update this table.
2.
So, I think it's better just removing this assertion. The assertion that we
really need is the next one(2nd one):
MOZ_ASSERT(!preTransformOverflows,
"GetVisualOverflowRect() won't return the pre-effects rect!");
Since hitting that assertion, the 2nd one, means caller will retrieve wrong
effect region. Hitting the first assertion only means we do not cache
PreEffectsBBoxProperty, it's pretty normal and not hurt anything. This is the
second reason that I think we should remvoe this assertion.
MozReview-Commit-ID: JfiYTiP2laG
