2017-10-27 20:33:53 +03:00
|
|
|
/* -*- Mode: C++; tab-width: 8; indent-tabs-mode: nil; c-basic-offset: 2 -*- */
|
|
|
|
/* vim: set ts=8 sts=2 et sw=2 tw=80: */
|
2012-05-21 15:12:37 +04:00
|
|
|
/* This Source Code Form is subject to the terms of the Mozilla Public
|
|
|
|
* License, v. 2.0. If a copy of the MPL was not distributed with this
|
|
|
|
* file, You can obtain one at http://mozilla.org/MPL/2.0/. */
|
2006-03-30 09:56:38 +04:00
|
|
|
|
|
|
|
/* constants for what needs to be recomputed in response to style changes */
|
|
|
|
|
2002-08-24 18:41:28 +04:00
|
|
|
#ifndef nsChangeHint_h___
|
|
|
|
#define nsChangeHint_h___
|
|
|
|
|
2015-08-04 13:20:20 +03:00
|
|
|
#include "mozilla/Types.h"
|
2020-02-11 04:22:15 +03:00
|
|
|
#include "mozilla/Assertions.h"
|
2015-08-05 15:42:20 +03:00
|
|
|
|
2002-08-24 18:41:28 +04:00
|
|
|
// Defines for various style related constants
|
|
|
|
|
2017-04-28 10:47:23 +03:00
|
|
|
enum nsChangeHint : uint32_t {
|
2016-11-02 09:11:24 +03:00
|
|
|
nsChangeHint_Empty = 0,
|
|
|
|
|
2009-08-04 05:22:13 +04:00
|
|
|
// change was visual only (e.g., COLOR=)
|
2012-02-08 02:52:59 +04:00
|
|
|
// Invalidates all descendant frames (including following
|
|
|
|
// placeholders to out-of-flow frames).
|
2015-12-01 08:37:08 +03:00
|
|
|
nsChangeHint_RepaintFrame = 1 << 0,
|
2009-08-04 05:22:13 +04:00
|
|
|
|
|
|
|
// For reflow, we want flags to give us arbitrary FrameNeedsReflow behavior.
|
2012-02-08 02:52:59 +04:00
|
|
|
// just do a FrameNeedsReflow.
|
2015-12-01 08:37:08 +03:00
|
|
|
nsChangeHint_NeedReflow = 1 << 1,
|
2009-08-04 05:22:13 +04:00
|
|
|
|
|
|
|
// Invalidate intrinsic widths on the frame's ancestors. Must not be set
|
|
|
|
// without setting nsChangeHint_NeedReflow.
|
2015-12-01 08:37:08 +03:00
|
|
|
nsChangeHint_ClearAncestorIntrinsics = 1 << 2,
|
2009-08-04 05:22:13 +04:00
|
|
|
|
|
|
|
// Invalidate intrinsic widths on the frame's descendants. Must not be set
|
2017-03-21 11:33:05 +03:00
|
|
|
// without also setting nsChangeHint_ClearAncestorIntrinsics,
|
|
|
|
// nsChangeHint_NeedDirtyReflow and nsChangeHint_NeedReflow.
|
2015-12-01 08:37:08 +03:00
|
|
|
nsChangeHint_ClearDescendantIntrinsics = 1 << 3,
|
2009-08-04 05:22:13 +04:00
|
|
|
|
|
|
|
// Force unconditional reflow of all descendants. Must not be set without
|
2017-03-21 11:33:05 +03:00
|
|
|
// setting nsChangeHint_NeedReflow, but can be set regardless of whether the
|
|
|
|
// Clear*Intrinsics flags are set.
|
2015-12-01 08:37:08 +03:00
|
|
|
nsChangeHint_NeedDirtyReflow = 1 << 4,
|
2009-08-04 05:22:13 +04:00
|
|
|
|
|
|
|
// The currently shown mouse cursor needs to be updated
|
2019-10-08 01:29:01 +03:00
|
|
|
nsChangeHint_UpdateCursor = 1 << 5,
|
2009-08-04 05:22:13 +04:00
|
|
|
|
2008-08-06 04:55:07 +04:00
|
|
|
/**
|
2012-11-07 17:54:55 +04:00
|
|
|
* Used when the computed value (a URI) of one or more of an element's
|
|
|
|
* filter/mask/clip/etc CSS properties changes, causing the element's frame
|
|
|
|
* to start/stop referencing (or reference different) SVG resource elements.
|
|
|
|
* (_Not_ used to handle changes to referenced resource elements.) Using this
|
2017-08-30 17:58:31 +03:00
|
|
|
* hint results in SVGObserverUtils::UpdateEffects being called on the
|
|
|
|
* element's frame.
|
2008-08-06 04:55:07 +04:00
|
|
|
*/
|
2019-10-08 01:29:01 +03:00
|
|
|
nsChangeHint_UpdateEffects = 1 << 6,
|
2009-08-04 05:22:13 +04:00
|
|
|
|
2010-07-16 01:07:53 +04:00
|
|
|
/**
|
|
|
|
* Visual change only, but the change can be handled entirely by
|
|
|
|
* updating the layer(s) for the frame.
|
2012-02-08 02:52:59 +04:00
|
|
|
* Updates all descendants (including following placeholders to out-of-flows).
|
2010-07-16 01:07:53 +04:00
|
|
|
*/
|
2019-10-08 01:29:01 +03:00
|
|
|
nsChangeHint_UpdateOpacityLayer = 1 << 7,
|
2012-02-08 02:52:59 +04:00
|
|
|
/**
|
|
|
|
* Updates all descendants. Any placeholder descendants' out-of-flows
|
|
|
|
* are also descendants of the transformed frame, so they're updated.
|
|
|
|
*/
|
2019-10-08 01:29:01 +03:00
|
|
|
nsChangeHint_UpdateTransformLayer = 1 << 8,
|
2010-07-16 01:07:53 +04:00
|
|
|
|
2012-02-08 02:52:59 +04:00
|
|
|
/**
|
|
|
|
* Change requires frame change (e.g., display:).
|
2016-07-06 08:06:13 +03:00
|
|
|
* Reconstructs all frame descendants, including following placeholders
|
|
|
|
* to out-of-flows.
|
|
|
|
*
|
|
|
|
* Note that this subsumes all the other change hints. (see
|
|
|
|
* RestyleManager::ProcessRestyledFrames for details).
|
2012-02-08 02:52:59 +04:00
|
|
|
*/
|
2019-10-08 01:29:01 +03:00
|
|
|
nsChangeHint_ReconstructFrame = 1 << 9,
|
2011-06-16 01:03:49 +04:00
|
|
|
|
2012-02-08 02:52:59 +04:00
|
|
|
/**
|
2014-06-17 18:32:00 +04:00
|
|
|
* The frame's overflow area has changed. Does not update any descendant
|
2014-03-14 04:21:25 +04:00
|
|
|
* frames.
|
2014-03-14 04:21:25 +04:00
|
|
|
*/
|
2019-10-08 01:29:01 +03:00
|
|
|
nsChangeHint_UpdateOverflow = 1 << 10,
|
2014-03-14 04:21:25 +04:00
|
|
|
|
2014-06-17 18:32:00 +04:00
|
|
|
/**
|
|
|
|
* The overflow area of the frame and all of its descendants has changed. This
|
|
|
|
* can happen through a text-decoration change.
|
|
|
|
*/
|
2019-10-08 01:29:01 +03:00
|
|
|
nsChangeHint_UpdateSubtreeOverflow = 1 << 11,
|
2014-06-17 18:32:00 +04:00
|
|
|
|
2014-03-14 04:21:25 +04:00
|
|
|
/**
|
|
|
|
* The frame's overflow area has changed, through a change in its transform.
|
2015-02-16 06:34:29 +03:00
|
|
|
* In other words, the frame's pre-transform overflow is unchanged, but
|
|
|
|
* its post-transform overflow has changed, and thus its effect on its
|
|
|
|
* parent's overflow has changed. If the pre-transform overflow has
|
|
|
|
* changed, see nsChangeHint_UpdateOverflow.
|
2014-03-14 04:21:25 +04:00
|
|
|
* Does not update any descendant frames.
|
|
|
|
*/
|
2019-10-08 01:29:01 +03:00
|
|
|
nsChangeHint_UpdatePostTransformOverflow = 1 << 12,
|
2014-03-14 04:21:25 +04:00
|
|
|
|
2015-02-16 06:34:29 +03:00
|
|
|
/**
|
|
|
|
* This frame's effect on its parent's overflow area has changed.
|
|
|
|
* (But neither its pre-transform nor post-transform overflow have
|
|
|
|
* changed; if those are the case, see
|
|
|
|
* nsChangeHint_UpdatePostTransformOverflow.)
|
|
|
|
*/
|
2019-10-08 01:29:01 +03:00
|
|
|
nsChangeHint_UpdateParentOverflow = 1 << 13,
|
2015-02-16 06:34:29 +03:00
|
|
|
|
2012-05-17 08:05:09 +04:00
|
|
|
/**
|
2018-11-23 18:58:27 +03:00
|
|
|
* The children-only transform of an SVG frame changed, requiring overflows to
|
|
|
|
* be updated.
|
2012-05-17 08:05:09 +04:00
|
|
|
*/
|
2019-10-08 01:29:01 +03:00
|
|
|
nsChangeHint_ChildrenOnlyTransform = 1 << 14,
|
Bug 157681 - Part 2: Optimize positioned frame offset changes by moving the frame as opposed to reflowing it in case we know that the size of the frame will not change; r=dbaron
This patch adds a change hint to signal that one of the offsets on a
frame has been changed. When processing the hint, we do one of the
following things based on the position property of the frame.
* For static frames, we ignore the offset changes completely, as they
will not change the layout.
* For relative positioned frames, this patch refactors the
nsHTMLReflowState::ComputeRelativeOffsets function so that it can be
used from other callers, and it uses that to compute the new relative
offsets, and uses the offsets computed previously to compute the new
position of the frame.
* For absolute positioned frames, we set up a fake parent reflow state
object, and then we create a new reflow state object for the frame in
question. This setup is similar to what nsFrame::BoxReflow does.
Once we have the new reflow state object, we use it to compute the
absolute offsets, and then we use the computed offsets to set the new
position of the frame. The offset computation is similar to what
nsAbsoluteContainingBlock::ReflowAbsoluteFrame does. In some cases
where it is possible for the dimensions of the frame to change based
on the offset changes, we fall back to a full reflow.
2012-06-06 08:53:48 +04:00
|
|
|
|
|
|
|
/**
|
|
|
|
* The frame's offsets have changed, while its dimensions might have
|
|
|
|
* changed as well. This hint is used for positioned frames if their
|
|
|
|
* offset changes. If we decide that the dimensions are likely to
|
|
|
|
* change, this will trigger a reflow.
|
|
|
|
*
|
|
|
|
* Note that this should probably be used in combination with
|
|
|
|
* nsChangeHint_UpdateOverflow in order to get the overflow areas of
|
|
|
|
* the ancestors updated as well.
|
|
|
|
*/
|
2019-10-08 01:29:01 +03:00
|
|
|
nsChangeHint_RecomputePosition = 1 << 15,
|
Bug 157681 - Part 2: Optimize positioned frame offset changes by moving the frame as opposed to reflowing it in case we know that the size of the frame will not change; r=dbaron
This patch adds a change hint to signal that one of the offsets on a
frame has been changed. When processing the hint, we do one of the
following things based on the position property of the frame.
* For static frames, we ignore the offset changes completely, as they
will not change the layout.
* For relative positioned frames, this patch refactors the
nsHTMLReflowState::ComputeRelativeOffsets function so that it can be
used from other callers, and it uses that to compute the new relative
offsets, and uses the offsets computed previously to compute the new
position of the frame.
* For absolute positioned frames, we set up a fake parent reflow state
object, and then we create a new reflow state object for the frame in
question. This setup is similar to what nsFrame::BoxReflow does.
Once we have the new reflow state object, we use it to compute the
absolute offsets, and then we use the computed offsets to set the new
position of the frame. The offset computation is similar to what
nsAbsoluteContainingBlock::ReflowAbsoluteFrame does. In some cases
where it is possible for the dimensions of the frame to change based
on the offset changes, we fall back to a full reflow.
2012-06-06 08:53:48 +04:00
|
|
|
|
2012-08-10 15:16:49 +04:00
|
|
|
/**
|
|
|
|
* Behaves like ReconstructFrame, but only if the frame has descendants
|
|
|
|
* that are absolutely or fixed position. Use this hint when a style change
|
|
|
|
* has changed whether the frame is a container for fixed-pos or abs-pos
|
|
|
|
* elements, but reframing is otherwise not needed.
|
2016-08-29 21:43:29 +03:00
|
|
|
*
|
2018-03-22 21:20:41 +03:00
|
|
|
* Note that ComputedStyle::CalcStyleDifference adjusts results
|
2016-08-29 21:43:29 +03:00
|
|
|
* returned by style struct CalcDifference methods to return this hint
|
|
|
|
* only if there was a change to whether the element's overall style
|
|
|
|
* indicates that it establishes a containing block.
|
2012-08-10 15:16:49 +04:00
|
|
|
*/
|
2019-10-08 01:29:01 +03:00
|
|
|
nsChangeHint_UpdateContainingBlock = 1 << 16,
|
2012-09-07 21:13:36 +04:00
|
|
|
|
|
|
|
/**
|
|
|
|
* This change hint has *no* change handling behavior. However, it
|
|
|
|
* exists to be a non-inherited hint, because when the border-style
|
|
|
|
* changes, and it's inherited by a child, that might require a reflow
|
|
|
|
* due to the border-width change on the child.
|
|
|
|
*/
|
2019-10-08 01:29:01 +03:00
|
|
|
nsChangeHint_BorderStyleNoneChange = 1 << 17,
|
2012-11-07 13:53:44 +04:00
|
|
|
|
2014-07-16 15:28:50 +04:00
|
|
|
/**
|
|
|
|
* This will schedule an invalidating paint. This is useful if something
|
|
|
|
* has changed which will be invalidated by DLBI.
|
|
|
|
*/
|
Bug 1585863 - Remove nsChangeHint_UpdateTextPath. r=longsonr
The only thing it does is asserting a bit and posting more async work to the
text frame. It seems we can just post all the async work early instead, and
remove the change hint.
This was only introduced to fix bug 779971, where a <textPath> element
references its parent SVG, which is obviously unsound if we allowed to render
it.
What we're doing right now is a bit silly... We're observing the <svg>, so when
we finish reflowing it and store its overflow, we invalidate its rendering
observers, but that invalidates a _descendant_, which makes no sense.
Fortunately we don't let the element affect its rendering, as it fails this
check:
* https://searchfox.org/mozilla-central/rev/35cc00a25c4471993fdaa5761952bd3afd4f1731/layout/svg/SVGObserverUtils.cpp#1390
But we still request reflow of the outer <text>, which is not amazing. We
shouldn't invalidate anything if the textpath doesn't reference a valid element
and that didn't change. This is roughly what the code tried to do when checking
mValid, except we always initialize mValid to true and thus always trigger at
least one bogus reflow call.
Differential Revision: https://phabricator.services.mozilla.com/D48008
--HG--
extra : moz-landing-system : lando
2019-10-09 00:52:47 +03:00
|
|
|
nsChangeHint_SchedulePaint = 1 << 18,
|
2014-09-05 07:48:43 +04:00
|
|
|
|
|
|
|
/**
|
|
|
|
* A hint reflecting that style data changed with no change handling
|
2016-07-17 17:20:21 +03:00
|
|
|
* behavior. We need to return this, rather than nsChangeHint(0),
|
2018-03-23 16:49:21 +03:00
|
|
|
* so that certain optimizations that manipulate the style tree are
|
2014-09-05 07:48:43 +04:00
|
|
|
* correct.
|
|
|
|
*
|
|
|
|
* nsChangeHint_NeutralChange must be returned by CalcDifference on a given
|
|
|
|
* style struct if the data in the style structs are meaningfully different
|
|
|
|
* and if no other change hints are returned. If any other change hints are
|
|
|
|
* set, then nsChangeHint_NeutralChange need not also be included, but it is
|
|
|
|
* safe to do so. (An example of style structs having non-meaningfully
|
|
|
|
* different data would be cached information that would be re-calculated
|
|
|
|
* to the same values, such as nsStyleBorder::mSubImages.)
|
|
|
|
*/
|
Bug 1585863 - Remove nsChangeHint_UpdateTextPath. r=longsonr
The only thing it does is asserting a bit and posting more async work to the
text frame. It seems we can just post all the async work early instead, and
remove the change hint.
This was only introduced to fix bug 779971, where a <textPath> element
references its parent SVG, which is obviously unsound if we allowed to render
it.
What we're doing right now is a bit silly... We're observing the <svg>, so when
we finish reflowing it and store its overflow, we invalidate its rendering
observers, but that invalidates a _descendant_, which makes no sense.
Fortunately we don't let the element affect its rendering, as it fails this
check:
* https://searchfox.org/mozilla-central/rev/35cc00a25c4471993fdaa5761952bd3afd4f1731/layout/svg/SVGObserverUtils.cpp#1390
But we still request reflow of the outer <text>, which is not amazing. We
shouldn't invalidate anything if the textpath doesn't reference a valid element
and that didn't change. This is roughly what the code tried to do when checking
mValid, except we always initialize mValid to true and thus always trigger at
least one bogus reflow call.
Differential Revision: https://phabricator.services.mozilla.com/D48008
--HG--
extra : moz-landing-system : lando
2019-10-09 00:52:47 +03:00
|
|
|
nsChangeHint_NeutralChange = 1 << 19,
|
2014-11-04 17:52:27 +03:00
|
|
|
|
|
|
|
/**
|
|
|
|
* This will cause rendering observers to be invalidated.
|
|
|
|
*/
|
Bug 1585863 - Remove nsChangeHint_UpdateTextPath. r=longsonr
The only thing it does is asserting a bit and posting more async work to the
text frame. It seems we can just post all the async work early instead, and
remove the change hint.
This was only introduced to fix bug 779971, where a <textPath> element
references its parent SVG, which is obviously unsound if we allowed to render
it.
What we're doing right now is a bit silly... We're observing the <svg>, so when
we finish reflowing it and store its overflow, we invalidate its rendering
observers, but that invalidates a _descendant_, which makes no sense.
Fortunately we don't let the element affect its rendering, as it fails this
check:
* https://searchfox.org/mozilla-central/rev/35cc00a25c4471993fdaa5761952bd3afd4f1731/layout/svg/SVGObserverUtils.cpp#1390
But we still request reflow of the outer <text>, which is not amazing. We
shouldn't invalidate anything if the textpath doesn't reference a valid element
and that didn't change. This is roughly what the code tried to do when checking
mValid, except we always initialize mValid to true and thus always trigger at
least one bogus reflow call.
Differential Revision: https://phabricator.services.mozilla.com/D48008
--HG--
extra : moz-landing-system : lando
2019-10-09 00:52:47 +03:00
|
|
|
nsChangeHint_InvalidateRenderingObservers = 1 << 20,
|
2015-06-21 07:54:15 +03:00
|
|
|
|
|
|
|
/**
|
|
|
|
* Indicates that the reflow changes the size or position of the
|
|
|
|
* element, and thus the reflow must start from at least the frame's
|
2018-11-14 13:16:01 +03:00
|
|
|
* parent. Must be not be set without also setting nsChangeHint_NeedReflow.
|
|
|
|
* And consider adding nsChangeHint_ClearAncestorIntrinsics if needed.
|
2015-06-21 07:54:15 +03:00
|
|
|
*/
|
Bug 1585863 - Remove nsChangeHint_UpdateTextPath. r=longsonr
The only thing it does is asserting a bit and posting more async work to the
text frame. It seems we can just post all the async work early instead, and
remove the change hint.
This was only introduced to fix bug 779971, where a <textPath> element
references its parent SVG, which is obviously unsound if we allowed to render
it.
What we're doing right now is a bit silly... We're observing the <svg>, so when
we finish reflowing it and store its overflow, we invalidate its rendering
observers, but that invalidates a _descendant_, which makes no sense.
Fortunately we don't let the element affect its rendering, as it fails this
check:
* https://searchfox.org/mozilla-central/rev/35cc00a25c4471993fdaa5761952bd3afd4f1731/layout/svg/SVGObserverUtils.cpp#1390
But we still request reflow of the outer <text>, which is not amazing. We
shouldn't invalidate anything if the textpath doesn't reference a valid element
and that didn't change. This is roughly what the code tried to do when checking
mValid, except we always initialize mValid to true and thus always trigger at
least one bogus reflow call.
Differential Revision: https://phabricator.services.mozilla.com/D48008
--HG--
extra : moz-landing-system : lando
2019-10-09 00:52:47 +03:00
|
|
|
nsChangeHint_ReflowChangesSizeOrPosition = 1 << 21,
|
2012-08-10 15:16:49 +04:00
|
|
|
|
2015-07-22 07:36:56 +03:00
|
|
|
/**
|
|
|
|
* Indicates that the style changes the computed BSize --- e.g. 'height'.
|
2017-03-21 11:33:05 +03:00
|
|
|
* Must not be set without also setting nsChangeHint_NeedReflow.
|
2015-07-22 07:36:56 +03:00
|
|
|
*/
|
Bug 1585863 - Remove nsChangeHint_UpdateTextPath. r=longsonr
The only thing it does is asserting a bit and posting more async work to the
text frame. It seems we can just post all the async work early instead, and
remove the change hint.
This was only introduced to fix bug 779971, where a <textPath> element
references its parent SVG, which is obviously unsound if we allowed to render
it.
What we're doing right now is a bit silly... We're observing the <svg>, so when
we finish reflowing it and store its overflow, we invalidate its rendering
observers, but that invalidates a _descendant_, which makes no sense.
Fortunately we don't let the element affect its rendering, as it fails this
check:
* https://searchfox.org/mozilla-central/rev/35cc00a25c4471993fdaa5761952bd3afd4f1731/layout/svg/SVGObserverUtils.cpp#1390
But we still request reflow of the outer <text>, which is not amazing. We
shouldn't invalidate anything if the textpath doesn't reference a valid element
and that didn't change. This is roughly what the code tried to do when checking
mValid, except we always initialize mValid to true and thus always trigger at
least one bogus reflow call.
Differential Revision: https://phabricator.services.mozilla.com/D48008
--HG--
extra : moz-landing-system : lando
2019-10-09 00:52:47 +03:00
|
|
|
nsChangeHint_UpdateComputedBSize = 1 << 22,
|
2015-07-22 07:36:56 +03:00
|
|
|
|
2015-12-01 08:25:54 +03:00
|
|
|
/**
|
|
|
|
* Indicates that the 'opacity' property changed between 1 and non-1.
|
|
|
|
*
|
|
|
|
* Used as extra data for handling UpdateOpacityLayer hints.
|
|
|
|
*
|
|
|
|
* Note that we do not send this hint if the non-1 value was 0.99 or
|
|
|
|
* greater, since in that case we send a RepaintFrame hint instead.
|
|
|
|
*/
|
Bug 1585863 - Remove nsChangeHint_UpdateTextPath. r=longsonr
The only thing it does is asserting a bit and posting more async work to the
text frame. It seems we can just post all the async work early instead, and
remove the change hint.
This was only introduced to fix bug 779971, where a <textPath> element
references its parent SVG, which is obviously unsound if we allowed to render
it.
What we're doing right now is a bit silly... We're observing the <svg>, so when
we finish reflowing it and store its overflow, we invalidate its rendering
observers, but that invalidates a _descendant_, which makes no sense.
Fortunately we don't let the element affect its rendering, as it fails this
check:
* https://searchfox.org/mozilla-central/rev/35cc00a25c4471993fdaa5761952bd3afd4f1731/layout/svg/SVGObserverUtils.cpp#1390
But we still request reflow of the outer <text>, which is not amazing. We
shouldn't invalidate anything if the textpath doesn't reference a valid element
and that didn't change. This is roughly what the code tried to do when checking
mValid, except we always initialize mValid to true and thus always trigger at
least one bogus reflow call.
Differential Revision: https://phabricator.services.mozilla.com/D48008
--HG--
extra : moz-landing-system : lando
2019-10-09 00:52:47 +03:00
|
|
|
nsChangeHint_UpdateUsesOpacity = 1 << 23,
|
2015-12-01 08:25:54 +03:00
|
|
|
|
2016-04-28 20:01:22 +03:00
|
|
|
/**
|
|
|
|
* Indicates that the 'background-position' property changed.
|
|
|
|
* Regular frames can invalidate these changes using DLBI, but
|
|
|
|
* for some frame types we need to repaint the whole frame because
|
|
|
|
* the frame does not build individual background image display items
|
|
|
|
* for each background layer.
|
|
|
|
*/
|
Bug 1585863 - Remove nsChangeHint_UpdateTextPath. r=longsonr
The only thing it does is asserting a bit and posting more async work to the
text frame. It seems we can just post all the async work early instead, and
remove the change hint.
This was only introduced to fix bug 779971, where a <textPath> element
references its parent SVG, which is obviously unsound if we allowed to render
it.
What we're doing right now is a bit silly... We're observing the <svg>, so when
we finish reflowing it and store its overflow, we invalidate its rendering
observers, but that invalidates a _descendant_, which makes no sense.
Fortunately we don't let the element affect its rendering, as it fails this
check:
* https://searchfox.org/mozilla-central/rev/35cc00a25c4471993fdaa5761952bd3afd4f1731/layout/svg/SVGObserverUtils.cpp#1390
But we still request reflow of the outer <text>, which is not amazing. We
shouldn't invalidate anything if the textpath doesn't reference a valid element
and that didn't change. This is roughly what the code tried to do when checking
mValid, except we always initialize mValid to true and thus always trigger at
least one bogus reflow call.
Differential Revision: https://phabricator.services.mozilla.com/D48008
--HG--
extra : moz-landing-system : lando
2019-10-09 00:52:47 +03:00
|
|
|
nsChangeHint_UpdateBackgroundPosition = 1 << 24,
|
2016-04-28 20:01:22 +03:00
|
|
|
|
2016-09-30 00:15:32 +03:00
|
|
|
/**
|
|
|
|
* Indicates that a frame has changed to or from having the CSS
|
|
|
|
* transform property set.
|
|
|
|
*/
|
Bug 1585863 - Remove nsChangeHint_UpdateTextPath. r=longsonr
The only thing it does is asserting a bit and posting more async work to the
text frame. It seems we can just post all the async work early instead, and
remove the change hint.
This was only introduced to fix bug 779971, where a <textPath> element
references its parent SVG, which is obviously unsound if we allowed to render
it.
What we're doing right now is a bit silly... We're observing the <svg>, so when
we finish reflowing it and store its overflow, we invalidate its rendering
observers, but that invalidates a _descendant_, which makes no sense.
Fortunately we don't let the element affect its rendering, as it fails this
check:
* https://searchfox.org/mozilla-central/rev/35cc00a25c4471993fdaa5761952bd3afd4f1731/layout/svg/SVGObserverUtils.cpp#1390
But we still request reflow of the outer <text>, which is not amazing. We
shouldn't invalidate anything if the textpath doesn't reference a valid element
and that didn't change. This is roughly what the code tried to do when checking
mValid, except we always initialize mValid to true and thus always trigger at
least one bogus reflow call.
Differential Revision: https://phabricator.services.mozilla.com/D48008
--HG--
extra : moz-landing-system : lando
2019-10-09 00:52:47 +03:00
|
|
|
nsChangeHint_AddOrRemoveTransform = 1 << 25,
|
2016-09-30 00:15:32 +03:00
|
|
|
|
2017-05-10 23:53:27 +03:00
|
|
|
/**
|
2018-08-03 09:42:33 +03:00
|
|
|
* Indicates that the presence of scrollbars might have changed.
|
|
|
|
*
|
|
|
|
* This happens when at least one of overflow-{x,y} properties changed.
|
2017-05-10 23:53:27 +03:00
|
|
|
*
|
|
|
|
* In most cases, this is equivalent to nsChangeHint_ReconstructFrame. But
|
|
|
|
* in some special cases where the change is really targeting the viewport's
|
|
|
|
* scrollframe, this is instead equivalent to nsChangeHint_AllReflowHints
|
|
|
|
* (because the viewport always has an associated scrollframe).
|
|
|
|
*/
|
Bug 1585863 - Remove nsChangeHint_UpdateTextPath. r=longsonr
The only thing it does is asserting a bit and posting more async work to the
text frame. It seems we can just post all the async work early instead, and
remove the change hint.
This was only introduced to fix bug 779971, where a <textPath> element
references its parent SVG, which is obviously unsound if we allowed to render
it.
What we're doing right now is a bit silly... We're observing the <svg>, so when
we finish reflowing it and store its overflow, we invalidate its rendering
observers, but that invalidates a _descendant_, which makes no sense.
Fortunately we don't let the element affect its rendering, as it fails this
check:
* https://searchfox.org/mozilla-central/rev/35cc00a25c4471993fdaa5761952bd3afd4f1731/layout/svg/SVGObserverUtils.cpp#1390
But we still request reflow of the outer <text>, which is not amazing. We
shouldn't invalidate anything if the textpath doesn't reference a valid element
and that didn't change. This is roughly what the code tried to do when checking
mValid, except we always initialize mValid to true and thus always trigger at
least one bogus reflow call.
Differential Revision: https://phabricator.services.mozilla.com/D48008
--HG--
extra : moz-landing-system : lando
2019-10-09 00:52:47 +03:00
|
|
|
nsChangeHint_ScrollbarChange = 1 << 26,
|
2017-05-10 23:53:27 +03:00
|
|
|
|
2017-09-05 23:30:40 +03:00
|
|
|
/**
|
|
|
|
* Indicates that there has been a colspan or rowspan attribute change
|
|
|
|
* on the cells of a table.
|
|
|
|
*/
|
2019-11-20 05:37:06 +03:00
|
|
|
nsChangeHint_UpdateTableCellSpans = 1 << 27,
|
2017-09-05 23:30:40 +03:00
|
|
|
|
2018-02-09 04:43:10 +03:00
|
|
|
/**
|
|
|
|
* Indicates that the visiblity property changed.
|
|
|
|
* This change hint is used for skip restyling for animations on
|
|
|
|
* visibility:hidden elements in the case where the elements have no visible
|
|
|
|
* descendants.
|
|
|
|
*/
|
2019-11-20 05:37:06 +03:00
|
|
|
nsChangeHint_VisibilityChange = 1u << 28,
|
2018-02-09 04:43:10 +03:00
|
|
|
|
2017-03-21 11:33:05 +03:00
|
|
|
// IMPORTANT NOTE: When adding a new hint, you will need to add it to
|
|
|
|
// one of:
|
|
|
|
//
|
|
|
|
// * nsChangeHint_Hints_NeverHandledForDescendants
|
|
|
|
// * nsChangeHint_Hints_AlwaysHandledForDescendants
|
|
|
|
// * nsChangeHint_Hints_SometimesHandledForDescendants
|
|
|
|
//
|
|
|
|
// and you also may need to handle it in NS_HintsNotHandledForDescendantsIn.
|
|
|
|
//
|
|
|
|
// Please also add it to RestyleManager::ChangeHintToString and
|
|
|
|
// modify nsChangeHint_AllHints below accordingly.
|
2016-08-15 13:52:05 +03:00
|
|
|
|
|
|
|
/**
|
|
|
|
* Dummy hint value for all hints. It exists for compile time check.
|
|
|
|
*/
|
2019-11-20 05:37:06 +03:00
|
|
|
nsChangeHint_AllHints = uint32_t((1ull << 29) - 1),
|
2002-08-24 18:41:28 +04:00
|
|
|
};
|
|
|
|
|
|
|
|
// Redefine these operators to return nothing. This will catch any use
|
|
|
|
// of these operators on hints. We should not be using these operators
|
|
|
|
// on nsChangeHints
|
|
|
|
inline void operator<(nsChangeHint s1, nsChangeHint s2) {}
|
|
|
|
inline void operator>(nsChangeHint s1, nsChangeHint s2) {}
|
|
|
|
inline void operator!=(nsChangeHint s1, nsChangeHint s2) {}
|
|
|
|
inline void operator==(nsChangeHint s1, nsChangeHint s2) {}
|
|
|
|
inline void operator<=(nsChangeHint s1, nsChangeHint s2) {}
|
|
|
|
inline void operator>=(nsChangeHint s1, nsChangeHint s2) {}
|
|
|
|
|
|
|
|
// Operators on nsChangeHints
|
|
|
|
|
|
|
|
// Returns true iff the second hint contains all the hints of the first hint
|
2011-09-29 10:19:26 +04:00
|
|
|
inline bool NS_IsHintSubset(nsChangeHint aSubset, nsChangeHint aSuperSet) {
|
2002-08-24 18:41:28 +04:00
|
|
|
return (aSubset & aSuperSet) == aSubset;
|
|
|
|
}
|
|
|
|
|
2015-06-21 07:54:15 +03:00
|
|
|
// The functions below need an integral type to cast to to avoid
|
|
|
|
// infinite recursion.
|
|
|
|
typedef decltype(nsChangeHint(0) + nsChangeHint(0)) nsChangeHint_size_t;
|
|
|
|
|
|
|
|
inline nsChangeHint constexpr operator|(nsChangeHint aLeft,
|
|
|
|
nsChangeHint aRight) {
|
|
|
|
return nsChangeHint(nsChangeHint_size_t(aLeft) | nsChangeHint_size_t(aRight));
|
|
|
|
}
|
|
|
|
|
|
|
|
inline nsChangeHint constexpr operator&(nsChangeHint aLeft,
|
|
|
|
nsChangeHint aRight) {
|
|
|
|
return nsChangeHint(nsChangeHint_size_t(aLeft) & nsChangeHint_size_t(aRight));
|
|
|
|
}
|
|
|
|
|
|
|
|
inline nsChangeHint& operator|=(nsChangeHint& aLeft, nsChangeHint aRight) {
|
|
|
|
return aLeft = aLeft | aRight;
|
|
|
|
}
|
|
|
|
|
|
|
|
inline nsChangeHint& operator&=(nsChangeHint& aLeft, nsChangeHint aRight) {
|
|
|
|
return aLeft = aLeft & aRight;
|
|
|
|
}
|
|
|
|
|
2016-07-09 00:39:53 +03:00
|
|
|
inline nsChangeHint constexpr operator~(nsChangeHint aArg) {
|
2015-06-21 07:54:15 +03:00
|
|
|
return nsChangeHint(~nsChangeHint_size_t(aArg));
|
|
|
|
}
|
|
|
|
|
|
|
|
inline nsChangeHint constexpr operator^(nsChangeHint aLeft,
|
|
|
|
nsChangeHint aRight) {
|
|
|
|
return nsChangeHint(nsChangeHint_size_t(aLeft) ^ nsChangeHint_size_t(aRight));
|
|
|
|
}
|
|
|
|
|
|
|
|
inline nsChangeHint operator^=(nsChangeHint& aLeft, nsChangeHint aRight) {
|
|
|
|
return aLeft = aLeft ^ aRight;
|
|
|
|
}
|
|
|
|
|
2012-09-07 21:13:36 +04:00
|
|
|
/**
|
|
|
|
* We have an optimization when processing change hints which prevents
|
|
|
|
* us from visiting the descendants of a node when a hint on that node
|
|
|
|
* is being processed. This optimization does not apply in some of the
|
|
|
|
* cases where applying a hint to an element does not necessarily result
|
|
|
|
* in the same hint being handled on the descendants.
|
|
|
|
*/
|
2012-09-07 21:13:36 +04:00
|
|
|
|
2017-03-21 11:33:05 +03:00
|
|
|
// The change hints that are always handled for descendants.
|
|
|
|
#define nsChangeHint_Hints_AlwaysHandledForDescendants \
|
|
|
|
(nsChangeHint_ClearDescendantIntrinsics | nsChangeHint_NeedDirtyReflow | \
|
|
|
|
nsChangeHint_NeutralChange | nsChangeHint_ReconstructFrame | \
|
|
|
|
nsChangeHint_RepaintFrame | nsChangeHint_SchedulePaint | \
|
2019-10-08 01:29:01 +03:00
|
|
|
nsChangeHint_UpdateCursor | nsChangeHint_UpdateSubtreeOverflow | \
|
Bug 1585863 - Remove nsChangeHint_UpdateTextPath. r=longsonr
The only thing it does is asserting a bit and posting more async work to the
text frame. It seems we can just post all the async work early instead, and
remove the change hint.
This was only introduced to fix bug 779971, where a <textPath> element
references its parent SVG, which is obviously unsound if we allowed to render
it.
What we're doing right now is a bit silly... We're observing the <svg>, so when
we finish reflowing it and store its overflow, we invalidate its rendering
observers, but that invalidates a _descendant_, which makes no sense.
Fortunately we don't let the element affect its rendering, as it fails this
check:
* https://searchfox.org/mozilla-central/rev/35cc00a25c4471993fdaa5761952bd3afd4f1731/layout/svg/SVGObserverUtils.cpp#1390
But we still request reflow of the outer <text>, which is not amazing. We
shouldn't invalidate anything if the textpath doesn't reference a valid element
and that didn't change. This is roughly what the code tried to do when checking
mValid, except we always initialize mValid to true and thus always trigger at
least one bogus reflow call.
Differential Revision: https://phabricator.services.mozilla.com/D48008
--HG--
extra : moz-landing-system : lando
2019-10-09 00:52:47 +03:00
|
|
|
nsChangeHint_VisibilityChange)
|
2017-03-21 11:33:05 +03:00
|
|
|
|
|
|
|
// The change hints that are never handled for descendants.
|
|
|
|
#define nsChangeHint_Hints_NeverHandledForDescendants \
|
|
|
|
(nsChangeHint_BorderStyleNoneChange | nsChangeHint_ChildrenOnlyTransform | \
|
|
|
|
nsChangeHint_ScrollbarChange | nsChangeHint_InvalidateRenderingObservers | \
|
|
|
|
nsChangeHint_RecomputePosition | nsChangeHint_UpdateBackgroundPosition | \
|
|
|
|
nsChangeHint_UpdateComputedBSize | nsChangeHint_UpdateContainingBlock | \
|
|
|
|
nsChangeHint_UpdateEffects | nsChangeHint_UpdateOpacityLayer | \
|
|
|
|
nsChangeHint_UpdateOverflow | nsChangeHint_UpdateParentOverflow | \
|
|
|
|
nsChangeHint_UpdatePostTransformOverflow | \
|
|
|
|
nsChangeHint_UpdateTableCellSpans | nsChangeHint_UpdateTransformLayer | \
|
2019-11-20 05:37:06 +03:00
|
|
|
nsChangeHint_UpdateUsesOpacity | nsChangeHint_AddOrRemoveTransform)
|
2017-03-21 11:33:05 +03:00
|
|
|
|
|
|
|
// The change hints that are sometimes considered to be handled for descendants.
|
|
|
|
#define nsChangeHint_Hints_SometimesHandledForDescendants \
|
|
|
|
(nsChangeHint_ClearAncestorIntrinsics | nsChangeHint_NeedReflow | \
|
|
|
|
nsChangeHint_ReflowChangesSizeOrPosition)
|
|
|
|
|
|
|
|
static_assert(!(nsChangeHint_Hints_AlwaysHandledForDescendants &
|
|
|
|
nsChangeHint_Hints_NeverHandledForDescendants) &&
|
|
|
|
!(nsChangeHint_Hints_AlwaysHandledForDescendants &
|
|
|
|
nsChangeHint_Hints_SometimesHandledForDescendants) &&
|
|
|
|
!(nsChangeHint_Hints_NeverHandledForDescendants &
|
|
|
|
nsChangeHint_Hints_SometimesHandledForDescendants) &&
|
|
|
|
!(nsChangeHint_AllHints ^
|
|
|
|
nsChangeHint_Hints_AlwaysHandledForDescendants ^
|
|
|
|
nsChangeHint_Hints_NeverHandledForDescendants ^
|
|
|
|
nsChangeHint_Hints_SometimesHandledForDescendants),
|
|
|
|
"change hints must be present in exactly one of "
|
|
|
|
"nsChangeHint_Hints_{Always,Never,Sometimes}"
|
|
|
|
"HandledForDescendants");
|
|
|
|
|
2012-09-07 21:13:36 +04:00
|
|
|
// The most hints that NS_HintsNotHandledForDescendantsIn could possibly return:
|
2017-03-21 11:33:05 +03:00
|
|
|
#define nsChangeHint_Hints_NotHandledForDescendants \
|
|
|
|
(nsChangeHint_Hints_NeverHandledForDescendants | \
|
|
|
|
nsChangeHint_Hints_SometimesHandledForDescendants)
|
2012-09-07 21:13:36 +04:00
|
|
|
|
2017-03-21 11:33:05 +03:00
|
|
|
// Redefine the old NS_STYLE_HINT constants in terms of the new hint structure
|
2019-10-08 01:29:01 +03:00
|
|
|
#define NS_STYLE_HINT_VISUAL \
|
|
|
|
nsChangeHint(nsChangeHint_RepaintFrame | nsChangeHint_SchedulePaint)
|
2017-03-21 11:33:05 +03:00
|
|
|
#define nsChangeHint_AllReflowHints \
|
|
|
|
nsChangeHint( \
|
|
|
|
nsChangeHint_NeedReflow | nsChangeHint_ReflowChangesSizeOrPosition | \
|
|
|
|
nsChangeHint_ClearAncestorIntrinsics | \
|
|
|
|
nsChangeHint_ClearDescendantIntrinsics | nsChangeHint_NeedDirtyReflow)
|
2017-05-26 00:46:03 +03:00
|
|
|
|
|
|
|
// Below are the change hints that we send for ISize & BSize changes.
|
|
|
|
// Each is similar to nsChangeHint_AllReflowHints with a few changes.
|
|
|
|
|
|
|
|
// * For an ISize change, we send nsChangeHint_AllReflowHints, with two bits
|
|
|
|
// excluded: nsChangeHint_ClearDescendantIntrinsics (because an ancestor's
|
|
|
|
// inline-size change can't affect descendant intrinsic sizes), and
|
|
|
|
// nsChangeHint_NeedDirtyReflow (because ISize changes don't need to *force*
|
|
|
|
// all descendants to reflow).
|
|
|
|
#define nsChangeHint_ReflowHintsForISizeChange \
|
|
|
|
nsChangeHint(nsChangeHint_AllReflowHints & \
|
|
|
|
~(nsChangeHint_ClearDescendantIntrinsics | \
|
|
|
|
nsChangeHint_NeedDirtyReflow))
|
|
|
|
|
|
|
|
// * For a BSize change, we send almost the same hints as for ISize changes,
|
|
|
|
// with one extra: nsChangeHint_UpdateComputedBSize. We need this hint because
|
|
|
|
// BSize changes CAN affect descendant intrinsic sizes, due to replaced
|
|
|
|
// elements with percentage BSizes in descendants which also have percentage
|
|
|
|
// BSizes. nsChangeHint_UpdateComputedBSize clears intrinsic sizes for frames
|
|
|
|
// that have such replaced elements. (We could instead send
|
|
|
|
// nsChangeHint_ClearDescendantIntrinsics, but that's broader than we need.)
|
|
|
|
//
|
|
|
|
// NOTE: You might think that BSize changes could exclude
|
|
|
|
// nsChangeHint_ClearAncestorIntrinsics (which is inline-axis specific), but we
|
|
|
|
// do need to send it, to clear cached results from CSS Flex measuring reflows.
|
|
|
|
#define nsChangeHint_ReflowHintsForBSizeChange \
|
2017-05-26 00:53:23 +03:00
|
|
|
nsChangeHint( \
|
|
|
|
(nsChangeHint_AllReflowHints | nsChangeHint_UpdateComputedBSize) & \
|
|
|
|
~(nsChangeHint_ClearDescendantIntrinsics | \
|
|
|
|
nsChangeHint_NeedDirtyReflow))
|
2017-05-26 00:46:03 +03:00
|
|
|
|
2019-10-22 02:47:31 +03:00
|
|
|
// For a change in whether a scrollframe displays or not scrollbars.
|
|
|
|
//
|
|
|
|
// When requesting this reflow, we send the exact same change hints that "width"
|
|
|
|
// and "height" would send (since conceptually, adding/removing scrollbars is
|
|
|
|
// like changing the available space).
|
|
|
|
//
|
|
|
|
// FIXME(emilio): Seems we could be a bit more efficient here, as adding or
|
|
|
|
// removing scrollbars doesn't change the size of the element itself, so maybe
|
|
|
|
// ClearAncestorIntrinsics or ReflowChangesSizeOrPosition are not needed... I
|
|
|
|
// think this ideally should be just nsChangehint_NeedReflow.
|
|
|
|
#define nsChangeHint_ReflowHintsForScrollbarChange \
|
|
|
|
nsChangeHint(nsChangeHint_ReflowHintsForBSizeChange | \
|
|
|
|
nsChangeHint_ReflowHintsForISizeChange)
|
|
|
|
|
2017-07-15 03:20:35 +03:00
|
|
|
// * For changes to the float area of an already-floated element, we need all
|
|
|
|
// reflow hints, but not the ones that apply to descendants.
|
|
|
|
// Our descendants aren't impacted when our float area only changes
|
|
|
|
// placement but not size/shape. (e.g. if we change which side we float to).
|
|
|
|
// But our ancestors/siblings are potentially impacted, so we need to send
|
|
|
|
// the non-descendant reflow hints.
|
|
|
|
#define nsChangeHint_ReflowHintsForFloatAreaChange \
|
|
|
|
nsChangeHint(nsChangeHint_AllReflowHints & \
|
|
|
|
~(nsChangeHint_ClearDescendantIntrinsics | \
|
|
|
|
nsChangeHint_NeedDirtyReflow))
|
|
|
|
|
2017-03-21 11:33:05 +03:00
|
|
|
#define NS_STYLE_HINT_REFLOW \
|
|
|
|
nsChangeHint(NS_STYLE_HINT_VISUAL | nsChangeHint_AllReflowHints)
|
|
|
|
|
2017-10-31 03:45:41 +03:00
|
|
|
#define nsChangeHint_Hints_CanIgnoreIfNotVisible \
|
|
|
|
nsChangeHint( \
|
|
|
|
NS_STYLE_HINT_VISUAL | nsChangeHint_NeutralChange | \
|
|
|
|
nsChangeHint_UpdateOpacityLayer | nsChangeHint_AddOrRemoveTransform | \
|
2018-06-25 12:29:09 +03:00
|
|
|
nsChangeHint_UpdateContainingBlock | nsChangeHint_UpdateOverflow | \
|
2017-10-31 03:45:41 +03:00
|
|
|
nsChangeHint_UpdatePostTransformOverflow | \
|
|
|
|
nsChangeHint_UpdateTransformLayer | nsChangeHint_UpdateUsesOpacity | \
|
2018-02-09 04:43:10 +03:00
|
|
|
nsChangeHint_VisibilityChange)
|
2017-03-21 11:33:05 +03:00
|
|
|
|
2018-07-30 04:51:18 +03:00
|
|
|
// Change hints for added or removed transform style.
|
|
|
|
//
|
|
|
|
// If we've added or removed the transform property, we need to reconstruct the
|
|
|
|
// frame to add or remove the view object, and also to handle abs-pos and
|
|
|
|
// fixed-pos containers.
|
|
|
|
//
|
|
|
|
// We do not need to apply nsChangeHint_UpdateTransformLayer since
|
|
|
|
// nsChangeHint_RepaintFrame will forcibly invalidate the frame area and
|
|
|
|
// ensure layers are rebuilt (or removed).
|
|
|
|
#define nsChangeHint_ComprehensiveAddOrRemoveTransform \
|
|
|
|
nsChangeHint(nsChangeHint_UpdateContainingBlock | \
|
|
|
|
nsChangeHint_AddOrRemoveTransform | \
|
|
|
|
nsChangeHint_UpdateOverflow | nsChangeHint_RepaintFrame)
|
|
|
|
|
2017-02-16 06:36:57 +03:00
|
|
|
// NB: Once we drop support for the old style system, this logic should be
|
|
|
|
// inlined in the Servo style system to eliminate the FFI call.
|
2012-09-07 21:13:36 +04:00
|
|
|
inline nsChangeHint NS_HintsNotHandledForDescendantsIn(
|
|
|
|
nsChangeHint aChangeHint) {
|
2017-03-21 11:33:05 +03:00
|
|
|
nsChangeHint result =
|
|
|
|
aChangeHint & nsChangeHint_Hints_NeverHandledForDescendants;
|
2012-09-07 21:13:36 +04:00
|
|
|
|
2015-06-21 07:54:15 +03:00
|
|
|
if (!NS_IsHintSubset(nsChangeHint_NeedDirtyReflow, aChangeHint)) {
|
|
|
|
if (NS_IsHintSubset(nsChangeHint_NeedReflow, aChangeHint)) {
|
|
|
|
// If NeedDirtyReflow is *not* set, then NeedReflow is a
|
|
|
|
// non-inherited hint.
|
2016-05-23 06:26:03 +03:00
|
|
|
result |= nsChangeHint_NeedReflow;
|
2015-06-21 07:54:15 +03:00
|
|
|
}
|
|
|
|
|
|
|
|
if (NS_IsHintSubset(nsChangeHint_ReflowChangesSizeOrPosition,
|
|
|
|
aChangeHint)) {
|
|
|
|
// If NeedDirtyReflow is *not* set, then ReflowChangesSizeOrPosition is a
|
|
|
|
// non-inherited hint.
|
2016-05-23 06:26:03 +03:00
|
|
|
result |= nsChangeHint_ReflowChangesSizeOrPosition;
|
2015-06-21 07:54:15 +03:00
|
|
|
}
|
2012-09-07 21:13:36 +04:00
|
|
|
}
|
|
|
|
|
|
|
|
if (!NS_IsHintSubset(nsChangeHint_ClearDescendantIntrinsics, aChangeHint) &&
|
|
|
|
NS_IsHintSubset(nsChangeHint_ClearAncestorIntrinsics, aChangeHint)) {
|
|
|
|
// If ClearDescendantIntrinsics is *not* set, then
|
|
|
|
// ClearAncestorIntrinsics is a non-inherited hint.
|
2016-05-23 06:26:03 +03:00
|
|
|
result |= nsChangeHint_ClearAncestorIntrinsics;
|
2012-09-07 21:13:36 +04:00
|
|
|
}
|
|
|
|
|
2015-02-10 01:34:50 +03:00
|
|
|
MOZ_ASSERT(
|
|
|
|
NS_IsHintSubset(result, nsChangeHint_Hints_NotHandledForDescendants),
|
|
|
|
"something is inconsistent");
|
2012-09-07 21:13:36 +04:00
|
|
|
|
2012-09-07 21:13:36 +04:00
|
|
|
return result;
|
|
|
|
}
|
|
|
|
|
2017-03-21 11:33:05 +03:00
|
|
|
inline nsChangeHint NS_HintsHandledForDescendantsIn(nsChangeHint aChangeHint) {
|
|
|
|
return aChangeHint & ~NS_HintsNotHandledForDescendantsIn(aChangeHint);
|
|
|
|
}
|
2002-08-24 18:41:28 +04:00
|
|
|
|
2017-03-21 11:33:05 +03:00
|
|
|
// Returns the change hints in aOurChange that are not subsumed by those
|
|
|
|
// in aHintsHandled (which are hints that have been handled by an ancestor).
|
|
|
|
inline nsChangeHint NS_RemoveSubsumedHints(nsChangeHint aOurChange,
|
|
|
|
nsChangeHint aHintsHandled) {
|
|
|
|
nsChangeHint result =
|
|
|
|
aOurChange & ~NS_HintsHandledForDescendantsIn(aHintsHandled);
|
|
|
|
|
|
|
|
if (result &
|
|
|
|
(nsChangeHint_ClearAncestorIntrinsics |
|
|
|
|
nsChangeHint_ClearDescendantIntrinsics | nsChangeHint_NeedDirtyReflow |
|
|
|
|
nsChangeHint_ReflowChangesSizeOrPosition |
|
|
|
|
nsChangeHint_UpdateComputedBSize)) {
|
|
|
|
result |= nsChangeHint_NeedReflow;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (result & (nsChangeHint_ClearDescendantIntrinsics)) {
|
|
|
|
MOZ_ASSERT(result & nsChangeHint_ClearAncestorIntrinsics);
|
|
|
|
result |= // nsChangeHint_ClearAncestorIntrinsics |
|
|
|
|
nsChangeHint_NeedDirtyReflow;
|
|
|
|
}
|
|
|
|
|
|
|
|
return result;
|
|
|
|
}
|
2016-05-24 06:57:42 +03:00
|
|
|
|
2019-03-14 14:47:50 +03:00
|
|
|
namespace mozilla {
|
2004-01-14 04:36:35 +03:00
|
|
|
|
2020-02-11 04:22:15 +03:00
|
|
|
struct StyleRestyleHint;
|
2014-09-13 17:17:36 +04:00
|
|
|
|
2020-02-11 04:22:15 +03:00
|
|
|
using RestyleHint = StyleRestyleHint;
|
2014-09-13 17:17:36 +04:00
|
|
|
|
2019-03-14 14:47:50 +03:00
|
|
|
} // namespace mozilla
|
2004-01-14 04:36:35 +03:00
|
|
|
|
2002-08-24 18:41:28 +04:00
|
|
|
#endif /* nsChangeHint_h___ */
|