Title: [194558] trunk
- Revision
- 194558
- Author
- svil...@igalia.com
- Date
- 2016-01-04 14:43:41 -0800 (Mon, 04 Jan 2016)
Log Message
REGRESSION(r194143): Float width incorrectly calculated on Wikipedia
https://bugs.webkit.org/show_bug.cgi?id=152644
Reviewed by Myles C. Maxfield.
Source/WebCore:
The min/max-content contribution computation for non replaced
blocks changed in r194143. The implementation was actually
more complex than it should be and actually incomplete as it
was not considering the case of out of flow elements (because
it was directly calling computeLogicalWidthInRegion()).
This new implementation simplifies a lot the code and makes it
more complete as it relies on min|maxPreferredLogicalWidth()
calls which already consider all the different types of boxes.
Test: fast/css-intrinsic-dimensions/inlinesize-contribution-floats.html
* rendering/RenderBlock.cpp:
(WebCore::RenderBlock::computeBlockPreferredLogicalWidths):
LayoutTests:
* fast/css-intrinsic-dimensions/inlinesize-contribution-floats-expected.html: Added.
* fast/css-intrinsic-dimensions/inlinesize-contribution-floats.html: Added.
Modified Paths
Added Paths
Diff
Modified: trunk/LayoutTests/ChangeLog (194557 => 194558)
--- trunk/LayoutTests/ChangeLog 2016-01-04 22:42:20 UTC (rev 194557)
+++ trunk/LayoutTests/ChangeLog 2016-01-04 22:43:41 UTC (rev 194558)
@@ -1,3 +1,13 @@
+2016-01-04 Sergio Villar Senin <svil...@igalia.com>
+
+ REGRESSION(r194143): Float width incorrectly calculated on Wikipedia
+ https://bugs.webkit.org/show_bug.cgi?id=152644
+
+ Reviewed by Myles C. Maxfield.
+
+ * fast/css-intrinsic-dimensions/inlinesize-contribution-floats-expected.html: Added.
+ * fast/css-intrinsic-dimensions/inlinesize-contribution-floats.html: Added.
+
2016-01-04 David Hyatt <hy...@apple.com>
word-wrap: break-word broken inside a flexbox
Added: trunk/LayoutTests/fast/css-intrinsic-dimensions/inlinesize-contribution-floats-expected.html (0 => 194558)
--- trunk/LayoutTests/fast/css-intrinsic-dimensions/inlinesize-contribution-floats-expected.html (rev 0)
+++ trunk/LayoutTests/fast/css-intrinsic-dimensions/inlinesize-contribution-floats-expected.html 2016-01-04 22:43:41 UTC (rev 194558)
@@ -0,0 +1,5 @@
+<!DOCTYPE html>
+
+<div style="word-wrap: break-word; max-width: 100%; width: 222px;">
+ The test PASS if this text is rendered horizontally.
+</div>
Added: trunk/LayoutTests/fast/css-intrinsic-dimensions/inlinesize-contribution-floats.html (0 => 194558)
--- trunk/LayoutTests/fast/css-intrinsic-dimensions/inlinesize-contribution-floats.html (rev 0)
+++ trunk/LayoutTests/fast/css-intrinsic-dimensions/inlinesize-contribution-floats.html 2016-01-04 22:43:41 UTC (rev 194558)
@@ -0,0 +1,7 @@
+<!DOCTYPE html>
+
+<div style="float:left;">
+ <div style="word-wrap: break-word; max-width: 100%; width: 222px;">
+ The test PASS if this text is rendered horizontally.
+ </div>
+</div>
Modified: trunk/Source/WebCore/ChangeLog (194557 => 194558)
--- trunk/Source/WebCore/ChangeLog 2016-01-04 22:42:20 UTC (rev 194557)
+++ trunk/Source/WebCore/ChangeLog 2016-01-04 22:43:41 UTC (rev 194558)
@@ -1,3 +1,25 @@
+2016-01-04 Sergio Villar Senin <svil...@igalia.com>
+
+ REGRESSION(r194143): Float width incorrectly calculated on Wikipedia
+ https://bugs.webkit.org/show_bug.cgi?id=152644
+
+ Reviewed by Myles C. Maxfield.
+
+ The min/max-content contribution computation for non replaced
+ blocks changed in r194143. The implementation was actually
+ more complex than it should be and actually incomplete as it
+ was not considering the case of out of flow elements (because
+ it was directly calling computeLogicalWidthInRegion()).
+
+ This new implementation simplifies a lot the code and makes it
+ more complete as it relies on min|maxPreferredLogicalWidth()
+ calls which already consider all the different types of boxes.
+
+ Test: fast/css-intrinsic-dimensions/inlinesize-contribution-floats.html
+
+ * rendering/RenderBlock.cpp:
+ (WebCore::RenderBlock::computeBlockPreferredLogicalWidths):
+
2016-01-04 David Hyatt <hy...@apple.com>
word-wrap: break-word broken inside a flexbox
Modified: trunk/Source/WebCore/rendering/RenderBlock.cpp (194557 => 194558)
--- trunk/Source/WebCore/rendering/RenderBlock.cpp 2016-01-04 22:42:20 UTC (rev 194557)
+++ trunk/Source/WebCore/rendering/RenderBlock.cpp 2016-01-04 22:43:41 UTC (rev 194558)
@@ -2751,26 +2751,15 @@
childBox.computeLogicalHeight(childBox.borderAndPaddingLogicalHeight(), 0, computedValues);
childMinPreferredLogicalWidth = childMaxPreferredLogicalWidth = computedValues.m_extent;
} else {
- if (is<RenderBlock>(*child) && !is<RenderTable>(*child)) {
+ childMinPreferredLogicalWidth = child->minPreferredLogicalWidth();
+ childMaxPreferredLogicalWidth = child->maxPreferredLogicalWidth();
+
+ if (is<RenderBlock>(*child)) {
const Length& computedInlineSize = child->style().logicalWidth();
- if (computedInlineSize.isFitContent() || computedInlineSize.isFillAvailable() || computedInlineSize.isAuto()
- || computedInlineSize.isPercentOrCalculated()) {
- // FIXME: we could do a lot better for percents (we're considering them always indefinite)
- // but we need https://bugs.webkit.org/show_bug.cgi?id=152262 to be fixed first
- childMinPreferredLogicalWidth = child->minPreferredLogicalWidth();
- childMaxPreferredLogicalWidth = child->maxPreferredLogicalWidth();
- } else {
- ASSERT(computedInlineSize.isMinContent() || computedInlineSize.isMaxContent() || computedInlineSize.isFixed());
- LogicalExtentComputedValues computedValues;
- downcast<RenderBlock>(*child).computeLogicalWidthInRegion(computedValues);
- childMinPreferredLogicalWidth = childMaxPreferredLogicalWidth = computedValues.m_extent;
- child->setPreferredLogicalWidthsDirty(false);
- }
- } else {
- // FIXME: we leave the original implementation as default fallback. Still need to specialcase
- // some other situations: https://drafts.csswg.org/css-sizing/#intrinsic
- childMinPreferredLogicalWidth = child->minPreferredLogicalWidth();
- childMaxPreferredLogicalWidth = child->maxPreferredLogicalWidth();
+ if (computedInlineSize.isMaxContent())
+ childMinPreferredLogicalWidth = childMaxPreferredLogicalWidth;
+ else if (computedInlineSize.isMinContent())
+ childMaxPreferredLogicalWidth = childMinPreferredLogicalWidth;
}
}
_______________________________________________
webkit-changes mailing list
webkit-changes@lists.webkit.org
https://lists.webkit.org/mailman/listinfo/webkit-changes
