[Webkit-unassigned] [Bug 179682] Incorrect bounds inside <mover>/<munder> when a stretchy operator is present

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=179682

--- Comment #24 from Minsheng Liu <lambda at liu.ms> ---
I have adopted the first part of your change of layout-ing sub-expressions with their stretchy operators’ widths set to zero. However, due to the <munderover> with narrow base case, the hacky last line cannot be removed. My code now looks like this:

void RenderMathMLUnderOver::stretchHorizontalOperatorsAndLayoutChildren()
{
    // The function will be called either "freshly"
    // or as a fix after an embellished operator is stretched.
    ASSERT(isValid());
    ASSERT(needsLayout());

    LayoutUnit stretchWidth = 0;
    // The loop layouts each sub-expression's width with
    // embellished stretchy operators' width set to 0,
    // unless the stretchy operators' stretch widths have been locked.
    for (auto* child = firstChildBox(); child; child = child->nextSiblingBox()) {
        auto renderOperator = toHorizontalStretchyOperator(child);
        if (renderOperator) {
            if (!renderOperator->isStretchWidthLocked())
                renderOperator->resetStretchSize();
            renderOperator->setStretchWidthLocked(true);
            // This step is necessary if the function is called as a fix,
            // since the embellished operator could be nested down deeply.
            // Each direct ancestor of the operator shall be re-layout-ed.
            child->setNeedsLayout();
        }
        child->layoutIfNeeded();
        if (renderOperator) {
            // Note that during the fix the embellished operator might be locked
            // and unlocked multiple times, so I change the implementation of the lock
            // to a counter.
            renderOperator->setStretchWidthLocked(false);
        }

        // Only base() will stretch to the maximum of all other elements,
        // but the spec requires base() to stretch to the maximum width of elements other than the base().
        // A decision is needed.
        // The behavior is not consistent with Firefox.
        stretchWidth = std::max(stretchWidth, child->logicalWidth());
    }

    for (auto* child = firstChildBox(); child; child = child->nextSiblingBox()) {
        if (auto renderOperator = toHorizontalStretchyOperator(child)) {
            if (!renderOperator->isStretchWidthLocked()) {
                renderOperator->resetStretchSize();
                renderOperator->stretchTo(stretchWidth);

                renderOperator->setStretchWidthLocked(true);
                child->setNeedsLayout();
                child->layout();
                renderOperator->setStretchWidthLocked(false);
            }
        }

        // For children other than base(), the stretchWidth is the width of base().
        // This is for <munderover> in which we have braces on the top and the bottom.
        // The braces should be stretched as wide as the base(), not lonoger or shorter.
        // The behavior is consistent with Firefox.
        stretchWidth = firstChildBox()->logicalWidth();
    }
}


And it causes no regression other than the expected one. I will provide a review-ready patch (without tests for now) soon.

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-unassigned/attachments/20171128/34dfb74a/attachment.html>