Bug 48537 – Fonts with no vertical metrics should synthesize baselines

• Status: RESOLVED
• Resolution: FIXED
• Component: Layout and Rendering

[Dave Hyatt]

Fonts with no vertical metrics should synthesize baselines when they appear on lines that do use fonts with vertical metrics.  Basically we want to make sure that English text behaves like vertical-align:middle, i.e., its baseline should be treated as though it cuts through the middle of the ascent+descent.

[Adele Peterson]

<rdar://problem/8612034>

[Dave Hyatt]

Created attachment 74096 [details]
Patch

[Eric Seidel (no email)]

Attachment 74096 [details] did not build on mac:
Build output: http://queues.webkit.org/results/6238014

[Early Warning System Bot]

Attachment 74096 [details] did not build on qt:
Build output: http://queues.webkit.org/results/6177018

[Build Bot]

Attachment 74096 [details] did not build on win:
Build output: http://queues.webkit.org/results/6165018

[Dave Hyatt]

Created attachment 74124 [details]
Patch

[Sam Weinig]

Comment on attachment 74124 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=74124&action=review

> WebCore/rendering/InlineFlowBox.cpp:484
> +            if (verticalAlign == SUB)
> +                verticalPosition += fontSize / 5 + 1;
> +            else if (verticalAlign == SUPER)
> +                verticalPosition -= fontSize / 3 + 1;
> +            else if (verticalAlign == TEXT_TOP)
> +                verticalPosition += renderer->baselinePosition(baselineType, firstLine, lineDirection) - font.ascent(baselineType);
> +            else if (verticalAlign == MIDDLE)
> +                verticalPosition += -static_cast<int>(font.xHeight() / 2) - renderer->lineHeight(firstLine, lineDirection) / 2 + renderer->baselinePosition(baselineType, firstLine, lineDirection);
> +            else if (verticalAlign == TEXT_BOTTOM) {
> +                verticalPosition += font.descent(baselineType);
> +                // lineHeight - baselinePosition is always 0 for replaced elements (except inline blocks), so don't bother wasting time in that case.
> +                if (!renderer->isReplaced() || renderer->isInlineBlockOrInlineTable())
> +                    verticalPosition -= (renderer->lineHeight(firstLine, lineDirection) - renderer->baselinePosition(baselineType, firstLine, lineDirection));
> +            } else if (verticalAlign == BASELINE_MIDDLE)
> +                verticalPosition += -renderer->lineHeight(firstLine, lineDirection) / 2 + renderer->baselinePosition(baselineType, firstLine, lineDirection);
> +            else if (verticalAlign == LENGTH)
> +                verticalPosition -= renderer->style()->verticalAlignLength().calcValue(renderer->lineHeight(firstLine, lineDirection));
> +        }

If you wanted, you could change this to a switch statement.

[Dave Hyatt]

Fixed in r72235.

[WebKit Review Bot]

http://trac.webkit.org/changeset/72235 might have broken SnowLeopard Intel Release (Tests)
The following tests are not passing:
fast/blockflow/vertical-font-fallback.html
fast/repaint/repaint-across-writing-mode-boundary.html

[Ryosuke Niwa]

It seems like this change requires rebaselines for the following tests on GTK:
fast/css/line-height-determined-by-primary-font.html
fast/css/negative-leading.html

+mrobinson for GTK rebaseline.

[Eric Seidel (no email)]

This broke snow leopard tests, and thus the commit-queue:

SUCCESS: Build 20999 (r72235) was the first to show failures: set([u'fast/blockflow/vertical-font-fallback.html', u'fast/repaint/repaint-across-writing-mode-boundary.html'])