Bug 34865 – REGRESSION (r50301): Problem selecting text in a Devanagari website

• Component: Text
• Status: RESOLVED FIXED

<rdar://problem/7609268>

Patch forthcoming.

Created attachment 48598 [details]
Account for non-monotonic character-to-glyph mapping

[WebKit Review Bot]

Attachment 48598 [details] did not pass style-queue:

Failed to run "WebKitTools/Scripts/check-webkit-style" exit_code: 1
WebCore/platform/graphics/mac/ComplexTextController.h:91:  More than one command on the same line  [whitespace/newline] [4]
Total errors found: 1


If any of these errors are false positives, please file a bug against check-webkit-style.

[Simon Fraser (smfr)]

Comment on attachment 48598 [details]
Account for non-monotonic character-to-glyph mapping

> Index: WebCore/ChangeLog
> ===================================================================
> +        * platform/graphics/mac/ComplexTextController.cpp:
> +        (WebCore::ComplexTextController::ComplexTextRun::ComplexTextRun): Initialize m_isMonotonic.
> +        (WebCore::ComplexTextController::ComplexTextRun::setIsMonotonic): Added. Sets m_isMonotonic,
> +        and if the run is not monotonic, populates m_lastInidices with the end offsets of each glyphâs

A typo and a garbled curly quote here (and one more lower down).

> Index: WebCore/platform/graphics/mac/ComplexTextController.cpp
> ===================================================================

> +void ComplexTextController::ComplexTextRun::setIsMonotonic(bool isMonotonic)
> +{
> +    m_isMonotonic = isMonotonic;
> +    if (isMonotonic)
> +        return;

Why doesn't this do the usual (if )isMonotonic == m_isMonotonic) return;?

Also, if changing from non-monotonic to monotonic, should you clear m_lastIndices?

> Index: WebCore/platform/graphics/mac/ComplexTextController.h
> ===================================================================

> +        Vector<CFIndex, 64> m_lastIndices;

The name of this member doesn't immediately communicate what it stores. Maybe m_glyphEndOffsets or something?

r=me

Fixed in <http://trac.webkit.org/changeset/54729>.

[Brian Weinstein]

This test fails on Windows, added to Skipped list in r54799.