[webkit-reviews] review requested: [Bug 50245] Add word-prefix search options to the text search without ICU : [Attachment 95002] word-prefix search for no-ICU code path (v1.0)

[bugzilla-daemon at webkit.org]

yi shen <yi.4.shen at nokia.com> has asked  for review:
Bug 50245: Add word-prefix search options to the text search without ICU
https://bugs.webkit.org/show_bug.cgi?id=50245

Attachment 95002: word-prefix search for no-ICU code path (v1.0)
https://bugs.webkit.org/attachment.cgi?id=95002&action=review

------- Additional Comments from yi shen <yi.4.shen at nokia.com>
For Qt, this patch makes the LayoutTests/editing/text-iterator/findString.html
pass all the tests except the one for the Thai words.

 var thaiWords = [
	"\u0e01\u0e23",
	"\u0e1b\u0e39\u0e40\u0e25",
	"\u0e01\u0e0a",
	"\u0e01\u0e0a\u0e01\u0e23", // thaiWords[2] + thaiWords[0]
	"\u0e01\u0e23\u0e01\u0e0a", // thaiWords[0] + thaiWords[2]
	"\u0e1a\u0e07\u0e01\u0e0a", // ends with thaiWords[2]
    ];

    testFindString(thaiWords.join(""), thaiWords[0], ["AtWordStarts"], [[0, 2],
[12, 14], []]); // Failed here

After traced down the problem in SearchBuffer::isWordStartMatch() function, I
found the issue is that findNextWordFromIndex() doesn't return a correct word
position for Thai words, which is implemented in TextBoundariesQt.cpp by using
QTextBoundaryFinder. In other word, this is a Qt issue and I will fire a Qt bug
for it.