647805 - ZWNJ and ZWJ in a page will disrupt search operation ("Find")

Status: UNCONFIRMED

(Core :: Find Backend, defect)

Description

[j. 'mach' wust]

User-Agent:       Mozilla/5.0 (Macintosh; Intel Mac OS X 10.6; rv:2.0) Gecko/20100101 Firefox/4.0
Build Identifier: 

According to The Unicode Standard, Version 6.0 (section 16.2 Layout Controls, paragraph "Filtering Joiner and Non-joiner", p. 528), the characters ZERO WIDTH JOINER and ZERO WIDTH NON-JOINER should be "ordinarily ignored by processes that analyze text content. For example, a spell-checker or a search operation should filter them out when checking for matches."

In Firefox, however, the search operation does not filter them out.

Reproducible: Always

Steps to Reproduce:
1. Go to: http://en.wikipedia.org/w/index.php?title=Zero-width_non-joiner&oldid=422495995
2. Select "Find" from the "Edit" menu and enter "Aufl"

Actual Results:  
Only the instances of "Auflösung" without ZWNJ are found.

Expected Results:  
All instances of "Auflösung" should be found, regardless of the ZWNJ.

This bug is probably related Bug 434044 which is about the spell-checker.

The SOFT HYPHEN (SHY) character, by comparison, does not disrupt the search operation. That is, a search for "hello" will match any instance of "hello", whether it is written with SHY or not. That is exactly how the search operation should behave be with regard to the ZWNJ and ZWJ.

Comment 1

[Masatoshi Kimura [:emk]]

Bug 640856 patch will fix this.