Add flags to handle 4 deviation characters between IDNA 2003 and IDNA 2008 separately

Description

Handling of the four deviation characters between IDNA 2003 and IDNA 2008 in the current UTS 46 implementation is controlled by one flag. One can choose either the transient mechanism or the non-transient mechanism. In the former, sharp-s and final sigma are mapped to 'ss' and regular sigma and ZWJ/ZWNJ are mapped away. In the latter, all of them are preserved.

As is known, ZWJ/ZWNJ in some languages (e.g. Farsi/Persian and some Indic languages) make a difference in the meaning. With ContextJ rules in place and the prevention of a script mixing, a security risk arising from them (e.g. between two Latin letters) is mitigated significantly.

So, Chrome is considering preserving ZWJ/ZWNJ with ContextJ rules ( Chrome:303404 ).

Activity

Show:
TracBot
June 30, 2018, 11:52 PM
Trac Comment 2 by —2013-11-20T18:36:26.510Z

Discuss with Jungshik & Mark.

Obsolete

Assignee

Markus Scherer

Reporter

Jungshik Shin

Components

Labels

Reviewer

None

Priority

assess

Time Needed

Hours

Fix versions

None
Configure