# New Ticket Created by  Zefram 
# Please include the string:  [perl #130922]
# in the subject line of all future correspondence about this issue. 
# <URL: https://rt.perl.org/Ticket/Display.html?id=130922 >


> "\x[111c2,111c4]".chars
2

The above is erroneous: these two codepoints make up a single extended
grapheme cluster.  U+111c2 "Sharada sign jihvamuliya" has the property
Indic_Syllabic_Category=Consonant_Prefixed, which per UAX #29 makes it a
Grapheme_Cluster_Break=Prepend, which means that it falls under rule GB9b,
so there should not be an extended grapheme cluster break following it
(except when what follows is a control character or end of text).

Strangely, the rule about Prepend characters doesn't seem to be being
entirely ignored.  It is taking effect if the following character is
also a Prepend:

> "\x[111c2,111c3]".chars
1

-zefram

Reply via email to