Re: alpha, print, graph, blank, etc.

• Previous message: Ben Dougall: "Re: regular expressions with unicode situation?"
• Maybe in reply to: Mark Davis: "alpha, print, graph, blank, etc."
• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ] [ attachment ]
• Mail actions: [ respond to this message ] [ mail a new topic ]

I updated the page based on some feedback and comparisons against TR 14652,
plus some more delving into the POSIX standard.

http://oss.software.ibm.com/cvs/icu/~checkout~/icuhtml/design/posix_classes.
html

I took Mario's suggestion, which is also followed by the TR, about titlecase
letters. Also expanded xdigit, since POSIX requires that it be a superset of
digit. And added Symbols to punct, since that is how the POSIX locale
handles them.

The main open issue is:

Posix requires that alpha include upper and lower, but in Unicode Alphabetic
does not include Lowercase and Uppercase. The differences are that Uppercase
includes U+24B6..U+24CF CIRCLED LATIN CAPITAL LETTER A..Z, and Lowercase
includes U+24D0..U+24E9 CIRCLED LATIN SMALL LETTER A..Z. None of these are
in Alphabetic. The three choices are:

1. Ignore the POSIX requirement
2. Add the missing characters to alpha
3. Subtract the excess characters from lower and upper respectively.

There are a couple of other issues listed on the page.

Märk Dāvĭs

• Next message: Michael Everson: "Private Use Area"
• Previous message: Ben Dougall: "Re: regular expressions with unicode situation?"
• Maybe in reply to: Mark Davis: "alpha, print, graph, blank, etc."
• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ] [ attachment ]
• Mail actions: [ respond to this message ] [ mail a new topic ]

This archive was generated by hypermail 2.1.5 : Wed Apr 23 2003 - 20:51:22 EDT