Re: proposal to add API to ubidi.h

Markus Scherer <[email protected]> Thu, 06 Jan 2005 09:59:07 -0800
Newsgroups gmane.comp.lib.icu.general
Organization IBM
Message-ID <[email protected]>
Doug Felt wrote:
> proposal to add multiple paragraph support to ubidi APIs.

Looks good to me.

Please make sure that paragraphs are recognized according to the Unicode Standard.

This was specified in UAX #13: Unicode Newline Guidlines,
but now it is in Unicode 4 (the book) section 5.8 Newline Guidelines (pp 116-119).

The guidelines are necessarily a bit fuzzy between line and paragraph separators, because ISO 
controls are not defined well. Suggestion for the bidi code (but open for discussion):

Treat as paragraph boundaries: CRLF, CR, LF, NEL, PS.
Ignore for paragraph boundaries: LS, FF, VT.
(See Table 5-3)

markus

-- 
Opinions expressed here may not reflect my company's positions unless otherwise noted.