[icu-design] Set of non-whitespace ignorables in number parsing

[Shane C. <sf...@go...>]

Dear ICU team & users,

I would like to propose the following for: ICU 62

Please provide feedback by: 2018-03-06

Designated API reviewer: Andy

Ticket: http://bugs.icu-project.org/trac/ticket/13084

A previous API proposal for ICU 60 set the set of allowable whitespace
characters in number parsing to "horizontal" whitespace.  The question of
which other characters besides whitespace to count as "ignorable" was
deferred.  I would now like to propose that the set of characters accepted
as "ignorable" in number formatting be horizontal whitespace plus the
Unicode set [:Default_Ignorable_Code_Point:], also known as [:DI:].

This would make the sets of characters ignored by number parsing equal to
the following:

Lenient Mode: [[:Zs:][\u0009][:DI:]]

Strict Mode: [[:DI:]]

Sincerely,

Shane