CVS difference for ais/ai-00285.txt

Differences between 1.10 and version 1.11
Log of other versions for file ais/ai-00285.txt

--- ais/ai-00285.txt	2003/06/13 02:09:09	1.10
+++ ais/ai-00285.txt	2003/08/01 01:40:07	1.11
@@ -3696,3 +3696,38 @@
+From: Robert I. Eachus
+Sent: Sunday, March 23, 2003  11:43 AM
+I'm confused.  I am certainly proposing using the IANA registry for the
+names of character sets, and a way for programmers to determine which
+set is in use.
+As understand the 16-bit character set issues, in addition to character
+names, there is characterization in terms of 2.2 Lexical Elements for
+non-Latin 1 characters. In other words which characters can be used in
+names and numeric literals.
+I suspect that what Robert is referring to the fact that if someone uses
+a non-Latin 1 eight-bit set by a command-line argument, that the names
+won't match the characters as displayed.  If so, I am actually
+recommending the permission for implementors to 'fix' more than that.
+For example, I don't think we should require that implementations
+support the Windows 1252 character set, but it would be nice to allow
+implementations which choose to do so to get it right.  Some of that
+will follow at compile time if implementations map Windows 1252 to the
+appropriate Unicode/BMP characters.  But it would also be nice to allow
+the Ada.Characters heirarchy and Ada.Text_IO features to be used with
+non-Latin 1 character sets.
+The keyword in the previous paragraph is "allow."  As I said, I think we
+can go a bit farther and provide a standard way to determine the current
+character set mapping.  But there is no reason for us to say you must
+support these character sets (other than Latin 1!), but must not support
+these other sets.  This really has to do with code points in the 00 to
+3F and 80 to BF ranges being printable characters instead of control

Questions? Ask the ACAA Technical Agent