Add the non-ASCII digit support, now #'parse-integer can handle it. lisp/ChangeLog addition: 2015-03-08 Aidan Kehoe <kehoea@parhasard.net> * mule/digit.el: New file, specifying integer weights for those decimal digits specified by Unicode, and for the Latin characters in ASCII. * dumped-lisp.el (preloaded-file-list): Dump this file, so those weights are available at runtime. tests/ChangeLog addition: 2015-03-16 Aidan Kehoe <kehoea@parhasard.net> * automated/lisp-tests.el: Extend the tests for #'parse-integer, #'digit-char, #'digit-char-p substantially, testing the Unicode decimal digits in detail. src/ChangeLog addition: 2015-03-08 Aidan Kehoe <kehoea@parhasard.net> * data.c (Fparse_integer): Document a complication that arises when a leading minus sign has a digit weight but is nonetheles interpreted as a minus.

--- a/src/data.c	Mon Mar 16 00:11:30 2015 +0000
+++ b/src/data.c	Mon Mar 16 00:28:18 2015 +0000
@@ -1940,6 +1940,13 @@
 possible values further, and the maximum RADIX available is always the largest
 positive value available RADIX-TABLE.
 
+If RADIX-TABLE (or `digit-fixnum-map') assigns a numeric value to `-', its
+digit value will be ignored if ?- is the first character in a string. The
+number will be treated as negative.  To work around this, double it, or assign
+another character the value zero and prefix non-negative strings with that.  A
+digit value for ?+ will be respected even if it is the first character in the
+representation of an integer.
+
 arguments: (STRING &key (START 0) end (RADIX 10) junk-allowed radix-table)
 */
        (int nargs, Lisp_Object *args))