Make #'parse-integer Lisp-visible, extend it, allowing non-ASCII digits. src/ChangeLog addition: 2015-02-25 Aidan Kehoe <kehoea@parhasard.net> * lread.c (read_atom): Use the new calling convention for parse_integer(). * lisp.h: Change the declaration of parse_integer (). * number.h (bignum_set_emacs_int, make_bignum_emacs_uint): New #defines, used in data.c. * lread.c (read_integer): Ditto. * lread.c (read1): Ditto. * data.c (find_highest_value): New. * data.c (fill_ichar_array): New. * data.c (build_fixnum_to_char_map): New. * data.c (Fset_digit_fixnum_map): New. * data.c (Fdigit_char_p): Moved from cl-extra.el. * data.c (Fdigit_char): Moved from cl-extra.el. * data.c (parse_integer): Moved from lread.c. * data.c (Fparse_integer): Made available to Lisp. * data.c (syms_of_data): Make the new subrs available. * data.c (vars_of_data): Make the new vars available. Expose parse_integer to Lisp, make it follow the Common Lisp API (with some extensions, to allow us to support non ASCII digit characters). lisp/ChangeLog addition: 2015-02-25 Aidan Kehoe <kehoea@parhasard.net> * cl-extra.el (digit-char-p): Moved to data.c. * cl-extra.el (digit-char): Moved to data.c. tests/ChangeLog addition: 2015-02-25 Aidan Kehoe <kehoea@parhasard.net> * automated/lisp-tests.el: parse_integer(), used in #'read, now signals invalid-argument rather than invalid-read-syntax, check for that. * automated/lisp-tests.el: Check #'parse-integer now it's available to Lisp, check #'digit-char, #'digit-char-p and the congruence in behaviour, check the XEmacs-specific RADIX-TABLE argument behaviour.