-*- mode: fundamental; coding: utf-8 -*-
Hack the Maxima parser to let non-utf8 aware Lisps to read (some)
utf8-encoded Maxima source files. It can also be used to extend the
alphabetic character set of utf8-aware Lisps, so that non-spacing
characters can be used, too (see below).
The hack works around the non-utf8 aware Lisps by interning the wide
characters as symbols and coercing the symbols to lists of (narrow)
characters. These narrow characters are used as hash keys
in *alpha-char-hash*, whose values are a list of conses of the form
("DESCRIPTION OF CHARACTER" . CHARACTER)
where the CDR is, of course, not the character per se, but the
character as an interned symbol.
The Maxima function `alphabetp' in src/nparse.lisp is modified to use
this hash table to verify if a character is alphabetical.
In the present configuration, utf8_hack admits only the Greek
alphabet (including accented characters) and all characters denoted as
mathematical. There is Emacs Lisp code in utf8-hack.el that allows
for the user to generate a wider or narrower default.
Finally, utf8_hack.lisp includes three Maxima top-level functions:
--set_alpha_char (char_sym, description)
Add a wide character, denoted by `char_sym', to the hashtable
of known alphabetical characters. The `char_sym' may be a
quoted character, a symbol or an integer; `description' should
be a string.
Example:
Maxima branch_5_36_base_198_g1a81b10 http://maxima.sourceforge.net
using Lisp SBCL 1.2.14.debian
Distributed under the GNU Public License. See the file COPYING.
Dedicated to the memory of William Schelter.
The function bug_report() provides bug reporting information.
(%i1) z̄ : conjugate(z);
incorrect syntax: Ì„ is not an infix operator
z̄
^
(%i1) load("utf8-hack.lisp")$
(%i2) set_alpha_char(772,"Non-spacing macron");
(%o2) done
(%i3) z̄ : conjugate(z);
(%o3) conjugate(z);
--utf8_hack(regexp):
Regexp may be MAXIMA-NREGEX regular expression, `all' or
empty. This function alters the range of accepted utf8
characters by choosing only those whose description matches
`regexp'.
If `regexp' is `all', the default is restored; if `regexp' is
empty (or `false'), then all non-ascii characters are made
non-alphabetical.
Examples:
./maxima-local --lisp=gcl
Maxima 5.28.0_179_g3ddde46 http://maxima.sourceforge.net
using Lisp GNU Common Lisp (GCL) GCL 2.6.7 (a.k.a. GCL)
Distributed under the GNU Public License. See the file COPYING.
Dedicated to the memory of William Schelter.
The function bug_report() provides bug reporting information.
(%i1) � : cos(θ);
incorrect syntax: � is not an infix operator
\317\201
^
(%i1) load("utf8-hack.lisp");
(%o1) utf8-hack.lisp
(%i2) � : cos(θ);
(%o2) cos(θ)
(%i3) utf8_hack("math");
(%o3) done
(%i4) � : cos(θ);
incorrect syntax: � is not an infix operator
\317\201Space
^
(%i4) � : cos(�);
(%o4) cos(�)
(%i5) utf8_hack("greek");
(%o5) done
(%i6) � : cos(�);
incorrect syntax: ��” is not an infix operator
\360\235\235\224Space
^
(%i6) � : cos(θ);
(%o6) cos(θ)
In (%i1), Maxima's parser flags an error due to the
wide utf8 character. After loading utf8-hack.lisp, the
parser parses (%i2) successfully.
In (%i3), utf8_hack redefines the alphabetical
characters to include only those extra characters that
have "math" in the description. (%i4) shows that the
parses treats the Greek characters as in the first
line. The second (%i4) is the same line, but with
mathematical rho and theta. In (%i5), we redefine the
alphabetical characters to include only the extra
Greek characters; in (%i6) we see the parser chokes on
the mathematical characters, but reads the Greek
characters successfully.
--print_alpha_char_hash()
Prints the current hash-table *alpha-char-hash*.
(%i7) print_alpha_char_hash();
#\\231 = ((GREEK CAPITAL LETTER IOTA WITH MACRON . á¿™)
(GREEK CAPITAL LETTER ETA WITH DASIA AND PROSGEGRAMMENI
. á¾™)
(GREEK CAPITAL LETTER UPSILON WITH DASIA . á½™)
(GREEK CAPITAL LETTER EPSILON WITH DASIA . á¼™)
(GREEK SMALL LETTER ARCHAIC KOPPA . Ï™)
(GREEK CAPITAL LETTER IOTA . Ι))
...
(%o7) done
WARNING: this is a HACK. Some wide characters will be
recognized as alphabetical, even if they are not explicitly
included. You have been warned.
utf8-hack-data.lisp contains the data the is read when utf8-hack.lisp
is read. To generate a new version of this file, evaluate the buffer
utf8-hack.el in Emacs and then evaluate
(utf8-hack <regexp>)
to generate the utf8-hack-data.lisp file.