xref: /386bsd/usr/local/lib/emacs/19.25/lisp/iso-cvt.el (revision a2142627) 
1;; iso-cvt.el -- translate to ISO 8859-1 from/to net/TeX conventions
2;; Copyright � 1993 Free Software Foundation, Inc.
3;; Was formerly called gm-lingo.el.
4
5;; Author: Michael Gschwind <mike@vlsivie.tuwien.ac.at>
6;; Keywords: tex, iso, latin, i18n
7
8;; This file is part of GNU Emacs.
9
10;; GNU Emacs is free software; you can redistribute it and/or modify
11;; it under the terms of the GNU General Public License as published by
12;; the Free Software Foundation; either version 2, or (at your option)
13;; any later version.
14
15;; GNU Emacs is distributed in the hope that it will be useful,
16;; but WITHOUT ANY WARRANTY; without even the implied warranty of
17;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
18;; GNU General Public License for more details.
19
20;; You should have received a copy of the GNU General Public License
21;; along with GNU Emacs; see the file COPYING.  If not, write to
22;; the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA.
23
24;;; Commentary:
25;; This lisp code server two purposes, both of which involve
26;; the translation of various conventions for representing European
27;; character sets to ISO 8859-1.
28
29; Net support:
30; Various conventions exist in Newsgroups on how to represent national
31; characters. The functions provided here translate these net conventions
32; to ISO.
33;
34; Calling `iso-german' will turn the net convention for umlauts ("a etc.)
35; into ISO latin1 umlaute for easy reading.
36; 'iso-spanish' will turn net conventions for representing spanish
37; to ISO latin1. (Note that accents are omitted in news posts most
38; of the time, only enye is escaped.)
39
40; TeX support
41; This mode installs hooks which change TeX files to ISO Latin-1 for
42; simplified editing. When the TeX file is saved, ISO latin1 characters are
43; translated back to escape sequences.
44;
45; An alternative is a TeX style that handles 8 bit ISO files
46; (available on ftp.vlsivie.tuwien.ac.at in /pub/8bit)
47; - but these files are difficult to transmit ... so while the net is
48; still @ 7 bit this may be useful
49
50;; TO DO:
51; The net support should install hooks (like TeX support does)
52; which recognizes certains news groups and translates all articles from
53; those groups.
54;
55; Cover more cases for translation (There is an infinite number of ways to
56; represent accented characters in TeX)
57
58;; SEE ALSO:
59; If you are interested in questions related to using the ISO 8859-1
60; characters set (configuring emacs, Unix, etc. to use ISO), then you
61; can get the ISO 8859-1 FAQ via anonymous ftp from
62; ftp.vlsivie.tuwien.ac.at in /pub/bit/FAQ-ISO-8859-1
63
64;;; Code:
65
66(provide 'iso-cvt)
67
68(defvar iso-spanish-trans-tab
69  '(
70    ("~n" "�")
71    ("\([a-zA-Z]\)#" "\\1�")
72    ("~N" "�")
73    ("\\([-a-zA-Z\"`]\\)\"u" "\\1�")
74    ("\\([-a-zA-Z\"`]\\)\"U" "\\1�")
75    ("\\([-a-zA-Z]\\)'o" "\\1�")
76    ("\\([-a-zA-Z]\\)'O" "\\�")
77    ("\\([-a-zA-Z]\\)'e" "\\1�")
78    ("\\([-a-zA-Z]\\)'E" "\\1�")
79    ("\\([-a-zA-Z]\\)'a" "\\1�")
80    ("\\([-a-zA-Z]\\)'A" "\\1A")
81    ("\\([-a-zA-Z]\\)'i" "\\1�")
82    ("\\([-a-zA-Z]\\)'I" "\\1�")
83    )
84  "Spanish translation table.")
85
86(defun iso-translate-conventions (trans-tab)
87  "Use the translation table argument to translate the current buffer."
88  (interactive)
89  (save-excursion
90    (widen)
91    (goto-char (point-min))
92    (let ((work-tab trans-tab)
93	  (buffer-read-only nil))
94      (while work-tab
95	(save-excursion
96	  (let ((trans-this (car work-tab)))
97	    (while (re-search-forward (car trans-this) nil t)
98	      (replace-match (car (cdr trans-this)) nil nil)))
99	  (setq work-tab (cdr work-tab)))))))
100
101(defun iso-spanish ()
102  "Translate net conventions for Spanish to ISO 8859-1."
103  (interactive)
104  (iso-translate-conventions iso-spanish-trans-tab))
105
106(defvar iso-aggressive-german-trans-tab
107  '(
108    ("\"a" "�")
109    ("\"A" "�")
110    ("\"o" "�")
111    ("\"O" "�")
112    ("\"u" "�")
113    ("\"U" "�")
114    ("\"s" "�")
115    ("\\\\3" "�")
116    )
117  "German translation table.
118This table uses an aggressive translation approach and may translate
119erroneously translate too much.")
120
121(defvar iso-conservative-german-trans-tab
122  '(
123    ("\\([-a-zA-Z\"`]\\)\"a" "\\1�")
124    ("\\([-a-zA-Z\"`]\\)\"A" "\\1�")
125    ("\\([-a-zA-Z\"`]\\)\"o" "\\1�")
126    ("\\([-a-zA-Z\"`]\\)\"O" "\\1�")
127    ("\\([-a-zA-Z\"`]\\)\"u" "\\1�")
128    ("\\([-a-zA-Z\"`]\\)\"U" "\\1�")
129    ("\\([-a-zA-Z\"`]\\)\"s" "\\1�")
130    ("\\([-a-zA-Z\"`]\\)\\\\3" "\\1�")
131    )
132  "German translation table.
133This table uses a conservative translation approach and may translate too
134little.")
135
136
137(defvar iso-german-trans-tab iso-aggressive-german-trans-tab
138  "Currently active translation table for German.")
139
140(defun iso-german ()
141 "Translate net conventions for German to ISO 8859-1."
142 (interactive)
143 (iso-translate-conventions iso-german-trans-tab))
144
145(defvar iso-iso2tex-trans-tab
146  '(
147    ("�" "{\\\\\"a}")
148    ("�" "{\\\\`a}")
149    ("�" "{\\\\'a}")
150    ("�" "{\\\\~a}")
151    ("�" "{\\\\^a}")
152    ("�" "{\\\\\"e}")
153    ("�" "{\\\\`e}")
154    ("�" "{\\\\'e}")
155    ("�" "{\\\\^e}")
156    ("�" "{\\\\\"\\\\i}")
157    ("�" "{\\\\`\\\\i}")
158    ("�" "{\\\\'\\\\i}")
159    ("�" "{\\\\^\\\\i}")
160    ("�" "{\\\\\"o}")
161    ("�" "{\\\\`o}")
162    ("�" "{\\\\'o}")
163    ("�" "{\\\\~o}")
164    ("�" "{\\\\^o}")
165    ("�" "{\\\\\"u}")
166    ("�" "{\\\\`u}")
167    ("�" "{\\\\'u}")
168    ("�" "{\\\\^u}")
169    ("�" "{\\\\\"A}")
170    ("�" "{\\\\`A}")
171    ("�" "{\\\\'A}")
172    ("�" "{\\\\~A}")
173    ("�" "{\\\\^A}")
174    ("�" "{\\\\\"E}")
175    ("�" "{\\\\`E}")
176    ("�" "{\\\\'E}")
177    ("�" "{\\\\^E}")
178    ("�" "{\\\\\"I}")
179    ("�" "{\\\\`I}")
180    ("�" "{\\\\'I}")
181    ("�" "{\\\\^I}")
182    ("�" "{\\\\\"O}")
183    ("�" "{\\\\`O}")
184    ("�" "{\\\\'O}")
185    ("�" "{\\\\~O}")
186    ("�" "{\\\\^O}")
187    ("�" "{\\\\\"U}")
188    ("�" "{\\\\`U}")
189    ("�" "{\\\\'U}")
190    ("�" "{\\\\^U}")
191    ("�" "{\\\\~n}")
192    ("�" "{\\\\~N}")
193    ("�" "{\\\\c c}")
194    ("�" "{\\\\c C}")
195    ("�" "{\\\\ss}")
196    ("�" "{?`}")
197    ("�" "{!`}")
198    )
199  "Translation table for translating ISO 8859-1 characters to TeX sequences.")
200
201
202
203
204(defun iso-iso2tex ()
205 "Translate ISO 8859-1 characters to TeX sequences."
206 (interactive)
207 (iso-translate-conventions iso-iso2tex-trans-tab))
208
209
210(defvar iso-tex2iso-trans-tab
211  '(
212    ("{\\\\\"a}" "�")
213    ("{\\\\`a}" "�")
214    ("{\\\\'a}" "�")
215    ("{\\\\~a}" "�")
216    ("{\\\\^a}" "�")
217    ("{\\\\\"e}" "�")
218    ("{\\\\`e}" "�")
219    ("{\\\\'e}" "�")
220    ("{\\\\^e}" "�")
221    ("{\\\\\"\\\\i}" "�")
222    ("{\\\\`\\\\i}" "�")
223    ("{\\\\'\\\\i}" "�")
224    ("{\\\\^\\\\i}" "�")
225    ("{\\\\\"i}" "�")
226    ("{\\\\`i}" "�")
227    ("{\\\\'i}" "�")
228    ("{\\\\^i}" "�")
229    ("{\\\\\"o}" "�")
230    ("{\\\\`o}" "�")
231    ("{\\\\'o}" "�")
232    ("{\\\\~o}" "�")
233    ("{\\\\^o}" "�")
234    ("{\\\\\"u}" "�")
235    ("{\\\\`u}" "�")
236    ("{\\\\'u}" "�")
237    ("{\\\\^u}" "�")
238    ("{\\\\\"A}" "�")
239    ("{\\\\`A}" "�")
240    ("{\\\\'A}" "�")
241    ("{\\\\~A}" "�")
242    ("{\\\\^A}" "�")
243    ("{\\\\\"E}" "�")
244    ("{\\\\`E}" "�")
245    ("{\\\\'E}" "�")
246    ("{\\\\^E}" "�")
247    ("{\\\\\"I}" "�")
248    ("{\\\\`I}" "�")
249    ("{\\\\'I}" "�")
250    ("{\\\\^I}" "�")
251    ("{\\\\\"O}" "�")
252    ("{\\\\`O}" "�")
253    ("{\\\\'O}" "�")
254    ("{\\\\~O}" "�")
255    ("{\\\\^O}" "�")
256    ("{\\\\\"U}" "�")
257    ("{\\\\`U}" "�")
258    ("{\\\\'U}" "�")
259    ("{\\\\^U}" "�")
260    ("{\\\\~n}" "�")
261    ("{\\\\~N}" "�")
262    ("{\\\\c c}" "�")
263    ("{\\\\c C}" "�")
264    ("\\\\\"{a}" "�")
265    ("\\\\`{a}" "�")
266    ("\\\\'{a}" "�")
267    ("\\\\~{a}" "�")
268    ("\\\\^{a}" "�")
269    ("\\\\\"{e}" "�")
270    ("\\\\`{e}" "�")
271    ("\\\\'{e}" "�")
272    ("\\\\^{e}" "�")
273    ("\\\\\"{\\\\i}" "�")
274    ("\\\\`{\\\\i}" "�")
275    ("\\\\'{\\\\i}" "�")
276    ("\\\\^{\\\\i}" "�")
277    ("\\\\\"{i}" "�")
278    ("\\\\`{i}" "�")
279    ("\\\\'{i}" "�")
280    ("\\\\^{i}" "�")
281    ("\\\\\"{o}" "�")
282    ("\\\\`{o}" "�")
283    ("\\\\'{o}" "�")
284    ("\\\\~{o}" "�")
285    ("\\\\^{o}" "�")
286    ("\\\\\"{u}" "�")
287    ("\\\\`{u}" "�")
288    ("\\\\'{u}" "�")
289    ("\\\\^{u}" "�")
290    ("\\\\\"{A}" "�")
291    ("\\\\`{A}" "�")
292    ("\\\\'{A}" "�")
293    ("\\\\~{A}" "�")
294    ("\\\\^{A}" "�")
295    ("\\\\\"{E}" "�")
296    ("\\\\`{E}" "�")
297    ("\\\\'{E}" "�")
298    ("\\\\^{E}" "�")
299    ("\\\\\"{I}" "�")
300    ("\\\\`{I}" "�")
301    ("\\\\'{I}" "�")
302    ("\\\\^{I}" "�")
303    ("\\\\\"{O}" "�")
304    ("\\\\`{O}" "�")
305    ("\\\\'{O}" "�")
306    ("\\\\~{O}" "�")
307    ("\\\\^{O}" "�")
308    ("\\\\\"{U}" "�")
309    ("\\\\`{U}" "�")
310    ("\\\\'{U}" "�")
311    ("\\\\^{U}" "�")
312    ("\\\\~{n}" "�")
313    ("\\\\~{N}" "�")
314    ("\\\\c{c}" "�")
315    ("\\\\c{C}" "�")
316    ("{\\\\ss}" "�")
317    ("?`" "�")
318    ("!`" "�")
319    ("{?`}" "�")
320    ("{!`}" "�")
321    )
322  "Translation table for translating TeX sequences to ISO 8859-1 characters.
323This table is not exhaustive (and due to TeX's power can never be). It only
324contains commonly used sequences.")
325
326(defun iso-tex2iso ()
327 "Translate TeX sequences to ISO 8859-1 characters."
328 (interactive)
329 (iso-translate-conventions iso-tex2iso-trans-tab))
330
331(defvar iso-gtex2iso-trans-tab
332  '(
333    ("{\\\\\"a}" "�")
334    ("{\\\\`a}" "�")
335    ("{\\\\'a}" "�")
336    ("{\\\\~a}" "�")
337    ("{\\\\^a}" "�")
338    ("{\\\\\"e}" "�")
339    ("{\\\\`e}" "�")
340    ("{\\\\'e}" "�")
341    ("{\\\\^e}" "�")
342    ("{\\\\\"\\\\i}" "�")
343    ("{\\\\`\\\\i}" "�")
344    ("{\\\\'\\\\i}" "�")
345    ("{\\\\^\\\\i}" "�")
346    ("{\\\\\"i}" "�")
347    ("{\\\\`i}" "�")
348    ("{\\\\'i}" "�")
349    ("{\\\\^i}" "�")
350    ("{\\\\\"o}" "�")
351    ("{\\\\`o}" "�")
352    ("{\\\\'o}" "�")
353    ("{\\\\~o}" "�")
354    ("{\\\\^o}" "�")
355    ("{\\\\\"u}" "�")
356    ("{\\\\`u}" "�")
357    ("{\\\\'u}" "�")
358    ("{\\\\^u}" "�")
359    ("{\\\\\"A}" "�")
360    ("{\\\\`A}" "�")
361    ("{\\\\'A}" "�")
362    ("{\\\\~A}" "�")
363    ("{\\\\^A}" "�")
364    ("{\\\\\"E}" "�")
365    ("{\\\\`E}" "�")
366    ("{\\\\'E}" "�")
367    ("{\\\\^E}" "�")
368    ("{\\\\\"I}" "�")
369    ("{\\\\`I}" "�")
370    ("{\\\\'I}" "�")
371    ("{\\\\^I}" "�")
372    ("{\\\\\"O}" "�")
373    ("{\\\\`O}" "�")
374    ("{\\\\'O}" "�")
375    ("{\\\\~O}" "�")
376    ("{\\\\^O}" "�")
377    ("{\\\\\"U}" "�")
378    ("{\\\\`U}" "�")
379    ("{\\\\'U}" "�")
380    ("{\\\\^U}" "�")
381    ("{\\\\~n}" "�")
382    ("{\\\\~N}" "�")
383    ("{\\\\c c}" "�")
384    ("{\\\\c C}" "�")
385    ("\\\\\"{a}" "�")
386    ("\\\\`{a}" "�")
387    ("\\\\'{a}" "�")
388    ("\\\\~{a}" "�")
389    ("\\\\^{a}" "�")
390    ("\\\\\"{e}" "�")
391    ("\\\\`{e}" "�")
392    ("\\\\'{e}" "�")
393    ("\\\\^{e}" "�")
394    ("\\\\\"{\\\\i}" "�")
395    ("\\\\`{\\\\i}" "�")
396    ("\\\\'{\\\\i}" "�")
397    ("\\\\^{\\\\i}" "�")
398    ("\\\\\"{i}" "�")
399    ("\\\\`{i}" "�")
400    ("\\\\'{i}" "�")
401    ("\\\\^{i}" "�")
402    ("\\\\\"{o}" "�")
403    ("\\\\`{o}" "�")
404    ("\\\\'{o}" "�")
405    ("\\\\~{o}" "�")
406    ("\\\\^{o}" "�")
407    ("\\\\\"{u}" "�")
408    ("\\\\`{u}" "�")
409    ("\\\\'{u}" "�")
410    ("\\\\^{u}" "�")
411    ("\\\\\"{A}" "�")
412    ("\\\\`{A}" "�")
413    ("\\\\'{A}" "�")
414    ("\\\\~{A}" "�")
415    ("\\\\^{A}" "�")
416    ("\\\\\"{E}" "�")
417    ("\\\\`{E}" "�")
418    ("\\\\'{E}" "�")
419    ("\\\\^{E}" "�")
420    ("\\\\\"{I}" "�")
421    ("\\\\`{I}" "�")
422    ("\\\\'{I}" "�")
423    ("\\\\^{I}" "�")
424    ("\\\\\"{O}" "�")
425    ("\\\\`{O}" "�")
426    ("\\\\'{O}" "�")
427    ("\\\\~{O}" "�")
428    ("\\\\^{O}" "�")
429    ("\\\\\"{U}" "�")
430    ("\\\\`{U}" "�")
431    ("\\\\'{U}" "�")
432    ("\\\\^{U}" "�")
433    ("\\\\~{n}" "�")
434    ("\\\\~{N}" "�")
435    ("\\\\c{c}" "�")
436    ("\\\\c{C}" "�")
437    ("{\\\\ss}" "�")
438    ("?`" "�")
439    ("!`" "�")
440    ("{?`}" "�")
441    ("{!`}" "�")
442    ("\"a" "�")
443    ("\"A" "�")
444    ("\"o" "�")
445    ("\"O" "�")
446    ("\"u" "�")
447    ("\"U" "�")
448    ("\"s" "�")
449    ("\\\\3" "�")
450    )
451  "Translation table for translating German TeX sequences to ISO 8859-1.
452This table is not exhaustive (and due to TeX's power can never be).  It only
453contains commonly used sequences.")
454
455(defvar iso-iso2gtex-trans-tab
456  '(
457    ("�" "\"a")
458    ("�" "{\\\\`a}")
459    ("�" "{\\\\'a}")
460    ("�" "{\\\\~a}")
461    ("�" "{\\\\^a}")
462    ("�" "{\\\\\"e}")
463    ("�" "{\\\\`e}")
464    ("�" "{\\\\'e}")
465    ("�" "{\\\\^e}")
466    ("�" "{\\\\\"\\\\i}")
467    ("�" "{\\\\`\\\\i}")
468    ("�" "{\\\\'\\\\i}")
469    ("�" "{\\\\^\\\\i}")
470    ("�" "\"o")
471    ("�" "{\\\\`o}")
472    ("�" "{\\\\'o}")
473    ("�" "{\\\\~o}")
474    ("�" "{\\\\^o}")
475    ("�" "\"u")
476    ("�" "{\\\\`u}")
477    ("�" "{\\\\'u}")
478    ("�" "{\\\\^u}")
479    ("�" "\"A")
480    ("�" "{\\\\`A}")
481    ("�" "{\\\\'A}")
482    ("�" "{\\\\~A}")
483    ("�" "{\\\\^A}")
484    ("�" "{\\\\\"E}")
485    ("�" "{\\\\`E}")
486    ("�" "{\\\\'E}")
487    ("�" "{\\\\^E}")
488    ("�" "{\\\\\"I}")
489    ("�" "{\\\\`I}")
490    ("�" "{\\\\'I}")
491    ("�" "{\\\\^I}")
492    ("�" "\"O")
493    ("�" "{\\\\`O}")
494    ("�" "{\\\\'O}")
495    ("�" "{\\\\~O}")
496    ("�" "{\\\\^O}")
497    ("�" "\"U")
498    ("�" "{\\\\`U}")
499    ("�" "{\\\\'U}")
500    ("�" "{\\\\^U}")
501    ("�" "{\\\\~n}")
502    ("�" "{\\\\~N}")
503    ("�" "{\\\\c c}")
504    ("�" "{\\\\c C}")
505    ("�" "\"s")
506    ("�" "{?`}")
507    ("�" "{!`}")
508    )
509  "Translation table for translating ISO 8859-1 characters to German TeX.")
510
511(defun iso-gtex2iso ()
512 "Translate German TeX sequences to ISO 8859-1 characters."
513 (interactive)
514 (iso-translate-conventions iso-gtex2iso-trans-tab))
515
516
517(defun iso-iso2gtex ()
518 "Translate ISO 8859-1 characters to German TeX sequences."
519 (interactive)
520 (iso-translate-conventions iso-iso2gtex-trans-tab))
521
522
523(defun iso-german-tex-p ()
524 "Check if tex buffer is German LaTeX."
525 (save-excursion
526   (widen)
527   (goto-char (point-min))
528   (re-search-forward "\\\\documentstyle\\[.*german.*\\]" nil t)))
529
530(defun iso-fix-iso2tex ()
531  "Turn ISO 8859-1 (aka. ISO Latin-1) buffer into TeX sequences.
532If German TeX is used, German TeX sequences are generated."
533  (if (or (equal major-mode 'latex-mode)
534	  (equal major-mode 'LaTeX-mode)) ; AucTeX wants this
535      (if (iso-german-tex-p)
536	  (iso-iso2gtex)
537	(iso-iso2tex)))
538  (if (or (equal major-mode 'tex-mode)
539	  (equal major-mode 'TeX-mode)) ; AucTeX wants this
540      (iso-iso2tex)))
541
542(defun iso-fix-tex2iso ()
543  "Turn TeX sequences into ISO 8859-1 (aka. ISO Latin-1) characters.
544This function recognices German TeX buffers."
545  (if (or (equal major-mode 'latex-mode)
546	  (equal major-mode 'Latex-mode)) ; AucTeX wants this
547      (if (iso-german-tex-p)
548	  (iso-gtex2iso)
549	(iso-tex2iso)))
550  (if (or (equal major-mode 'tex-mode)
551	  (equal major-mode 'TeX-mode))  ;; AucTeX wants this
552      (iso-tex2iso)))
553
554(defun iso-cvt-ffh ()
555  "find-file-hook for iso-cvt-cvt.el."
556  (iso-fix-tex2iso)
557  (set-buffer-modified-p nil))
558
559(defun iso-cvt-wfh ()
560  "write file hook for iso-cvt-cvt.el."
561  (iso-fix-iso2tex))
562
563(defun iso-cvt-ash ()
564  "after save hook for iso-cvt-cvt.el."
565  (iso-fix-tex2iso)
566  (set-buffer-modified-p nil))
567
568(add-hook 'find-file-hooks 'iso-cvt-ffh)
569(add-hook 'write-file-hooks 'iso-cvt-wfh)
570(add-hook 'after-save-hook 'iso-cvt-ash)
571
572;;; iso-cvt.el ends here
573