[xemacs-hg @ 2006-12-29 18:09:38 by aidan] etc/ChangeLog addition: 2006-12-21 Aidan Kehoe <kehoea@parhasard.net> * unicode/unicode-consortium/8859-7.TXT: Update the mapping to the 2003 version of ISO 8859-7. lisp/ChangeLog addition: 2006-12-21 Aidan Kehoe <kehoea@parhasard.net> * mule/cyrillic.el: * mule/cyrillic.el (iso-8859-5): * mule/cyrillic.el (cyrillic-koi8-r-encode-table): Add syntax, case support for Cyrillic; make some parentheses more Lispy. * mule/european.el: Content moved to latin.el, file deleted. * mule/general-late.el: If Unicode tables are to be loaded at dump time, do it here, not in loadup.el. * mule/greek.el: Add syntax, case support for Greek. * mule/latin.el: Move the content of european.el here. Change the case table mappings to use hexadecimal codes, to make cross reference to the standards easier. In all cases, take character syntax from similar characters in Latin-1 , rather than deciding separately what syntax they should take. Add (incomplete) support for case with Turkish. Remove description of the character sets used from the language environments' doc strings, since now that we create variant language environments on the fly, such descriptions will often be inaccurate. Set the native-coding-system language info property while setting the other coding-system properties of the language. * mule/misc-lang.el (ipa): Remove the language environment. The International Phonetic _Alphabet_ is not a language, it's inane to have a corresponding language environment in XEmacs. * mule/mule-cmds.el (create-variant-language-environment): Also modify the coding-priority when creating a new language environment; document that. * mule/mule-cmds.el (get-language-environment-from-locale): Recognise that the 'native-coding-system language-info property can be a list, interpret it correctly when it is one. 2006-12-21 Aidan Kehoe <kehoea@parhasard.net> * coding.el (coding-system-category): Use the new 'unicode-type property for finding what sort of Unicode coding system subtype a coding system is, instead of the overshadowed 'type property. * dumped-lisp.el (preloaded-file-list): mule/european.el has been removed. * loadup.el (really-early-error-handler): Unicode tables loaded at dump time are now in mule/general-late.el. * simple.el (count-lines): Add some backslashes to to parentheses in docstrings to help fontification along. * simple.el (what-cursor-position): Wrap a line to fit in 80 characters. * unicode.el: Use the 'unicode-type property, not 'type, for setting the Unicode coding-system subtype. src/ChangeLog addition: 2006-12-21 Aidan Kehoe <kehoea@parhasard.net> * file-coding.c: Update the make-coding-system docstring to reflect unicode-type * general-slots.h: New symbol, unicode-type, since 'type was being overridden when accessing a coding system's Unicode subtype. * intl-win32.c: Backslash a few parentheses, to help fontification along. * intl-win32.c (complex_vars_of_intl_win32): Use the 'unicode-type symbol, not 'type, when creating the Microsoft Unicode coding system. * unicode.c (unicode_putprop): * unicode.c (unicode_getprop): * unicode.c (unicode_print): Using 'type as the property name when working out what Unicode subtype a given coding system is was broken, since there's a general coding system property called 'type. Change the former to use 'unicode-type instead.

--- a/lisp/mule/cyrillic.el	Thu Dec 28 22:50:55 2006 +0000
+++ b/lisp/mule/cyrillic.el	Fri Dec 29 18:09:51 2006 +0000
@@ -29,16 +29,19 @@
 ;; The character set ISO8859-5 is supported.  KOI-8 and ALTERNATIVNYJ are
 ;; converted to ISO8859-5 internally.
 
-;; Windows-1251 support deleted because XEmacs has automatic support.
+;; [Windows-1251 support deleted because XEmacs has automatic support.]
+
+;; #### We only have automatic support on Windows; that needs to be put
+;; back. Also, the Russian Wikipedia articles on KOI-8 list several other
+;; related encodings--KOI8-U (Ukrainian), KOI8-RU (simultaneous support for
+;; Russian, Belorussian, and Ukrainian), KOI8-C (for languages of the
+;; Caucasus), KOI8-O (Old Church Slavonic)--and it would be nice to have
+;; them. Beyond that, we're currently trashing lots of code points with
+;; KOI-8 R; it would be nice to leverage the Unicode support to not do that. 
 
 ;;; Code:
 
-;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
-;;; CYRILLIC
-;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
-
-;; ISO-8859-5
-
+;; Case table:
 (loop
   for (upper lower)
   in '((#xcf #xef) ; YA
@@ -94,14 +97,22 @@
 		       case-table))
 
 ;; The default character syntax is now word. Pay attention to the
-;; exceptions in ISO-8859-5. 
-(dolist (code '(#xAD	;; SOFT HYPHEN
-		#xF0	;; NUMERO SIGN
-		#xFD))  ;; SECTION SIGN
-  (modify-syntax-entry (make-char 'cyrillic-iso8859-5 code) "."))
-
-;; NO-BREAK SPACE
-(modify-syntax-entry (make-char 'cyrillic-iso8859-5 #xA0) " ")
+;; exceptions in ISO-8859-5, copying them from ISO-8859-1. 
+(loop
+  for (latin-1 cyrillic) 
+  in '((#xAD #xAD)  ;; SOFT HYPHEN
+       (#xA7 #xFD)  ;; SECTION SIGN
+       (#xA0 #xA0)) ;; NO BREAK SPACE
+  with syntax-table = (standard-syntax-table)
+  do (modify-syntax-entry
+      (make-char 'cyrillic-iso8859-5 cyrillic)
+      (string (char-syntax (make-char 'latin-iso8859-1 latin-1)))
+      syntax-table))
+  
+;; Take NUMERO SIGN's syntax from #. 
+(modify-syntax-entry (make-char 'cyrillic-iso8859-5 #xF0)
+                     (string (char-syntax ?\# (standard-syntax-table)))
+                     (standard-syntax-table))
 
 (make-coding-system
  'iso-8859-5 'iso2022
@@ -110,8 +121,7 @@
    charset-g1 cyrillic-iso8859-5
    charset-g2 t
    charset-g3 t
-   mnemonic "ISO8/Cyr"
-   ))
+   mnemonic "ISO8/Cyr"))
 
 (set-language-info-alist
  "Cyrillic-ISO" '((charset cyrillic-iso8859-5)
@@ -155,12 +165,10 @@
       (let* ((ch (aref cyrillic-koi8-r-decode-table i))
 	     (split (split-char ch)))
 	(cond ((eq (car split) 'cyrillic-iso8859-5)
-	       (aset table (logior (nth 1 split) 128) i)
-	       )
+	       (aset table (logior (nth 1 split) 128) i))
 	      ((eq ch 32))
 	      ((eq (car split) 'ascii)
-	       (aset table ch i)
-	       )))
+	       (aset table ch i))))
       (setq i (1+ i)))
     table)
   "Cyrillic KOI8-R encoding table.")