+(defvar non-iso-charset-alist
+ `((mac-roman
+ (ascii latin-iso8859-1 mule-unicode-2500-33ff
+ mule-unicode-0100-24ff mule-unicode-e000-ffff)
+ mac-roman-decoder
+ ((0 255)))
+ (viscii
+ (ascii vietnamese-viscii-lower vietnamese-viscii-upper)
+ viet-viscii-nonascii-translation-table
+ ((0 255)))
+ (koi8-r
+ (ascii cyrillic-iso8859-5)
+ cyrillic-koi8-r-nonascii-translation-table
+ ((32 255)))
+ (alternativnyj
+ (ascii cyrillic-iso8859-5)
+ cyrillic-alternativnyj-nonascii-translation-table
+ ((32 255)))
+ (koi8-u
+ (ascii cyrillic-iso8859-5 mule-unicode-0100-24ff)
+ cyrillic-koi8-u-nonascii-translation-table
+ ((32 255)))
+ (big5
+ (ascii chinese-big5-1 chinese-big5-2)
+ decode-big5-char
+ ((32 127)
+ ((?\xA1 ?\xFE) . (?\x40 ?\x7E ?\xA1 ?\xFE))))
+ (sjis
+ (ascii katakana-jisx0201 japanese-jisx0208)
+ decode-sjis-char
+ ((32 127 ?\xA1 ?\xDF)
+ ((?\x81 ?\x9F ?\xE0 ?\xEF) . (?\x40 ?\x7E ?\x80 ?\xFC)))))
+ "Alist of charset names vs the corresponding information.
+This is mis-named for historical reasons. The charsets are actually
+non-built-in ones. They correspond to Emacs coding systems, not Emacs
+charsets, i.e. what Emacs can read (or write) by mapping to (or
+from) Emacs internal charsets that typically correspond to a limited
+set of ISO charsets.
+
+Each element has the following format:
+ (CHARSET CHARSET-LIST TRANSLATION-METHOD [ CODE-RANGE ])
+
+CHARSET is the name (symbol) of the charset.
+
+CHARSET-LIST is a list of Emacs charsets into which characters of
+CHARSET are mapped.
+
+TRANSLATION-METHOD is a translation table (symbol) to translate a
+character code of CHARSET to the corresponding Emacs character
+code. It can also be a function to call with one argument, a
+character code in CHARSET.
+
+CODE-RANGE specifies the valid code ranges of CHARSET.
+It is a list of RANGEs, where each RANGE is of the form:
+ (FROM1 TO1 FROM2 TO2 ...)
+or
+ ((FROM1-1 TO1-1 FROM1-2 TO1-2 ...) . (FROM2-1 TO2-1 FROM2-2 TO2-2 ...))
+In the first form, valid codes are between FROM1 and TO1, or FROM2 and
+TO2, or...
+The second form is used for 2-byte codes. The car part is the ranges
+of the first byte, and the cdr part is the ranges of the second byte.")