NOTE: This patch has been committed. APPROVE COMMIT 21.5 RECOMMEND 21.4 This has been in my queue for far too long. man/ChangeLog addition: 2005-03-26 Aidan Kehoe <kehoea(a)parhasard.net&gt; * lispref/mule.texi (CCL Example): char-int -> char-to-int, and hex 41 is decimal 64, both problems with my previous patch pointed out by Stephen. * lispref/mule.texi (The actual coding system): Give information on the make-coding-system call, and where the actual package can be found. XEmacs Trunk source patch: Diff command: cvs -q diff -u Files affected: man/lispref/mule.texi Index: man/lispref/mule.texi =================================================================== RCS file: /pack/xemacscvs/XEmacs/xemacs/man/lispref/mule.texi,v retrieving revision 1.12 diff -u -u -r1.12 mule.texi --- man/lispref/mule.texi 2005/03/09 05:36:50 1.12 +++ man/lispref/mule.texi 2005/03/26 16:16:17 ＠＠ -2097,7 +2097,7 ＠＠ * Characters to be preserved:: No transformation needed for these characters. * The program to decode to internal format:: . * The program to encode from internal format:: . - +* The actual coding system:: . ＠end menu ＠node Four bits to ASCII, URI Encoding constants, , CCL Example ＠＠ -2117,7 +2117,7 ＠＠ (let ((val (make-vector 256 0)) (i 0)) (while (< i (length val)) - (aset val i (char-int (aref (format "%02X" i) 0))) + (aset val i (char-to-int (aref (format "%02X" i) 0))) (setq i (1+ i))) val) "Table to find an ASCII version of an octet's most significant 4 bits.") ＠＠ -2125,7 +2125,7 ＠＠ The next table, ＠code{url-coding-low-order-nybble-as-ascii} is almost the same thing, but this time it has a map for the hex encoding of the -low-order four bits. So the sixty-fifth entry (offset ＠samp{#x51}) is +low-order four bits. So the sixty-fifth entry (offset ＠samp{#x41}) is the ASCII encoding of `1', the hundred-and-twenty-second (offset ＠samp{#x7a}) is the ASCII encoding of `A'. ＠＠ -2134,7 +2134,7 ＠＠ (let ((val (make-vector 256 0)) (i 0)) (while (< i (length val)) - (aset val i (char-int (aref (format "%02X" i) 1))) + (aset val i (char-to-int (aref (format "%02X" i) 1))) (setq i (1+ i))) val) "Table to find an ASCII version of an octet's least significant 4 bits.") ＠＠ -2154,14 +2154,14 ＠＠ ＠code{url-coding-escaped-space-code} variable. ＠example -(defvar url-coding-escape-character-code (char-int ?%) +(defvar url-coding-escape-character-code (char-to-int ?%) "The code point for the percentage sign, in ASCII.") -(defvar url-coding-escaped-space-code (char-int ?+) +(defvar url-coding-escaped-space-code (char-to-int ?+) "The URL-encoded value of the space character, that is, +.") ＠end example -＠node Numeric to ASCII-hexadecimal conversion +＠node Numeric to ASCII-hexadecimal conversion, Characters to be preserved, URI Encoding constants, CCL Example ＠subsubsection Numeric to ASCII-hexadecimal conversion Now, we have a couple of utility tables that wouldn't be necessary in ＠＠ -2177,7 +2177,7 ＠＠ (let ((i 0) (val (make-vector 16 0))) (while (< i 16) - (aset val i (char-int (aref (format "%X" i) 0))) + (aset val i (char-to-int (aref (format "%X" i) 0))) (setq i (1+ i))) val) "A map from a hexadecimal digit's numeric value to its encoding in ASCII.") ＠＠ -2193,7 +2193,7 ＠＠ "A map from Latin 1 code points to their values as hexadecimal digits.") ＠end example -＠node Characters to be preserved +＠node Characters to be preserved, The program to decode to internal format, Numeric to ASCII-hexadecimal conversion, CCL Example ＠subsubsection Characters to be preserved And finally, the last of these tables. URL encoding says that ＠＠ -2227,7 +2227,7 ＠＠ octet as its ASCII encoding.") ＠end example -＠node The program to decode to internal format +＠node The program to decode to internal format, The program to encode from internal format, Characters to be preserved, CCL Example ＠subsubsection The program to decode to internal format After the almost interminable tables, we get to the CCL. The first ＠＠ -2288,7 +2288,7 ＠＠ internal encoding. ") ＠end example -＠node The program to encode from internal format +＠node The program to encode from internal format, The actual coding system, The program to decode to internal format, CCL Example ＠subsubsection The program to encode from internal format Next, we see the CCL program to encode ASCII text as URL coded text. ＠＠ -2323,6 +2323,36 ＠＠ (repeat)))) "CCL program to encode octets (almost) according to RFC 1738") ＠end example + +＠node The actual coding system, , The program to encode from internal format, CCL Example +＠subsubsection The actual coding system + +To actually create the coding system, we call +＠samp{make-coding-system}. The first argument is the symbol that is to +be the name of the coding system, in our case ＠samp{url-coding}. The +second specifies that the coding system is to be of type +＠samp{ccl}---there are several other coding system types available, +including, see the documentation for ＠samp{make-coding-system} for the +full list. Then there's a documentation string describing the wherefore +and caveats of the coding system, and the final argument is a property +list giving information about the CCL programs and the coding system's +mnemonic. + +＠example +(make-coding-system + 'url-coding 'ccl + "The coding used by application/x-www-form-urlencoded HTTP applications. +This coding form doesn't specify anything about non-ASCII characters, so +make sure you've transformed to a seven-bit coding system first." + '(decode ccl-decode-urlcoding + encode ccl-encode-urlcoding + mnemonic "URLenc")) +＠end example + +If you're lucky, the ＠samp{url-coding} coding system describe here +should be available in the XEmacs package system. Otherwise, downloading +it from ＠samp{http://www.parhasard.net/url-coding.el} should work for +the foreseeable future. ＠node Category Tables, Unicode Support, CCL, MULE ＠section Category Tables -- “I, for instance, am gung-ho about open source because my family is being held hostage in Rob Malda’s basement. But who fact-checks me, or Enderle, when we say something in public? No-one!” -- Danny O’Brien