annotate admin/charsets/mapconv @ 102120:b39c9d4ea068

(compilation-error-regexp-alist-alist): Move entry for maven (alphabetical order).
author Nick Roberts <nickrob@snap.net.nz>
date Thu, 19 Feb 2009 06:54:22 +0000
parents ce88a631c161
children 24af3ac0cac1
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
88123
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
1 #!/bin/sh
94832
eb2d9dfc8486 Switch to recommended form of GPLv3 permissions notice.
Glenn Morris <rgm@gnu.org>
parents: 91418
diff changeset
2
100971
ce88a631c161 Add 2009 to copyright years.
Glenn Morris <rgm@gnu.org>
parents: 94832
diff changeset
3 # Copyright (C) 2003, 2004, 2005, 2006, 2007, 2008, 2009
88123
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
4 # National Institute of Advanced Industrial Science and Technology (AIST)
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
5 # Registration Number H13PRO009
94832
eb2d9dfc8486 Switch to recommended form of GPLv3 permissions notice.
Glenn Morris <rgm@gnu.org>
parents: 91418
diff changeset
6
88123
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
7 # This file is part of GNU Emacs.
94832
eb2d9dfc8486 Switch to recommended form of GPLv3 permissions notice.
Glenn Morris <rgm@gnu.org>
parents: 91418
diff changeset
8
eb2d9dfc8486 Switch to recommended form of GPLv3 permissions notice.
Glenn Morris <rgm@gnu.org>
parents: 91418
diff changeset
9 # GNU Emacs is free software: you can redistribute it and/or modify
88123
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
10 # it under the terms of the GNU General Public License as published by
94832
eb2d9dfc8486 Switch to recommended form of GPLv3 permissions notice.
Glenn Morris <rgm@gnu.org>
parents: 91418
diff changeset
11 # the Free Software Foundation, either version 3 of the License, or
eb2d9dfc8486 Switch to recommended form of GPLv3 permissions notice.
Glenn Morris <rgm@gnu.org>
parents: 91418
diff changeset
12 # (at your option) any later version.
eb2d9dfc8486 Switch to recommended form of GPLv3 permissions notice.
Glenn Morris <rgm@gnu.org>
parents: 91418
diff changeset
13
88123
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
14 # GNU Emacs is distributed in the hope that it will be useful,
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
15 # but WITHOUT ANY WARRANTY; without even the implied warranty of
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
16 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
17 # GNU General Public License for more details.
94832
eb2d9dfc8486 Switch to recommended form of GPLv3 permissions notice.
Glenn Morris <rgm@gnu.org>
parents: 91418
diff changeset
18
88123
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
19 # You should have received a copy of the GNU General Public License
94832
eb2d9dfc8486 Switch to recommended form of GPLv3 permissions notice.
Glenn Morris <rgm@gnu.org>
parents: 91418
diff changeset
20 # along with GNU Emacs. If not, see <http://www.gnu.org/licenses/>.
88123
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
21
94832
eb2d9dfc8486 Switch to recommended form of GPLv3 permissions notice.
Glenn Morris <rgm@gnu.org>
parents: 91418
diff changeset
22 # Commentary:
eb2d9dfc8486 Switch to recommended form of GPLv3 permissions notice.
Glenn Morris <rgm@gnu.org>
parents: 91418
diff changeset
23
88123
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
24 # Convert charset map of various format into this:
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
25 # 0xXX 0xYYYY
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
26 # where,
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
27 # XX is a code point of the charset in hexa-decimal,
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
28 # YYYY is the corresponding Unicode character code in hexa-decimal.
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
29 # Arguments are:
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
30 # $1: source map file
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
31 # $2: address pattern for sed (optionally with substitution command)
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
32 # $3: format of source map file
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
33 # GLIBC-1 GLIBC-2 GLIBC-2-7 CZYBORRA IANA UNICODE YASUOKA MICROSOFT
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
34 # $4: awk script
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
35
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
36 BASE=`basename $1`
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
37
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
38 case "$3" in
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
39 GLIBC*)
89823
4dbae0eca96d Fix previous change.
Kenichi Handa <handa@m17n.org>
parents: 89821
diff changeset
40 SOURCE="glibc-2.3.2/localedata/charmaps/${BASE}";;
88123
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
41 CZYBORRA)
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
42 SOURCE="http://czyborra.com/charsets/${BASE}";;
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
43 IANA)
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
44 SOURCE="http://www.iana.org/assignments/charset-reg/${BASE}";;
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
45 UNICODE)
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
46 SOURCE="http://www.unicode.org/Public/MAPPINGS/.../${BASE}";;
89742
a750cf193ec6 Handle UNICODE2.
Kenichi Handa <handa@m17n.org>
parents: 88123
diff changeset
47 UNICODE2)
a750cf193ec6 Handle UNICODE2.
Kenichi Handa <handa@m17n.org>
parents: 88123
diff changeset
48 SOURCE="http://www.unicode.org/Public/MAPPINGS/.../${BASE}";;
88123
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
49 YASUOKA)
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
50 SOURCE="http://kanji.zinbun.kyoto-u.ac.jp/~yasuoka/.../${BASE}";;
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
51 MICROSOFT)
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
52 SOURCE="http://www.microsoft.com/globaldev/reference/oem/${BASE}";;
89815
59bbae51ad73 Add code for handling KANJI-DATABASE format.
Kenichi Handa <handa@m17n.org>
parents: 89746
diff changeset
53 KANJI-DATABASE)
59bbae51ad73 Add code for handling KANJI-DATABASE format.
Kenichi Handa <handa@m17n.org>
parents: 89746
diff changeset
54 SOURCE="data at http://sourceforge.net/cvs/?group_id=26261";;
88123
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
55 *)
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
56 echo "Unknown file type: $3";
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
57 exit 1;;
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
58 esac
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
59
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
60 echo "# Generated from $SOURCE"
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
61
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
62 if [ -n "$4" ] ; then
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
63 if [ -f "$4" ] ; then
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
64 AWKPROG="gawk -f $4"
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
65 else
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
66 echo "Awk program does not exist: $4"
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
67 exit 1
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
68 fi
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
69 else
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
70 AWKPROG=cat
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
71 fi
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
72
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
73 if [ "$3" == "GLIBC-1" ] ; then
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
74 # Source format is:
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
75 # <UYYYY> /xXX
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
76 sed -n -e "$2 p" < $1 \
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
77 | sed -e 's,<U\([^>]*\)>[ ]*/x\(..\).*,0x\2 0x\1,' \
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
78 | sort | ${AWKPROG}
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
79 elif [ "$3" == "GLIBC-2" ] ; then
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
80 # Source format is:
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
81 # <UYYYY> /xXX/xZZ
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
82 sed -n -e "$2 p" < $1 \
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
83 | sed -e 's,<U\([^>]*\)>[ ]*/x\(..\)/x\(..\).*,0x\2\3 0x\1,' \
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
84 | sort | ${AWKPROG}
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
85 elif [ "$3" == "GLIBC-2-7" ] ; then
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
86 # Source format is:
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
87 # <UYYYY> /xXX/xZZ
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
88 # We must drop MSBs of XX and ZZ
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
89 sed -n -e "$2 p" < $1 \
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
90 | sed -e 's/xa/x2/g' -e 's/xb/x3/g' -e 's/xc/x4/g' \
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
91 -e 's/xd/x5/g' -e 's/xe/x6/g' -e 's/xf/x7/g' \
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
92 -e 's,<U\([^>]*\)>[ ]*/x\(..\)/x\(..\).*,0x\2\3 0x\1,' \
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
93 | tee temp \
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
94 | sort | ${AWKPROG}
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
95 elif [ "$3" == "CZYBORRA" ] ; then
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
96 # Source format is:
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
97 # =XX U+YYYY
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
98 zcat $1 | sed -n -e "$2 p" \
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
99 | sed -e 's/=\(..\)[^U]*U+\([0-9A-F]*\).*/0x\1 0x\2/' \
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
100 | sort | ${AWKPROG}
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
101 elif [ "$3" == "IANA" ] ; then
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
102 # Source format is:
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
103 # 0xXX 0xYYYY
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
104 sed -n -e "$2 p" < $1 \
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
105 | sed -e 's/\(0x[0-9A-Fa-f]*\)[^0]*\(0x[0-9A-Fa-f]*\).*/\1 \2/' \
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
106 | sort | ${AWKPROG}
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
107 elif [ "$3" == "UNICODE" ] ; then
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
108 # Source format is:
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
109 # YYYY XX
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
110 sed -n -e "$2 p" < $1 \
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
111 | sed -e 's/\([0-9A-F]*\)[^0-9A-F]*\([0-9A-F]*\).*/0x\2 0x\1/' \
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
112 | sort | ${AWKPROG}
89742
a750cf193ec6 Handle UNICODE2.
Kenichi Handa <handa@m17n.org>
parents: 88123
diff changeset
113 elif [ "$3" == "UNICODE2" ] ; then
a750cf193ec6 Handle UNICODE2.
Kenichi Handa <handa@m17n.org>
parents: 88123
diff changeset
114 # Source format is:
a750cf193ec6 Handle UNICODE2.
Kenichi Handa <handa@m17n.org>
parents: 88123
diff changeset
115 # 0xXXXX 0xYYYY # ...
a750cf193ec6 Handle UNICODE2.
Kenichi Handa <handa@m17n.org>
parents: 88123
diff changeset
116 sed -n -e "$2 p" < $1 \
a750cf193ec6 Handle UNICODE2.
Kenichi Handa <handa@m17n.org>
parents: 88123
diff changeset
117 | sed -e 's/\([0-9A-Fx]*\)[^0]*\([0-9A-Fx]*\).*/\1 \2/' \
89746
b7afe7c870d3 In UNICODE2 case, sort by 4th field after
Kenichi Handa <handa@m17n.org>
parents: 89742
diff changeset
118 | ${AWKPROG} | sort -n -k 4,4
88123
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
119 elif [ "$3" == "YASUOKA" ] ; then
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
120 # Source format is:
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
121 # YYYY 0-XXXX (XXXX is a Kuten code)
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
122 sed -n -e "$2 p" < $1 \
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
123 | sed -e 's/\([0-9A-F]*\)[^0]*0-\([0-9]*\).*/0x\2 0x\1/' \
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
124 | sort | ${AWKPROG}
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
125 elif [ "$3" == "MICROSOFT" ] ; then
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
126 # Source format is:
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
127 # XX = U+YYYY
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
128 sed -n -e "$2 p" < $1 \
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
129 | sed -e 's/\([0-9A-F]*\).*U+\([0-9A-F]*\).*/0x\1 0x\2/' \
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
130 | sort | ${AWKPROG}
89815
59bbae51ad73 Add code for handling KANJI-DATABASE format.
Kenichi Handa <handa@m17n.org>
parents: 89746
diff changeset
131 elif [ "$3" == "KANJI-DATABASE" ] ; then
59bbae51ad73 Add code for handling KANJI-DATABASE format.
Kenichi Handa <handa@m17n.org>
parents: 89746
diff changeset
132 # Source format is:
59bbae51ad73 Add code for handling KANJI-DATABASE format.
Kenichi Handa <handa@m17n.org>
parents: 89746
diff changeset
133 # C?-XXXX U+YYYYY .....
59bbae51ad73 Add code for handling KANJI-DATABASE format.
Kenichi Handa <handa@m17n.org>
parents: 89746
diff changeset
134 sed -n -e "$2 p" < $1 \
59bbae51ad73 Add code for handling KANJI-DATABASE format.
Kenichi Handa <handa@m17n.org>
parents: 89746
diff changeset
135 | sed -e 's/...\(....\) U+\([0-9A-F]*\).*/0x\1 0x\2/' \
59bbae51ad73 Add code for handling KANJI-DATABASE format.
Kenichi Handa <handa@m17n.org>
parents: 89746
diff changeset
136 | sort | ${AWKPROG}
88123
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
137 else
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
138 echo "Invalid arguments"
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
139 exit 1
375f2633d815 New directory
Kenichi Handa <handa@m17n.org>
parents:
diff changeset
140 fi
89916
e0e4e6a0599f Changes from arch/CVS synchronization
Miles Bader <miles@gnu.org>
parents: 89823
diff changeset
141
e0e4e6a0599f Changes from arch/CVS synchronization
Miles Bader <miles@gnu.org>
parents: 89823
diff changeset
142 # arch-tag: c33acb47-7eb6-4872-b871-15e1447e8f0e