This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
Don't generate EBCDIC POSIX-BC tables
[perl5.git] / regen / charset_translations.pl
CommitLineData
6ff677df
KW
1#!/usr/bin/perl -w
2use strict;
3use warnings;
4
c0236afe
KW
5# WARNING: This must be kept in sync with the UTF8_MAXBYTES value in
6# utfebcdic.h
7$CHARSET_TRANSLATIONS::UTF_EBCDIC_MAXBYTES = 14;
8
6ff677df
KW
9# Utilities for various character set issues. Currently handles ASCII and
10# EBCDIC only. It is trivial to add support for new EBCDIC code pages (unless
11# they have identical variant character signatures as existing ones, and there
12# aren't other glitches that arise): just add a mapping table to
13# %ebcdic_translations and regen everything that uses this.
14
15my %ebcdic_translations = (
16 # Keys are code page name; values are arrays that map ASCII ordinals to
17 # the code page's ordinals
18
19 'EBCDIC 1047' =>
20 [ 0x00, 0x01, 0x02, 0x03, 0x37, 0x2D, 0x2E, 0x2F, 0x16, 0x05, 0x15, 0x0B, 0x0C, 0x0D, 0x0E, 0x0F,
21 0x10, 0x11, 0x12, 0x13, 0x3C, 0x3D, 0x32, 0x26, 0x18, 0x19, 0x3F, 0x27, 0x1C, 0x1D, 0x1E, 0x1F,
22 0x40, 0x5A, 0x7F, 0x7B, 0x5B, 0x6C, 0x50, 0x7D, 0x4D, 0x5D, 0x5C, 0x4E, 0x6B, 0x60, 0x4B, 0x61,
23 0xF0, 0xF1, 0xF2, 0xF3, 0xF4, 0xF5, 0xF6, 0xF7, 0xF8, 0xF9, 0x7A, 0x5E, 0x4C, 0x7E, 0x6E, 0x6F,
24 0x7C, 0xC1, 0xC2, 0xC3, 0xC4, 0xC5, 0xC6, 0xC7, 0xC8, 0xC9, 0xD1, 0xD2, 0xD3, 0xD4, 0xD5, 0xD6,
25 0xD7, 0xD8, 0xD9, 0xE2, 0xE3, 0xE4, 0xE5, 0xE6, 0xE7, 0xE8, 0xE9, 0xAD, 0xE0, 0xBD, 0x5F, 0x6D,
26 0x79, 0x81, 0x82, 0x83, 0x84, 0x85, 0x86, 0x87, 0x88, 0x89, 0x91, 0x92, 0x93, 0x94, 0x95, 0x96,
27 0x97, 0x98, 0x99, 0xA2, 0xA3, 0xA4, 0xA5, 0xA6, 0xA7, 0xA8, 0xA9, 0xC0, 0x4F, 0xD0, 0xA1, 0x07,
28 0x20, 0x21, 0x22, 0x23, 0x24, 0x25, 0x06, 0x17, 0x28, 0x29, 0x2A, 0x2B, 0x2C, 0x09, 0x0A, 0x1B,
29 0x30, 0x31, 0x1A, 0x33, 0x34, 0x35, 0x36, 0x08, 0x38, 0x39, 0x3A, 0x3B, 0x04, 0x14, 0x3E, 0xFF,
30 0x41, 0xAA, 0x4A, 0xB1, 0x9F, 0xB2, 0x6A, 0xB5, 0xBB, 0xB4, 0x9A, 0x8A, 0xB0, 0xCA, 0xAF, 0xBC,
31 0x90, 0x8F, 0xEA, 0xFA, 0xBE, 0xA0, 0xB6, 0xB3, 0x9D, 0xDA, 0x9B, 0x8B, 0xB7, 0xB8, 0xB9, 0xAB,
32 0x64, 0x65, 0x62, 0x66, 0x63, 0x67, 0x9E, 0x68, 0x74, 0x71, 0x72, 0x73, 0x78, 0x75, 0x76, 0x77,
33 0xAC, 0x69, 0xED, 0xEE, 0xEB, 0xEF, 0xEC, 0xBF, 0x80, 0xFD, 0xFE, 0xFB, 0xFC, 0xBA, 0xAE, 0x59,
34 0x44, 0x45, 0x42, 0x46, 0x43, 0x47, 0x9C, 0x48, 0x54, 0x51, 0x52, 0x53, 0x58, 0x55, 0x56, 0x57,
35 0x8C, 0x49, 0xCD, 0xCE, 0xCB, 0xCF, 0xCC, 0xE1, 0x70, 0xDD, 0xDE, 0xDB, 0xDC, 0x8D, 0x8E, 0xDF
36 ],
37
01ffcbd4
KW
38# 'EBCDIC POSIX-BC' =>
39# [
40# 0x00, 0x01, 0x02, 0x03, 0x37, 0x2D, 0x2E, 0x2F, 0x16, 0x05, 0x15, 0x0B, 0x0C, 0x0D, 0x0E, 0x0F,
41# 0x10, 0x11, 0x12, 0x13, 0x3C, 0x3D, 0x32, 0x26, 0x18, 0x19, 0x3F, 0x27, 0x1C, 0x1D, 0x1E, 0x1F,
42# 0x40, 0x5A, 0x7F, 0x7B, 0x5B, 0x6C, 0x50, 0x7D, 0x4D, 0x5D, 0x5C, 0x4E, 0x6B, 0x60, 0x4B, 0x61,
43# 0xF0, 0xF1, 0xF2, 0xF3, 0xF4, 0xF5, 0xF6, 0xF7, 0xF8, 0xF9, 0x7A, 0x5E, 0x4C, 0x7E, 0x6E, 0x6F,
44# 0x7C, 0xC1, 0xC2, 0xC3, 0xC4, 0xC5, 0xC6, 0xC7, 0xC8, 0xC9, 0xD1, 0xD2, 0xD3, 0xD4, 0xD5, 0xD6,
45# 0xD7, 0xD8, 0xD9, 0xE2, 0xE3, 0xE4, 0xE5, 0xE6, 0xE7, 0xE8, 0xE9, 0xBB, 0xBC, 0xBD, 0x6A, 0x6D,
46# 0x4A, 0x81, 0x82, 0x83, 0x84, 0x85, 0x86, 0x87, 0x88, 0x89, 0x91, 0x92, 0x93, 0x94, 0x95, 0x96,
47# 0x97, 0x98, 0x99, 0xA2, 0xA3, 0xA4, 0xA5, 0xA6, 0xA7, 0xA8, 0xA9, 0xFB, 0x4F, 0xFD, 0xFF, 0x07,
48# 0x20, 0x21, 0x22, 0x23, 0x24, 0x25, 0x06, 0x17, 0x28, 0x29, 0x2A, 0x2B, 0x2C, 0x09, 0x0A, 0x1B,
49# 0x30, 0x31, 0x1A, 0x33, 0x34, 0x35, 0x36, 0x08, 0x38, 0x39, 0x3A, 0x3B, 0x04, 0x14, 0x3E, 0x5F,
50# 0x41, 0xAA, 0xB0, 0xB1, 0x9F, 0xB2, 0xD0, 0xB5, 0x79, 0xB4, 0x9A, 0x8A, 0xBA, 0xCA, 0xAF, 0xA1,
51# 0x90, 0x8F, 0xEA, 0xFA, 0xBE, 0xA0, 0xB6, 0xB3, 0x9D, 0xDA, 0x9B, 0x8B, 0xB7, 0xB8, 0xB9, 0xAB,
52# 0x64, 0x65, 0x62, 0x66, 0x63, 0x67, 0x9E, 0x68, 0x74, 0x71, 0x72, 0x73, 0x78, 0x75, 0x76, 0x77,
53# 0xAC, 0x69, 0xED, 0xEE, 0xEB, 0xEF, 0xEC, 0xBF, 0x80, 0xE0, 0xFE, 0xDD, 0xFC, 0xAD, 0xAE, 0x59,
54# 0x44, 0x45, 0x42, 0x46, 0x43, 0x47, 0x9C, 0x48, 0x54, 0x51, 0x52, 0x53, 0x58, 0x55, 0x56, 0x57,
55# 0x8C, 0x49, 0xCD, 0xCE, 0xCB, 0xCF, 0xCC, 0xE1, 0x70, 0xC0, 0xDE, 0xDB, 0xDC, 0x8D, 0x8E, 0xDF
56# ],
6ff677df
KW
57
58 'EBCDIC 037' =>
59 [
60 0x00, 0x01, 0x02, 0x03, 0x37, 0x2D, 0x2E, 0x2F, 0x16, 0x05, 0x25, 0x0B, 0x0C, 0x0D, 0x0E, 0x0F,
61 0x10, 0x11, 0x12, 0x13, 0x3C, 0x3D, 0x32, 0x26, 0x18, 0x19, 0x3F, 0x27, 0x1C, 0x1D, 0x1E, 0x1F,
62 0x40, 0x5A, 0x7F, 0x7B, 0x5B, 0x6C, 0x50, 0x7D, 0x4D, 0x5D, 0x5C, 0x4E, 0x6B, 0x60, 0x4B, 0x61,
63 0xF0, 0xF1, 0xF2, 0xF3, 0xF4, 0xF5, 0xF6, 0xF7, 0xF8, 0xF9, 0x7A, 0x5E, 0x4C, 0x7E, 0x6E, 0x6F,
64 0x7C, 0xC1, 0xC2, 0xC3, 0xC4, 0xC5, 0xC6, 0xC7, 0xC8, 0xC9, 0xD1, 0xD2, 0xD3, 0xD4, 0xD5, 0xD6,
65 0xD7, 0xD8, 0xD9, 0xE2, 0xE3, 0xE4, 0xE5, 0xE6, 0xE7, 0xE8, 0xE9, 0xBA, 0xE0, 0xBB, 0xB0, 0x6D,
66 0x79, 0x81, 0x82, 0x83, 0x84, 0x85, 0x86, 0x87, 0x88, 0x89, 0x91, 0x92, 0x93, 0x94, 0x95, 0x96,
67 0x97, 0x98, 0x99, 0xA2, 0xA3, 0xA4, 0xA5, 0xA6, 0xA7, 0xA8, 0xA9, 0xC0, 0x4F, 0xD0, 0xA1, 0x07,
68 0x20, 0x21, 0x22, 0x23, 0x24, 0x15, 0x06, 0x17, 0x28, 0x29, 0x2A, 0x2B, 0x2C, 0x09, 0x0A, 0x1B,
69 0x30, 0x31, 0x1A, 0x33, 0x34, 0x35, 0x36, 0x08, 0x38, 0x39, 0x3A, 0x3B, 0x04, 0x14, 0x3E, 0xFF,
70 0x41, 0xAA, 0x4A, 0xB1, 0x9F, 0xB2, 0x6A, 0xB5, 0xBD, 0xB4, 0x9A, 0x8A, 0x5F, 0xCA, 0xAF, 0xBC,
71 0x90, 0x8F, 0xEA, 0xFA, 0xBE, 0xA0, 0xB6, 0xB3, 0x9D, 0xDA, 0x9B, 0x8B, 0xB7, 0xB8, 0xB9, 0xAB,
72 0x64, 0x65, 0x62, 0x66, 0x63, 0x67, 0x9E, 0x68, 0x74, 0x71, 0x72, 0x73, 0x78, 0x75, 0x76, 0x77,
73 0xAC, 0x69, 0xED, 0xEE, 0xEB, 0xEF, 0xEC, 0xBF, 0x80, 0xFD, 0xFE, 0xFB, 0xFC, 0xAD, 0xAE, 0x59,
74 0x44, 0x45, 0x42, 0x46, 0x43, 0x47, 0x9C, 0x48, 0x54, 0x51, 0x52, 0x53, 0x58, 0x55, 0x56, 0x57,
75 0x8C, 0x49, 0xCD, 0xCE, 0xCB, 0xCF, 0xCC, 0xE1, 0x70, 0xDD, 0xDE, 0xDB, 0xDC, 0x8D, 0x8E, 0xDF
76 ],
77);
78
79my $ascii_key = 'ASCII/Latin1';
80
81my %I8_TO_NATIVE_UTF8; # Maps I8 UTF to final UTF-EBCDIC
82 # See http://www.unicode.org/reports/tr16/
83
84sub get_supported_code_pages() {
85 # Returns an ordered array of the currently supported code pages,
86 # including ASCII as the 0th element, 1047 as the 1th, and the others
87 # sorted lexically by code page name.
88
89 # Create an ASCII table.
90 unless (exists $ebcdic_translations{$ascii_key}) {
91 for my $i (0 .. 255) {
92 $ebcdic_translations{$ascii_key}->[$i] = $i;
93 }
94 }
95
96 return sort {
97 $a eq $ascii_key
98 ? -1
99 : $b eq $ascii_key
100 ? 1
101 : $a =~ /1047/
102 ? -1
103 : $b =~ /1047/
104 ? 1
105 : $a cmp $b
106 } keys %ebcdic_translations;
107}
108
109sub get_a2n($) {
110 # Returns the mapping array for ASCII to code page for the code page named
111 # by the input parameter.
112
113 my $charset = shift;
114
115 if (! exists $ebcdic_translations{$charset}) {
116 die "Unknown character set '$charset'";
117 }
118
c30a0cf2 119 return $ebcdic_translations{$charset};
6ff677df
KW
120}
121
122sub get_I8_2_utf($) {
123 # Returns the mapping array for I8 to code page UTF-EBCDIC for the code
124 # page named by the input parameter. This is Table 2 of TR16 customized
125 # for the code page. See utfebcdic.h for why, contrary to TR16, it has to
126 # be code-page-specific.
127
128 my $charset = shift;
129
130 die "I8 not a valid concept for ASCII" if $charset eq $ascii_key;
131 die "'$charset' unknown" unless exists $ebcdic_translations{$charset};
132
133 # Generate the table if not already present
134 if (! exists $I8_TO_NATIVE_UTF8{$charset}) {
135
136 # The code points not used for invariants. Initialized to everything,
137 # then entries are removed as we go along.
138 my %unused_cps;
139 for my $i (0 .. 255) {
140 $unused_cps{$i} = 1;
141 }
142
143 # These are the invariants. The output has them mapped to the
144 # original EBCDIC code point.
145 for my $i (0 .. 0x9F) {
146 use charnames ();
147 my $ebcdic_value = $ebcdic_translations{$charset}[$i];
148 #printf "$charset: using %02x which is %02x ascii, %s\n", $ebcdic_value, $i, charnames::viacode($i);
149 $I8_TO_NATIVE_UTF8{$charset}[$i] = $ebcdic_value;
150 if (! defined delete $unused_cps{$ebcdic_value}) {
151 die "Two code points map to $ebcdic_value; one is $i";
152 }
153 }
154
155 # Put the unused code points in order
156 my @unused_cps = sort { $a <=> $b } keys %unused_cps;
157
158 # Fill in the rest of the map with these ordered code points, as TR16
159 # specifies
160 for my $i (0xA0 .. 255) {
161 $I8_TO_NATIVE_UTF8{$charset}[$i] = shift @unused_cps;
162 #printf "$charset: filling in %02x which is %02x ascii, %s\n", $I8_TO_NATIVE_UTF8{$charset}[$i], $i, charnames::viacode($i);
163 }
164
165 if (@unused_cps) {
166 die "Left-over code points";
167 }
168 }
169
e0dcdb0a 170 return $I8_TO_NATIVE_UTF8{$charset};
6ff677df
KW
171}
172
173{ # Closure
174
175 my $charset; # We use these to do some error checking that the #if and
176 # #endif are matched.
177 my $indent;
178
179 sub get_conditional_compile_line_start($;$) {
180 # Returns the '#if' line to put into C code to compile for the code
181 # page given by the first parameter. The second parameter, if
182 # present, is the indentation level, like '# if ...'
183
184 if (defined $charset || defined $indent) {
185 die "Missing call to get_conditional_compile_line_end()"
186 }
187
188 $charset = shift;
189 my $indent_level = shift // 0;
190
191 die "This is designed to run only on an ASCII platform" unless ord "A" == 65;
192
193 if ($indent_level == 0) {
194 $indent = "";
195 }
196 else {
197 $indent = " " x (($indent_level * 4) - 1);
198 }
199
200 die "Unknown character set '$charset'" unless exists $ebcdic_translations{$charset};
201
202 my $return = "";
203 {
204 no warnings 'qw';
205 my $count = -1;
206
207 # We use all the typical variant characters to construct the #if,
208 # so that it is unlikely that a different code page will match
209 # this #if
210 for my $char (qw/A \\\ [ ] { } ^ ~ ! # | $ @ `/) {
211 my $compare;
212 my $ascii_ord = ord $char;
213 my $first_time = $return eq "";
214
215 $compare = $ebcdic_translations{$charset}[$ascii_ord];
216 $return .= " && " unless $first_time;
217 $return .= "'$char' == $compare";
218 $return .= " /* $charset */" if $first_time;
219 last if $charset eq $ascii_key;
220 $count++;
221 $return .= " \\\n " if $first_time || $count % 5 == 0;
222 }
223 }
224
225 return "#${indent}if $return\n";
226 }
227
228 sub get_conditional_compile_line_end () {
229 # Returns the #endif for the currently open #if
230
231 my $return = "#${indent}endif\t/* $charset */\n";
232 undef $charset;
233 undef $indent;
234 return $return;
235 }
236}
237
238sub _UTF_START_MASK($) {
239 # Internal
240 my $len = shift;
c0236afe 241 return (($len >= 7) ? 0x00 : (0x1F >> ($len - 2)));
6ff677df
KW
242}
243
244sub _UTF_START_MARK($) {
245 # Internal
c0236afe
KW
246 my $len = shift;
247 return (($len > 7) ? 0xFF : (0xFF & (0xFE << (7- $len))));
6ff677df
KW
248}
249
250sub cp_2_utfbytes($$) {
251 # Returns a string consisting of the UTF-EBCDIC for the code page given by
7799b1c9
KW
252 # the 2nd parameter, of the Unicode code point given by the first
253 # parameter, using the UTF-MOD algorithm published in TR16. (If the "code
254 # page" is ASCII, straight UTF-8 is returned.)
6ff677df
KW
255
256 my ($ucp, $charset) = @_;
257
258 if ($charset eq $ascii_key) {
259 my $str = chr $ucp;
260 utf8::upgrade($str);
261 utf8::encode($str);
262 return $str;
263 }
264 elsif (exists $ebcdic_translations{$charset}) {
265
266 if ($ucp < 0xA0) {
267 return chr $ebcdic_translations{$charset}[$ucp];
268 }
269
e0dcdb0a 270 my $I8_2_utf = get_I8_2_utf($charset);
6ff677df
KW
271
272 my $len = $ucp < 0xA0 ? 1 :
273 $ucp < 0x400 ? 2 :
274 $ucp < 0x4000 ? 3 :
275 $ucp < 0x40000 ? 4 :
276 $ucp < 0x400000 ? 5 :
c0236afe
KW
277 $ucp < 0x4000000 ? 6 :
278 $ucp < 0x40000000? 7 :
279 $CHARSET_TRANSLATIONS::UTF_EBCDIC_MAXBYTES;
6ff677df
KW
280
281 my @str;
282 for (1 .. $len - 1) {
e0dcdb0a 283 unshift @str, chr $I8_2_utf->[($ucp & 0x1f) | 0xA0];
6ff677df
KW
284 $ucp >>= 5;
285 }
286
e0dcdb0a 287 unshift @str, chr $I8_2_utf->[($ucp & _UTF_START_MASK($len)) | _UTF_START_MARK($len)];
6ff677df
KW
288
289 return join "", @str;
290 }
291 else {
292 die "Unknown character set '$charset'";
293 }
294}
295
2961;