This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
regen/unicode_constants.pl: Remove unused constant
[perl5.git] / regen / unicode_constants.pl
CommitLineData
61dad979
KW
1use v5.16.0;
2use strict;
3use warnings;
4require 'regen/regen_lib.pl';
5use charnames qw(:loose);
6
1b0f46bf 7my $out_fh = open_new('unicode_constants.h', '>',
61dad979
KW
8 {style => '*', by => $0,
9 from => "Unicode data"});
10
11print $out_fh <<END;
d10c72f2 12
1b0f46bf
KW
13#ifndef H_UNICODE_CONSTANTS /* Guard against nested #includes */
14#define H_UNICODE_CONSTANTS 1
d10c72f2 15
61dad979 16/* This file contains #defines for various Unicode code points. The values
525b6419
KW
17 * the macros expand to are the native Unicode code point, or all or portions
18 * of the UTF-8 encoding for the code point. In the former case, the macro
19 * name has the suffix "_NATIVE"; otherwise, the suffix "_UTF8".
61dad979 20 *
525b6419
KW
21 * The macros that have the suffix "_UTF8" may have further suffixes, as
22 * follows:
23 * "_FIRST_BYTE" if the value is just the first byte of the UTF-8
24 * representation; the value will be a numeric constant.
25 * "_TAIL" if instead it represents all but the first byte. This, and
26 * with no additional suffix are both string constants */
61dad979
KW
27
28END
29
76837d21 30# The data are at the end of this file. A blank line is output as-is.
5a731a17
KW
31# Comments (lines whose first non-blank is a '#') are converted to C-style,
32# though empty comments are converted to blank lines. Otherwise, each line
33# represents one #define, and begins with either a Unicode character name with
34# the blanks and dashes in it squeezed out or replaced by underscores; or it
35# may be a hexadecimal Unicode code point. In the latter
76837d21 36# case, the name will be looked-up to use as the name of the macro. In either
e9cddfae
KW
37# case, the macro name will have suffixes as listed above, and all blanks and
38# dashes will be replaced by underscores.
61dad979
KW
39#
40# Each line may optionally have one of the following flags on it, separated by
41# white space from the initial token.
5f1720e9 42# string indicates that the output is to be of the string form
61dad979 43# described in the comments above that are placed in the file.
632c9f80
KW
44# string_skip_ifundef is the same as 'string', but instead of dying if the
45# code point doesn't exist, the line is just skipped: no output is
46# generated for it
5f1720e9 47# first indicates that the output is to be of the FIRST_BYTE form.
61dad979 48# tail indicates that the output is of the _TAIL form.
525b6419
KW
49# native indicates that the output is the code point, converted to the
50# platform's native character set if applicable
61dad979 51#
765ec46c
KW
52# If the code point has no official name, the desired name may be appended
53# after the flag, which will be ignored if there is an official name.
54#
61dad979
KW
55# This program is used to make it convenient to create compile time constants
56# of UTF-8, and to generate proper EBCDIC as well as ASCII without manually
57# having to figure things out.
58
59while ( <DATA> ) {
5a731a17
KW
60 chomp;
61
62 # Convert any '#' comments to /* ... */; empty lines and comments are
63 # output as blank lines
64 if ($_ =~ m/ ^ \s* (?: \# ( .* ) )? $ /x) {
65 my $comment_body = $1 // "";
66 if ($comment_body ne "") {
67 print $out_fh "/* $comment_body */\n";
68 }
69 else {
70 print $out_fh "\n";
71 }
76837d21
KW
72 next;
73 }
74
61dad979 75 unless ($_ =~ m/ ^ ( [^\ ]* ) # Name or code point token
765ec46c
KW
76 (?: [\ ]+ ( [^ ]* ) )? # optional flag
77 (?: [\ ]+ ( .* ) )? # name if unnamed; flag is required
61dad979
KW
78 /x)
79 {
80 die "Unexpected syntax at line $.: $_\n";
81 }
82
83 my $name_or_cp = $1;
84 my $flag = $2;
765ec46c 85 my $desired_name = $3;
61dad979
KW
86
87 my $name;
88 my $cp;
632c9f80 89 my $undef_ok = $desired_name || $flag =~ /skip_if_undef/;
61dad979
KW
90
91 if ($name_or_cp =~ /[^[:xdigit:]]/) {
92
93 # Anything that isn't a hex value must be a name.
94 $name = $name_or_cp;
95 $cp = charnames::vianame($name =~ s/_/ /gr);
96 die "Unknown name '$name' at line $.: $_\n" unless defined $name;
97 }
98 else {
99 $cp = $name_or_cp;
632c9f80
KW
100 $name = charnames::viacode("0$cp"); # viacode requires a leading zero
101 # to be sure that the argument is
102 # hex
103 if (! defined $name) {
104 die "Unknown code point '$cp' at line $.: $_\n" unless $undef_ok;
105 $name = "";
106 }
61dad979
KW
107 }
108
632c9f80 109 $name = $desired_name if $name eq "" && $desired_name;
e9cddfae 110 $name =~ s/[- ]/_/g; # The macro name can have no blanks nor dashes
61dad979
KW
111
112 my $str = join "", map { sprintf "\\x%02X", $_ }
113 unpack("U0C*", pack("U", hex $cp));
114
115 my $suffix = '_UTF8';
632c9f80 116 if (! defined $flag || $flag =~ /^ string (_skip_if_undef)? $/x) {
61dad979
KW
117 $str = "\"$str\""; # Will be a string constant
118 } elsif ($flag eq 'tail') {
119 $str =~ s/\\x..//; # Remove the first byte
120 $suffix .= '_TAIL';
121 $str = "\"$str\""; # Will be a string constant
122 }
123 elsif ($flag eq 'first') {
124 $str =~ s/ \\x ( .. ) .* /$1/x; # Get the two nibbles of the 1st byte
125 $suffix .= '_FIRST_BYTE';
126 $str = "0x$str"; # Is a numeric constant
127 }
525b6419
KW
128 elsif ($flag eq 'native') {
129 die "Are you sure you want to run this on an above-Latin1 code point?" if hex $cp > 0xff;
130 $suffix = '_NATIVE';
131 $str = utf8::unicode_to_native(hex $cp);
132 $str = "0x$cp"; # Is a numeric constant
133 }
61dad979
KW
134 else {
135 die "Unknown flag at line $.: $_\n";
136 }
137 print $out_fh "#define ${name}$suffix $str /* U+$cp */\n";
138}
139
1b0f46bf 140print $out_fh "\n#endif /* H_UNICODE_CONSTANTS */\n";
d10c72f2 141
61dad979
KW
142read_only_bottom_close_and_rename($out_fh);
143
144__DATA__
5f1720e9
KW
1450300 string
1460301 string
1470308 string
76837d21 148
0a982f06 14903B9 string
76837d21 150
0a982f06 15103C5 string
76837d21 152
5f1720e9 1532010 string
765ec46c 154D800 first FIRST_SURROGATE
525b6419
KW
155
156007F native
15700DF native
15800E5 native
15900C5 native
16000FF native
16100B5 native