5 require 'regen/regen_lib.pl';
7 # This program outputs l1_charclass_tab.h, which defines the guts of the
8 # PL_charclass table. Each line is a bit map of properties that the Unicode
9 # code point at the corresponding position in the table array has. The first
10 # line corresponds to code point U+0000, NULL, the last line to U+00FF. For
11 # an application to see if the code point "i" has a particular property, it
13 # 'PL_charclass[i] & BIT'
14 # The bit names are of the form '_CC_property_suffix', where 'CC' stands for
15 # character class, and 'property' is the corresponding property, and 'suffix'
16 # is one of '_A' to mean the property is true only if the corresponding code
17 # point is ASCII, and '_L1' means that the range includes any Latin1
18 # character (ISO-8859-1 including the C0 and C1 controls). A property without
19 # these suffixes does not have different forms for both ranges.
21 # This program need be run only when adding new properties to it, or upon a
22 # new Unicode release, to make sure things haven't been changed by it.
49 # Read in the case fold mappings.
51 my @hex_non_final_folds;
55 BEGIN { # Have to do this at compile time because using user-defined \p{property}
57 # Use the Unicode data file if we are on an ASCII platform (which its data
58 # is for), and it is in the modern format (starting in Unicode 3.1.0) and
59 # it is available. This avoids being affected by potential bugs
60 # introduced by other layers of Perl
61 my $file="lib/unicore/CaseFolding.txt";
64 && pack("C*", split /\./, Unicode::UCD::UnicodeVersion()) ge v3.1.0
65 && open my $fh, "<", $file)
70 my ($invlist_ref, $invmap_ref, undef, $default)
71 = Unicode::UCD::prop_invmap('Case_Folding');
72 for my $i (0 .. @$invlist_ref - 1 - 1) {
73 next if $invmap_ref->[$i] == $default;
75 for my $j ($invlist_ref->[$i] .. $invlist_ref->[$i+1] -1) {
78 # Single-code point maps go to a 'C' type
79 if (! ref $invmap_ref->[$i]) {
80 push @folds, sprintf("%04X; C; %04X\n",
82 $invmap_ref->[$i] + $adjust);
84 else { # Multi-code point maps go to 'F'. prop_invmap()
85 # guarantees that no adjustment is needed for these,
86 # as the range will contain just one element
87 push @folds, sprintf("%04X; F; %s\n",
89 join " ", map { sprintf "%04X", $_ }
90 @{$invmap_ref->[$i]});
99 # Lines look like (without the initial '#'
100 #0130; F; 0069 0307; # LATIN CAPITAL LETTER I WITH DOT ABOVE
101 # Get rid of comments, ignore blank or comment-only lines
102 my $line = $_ =~ s/ (?: \s* \# .* )? $ //rx;
103 next unless length $line;
104 my ($hex_from, $fold_type, @folded) = split /[\s;]+/, $line;
106 my $from = hex $hex_from;
108 # Perl only deals with C and F folds
109 next if $fold_type ne 'C' and $fold_type ne 'F';
111 # Get each code point in the range that participates in this line's fold.
112 # The hash has keys of each code point in the range, and values of what it
113 # folds to and what folds to it
114 for my $i (0 .. @folded - 1) {
115 my $hex_fold = $folded[$i];
116 my $fold = hex $hex_fold;
117 push @{$folded_closure{$fold}}, $from if $fold < 256;
118 push @{$folded_closure{$from}}, $fold if $from < 256;
122 && ! grep { $_ eq $hex_fold } @hex_non_final_folds)
124 push @hex_non_final_folds, $hex_fold;
126 # Also add the upper case, which in the latin1 range folds to
128 push @hex_non_final_folds, sprintf "%04X", ord uc chr $fold;
133 # Now having read all the lines, combine them into the full closure of each
134 # code point in the range by adding lists together that share a common
136 foreach my $folded (keys %folded_closure) {
137 foreach my $from (grep { $_ < 256 } @{$folded_closure{$folded}}) {
138 push @{$folded_closure{$from}}, @{$folded_closure{$folded}};
143 sub Is_Non_Latin1_Fold {
146 foreach my $folded (keys %folded_closure) {
147 push @return, sprintf("%X", $folded), if grep { $_ > 255 }
148 @{$folded_closure{$folded}};
150 return join("\n", @return) . "\n";
153 sub Is_Non_Final_Fold {
154 return join("\n", @hex_non_final_folds) . "\n";
157 my @bits; # Bit map for each code point
159 # For each character, calculate which properties it matches.
160 for my $ord (0..255) {
161 my $char = chr($ord);
162 utf8::upgrade($char); # Important to use Unicode semantics!
164 # Look at all the properties we care about here.
165 for my $property (@properties) {
166 my $name = $property;
168 # Remove the suffix to get the actual property name.
169 # Currently the suffixes are '_L1', '_A', and none.
170 # If is a latin1 version, no further checking is needed.
171 if (! ($name =~ s/_L1$//)) {
173 # Here, isn't an _L1. If its _A, it's automatically false for
174 # non-ascii. The only current ones (besides ASCII) without a
175 # suffix are valid over the whole range.
176 next if $name =~ s/_A$// && $ord >= 128;
180 if ($name eq 'PUNCT') {;
182 # Sadly, this is inconsistent: \pP and \pS for the ascii range,
183 # just \pP outside it.
184 $re = qr/\p{Punct}|[^\P{Symbol}\P{ASCII}]/;
185 } elsif ($name eq 'CHARNAME_CONT') {;
186 $re = qr/\p{_Perl_Charname_Continue}/,
187 } elsif ($name eq 'SPACE') {;
188 $re = qr/\p{XPerlSpace}/;
189 } elsif ($name eq 'IDFIRST') {
190 $re = qr/[_\p{Alpha}]/;
191 } elsif ($name eq 'PSXSPC') {
192 $re = qr/[\v\p{Space}]/;
193 } elsif ($name eq 'WORDCHAR') {
194 $re = qr/\p{XPosixWord}/;
195 } elsif ($name eq 'ALNUMC') {
196 # Like \w, but no underscore
198 } elsif ($name eq 'QUOTEMETA') {
199 $re = qr/\p{_Perl_Quotemeta}/;
200 } elsif ($name eq 'NONLATIN1_FOLD') {
201 $re = qr/\p{Is_Non_Latin1_Fold}/;
202 } elsif ($name eq 'NON_FINAL_FOLD') {
203 $re = qr/\p{Is_Non_Final_Fold}/;
204 } elsif ($name eq 'IS_IN_SOME_FOLD') {
205 $re = qr/\p{_Perl_Any_Folds}/;
206 } else { # The remainder have the same name and values as Unicode
207 $re = eval "qr/\\p{$name}/";
209 carp $@ if ! defined $re;
211 #print "$ord, $name $property, $re\n";
212 if ($char =~ $re) { # Add this property if matches
213 $bits[$ord] .= '|' if $bits[$ord];
214 $bits[$ord] .= "(1U<<_CC_$property)";
217 #print __LINE__, " $ord $char $bits[$ord]\n";
220 # Names of C0 controls
256 # Names of C1 controls, plus the adjacent DEL
293 my $out_fh = open_new('l1_char_class_tab.h', '>',
294 {style => '*', by => $0,
295 from => "property definitions"});
297 # Output the table using fairly short names for each char.
298 for my $ord (0..255) {
300 if ($ord < 32) { # A C0 control
302 } elsif ($ord > 32 && $ord < 127) { # Graphic
303 $name = "'" . chr($ord) . "'";
304 } elsif ($ord >= 127 && $ord <= 0x9f) {
305 $name = $C1[$ord - 127]; # A C1 control + DEL
306 } else { # SPACE, or, if Latin1, shorten the name */
308 $name = charnames::viacode($ord);
309 $name =~ s/LATIN CAPITAL LETTER //
310 || $name =~ s/LATIN SMALL LETTER (.*)/\L$1/;
312 printf $out_fh "/* U+%02X %s */ %s,\n", $ord, $name, $bits[$ord];
315 read_only_bottom_close_and_rename($out_fh)