package utf8;
use strict;
use warnings;
+use re "/aa"; # So we won't even try to look at above Latin1, potentially
+ # resulting in a recursive call
sub DEBUG () { 0 }
$|=1 if DEBUG;
# out blanks, underscores and dashes. The complication stems from the
# grandfathered-in 'L_', which retains a single trailing underscore.
- my $loose = $_[0] =~ s/[-\s_]//rg;
+ (my $loose = $_[0]) =~ tr/-_ \t//d;
return $loose if $loose !~ / ^ (?: is | to )? l $/x;
return 'l_' if $_[0] =~ / l .* _ /x; # If original had a trailing '_'
# keys TYPE, BITS, EXTRAS, LIST, and NONE with values having the
# same meanings as the input parameters.
# SPECIALS contains a reference to any special-treatment hash in the
+ # property.
# INVERT_IT is non-zero if the result should be inverted before use
# USER_DEFINED is non-zero if the result came from a user-defined
- # property.
my $file; ## file to load data from, and also part of the %Cache key.
# Change this to get a different set of Unicode tables
my $caller0 = caller(0);
- my $caller1 = $type =~ s/(.+)::// ? $1 : $caller0 eq 'main' ?
- 'main' : caller(1);
+ my $caller1 = $type =~ s/(.+):://
+ ? $1
+ : $caller0 eq 'main'
+ ? 'main'
+ : caller(1);
if (defined $caller1 && $type =~ /^I[ns]\w+$/) {
my $prop = "${caller1}::$type";
}
if (miniperl) {
eval "require '$unicore_dir/Heavy.pl'";
- last GETFILE if $@;
+ if ($@) {
+ print STDERR __LINE__, ": '$@'\n" if DEBUG;
+ pop @recursed if @recursed;
+ return $type;
+ }
}
else {
require "$unicore_dir/Heavy.pl";
# If the rhs looks like it is a number...
print STDERR __LINE__, ": table=$table\n" if DEBUG;
- if ($table =~ qr{ ^ [ \s 0-9 _ + / . -]+ $ }x) {
+ if ($table =~ m{ ^ [ \s 0-9 _ + / . -]+ $ }x) {
print STDERR __LINE__, ": table=$table\n" if DEBUG;
# Don't allow leading nor trailing slashes
}
# Split on slash, in case it is a rational, like \p{1/5}
- my @parts = split qr{ \s* / \s* }x, $table, -1;
+ my @parts = split m{ \s* / \s* }x, $table, -1;
print __LINE__, ": $type\n" if @parts > 2 && DEBUG;
# Can have maximum of one slash
# minus
# Remove underscores between digits.
- $part =~ s/( ?<= [0-9] ) _ (?= [0-9] ) //xg;
+ $part =~ s/(?<= [0-9] ) _ (?= [0-9] ) //xg;
# No leading zeros (but don't make a single '0'
# into a null string)
# Add the constant and go fetch it in.
if (defined $file) {
- # A beginning ! means to invert. The 0+ makes sure is
- # numeric
- $invert_it = 0 + $file =~ s/^!//;
+ # If the file name contains a !, it means to invert. The
+ # 0+ makes sure result is numeric
+ $invert_it = 0 + $file =~ s/!//;
if ($utf8::why_deprecated{$file}) {
warnings::warnif('deprecated', "Use of '$type' in \\p{} or \\P{} is deprecated because: $utf8::why_deprecated{$file};");
{
$file = $utf8::caseless_equivalent{$property_and_table};
}
- $file= "$unicore_dir/lib/$file.pl";
+
+ # The pseudo-directory '#' means that there really isn't a
+ # file to read, the data is in-line as part of the string;
+ # we extract it below.
+ $file = "$unicore_dir/lib/$file.pl" unless $file =~ m!^#/!;
last GETFILE;
}
print STDERR __LINE__, ": didn't find $property_and_table\n" if DEBUG;
$minbits = 1;
# The 0+ makes sure is numeric
- $invert_it = 0 + $file =~ s/^!//;
- $file = "$unicore_dir/lib/$file.pl";
+ $invert_it = 0 + $file =~ s/!//;
+ $file = "$unicore_dir/lib/$file.pl" unless $file =~ m!^#/!;
last GETFILE;
}
} }
## If we reach here, it was due to a 'last GETFILE' above
## (exception: user-defined properties and mappings), so we
## have a filename, so now we load it if we haven't already.
- ## If we have, return the cached results. The cache key is the
- ## class and file to load, and whether the results need to be
- ## inverted.
- ##
- my $found = $Cache{$class, $file, $invert_it};
- if ($found and ref($found) eq $class) {
- print STDERR __LINE__, ": Returning cached swash for '$class,$file,$invert_it' for \\p{$type}\n" if DEBUG;
- pop @recursed if @recursed;
- return $found;
+
+ # The pseudo-directory '#' means the result isn't really a
+ # file, but is in-line, with semi-colons to be turned into
+ # new-lines. Since it is in-line there is no advantage to
+ # caching the result
+ if ($file =~ s!^#/!!) {
+ $list = $utf8::inline_definitions[$file];
+ }
+ else {
+ # Here, we have an actual file to read in and load, but it
+ # may already have been read-in and cached. The cache key
+ # is the class and file to load, and whether the results
+ # need to be inverted.
+ my $found = $Cache{$class, $file, $invert_it};
+ if ($found and ref($found) eq $class) {
+ print STDERR __LINE__, ": Returning cached swash for '$class,$file,$invert_it' for \\p{$type}\n" if DEBUG;
+ pop @recursed if @recursed;
+ return $found;
+ }
+
+ local $@;
+ local $!;
+ $list = do $file; die $@ if $@;
}
- local $@;
- local $!;
- $list = do $file; die $@ if $@;
$list_is_from_mktables = 1;
}
} # End of $type is non-null
$list = join '', $taint,
map { $_->[1] }
sort { $a->[0] <=> $b->[0] }
- map { /^([0-9a-fA-F]+)/; [ CORE::hex($1), $_ ] }
- grep { /^([0-9a-fA-F]+)/ and not $seen{$1}++ } @tmp; # XXX doesn't do ranges right
+ map { /^([0-9a-fA-F]+)/ && !$seen{$1}++ ? [ CORE::hex($1), $_ ] : () }
+ @tmp; # XXX doesn't do ranges right
}
else {
# mktables has gone to some trouble to make non-user defined