package utf8;
+use strict;
+use warnings;
sub DEBUG () { 0 }
sub DESTROY {}
+my %Cache;
+
sub croak { require Carp; Carp::croak(@_) }
+##
+## "SWASH" == "SWATCH HASH". A "swatch" is a swatch of the Unicode landscape
+##
+
sub SWASHNEW {
my ($class, $type, $list, $minbits, $none) = @_;
local $^D = 0 if $^D;
print STDERR "SWASHNEW @_\n" if DEBUG;
- my $file;
+ ##
+ ## Get the list of codepoints for the type.
+ ## Called from utf8.c
+ ##
+ ## Given a $type, our goal is to fill $list with the set of codepoint
+ ## ranges.
+ ##
+ ## To make the parsing of $type clear, this code takes the a rather
+ ## unorthadox approach of last'ing out of the block once we have the
+ ## info we need. Were this to be a subroutine, the 'last' would just
+ ## be a 'return'.
+ ##
+ my $file; ## file to load data from, and also part of the %Cache key.
+ my $ListSorted = 0;
- if ($type and ref ${"${class}::{$type}"} eq $class) {
- warn qq/Found \${"${class}::{$type}"}\n/ if DEBUG;
- return ${"${class}::{$type}"}; # Already there...
- }
+ if ($type)
+ {
+ $type =~ s/^\s+//;
+ $type =~ s/\s+$//;
- if ($type) {
+ print "type = $type\n" if DEBUG;
- defined %utf8::In || do "unicore/In.pl";
+ GETFILE:
+ {
+ ##
+ ## 'Is' is always optional, so if it's there, remove it.
+ ## Same with 'Category=' and 'Script='.
+ ##
+ ## 'Block=' is replaced by 'In'.
+ ##
+ $type =~ s/^Is(?:\s+|[-_])?//i
+ or
+ $type =~ s/^Category\s*=\s*//i
+ or
+ $type =~ s/^Script\s*=\s*//i
+ or
+ $type =~ s/^Block\s*=\s*/In/i;
- $type =~ s/^In(?:[-_]|\s+)?(?!herited$)//i;
- $type =~ s/\s+$//;
+ ##
+ ## See if it's in the direct mapping table.
+ ##
+ require "unicore/Exact.pl";
+ if (my $base = $utf8::Exact{$type}) {
+ $file = "unicore/lib/$base.pl";
+ last GETFILE;
+ }
- my $inprefix = substr(lc($type), 0, 3);
- if (exists $utf8::InPat{$inprefix}) {
- my $In = $type;
- for my $k (keys %{$utf8::InPat{$inprefix}}) {
- if ($In =~ /^$k$/i) {
- $In = $utf8::InPat{$inprefix}->{$k};
- if (exists $utf8::In{$In}) {
- $file = "unicore/In/$utf8::In{$In}";
- print "inprefix = $inprefix, In = $In, k = $k, file = $file\n" if DEBUG;
- last;
- }
- }
- }
- }
+ ##
+ ## If not there exactly, try the canonical form. The canonical
+ ## form is lowercased, with any separators (\s+|[-_]) removed.
+ ##
+ my $canonical = lc $type;
+ $canonical =~ s/(?<=[a-z\d])(?:\s+|[-_])(?=[a-z\d])//g;
+ print "canonical = $canonical\n" if DEBUG;
+
+ require "unicore/Canonical.pl";
+ if (my $base = $utf8::Canonical{$canonical}) {
+ $file = "unicore/lib/$base.pl";
+ last GETFILE;
+ }
+
+ ##
+ ## It could be a user-defined property.
+ ##
+
+ if ($type =~ /^I[ns](\w+)$/) {
+ my @caller = caller(1);
+
+ if (defined $caller[0]) {
+ my $prop = $caller[0] . "::" . $type;
+
+ if (exists &{$prop}) {
+ no strict 'refs';
+
+ $list = &{$prop};
+ last GETFILE;
+ }
+ }
+ }
- unless (defined $file) {
- # This is separate from 'To' in preparation of Is.pl (a la In.pl).
- if ($type =~ /^Is([A-Z][A-Za-z]*|L&)$/ || $type =~ /^(L&)$/ ) {
- my $cat = $1 eq 'L&' ? 'L' : $1;
- $file = "unicore/Is/$cat";
- } elsif ((not defined $file) && $type =~ /^To([A-Z][A-Za-z]*)$/) {
- $file = "unicore/To/$1";
+ ##
+ ## Last attempt -- see if it's a "To" name (e.g. "ToLower")
+ ##
+ if ($type =~ /^To([A-Z][A-Za-z]+)$/)
+ {
+ $file = "unicore/To/$1.pl";
+ ## would like to test to see if $file actually exists....
+ last GETFILE;
+ }
+
+ ##
+ ## If we reach this line, it's because we couldn't figure
+ ## out what to do with $type. Ouch.
+ ##
+
+ return $type;
+ }
+
+ if (defined $file) {
+ print "found it (file='$file')\n" if DEBUG;
+
+ ##
+ ## If we reach here, it was due to a 'last GETFILE' above
+ ## (exception: user-defined properties), 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
+ ## file to load.
+ ##
+ if ($Cache{$file} and ref($Cache{$file}) eq $class)
+ {
+ print "Returning cached '$file' for \\p{$type}\n" if DEBUG;
+ return $Cache{$class, $file};
}
+
+ $list = do $file;
}
- }
- {
- $list ||= do "$file.pl"
- || do "unicore/Is/$type.pl"
- || croak("Can't find Unicode character property \"$type\"");
+ $ListSorted = 1; ## we know that these lists are sorted
}
my $extras;
my $bits;
-
+
+ my $ORIG = $list;
if ($list) {
my @tmp = split(/^/m, $list);
my %seen;
while ($x =~ /^([^0-9a-fA-F\n])(.*)/mg) {
my $char = $1;
my $name = $2;
-# print STDERR "$1 => $2\n" if DEBUG;
+ print STDERR "$1 => $2\n" if DEBUG;
if ($char =~ /[-+!]/) {
my ($c,$t) = split(/::/, $name, 2); # bogus use of ::, really
- my $subobj = $c->SWASHNEW($t, "", 0, 0, 0);
+ my $subobj;
+ if ($c eq 'utf8') {
+ $subobj = $c->SWASHNEW($t, "", 0, 0, 0);
+ }
+ elsif ($c =~ /^([0-9a-fA-F]+)/) {
+ $subobj = utf8->SWASHNEW("", $c, 0, 0, 0);
+ }
+ return $subobj unless ref $subobj;
push @extras, $name => $subobj;
$bits = $subobj->{BITS} if $bits < $subobj->{BITS};
}
print STDERR "CLASS = $class, TYPE => $type, BITS => $bits, NONE => $none\nEXTRAS =>\n$extras\nLIST =>\n$list\n" if DEBUG;
- ${"${class}::{$type}"} = bless {
+ my $SWASH = bless {
TYPE => $type,
BITS => $bits,
EXTRAS => $extras,
NONE => $none,
@extras,
} => $class;
+
+ if ($file) {
+ $Cache{$class, $file} = $SWASH;
+ }
+
+ return $SWASH;
}
# NOTE: utf8.c:swash_init() assumes entries are never modified once generated.
sub SWASHGET {
+ # See utf8.c:Perl_swash_fetch for problems with this interface.
my ($self, $start, $len) = @_;
local $^D = 0 if $^D;
my $type = $self->{TYPE};
my $max = (defined $2 ? hex $2 : $min);
my $val = hex $3;
next if $max < $start;
-# print "$min $max $val\n";
+ print "$min $max $val\n" if DEBUG;
if ($none) {
if ($min < $start) {
$val += $start - $min if $val < $none;
}
for ($key = $min; $key <= $max; $key++) {
last LINE if $key >= $end;
-# print STDERR "$key => $val\n" if DEBUG;
+ print STDERR "$key => $val\n" if DEBUG;
vec($swatch, $key - $start, $bits) = $val;
++$val if $val < $none;
}
}
for ($key = $min; $key <= $max; $key++, $val++) {
last LINE if $key >= $end;
-# print STDERR "$key => $val\n" if DEBUG;
+ print STDERR "$key => $val\n" if DEBUG;
vec($swatch, $key - $start, $bits) = $val;
}
}
}
else {
LINE:
- while (/^([0-9a-fA-F]+)(?:\t([0-9a-fA-F]+))?/mg) {
+ while (/^([0-9a-fA-F]+)(?:[ \t]+([0-9a-fA-F]+))?/mg) {
my $min = hex $1;
my $max = (defined $2 ? hex $2 : $min);
next if $max < $start;
}
for ($key = $min; $key <= $max; $key++) {
last LINE if $key >= $end;
-# print STDERR "$key => 1\n" if DEBUG;
+ print STDERR "$key => 1\n" if DEBUG;
vec($swatch, $key - $start, 1) = 1;
}
}