X-Git-Url: http://git.shadowcat.co.uk/gitweb/gitweb.cgi?a=blobdiff_plain;f=lib%2Futf8_heavy.pl;h=b6fdeb997b3407bb003b5d3d884cc86c4ed47270;hb=8af3c3489f0c9ca8d37ebae27991ac4f34dfdb2f;hp=4fbb112c230dba9eefa3fdc198e7c0f0e285761e;hpb=fa689c2334baa95cf3f3faa003c5d50ee4cd9b7a;p=p5sagit%2Fp5-mst-13.2.git diff --git a/lib/utf8_heavy.pl b/lib/utf8_heavy.pl index 4fbb112..b6fdeb9 100644 --- a/lib/utf8_heavy.pl +++ b/lib/utf8_heavy.pl @@ -1,119 +1,204 @@ package utf8; +use strict; +use warnings; sub DEBUG () { 0 } sub DESTROY {} +my %Cache; + +our (%PropertyAlias, %PA_reverse, %PropValueAlias, %PVA_reverse, %PVA_abbr_map); + sub croak { require Carp; Carp::croak(@_) } +## +## "SWASH" == "SWATCH HASH". A "swatch" is a swatch of the Unicode landscape. +## It's a data structure that encodes a set of Unicode characters. +## + sub SWASHNEW { my ($class, $type, $list, $minbits, $none) = @_; local $^D = 0 if $^D; print STDERR "SWASHNEW @_\n" if DEBUG; - if ($type and ref ${"${class}::{$type}"} eq $class) { - warn qq/Found \${"${class}::{$type}"}\n/ if DEBUG; - return ${"${class}::{$type}"}; # Already there... - } + ## + ## 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 + ## unorthodox 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) { - $type =~ s/^\s+//; - $type =~ s/\s+$//; - - print "type = $type\n" if DEBUG; - - my $file; - - unless (defined $file) { - defined %utf8::Is || do "unicore/Is.pl"; - if ($type =~ /^(?:Is|Category\s*=\s*)?[- _]?([A-Z].*)$/i) { - my $istype = $1; - print "istype = $istype\n" if DEBUG; - unless ($list = do "unicore/Is/$istype.pl") { - if (exists $utf8::Is{$istype}) { - $file = "unicore/Is/$utf8::Is{$istype}"; - } else { - my $isprefix = substr(lc($istype), 0, 2); - print "isprefix = $isprefix\n" if DEBUG; - if (exists $utf8::IsPat{$isprefix}) { - my $Is = $istype; - print "isprefix = $isprefix, Is = $Is\n" if DEBUG; - for my $k (keys %{$utf8::IsPat{$isprefix}}) { - print "isprefix = $isprefix, Is = $Is, k = $k\n" if DEBUG; - if ($Is =~ /^$k$/i) { - $file = "unicore/Is/$utf8::IsPat{$isprefix}->{$k}"; - print "isprefix = $isprefix, Is = $Is, k = $k, file = $file\n" if DEBUG; - last; - } - } - } - } + if ($type) + { + $type =~ s/^\s+//; + $type =~ s/\s+$//; + + print "type = $type\n" if DEBUG; + + GETFILE: + { + ## + ## It could be a user-defined property. + ## + + my $caller1 = $type =~ s/(.+)::// ? $1 : caller(1); + + if (defined $caller1 && $type =~ /^(?:\w+)$/) { + my $prop = "${caller1}::$type"; + if (exists &{$prop}) { + no strict 'refs'; + + $list = &{$prop}; + last GETFILE; } } - unless (defined $file) { - defined %utf8::In || do "unicore/In.pl"; - $type = 'Lampersand' if $type =~ /^(?:Is)?L&$/; - $type = 'Assigned' if $type =~ /^(?:Is)?Assigned$/i; - $type = 'Unassigned' if $type =~ /^(?:Is)?Unassigned$/i; - if ($type =~ /^(In|(?:Script|Block)\s*=\s*)?[- _]?(?!herited$)(.+)/i) { - my $incat = $1 || ''; - my $intype = $2; - print "incat = $incat, intype = $intype\n" if DEBUG; - if (exists $utf8::In{$intype}) { - $file = "unicore/In/$utf8::In{$intype}"; - } else { - my $inprefix = substr(lc($intype), 0, 2); - print "inprefix = $inprefix\n" if DEBUG; - if (exists $utf8::InPat{$inprefix}) { - my $In = $intype; - print "inprefix = $inprefix, In = $In\n" if DEBUG; - for my $k (keys %{$utf8::InPat{$inprefix}}) { - print "inprefix = $inprefix, In = $In, k = $k\n" if DEBUG; - if ($In =~ /^$k$/i) { - my $i = $utf8::InPat{$inprefix}->{$k}; - print "inprefix = $inprefix, In = $In, k = $k, i = $i\n" if DEBUG; - next if $incat =~ /^S/ && - !exists $utf8::InScript{$i}; - next if $incat =~ /^B/ && - !exists $utf8::InBlock{$i}; - $file = "unicore/In/$i"; - print "inprefix = $inprefix, In = $In, k = $k, file = $file\n" if DEBUG; - last; - } - } - } - } + my $wasIs; + + ($wasIs = $type =~ s/^Is(?:\s+|[-_])?//i) + or + $type =~ s/^(?:(?:General(?:\s+|_)?)?Category|gc)\s*[:=]\s*//i + or + $type =~ s/^(?:Script|sc)\s*[:=]\s*//i + or + $type =~ s/^Block\s*[:=]\s*/In/i; + + + ## + ## See if it's in some enumeration. + ## + require "unicore/PVA.pl"; + if ($type =~ /^([\w\s]+)[:=]\s*(.*)/) { + my ($enum, $val) = (lc $1, lc $2); + $enum =~ tr/ _-//d; + $val =~ tr/ _-//d; + + my $pa = $PropertyAlias{$enum} ? $enum : $PA_reverse{$enum}; + my $f = $PropValueAlias{$pa}{$val} ? $val : $PVA_reverse{$pa}{lc $val}; + + if ($pa and $f) { + $pa = "gc_sc" if $pa eq "gc" or $pa eq "sc"; + $file = "unicore/lib/$pa/$PVA_abbr_map{$pa}{lc $f}.pl"; + last GETFILE; } } + else { + my $t = lc $type; + $t =~ tr/ _-//d; - unless (defined $file) { - if ($type =~ /^To([A-Z][A-Za-z]+)$/) { - $file = "unicore/To/$1"; + if ($PropValueAlias{gc}{$t} or $PropValueAlias{sc}{$t}) { + $file = "unicore/lib/gc_sc/$PVA_abbr_map{gc_sc}{$t}.pl"; + last GETFILE; } } - } + + ## + ## See if it's in the direct mapping table. + ## + require "unicore/Exact.pl"; + if (my $base = $utf8::Exact{$type}) { + $file = "unicore/lib/gc_sc/$base.pl"; + last GETFILE; + } + + ## + ## 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} || $utf8::Canonical{ lc $utf8::PropertyAlias{$canonical} })) { + $file = "unicore/lib/gc_sc/$base.pl"; + last GETFILE; + } + + ## + ## See if it's a user-level "To". + ## + + my $caller0 = caller(0); + + if (defined $caller0 && $type =~ /^To(?:\w+)$/) { + my $map = $caller0 . "::" . $type; + + if (exists &{$map}) { + no strict 'refs'; + + $list = &{$map}; + last GETFILE; + } + } + + ## + ## Last attempt -- see if it's a standard "To" name + ## (e.g. "ToLower") ToTitle is used by ucfirst(). + ## The user-level way to access ToDigit() and ToFold() + ## is to use Unicode::UCD. + ## + if ($type =~ /^To(Digit|Fold|Lower|Title|Upper)$/) + { + $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) { - $list = do "$file.pl"; + print "found it (file='$file')\n" if DEBUG; + + ## + ## If we reach here, it was due to a 'last GETFILE' above + ## (exception: user-defined properties and mappings), so we + ## have a filename, so now we load it if we haven't already. + ## If we have, return the cached results. The cache key is the + ## file to load. + ## + if ($Cache{$file} and ref($Cache{$file}) eq $class) + { + print "Returning cached '$file' for \\p{$type}\n" if DEBUG; + return $Cache{$class, $file}; + } + + $list = do $file; } - croak("Can't find Unicode character property \"$type\"") - unless $list; + $ListSorted = 1; ## we know that these lists are sorted } my $extras; - my $bits; - + my $bits = 0; + + my $ORIG = $list; if ($list) { my @tmp = split(/^/m, $list); my %seen; no warnings; $extras = join '', grep /^[^0-9a-fA-F]/, @tmp; $list = join '', - sort { hex $a <=> hex $b } - grep {/^([0-9a-fA-F]+)/ and not $seen{$1}++} @tmp; # XXX doesn't do ranges right + map { $_->[1] } + sort { $a->[0] <=> $b->[0] } + map { /^([0-9a-fA-F]+)/; [ hex($1), $_ ] } + grep { /^([0-9a-fA-F]+)/ and not $seen{$1}++ } @tmp; # XXX doesn't do ranges right } if ($none) { @@ -123,10 +208,10 @@ sub SWASHNEW { if ($minbits < 32) { my $top = 0; - while ($list =~ /^([0-9a-fA-F]+)(?:\t([0-9a-fA-F]+)?)(?:\t([0-9a-fA-F]+))?/mg) { + while ($list =~ /^([0-9a-fA-F]+)(?:[\t]([0-9a-fA-F]+)?)(?:[ \t]([0-9a-fA-F]+))?/mg) { my $min = hex $1; - my $max = hex(defined $2 ? $2 : $1); - my $val = hex(defined $3 ? $3 : ""); + my $max = defined $2 ? hex $2 : $min; + my $val = defined $3 ? hex $3 : 0; $val += $max - $min if defined $3; $top = $val if $val > $top; } @@ -144,9 +229,19 @@ sub SWASHNEW { my $char = $1; my $name = $2; print STDERR "$1 => $2\n" if DEBUG; - if ($char =~ /[-+!]/) { + 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') { + $subobj = utf8->SWASHNEW($t, "", 0, 0, 0); + } + elsif (exists &$name) { + $subobj = utf8->SWASHNEW($name, "", 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}; } @@ -155,7 +250,7 @@ sub SWASHNEW { 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, @@ -163,6 +258,12 @@ sub SWASHNEW { NONE => $none, @extras, } => $class; + + if ($file) { + $Cache{$class, $file} = $SWASH; + } + + return $SWASH; } # NOTE: utf8.c:swash_init() assumes entries are never modified once generated. @@ -187,10 +288,15 @@ sub SWASHGET { pos $_ = 0; if ($bits > 1) { LINE: - while (/^([0-9a-fA-F]+)(?:\t([0-9a-fA-F]+)?)(?:\t([0-9a-fA-F]+))?/mg) { - my $min = hex $1; - my $max = (defined $2 ? hex $2 : $min); - my $val = hex $3; + while (/^([0-9a-fA-F]+)(?:[ \t]([0-9a-fA-F]+)?)?(?:[ \t]([0-9a-fA-F]+))?/mg) { + chomp; + my ($a, $b, $c) = ($1, $2, $3); + croak "$type: illegal mapping '$_'" + if $type =~ /^To/ && + !(defined $a && defined $c); + my $min = hex $a; + my $max = defined $b ? hex $b : $min; + my $val = defined $c ? hex $c : 0; next if $max < $start; print "$min $max $val\n" if DEBUG; if ($none) { @@ -220,9 +326,10 @@ sub SWASHGET { } else { LINE: - while (/^([0-9a-fA-F]+)(?:\t([0-9a-fA-F]+))?/mg) { + while (/^([0-9a-fA-F]+)(?:[ \t]+([0-9a-fA-F]+))?/mg) { + chomp; my $min = hex $1; - my $max = (defined $2 ? hex $2 : $min); + my $max = defined $2 ? hex $2 : $min; next if $max < $start; if ($min < $start) { $min = $start; @@ -237,7 +344,7 @@ sub SWASHGET { } for my $x ($self->{EXTRAS}) { pos $x = 0; - while ($x =~ /^([-+!])(.*)/mg) { + while ($x =~ /^([-+!&])(.*)/mg) { my $char = $1; my $name = $2; print STDERR "INDIRECT $1 $2\n" if DEBUG; @@ -278,6 +385,18 @@ sub SWASHGET { } } } + elsif ($char eq '&') { + if ($bits == 1 and $otherbits == 1) { + $swatch &= $other; + } + else { + for ($key = 0; $key < $len; $key++) { + if (!vec($other, $key, $otherbits)) { + vec($swatch, $key, $bits) = 0; + } + } + } + } } } if (DEBUG) {