parse-wormedit: distinguish newer v95 level strings
[wormy.git] / parse-wormedit
index 2eb9e277e417596d6d9a72204e5caba48a2815dd..84e4e31c0140810983205359b0b05df2d3a54283 100755 (executable)
@@ -1,17 +1,30 @@
 #!/usr/bin/env perl
 use strict;
 use warnings;
+use 5.010;
 
 use Data::Dumper;
 use Getopt::Long 2.33 qw(HelpMessage :config bundling);
 
-our $VERSION = '1.01';
+our $VERSION = '1.03';
 
 GetOptions(\my %opt,
-       'raw|r',
+       'raw|r',  # full output
+       'version=i',  # force version
 ) or HelpMessage(-exitval => 2);
 
-my $MAGICID = "WormEdit053\000LVL";
+
+package Shiar_Parse::WormEdit;
+
+use strict;
+use warnings;
+
+our %MAGICID = (
+       "WormEdit053\000LVL" => 53,
+       "WormEdit\34195\000LVL" => 95,
+       "WormEdit\34194\000LVL" => 94,
+       "WormEdit\34193\000LVL" => 93,
+);
 
 my @FORMAT = (
        magic       => 'a15',
@@ -41,9 +54,11 @@ my @FORMAT = (
        sprite     => ['8C',
                line => 'B8',
        ],
-       endtype     => 's',
-       endstr      => 'Ca255',
-       enddata     => 'Ca255x256',
+       finish      => [1,
+               type    => 's',
+               message => 'Ca255',
+               code    => 'Ca255x256',
+       ],
        hiname      => 'a3',
        levels      => ['*', # levelcount->total actually
                id         => 'Ca22',
@@ -82,81 +97,320 @@ my @FORMAT = (
        ],
 );
 
-my @OBJTYPE = ('none', 'line', 'fat line', 'bar', 'circle');
-my @ENDTYPE = ('none', 'message', 'small message');
+sub read {
+       my ($self, $input) = @_;
+       my ($id, $version) = (substr($input, 0, 15), ord substr($input, 15, 1));
+       my $fileversion = $MAGICID{$id}
+               or die "File does not match any known WormEdit level header\n";
 
-sub objsummary {
-       my ($objects) = @_;
-       my @objtypes = map { $_->{type} } @$objects;
-       my %count;
-       $count{$_}++ for @objtypes;
-       return (@objtypes > 1 && keys %count == 1 && 'all ') . join(', ',
-               map { $OBJTYPE[$_] ? $OBJTYPE[$_] . ($count{$_} > 1 && 's') : $_ }
-               sort keys %count
-       );
+       if ($opt{version}) {
+               warn "Override version $version to $opt{version}\n";
+               $version = $opt{version};
+       }
+       elsif ($version != $fileversion) {
+               warn "Unexpected version $version (expecting $fileversion)\n";
+       }
+       elsif ($version == 95) {
+               # auto-detect exact variant
+               if (ord substr($input, 70, 1) ~~ [1 .. 8]) {
+                       # valid sprite length instead of description byte
+                       # (which is usually a letter or nul)
+                       $version = 94;
+               }
+               elsif (ord substr($input, 147, 1) == 0) {
+                       # nul of finish type is 2 bytes later (unlike first char of message)
+                       $version = 96;
+               }
+               warn "Ambiguous file version 95; guessing subversion $version\n";
+       };
+
+       $fileversion += 100 if $fileversion < 90;  # 93..95 came before 50..53
+       given ($fileversion) {
+               when (153) { } # current @FORMAT
+                       $FORMAT[7] = 'Ca64'; # no reserved space after description
+                       $FORMAT[15]->[-1] = 'Ca255'; # enddata
+                       $FORMAT[-1]->[-1]->[0] = '32C'; # less objects
+               when ($version == 96) {
+                       ref $_ and $_->[-1] = 'C' for @{ $FORMAT[11] }; # 9 moderefs
+               }
+                       ref $_ and pop @$_ for @{ $FORMAT[11] }; # only 8 moderefs
+                       splice @FORMAT, 6, 2 if $version <= 94;  # earlier version without description
+               when (95) { }
+                       splice @{ $FORMAT[7] }, 4, 2;  # no race
+                       splice @{ $FORMAT[13] }, 4, 2; # no enddata
+                       splice @{ $FORMAT[-1] }, 1, 2; # no name
+               when (94) { }
+                       splice @FORMAT, 14, 2; # no hiname
+                       $FORMAT[-1]->[0] = 64; # constant amount of levels
+               when (93) { }
+               default {
+                       die "Cannot parse data for Wormedit $fileversion/$version\n";
+               }
+       }
+
+       # convert to an easily accessible hash
+       my @values = unpack Shiar_Parse::Nested->template(\@FORMAT).'a*', $input;
+       my $data = Shiar_Parse::Nested->convert(\@FORMAT, \@values);
+       warn "Trailing data left unparsed\n" if grep {length} @values;
+       $data->{format} = 'WormEdit';
+       return $data;
 }
 
-# read and parse all input data
-local $/;
-my @rawdata = unpack Shiar_Parse::Nested->template(\@FORMAT).'a*', readline;
-$rawdata[0] eq $MAGICID
-       or die "File does not match WormEdit level header\n";
-$rawdata[1] == 53
-       or warn "Unsupported version $rawdata[1] (expecting 53)\n";
 
-# convert to an easily accessible hash
-my $data = Shiar_Parse::Nested->convert(\@FORMAT, \@rawdata);
-warn "Trailing data left unparsed\n" if grep {length} @rawdata;
+package Shiar_Parse::WormyLevel;
 
-# output with user-preferred formatting
-if ($opt{raw}) {
-       require JSON::XS;
-       my $output = JSON::XS->new->ascii->canonical->pretty->allow_nonref;
-       print $output->encode($data), "\n";
-}
-else {
-       print "$data->{name} ($data->{description})\n";
-       printf "File version: %s\n", "WormEdit v$data->{version}";
-       printf "Defaults: %s\n", join('; ',
-               'sprite ' . scalar @{ $data->{sprite} },
-               'hiscore by ' . $data->{hiname},
+use strict;
+use warnings;
+
+use List::Util qw(sum min max);
+use Data::Dumper;
+
+sub read {
+       my ($self, $input) = @_;
+       my ($psize, $ptype, $size, $type, $vsize, $dsize, $id, $version) = unpack q{
+               x11 x42    # file signature and comment
+               S a2 S a2  # file size, type; data size, type
+               x8         # var name
+               S S        # var size; content size
+               CC         # wormy header
+       }, $input;
+       $ptype eq "\014\000"
+               or die "Not a calculator string, thus cannot be a Wormy level file\n";
+       $size == $psize - 16
+               or warn "File size ($size) does not correspond with data size ($psize)\n";
+       $type eq "\014\010"
+               or die "Not a calculator string, thus cannot be a Wormy level file\n";
+       $size == $vsize and $vsize == $dsize+2
+               or warn "Mismatch in string data size declarations\n";
+#      substr($input, -2) eq $CHECKSUM
+
+       $input = substr $input, 73, -2;
+       $id eq ord 'w'
+               or die "Wormy level identifier not found\n";
+
+       if ($opt{version}) {
+               warn "Override version $version to $opt{version}\n";
+               $version = $opt{version};
+       }
+       elsif ($version == 95) {
+               # level offset instead of description byte
+               $version-- if (unpack('x2Z*x2xC', $input))[1] == 0xF4;
+               warn "Ambiguous file version 95; guessing subversion $version\n";
+       }
+
+       my @FORMAT = (
+               magic       => 'a1',
+               version     => 'C',
+               name        => 'Z*',
+               description => 'Z*',
+               levelcount  => [1,
+                       total  => 'S',
+               ],
+               moderef     => [1,
+                       map { (
+                               offset => [1, map {$_ => 'S'} @$_], # byte location of start
+                               end    => [1, map {$_ => 'C'} @$_],
+                       ) }
+                       [qw/single peaworm tron deathmatch foodmatch multifood timematch race ctf/]
+               ],
+               theanswer => 'C', # 42
+               sprite     => ['C',
+                       line => 'B8',
+               ],
+               leveldata => 'a*',
+       );
+       my @LEVELFORM = (
+               peas       => 'C',
+               delay      => 'C',
+               growth     => 'C',
+               bsize      => 'C',
+               sprite     => ['C',
+                       line => 'B8',
+               ],
+               balls      => ['C',
+                       y   => 'C',
+                       x   => 'C',
+                       dir => 'C',
+               ],
+               worms      => [1,
+                       d => 'C',
+                       y => 'C',
+                       x => 'C',
+               ],
+               width      => 'C',
+               height     => 'C',
+               flags      => [0,
+                       y => 'C',
+                       x => 'C',
+               ],
+               #levels
+               #finish code
+               #levels-multi
+               #hinames
+       );
+       my @OBJECTFORM = (
+                       type => 'C',
+                       x1   => 'C',
+                       y1   => 'C',
+                       x2   => 'C',
+                       y2   => 'C',
        );
+       my $offsetbase = 0xF080;
 
-       my $startnr = 0;
-       for my $variant (qw/single multi race ctf/) {
-               print "\n";
-               my $count = $data->{levelcount}->{$variant};
-               printf "\u$variant ($count)";
-               $count or next;
-               print ":";
-               printf("\n- %-22s%4s:%3s+%2s%3s %3sx%-3s%s",
-                       $_->{id} || $_->{name},
-                       @$_{qw/size bsize growth/},
-                       $variant eq 'single' && "x$_->{peas}",
-                       @$_{qw/width height/},
-                       join(';', map {" $_"} grep {$_}
-                               @{$_->{objects}} && sprintf('%2d object%s (%s)',
-                                       scalar @{$_->{objects}}, @{$_->{objects}} != 1 && 's',
-                                       objsummary($_->{objects}),
+       given ($version) {
+               when (97) {
+                       # current @FORMAT
+               }
+                       $offsetbase = 0xF400;
+               when (96) {}
+                       ref $_ and splice(@$_, -8, 2) for @{ $FORMAT[11] }; # no multifood
+                       splice @FORMAT, 12, 2;  # no reserved byte
+               when (95) {}
+                       splice @FORMAT, 6, 2;  # no description
+               when (94) {}
+               when (90) {
+                       $FORMAT[5] = 'C/a';  # length-preceding name
+                       splice @FORMAT, 10, 2;  # no default sprite
+                       ref $_ and do {
+                               $_->[5] = $_->[7];  # no tron; deathmatch instead
+                               $_->[7] = $_->[9];  # foodmatch instead
+                               $_->[9] = 'linkmatch';  # replaces timematch
+                               $_->[11] = $_->[13];  # race
+                               $_->[13] = $_->[15];  # ctf
+                               $_->[15] = 'domination';
+                       } for @{ $FORMAT[9] }; # no multifood
+                       push @LEVELFORM, "object$_" => ['C',
+                               map {$_ => 'C'} qw(x1 y1 x2 y2)
+                       ] for qw(lines boxes);
+               }
+               default {
+                       die "Unsupported level version $version\n";
+               }
+       }
+
+       my $data = Shiar_Parse::Nested->unpack(\@FORMAT, $input);
+       my $offset = 0;
+       $offsetbase += 1 + @{ $data->{sprite} } if $data->{sprite};
+       $data->{moderef}->{offset}->{single} == $offsetbase
+               or warn "First singleplayer level is not in front\n";
+
+       my $slots = sum(
+               $data->{moderef}->{end}->{single} > 0,  # singleplayer slot if any levels
+               $data->{moderef}->{end}->{peaworm},     # one for each peaworm arena
+               $data->{moderef}->{end}->{tron},        # idem for tron
+       );
+       $data->{hinames} = [ unpack '(x2a3)*', substr($data->{leveldata}, -5 * $slots) ];
+       $data->{format} = '86s';
+
+       my @VARMODES = (
+               [qw'single  single'],
+               [qw'multi   peaworm tron deathmatch foodmatch multifood timematch'],
+               [qw'race    race'],
+               [qw'ctf     ctf'],
+       );
+
+       $data->{levels} = [];
+       for my $modes (@VARMODES) {
+               my $variant = shift @$modes;
+               my @modeoffsets = grep {defined} #TODO: comment
+                       map { $data->{moderef}->{offset}->{$_} } @$modes;
+               @modeoffsets or next;
+               $data->{levelcount}->{$variant} = 0;
+               $offset = min(grep {$_} @modeoffsets) or next;
+               $offset -= $offsetbase;
+               my $amount = $variant eq 'single' ? 100
+                       : max(grep {defined} map { $data->{moderef}->{end}->{$_} } @$modes);
+
+               my @varform = @LEVELFORM;
+               $varform[13]->[0] = $variant eq 'single' ? 1 : 4;
+               unshift @varform, name => 'Z*' unless $variant eq 'single' or $version <= 91;
+               $varform[-1]->[0] = 1 if $variant eq 'race' and $version > 91;
+               $varform[-1]->[0] = 2 if $variant eq 'ctf';
+
+               while ($offset < length $data->{leveldata}) {
+                       last if substr($data->{leveldata}, $offset, 1) eq chr(255);
+
+                       # find references to this level offset, and set start number to matching modes
+                       while (my ($mode, $location) = each %{ $data->{moderef}->{offset} }) {
+                               $location == $offset + $offsetbase or next;
+                               $data->{moderef}->{start}->{$mode} = 1 + scalar @{ $data->{levels} };
+                       }
+
+                       my $level = Shiar_Parse::Nested->unpack(
+                               [@varform], substr $data->{leveldata}, $offset
+                       );
+                       my $size = 8  # unpack length (ugh, ugly recalculation)
+                               + (defined $level->{name} ? 1 + length $level->{name} : 0)
+                               + 3 * (ref $level->{worms} eq 'ARRAY' ? scalar @{$level->{worms}} : 1)
+                               + 2 * ($level->{flags} ? ref $level->{flags} eq 'ARRAY' ? scalar @{$level->{flags}} : 1 : 0)
+                               + ($level->{sprite} ? scalar @{$level->{sprite}} : 0)
+                               + ($level->{balls} ? 3 * scalar @{$level->{balls}} : 0);
+                       $level->{size} = $size;
+                       $level->{offset} = $offset + $offsetbase;
+
+                       # add objects until terminator
+                       $level->{objects} = [];
+               if ($version <= 91) {
+                       push @{ $level->{objects} }, { %$_, type => 2 } for map { $level->{$_} ? @{ $level->{$_} } : () } qw(objectlines);
+                       push @{ $level->{objects} }, { %$_, type => 3 } for map { $level->{$_} ? @{ $level->{$_} } : () } qw(objectboxes);
+                       $size += 1 + 4 * scalar @{ $level->{objects} };
+               }
+               else {
+                       while (my $object = ord substr($data->{leveldata}, $offset+$size, 1)) {
+                               push @{ $level->{objects} }, Shiar_Parse::Nested->unpack(
+                                       [@OBJECTFORM], substr($data->{leveldata}, $offset+$size, 5)
+                               );
+                               $size += 5;
+                       }
+               }
+
+                       # add parsed level and advance
+                       push @{ $data->{levels} }, $level;
+                       $offset += ++$size;
+                       last if ++$data->{levelcount}->{$variant} >= $amount;
+               }
+
+               if ($variant eq 'single') {
+                       $offset++;
+                       $data->{finish}->{code} =
+                       my $code = substr $data->{leveldata}, $offset, -5*$slots;
+
+                       my %FINISHCODE = (
+                               0 => chr 0xC9, # ret
+                               1 => join('',
+                                       chr 0x21,  # ld hl, MESSAGE
+                                       pack('v', $offsetbase + $offset + 9),
+                                       (map {chr}
+                                               0xCD, 0x37, 0x4A,  # call _puts
+                                               0xC3, 0xAA, 0x55,  # jp _getkey
+                                       ),
                                ),
-                               @{$_->{sprite}} && sprintf('sprite %d',
-                                       scalar @{$_->{sprite}},
+                               2 => join('',
+                                       (map {chr}
+                                               0x21, 0, 0x1C,  # ld hl, $POS
+                                               0x22, 0x7C, 0xC3, # ld (_penCol), hl
+                                               0x21,  # ld hl, MESSAGE
+                                       ),
+                                       pack('v', $offsetbase + $offset + 15),
+                                       (map {chr}
+                                               0xCD, 0xA5, 0x4A, # call _vputs
+                                               0xC3, 0xAA, 0x55, # jp _getkey
+                                       ),
                                ),
-                       ),
-               ) for map { $data->{levels}->[$_ + $startnr] }
-                       0 .. $count - 1;
-               $startnr += $count;
-       }
-       continue {
-               print "\n";
-               printf("-- %-21s%4d: %s (%s)\n",
-                       '(ending)',
-                       length $data->{enddata},
-                       $ENDTYPE[$data->{endtype}] || 'unknown', $data->{endstr},
-               ) if $variant eq 'single';
+                       );
+                       while (my ($finish, $match) = each %FINISHCODE) {
+                               $match eq substr $code, 0, length $match or next;
+                               $data->{finish}->{type} = $finish and
+                               $data->{finish}->{message} = unpack 'Z*', substr($code, length $match);
+                               last;
+                       }
+               }
        }
+
+       return $data;
 }
 
+
 package Shiar_Parse::Nested;
 
 sub template {
@@ -187,25 +441,139 @@ sub convert {
                        my $max = $count =~ s/^(\d+)// ? $1 : 0;
                        $count = !$count ? $max
                                : $count eq '*' ? $res{levelcount}->{total} : shift @$data;
-                       $max ||= $count;
-                       $res{$field}->[$_] = $self->convert([@subformat], $data) for 0 .. $max-1;
+                       $res{$field}->[$_] = $self->convert([@subformat], $data) for 0 .. ($max || $count)-1;
                        splice @{ $res{$field} }, $count if $max > $count;
                        $res{$field} = $res{$field}->[0] if $max == 1;
                        next;
                }
                elsif ($template =~ /^Ca/) {
-                       $data->[0] = unpack 'C/a', $data->[0];
+                       $data->[0] = CORE::unpack 'C/a', $data->[0];
                }
                $res{$field} = shift @$data;
        }
        return \%res;
 }
 
+sub unpack {
+       my ($self, $format, $input) = @_;
+       my @data = CORE::unpack $self->template($format), $input;
+       return $self->convert($format, \@data);
+}
+
+
+package main;
+
+my @OBJTYPE = ('none', 'line', 'fat line', 'bar', 'circle');
+my @ENDTYPE = ('none', 'message', 'small message');
+
+sub objsummary {
+       my ($objects) = @_;
+       my @objtypes = map { $_->{type} } @$objects;
+       my %count;
+       $count{$_}++ for @objtypes;
+       return (@objtypes > 1 && keys %count == 1 && 'all ') . join(', ',
+               map { $OBJTYPE[$_] ? $OBJTYPE[$_] . ($count{$_} > 1 && 's') : $_ }
+               sort keys %count
+       );
+}
+
+# read and parse all input data
+my $data;
+local $/;
+my $rawdata = readline;
+if (substr($rawdata, 0, 11) eq "**TI86**\032\012\000") {
+       # compiled calculator file
+       $data = Shiar_Parse::WormyLevel->read($rawdata);
+}
+elsif (substr($rawdata, 0, 8) eq 'WormEdit') {
+       # original wormedit source
+       $data = Shiar_Parse::WormEdit->read($rawdata);
+}
+else {
+       die "Unrecognised file type\n";
+}
+
+# output with user-preferred formatting
+if ($opt{raw}) {
+       # full data in yaml (human-readable) formatting
+       require YAML;
+       local $YAML::CompressSeries;
+             $YAML::CompressSeries = 0;
+       my $yml = "# Wormy levelset\n" . YAML::Dump($data);
+
+       # inline format of short hashes
+       $yml =~ s{
+               ^(\ *) - \n                          # array indicator
+               ((?:\1\ \ [a-z0-9]{1,5}:\ *\d+\n)+)  # simple hash declaration
+               (?!\1\ )                             # no further children
+       }[
+               my ($indent, $value) = ($1, $2);
+               chop $value;
+               $value =~ s/^ +//gm;
+               $value =~ s/\n/, /g;
+               "$indent- {$value}\n";
+       ]egmx;
+
+       print $yml;
+}
+else {
+       print $data->{name};
+       print " ($data->{description})" if defined $data->{description};
+       print "\n";
+       printf "File version: %s\n", "$data->{format} v$data->{version}";
+       printf "Defaults: %s\n", join('; ',
+               $data->{sprite} ? 'sprite ' . scalar @{ $data->{sprite} } : (),
+               defined $data->{hiname} ? 'hiscore by ' . $data->{hiname} : (),
+       );
+
+       my $startnr = 0;
+       for my $variant (qw/single multi race ctf/) {
+               my $count = $data->{levelcount}->{$variant};
+               print "\n";
+               printf '%s (%s)', ucfirst $variant, $count // 'invalid';
+               $count or next;
+               print ":";
+               for (0 .. $count - 1) {
+                       my $level = $data->{levels}->[$_ + $startnr];
+                       printf("\n- %-22s%4s:%3s+%2s%3s %3sx%-3s%s",
+                               $level->{id} || $level->{name} || '#'.($_+1),
+                               @$level{qw/size bsize growth/},
+                               $variant eq 'single' && "x$level->{peas}",
+                               @$level{qw/width height/},
+                               join(';', map {" $_"} grep {$_}
+                                       @{$level->{objects}} && sprintf('%2d object%s (%s)',
+                                               scalar @{$level->{objects}}, @{$level->{objects}} != 1 && 's',
+                                               objsummary($level->{objects}),
+                                       ),
+                                       $level->{sprite} && @{$level->{sprite}} && sprintf('sprite %d',
+                                               scalar @{$level->{sprite}},
+                                       ),
+                                       $level->{balls} && @{$level->{balls}} && sprintf('%d bounc%s',
+                                               scalar @{$level->{balls}}, @{$level->{balls}} == 1 ? 'y' : 'ies',
+                                       ),
+                               ),
+                       );
+               }
+               $startnr += $count;
+       }
+       continue {
+               print "\n";
+               printf("-- %-21s%4s: %s (%s)\n",
+                       '(ending)',
+                       defined $data->{finish}->{code}
+                               ? length $data->{finish}->{code} : '?',
+                       defined $data->{finish}->{type}
+                               ? $ENDTYPE[$data->{finish}->{type}] || 'unknown' : 'code',
+                       $data->{finish}->{message} // '?',
+               ) if $variant eq 'single';
+       }
+}
+
 __END__
 
 =head1 NAME
 
-parse-wormedit - WormEdit level data parser
+parse-wormedit - Wormy level data parser
 
 =head1 SYNOPSIS
 
@@ -213,8 +581,8 @@ parse-wormedit - WormEdit level data parser
 
 =head1 DESCRIPTION
 
-Reads WormEdit v0.53 levels from STDIN or given file,
-and outputs contents, summarised or in full.
+Reads Wormy levels (either original WormEdit source or compiled TI-86 string)
+from STDIN or given file, and outputs contents, summarised or in full.
 
 =head1 AUTHOR