create_hash_table

   1 #! /usr/bin/perl -w
   2 #
   3 # Static Hashtable Generator
   4 #
   5 # (c) 2000-2002 by Harri Porten <porten@kde.org> and
   6 #                  David Faure <faure@kde.org>
   7 # Modified (c) 2004 by Nikolas Zimmermann <wildfox@kde.org>
   8 # Copyright (C) 2007, 2008, 2009 Apple Inc. All rights reserved.
   9 #
  10 # This library is free software; you can redistribute it and/or
  11 # modify it under the terms of the GNU Lesser General Public
  12 # License as published by the Free Software Foundation; either
  13 # version 2 of the License, or (at your option) any later version.
  14 #
  15 # This library is distributed in the hope that it will be useful,
  16 # but WITHOUT ANY WARRANTY; without even the implied warranty of
  17 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  18 # Lesser General Public License for more details.
  19 #
  20 # You should have received a copy of the GNU Lesser General Public
  21 # License along with this library; if not, write to the Free Software
  22 # Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301  USA
  23 #
  24
  25 use strict;
  26
  27 my $file = $ARGV[0];
  28 shift;
  29 my $includelookup = 0;
  30
  31 # Use -i as second argument to make it include "Lookup.h"
  32 $includelookup = 1 if (defined($ARGV[0]) && $ARGV[0] eq "-i");
  33
  34 # Use -n as second argument to make it use the third argument as namespace parameter ie. -n KDOM
  35 my $useNameSpace = $ARGV[1] if (defined($ARGV[0]) && $ARGV[0] eq "-n");
  36
  37 print STDERR "Creating hashtable for $file\n";
  38 open(IN, $file) or die "No such file $file";
  39
  40 my @keys = ();
  41 my @attrs = ();
  42 my @values = ();
  43 my @hashes = ();
  44 my @table = ();
  45 my @links = ();
  46
  47 my $hasSetter = "false";
  48
  49 my $inside = 0;
  50 my $name;
  51 my $pefectHashSize;
  52 my $compactSize;
  53 my $compactHashSizeMask;
  54 my $banner = 0;
  55 sub calcPerfectHashSize();
  56 sub calcCompactHashSize();
  57 sub output();
  58 sub jsc_ucfirst($);
  59 sub hashValue($);
  60
  61 while (<IN>) {
  62     chomp;
  63     s/^\s+//;
  64     next if /^\#|^$/; # Comment or blank line. Do nothing.
  65     if (/^\@begin/ && !$inside) {
  66         if (/^\@begin\s*([:_\w]+)\s*\d*\s*$/) {
  67             $inside = 1;
  68             $name = $1;
  69         } else {
  70             print STDERR "WARNING: \@begin without table name, skipping $_\n";
  71         }
  72     } elsif (/^\@end\s*$/ && $inside) {
  73         calcPerfectHashSize();
  74         calcCompactHashSize();
  75         output();
  76
  77         @keys = ();
  78         @attrs = ();
  79         @values = ();
  80         @hashes = ();
  81         @table = ();
  82         @links = ();
  83
  84         $inside = 0;
  85     } elsif (/^(\S+)\s*(\S+)\s*([\w\|]*)\s*(\w*)\s*$/ && $inside) {
  86         my $key = $1;
  87         my $val = $2;
  88         my $att = $3;
  89         my $param = $4;
  90
  91         push(@keys, $key);
  92         push(@attrs, length($att) > 0 ? $att : "0");
  93
  94         if ($att =~ m/Function/) {
  95             push(@values, { "type" => "Function", "function" => $val, "params" => (length($param) ? $param : "") });
  96             #printf STDERR "WARNING: Number of arguments missing for $key/$val\n" if (length($param) == 0);
  97         } elsif (length($att)) {
  98             my $get = $val;
  99             my $put = "0";
 100             if (!($att =~ m/ReadOnly/)) {
 101                 $put = "set" . jsc_ucfirst($val);
 102             }
 103             $hasSetter = "true";
 104             push(@values, { "type" => "Property", "get" => $get, "put" => $put });
 105         } else {
 106             push(@values, { "type" => "Lexer", "value" => $val });
 107         }
 108         push(@hashes, hashValue($key));
 109     } elsif ($inside) {
 110         die "invalid data {" . $_ . "}";
 111     }
 112 }
 113
 114 die "missing closing \@end" if ($inside);
 115
 116 sub jsc_ucfirst($)
 117 {
 118     my ($value) = @_;
 119
 120     if ($value =~ /js/) {
 121         $value =~ s/js/JS/;
 122         return $value;
 123     }
 124
 125     return ucfirst($value);
 126 }
 127
 128
 129 sub ceilingToPowerOf2
 130 {
 131     my ($pefectHashSize) = @_;
 132
 133     my $powerOf2 = 1;
 134     while ($pefectHashSize > $powerOf2) {
 135         $powerOf2 <<= 1;
 136     }
 137
 138     return $powerOf2;
 139 }
 140
 141 sub calcPerfectHashSize()
 142 {
 143 tableSizeLoop:
 144     for ($pefectHashSize = ceilingToPowerOf2(scalar @keys); ; $pefectHashSize += $pefectHashSize) {
 145         my @table = ();
 146         foreach my $key (@keys) {
 147             my $h = hashValue($key) % $pefectHashSize;
 148             next tableSizeLoop if $table[$h];
 149             $table[$h] = 1;
 150         }
 151         last;
 152     }
 153 }
 154
 155 sub leftShift($$) {
 156     my ($value, $distance) = @_;
 157     return (($value << $distance) & 0xFFFFFFFF);
 158 }
 159
 160 sub calcCompactHashSize()
 161 {
 162     my $compactHashSize = ceilingToPowerOf2(2 * @keys);
 163     $compactHashSizeMask = $compactHashSize - 1;
 164     $compactSize = $compactHashSize;
 165     my $collisions = 0;
 166     my $maxdepth = 0;
 167     my $i = 0;
 168     foreach my $key (@keys) {
 169         my $depth = 0;
 170         my $h = hashValue($key) % $compactHashSize;
 171         while (defined($table[$h])) {
 172             if (defined($links[$h])) {
 173                 $h = $links[$h];
 174                 $depth++;
 175             } else {
 176                 $collisions++;
 177                 $links[$h] = $compactSize;
 178                 $h = $compactSize;
 179                 $compactSize++;
 180             }
 181         }
 182         $table[$h] = $i;
 183         $i++;
 184         $maxdepth = $depth if ( $depth > $maxdepth);
 185     }
 186 }
 187
 188 # Paul Hsieh's SuperFastHash
 189 # http://www.azillionmonkeys.com/qed/hash.html
 190 sub hashValue($) {
 191   my @chars = split(/ */, $_[0]);
 192
 193   # This hash is designed to work on 16-bit chunks at a time. But since the normal case
 194   # (above) is to hash UTF-16 characters, we just treat the 8-bit chars as if they
 195   # were 16-bit chunks, which should give matching results
 196
 197   my $EXP2_32 = 4294967296;
 198
 199   my $hash = 0x9e3779b9;
 200   my $l    = scalar @chars; #I wish this was in Ruby --- Maks
 201   my $rem  = $l & 1;
 202   $l = $l >> 1;
 203
 204   my $s = 0;
 205
 206   # Main loop
 207   for (; $l > 0; $l--) {
 208     $hash   += ord($chars[$s]);
 209     my $tmp = leftShift(ord($chars[$s+1]), 11) ^ $hash;
 210     $hash   = (leftShift($hash, 16)% $EXP2_32) ^ $tmp;
 211     $s += 2;
 212     $hash += $hash >> 11;
 213     $hash %= $EXP2_32;
 214   }
 215
 216   # Handle end case
 217   if ($rem != 0) {
 218     $hash += ord($chars[$s]);
 219     $hash ^= (leftShift($hash, 11)% $EXP2_32);
 220     $hash += $hash >> 17;
 221   }
 222
 223   # Force "avalanching" of final 127 bits
 224   $hash ^= leftShift($hash, 3);
 225   $hash += ($hash >> 5);
 226   $hash = ($hash% $EXP2_32);
 227   $hash ^= (leftShift($hash, 2)% $EXP2_32);
 228   $hash += ($hash >> 15);
 229   $hash = $hash% $EXP2_32;
 230   $hash ^= (leftShift($hash, 10)% $EXP2_32);
 231
 232   # Save 8 bits for StringImpl to use as flags.
 233   $hash &= 0xffffff;
 234
 235   # This avoids ever returning a hash code of 0, since that is used to
 236   # signal "hash not computed yet". Setting the high bit maintains
 237   # reasonable fidelity to a hash code of 0 because it is likely to yield
 238   # exactly 0 when hash lookup masks out the high bits.
 239   $hash = (0x80000000 >> 8) if ($hash == 0);
 240
 241   return $hash;
 242 }
 243
 244 sub output() {
 245     if (!$banner) {
 246         $banner = 1;
 247         print "// Automatically generated from $file using $0. DO NOT EDIT!\n";
 248     }
 249
 250     my $nameEntries = "${name}Values";
 251     $nameEntries =~ s/:/_/g;
 252     my $nameIndex = "${name}Index";
 253     $nameIndex =~ s/:/_/g;
 254
 255     print "\n#include \"JSCBuiltins.h\"\n";
 256     print "\n#include \"Lookup.h\"\n" if ($includelookup);
 257
 258     if ($useNameSpace) {
 259         print "\nnamespace ${useNameSpace} {\n";
 260         print "\nusing namespace JSC;\n";
 261     } else {
 262         print "\nnamespace JSC {\n";
 263     }
 264
 265     print "\nstatic const struct CompactHashIndex ${nameIndex}\[$compactSize\] = {\n";
 266     for (my $i = 0; $i < $compactSize; $i++) {
 267         my $T = -1;
 268         if (defined($table[$i])) { $T = $table[$i]; }
 269         my $L = -1;
 270         if (defined($links[$i])) { $L = $links[$i]; }
 271         print "    { $T, $L },\n";
 272     }
 273     print "};\n\n";
 274
 275     my $packedSize = scalar @keys;
 276     print "\nstatic const struct HashTableValue ${nameEntries}\[$packedSize\] = {\n";
 277     my $i = 0;
 278     foreach my $key (@keys) {
 279         my $firstValue = "";
 280         my $secondValue = "";
 281         my $firstCastStr = "";
 282         my $secondCastStr = "";
 283
 284         if ($values[$i]{"type"} eq "Function") {
 285             $firstCastStr = "static_cast<NativeFunction>";
 286             $firstValue = $values[$i]{"function"};
 287             $secondValue = $values[$i]{"params"};
 288         } elsif ($values[$i]{"type"} eq "Property") {
 289             $firstCastStr = "static_cast<PropertySlot::GetValueFunc>";
 290             $secondCastStr = "static_cast<PutPropertySlot::PutValueFunc>";
 291             $firstValue = $values[$i]{"get"};
 292             $secondValue = $values[$i]{"put"};
 293         } elsif ($values[$i]{"type"} eq "Lexer") {
 294             $firstValue = $values[$i]{"value"};
 295             $secondValue = "0";
 296         }
 297
 298         my $intrinsic = "NoIntrinsic";
 299         $intrinsic = "FromCharCodeIntrinsic" if ($key eq "fromCharCode");
 300         if ($name eq "arrayPrototypeTable") {
 301             $intrinsic = "ArrayPushIntrinsic" if ($key eq "push");
 302             $intrinsic = "ArrayPopIntrinsic" if ($key eq "pop");
 303         }
 304         if ($name eq "regExpPrototypeTable") {
 305             $intrinsic = "RegExpExecIntrinsic" if ($key eq "exec");
 306             $intrinsic = "RegExpTestIntrinsic" if ($key eq "test");
 307         }
 308
 309         if ($values[$i]{"type"} eq "Function")  {
 310             my $tableHead = $name;
 311             $tableHead =~ s/Table$//;
 312             print " #if JSC_BUILTIN_EXISTS(" . uc($tableHead . $key) .")\n";
 313             print "   { \"$key\", (($attrs[$i]) & ~Function) | Builtin, $intrinsic, (intptr_t)static_cast<BuiltinGenerator>(" . $tableHead . ucfirst($key) . "CodeGenerator), (intptr_t)$secondValue },\n";
 314             print " #else\n"
 315         }
 316         print "   { \"$key\", $attrs[$i], $intrinsic, (intptr_t)" . $firstCastStr . "($firstValue), (intptr_t)" . $secondCastStr . "($secondValue) },\n";
 317         if ($values[$i]{"type"} eq "Function")  {
 318             print " #endif\n"
 319         }
 320         $i++;
 321     }
 322     print "};\n\n";
 323     print "extern const struct HashTable $name =\n";
 324     print "    \{ $packedSize, $compactHashSizeMask, $hasSetter, $nameEntries, 0, $nameIndex \};\n";
 325     print "} // namespace\n";
 326 }