| 1 | #====================================================================================== |
|---|
| 2 | # |
|---|
| 3 | # Project name : XTIDE Universal BIOS |
|---|
| 4 | # |
|---|
| 5 | # Authors : Greg Lindhorst |
|---|
| 6 | # gregli@hotmail.com |
|---|
| 7 | # |
|---|
| 8 | # Description : Script for compiling and compressing strings for |
|---|
| 9 | # use by DisplayFormatCompressed.asm. See the header of that file |
|---|
| 10 | # for a description of the compression scheme. |
|---|
| 11 | # |
|---|
| 12 | # Usage : stdin: Listing of strings.asm, |
|---|
| 13 | # assembled with MODULE_STRINGS_COMPRESSED_PRECOMPRESS. |
|---|
| 14 | # We used the listing so that the assembler can take care of |
|---|
| 15 | # resolving %define and EQU symbol definitions. |
|---|
| 16 | # |
|---|
| 17 | # stdout: StringsCompressed.asm, |
|---|
| 18 | # plug replacement for Strings.asm (included by Main.asm) |
|---|
| 19 | # |
|---|
| 20 | # Also see the XTIDE makefile for building StringsCompressed.asm |
|---|
| 21 | # |
|---|
| 22 | |
|---|
| 23 | #---------------------------------------------------------------------- |
|---|
| 24 | # |
|---|
| 25 | # Translated, Format, and "Normal" characters |
|---|
| 26 | # |
|---|
| 27 | # DisplayFormatCompressed can only deal with characters in one of the following categories: |
|---|
| 28 | # 1. Those in the Translate associative array |
|---|
| 29 | # 2. Those in the Format associative array |
|---|
| 30 | # 3. Characters between $normal_base and $normal_base+0x40 |
|---|
| 31 | # (typically covers upper and lower case alphabets) |
|---|
| 32 | # 4. Null characters (marking the end of strings) |
|---|
| 33 | # 5. The special string LF,CR |
|---|
| 34 | # |
|---|
| 35 | # If a character or format read at the input cannot be found in one of the above categories, |
|---|
| 36 | # it must be added here before this script will accept it (and DisplayFormatCompressed can |
|---|
| 37 | # display it). |
|---|
| 38 | # |
|---|
| 39 | # Tables for the above categories are expected in the input stream, before string to be |
|---|
| 40 | # compressed are provided. Note that these tables are not present in DisplayFormatCompressed, |
|---|
| 41 | # and do not need to be updated there. Needed information is put in the compression output |
|---|
| 42 | # that it reads. |
|---|
| 43 | # |
|---|
| 44 | |
|---|
| 45 | # |
|---|
| 46 | # High order code bits, determining which type of character we have (translated or not) and |
|---|
| 47 | # if a space or null should come after this character. |
|---|
| 48 | # |
|---|
| 49 | $code_space = 0xc0; |
|---|
| 50 | $code_null = 0x80; |
|---|
| 51 | $code_normal = 0x40; |
|---|
| 52 | $code_translate = 0x00; |
|---|
| 53 | |
|---|
| 54 | # |
|---|
| 55 | # Bit used if it is a translated byte |
|---|
| 56 | # |
|---|
| 57 | $code_translate_null = 0x00; |
|---|
| 58 | $code_translate_normal = 0x20; |
|---|
| 59 | |
|---|
| 60 | print ";;;======================================================================\n"; |
|---|
| 61 | print ";;;\n"; |
|---|
| 62 | print ";;; This file is generated by StringsCompress.pl from source in Strings.asm\n"; |
|---|
| 63 | print ";;; DO NOT EDIT DIRECTLY - See the makefile for how to rebuild this file.\n"; |
|---|
| 64 | print ";;; This file only needs to be rebuilt if Strings.asm is changed.\n"; |
|---|
| 65 | print ";;;\n"; |
|---|
| 66 | print ";;;======================================================================\n\n"; |
|---|
| 67 | |
|---|
| 68 | |
|---|
| 69 | # |
|---|
| 70 | # On a first pass, look for our table directives. $translate{...}, $format{...}, etc. |
|---|
| 71 | # are expected in the input stream. |
|---|
| 72 | # |
|---|
| 73 | $processed = " [StringsCompress Processed]"; |
|---|
| 74 | while(<>) |
|---|
| 75 | { |
|---|
| 76 | chop; |
|---|
| 77 | $o = $_; |
|---|
| 78 | |
|---|
| 79 | # |
|---|
| 80 | # Table entries for this script |
|---|
| 81 | # |
|---|
| 82 | if( /^\s*\d+\s*(\;\$translate\{\s*ord\(\s*'(.)'\s*\)\s*\}\s*=\s*([0-9]+).*$)/ ) |
|---|
| 83 | { |
|---|
| 84 | $translate{ord($2)} = int($3); |
|---|
| 85 | $o .= $processed; |
|---|
| 86 | } |
|---|
| 87 | elsif( /^\s*\d+\s*(\;\$translate\{\s*([0-9]+)\s*\}\s*=\s*([0-9]+).*$)/ ) |
|---|
| 88 | { |
|---|
| 89 | $translate{int($2)} = int($3); |
|---|
| 90 | $o .= $processed; |
|---|
| 91 | } |
|---|
| 92 | elsif( /^\s*\d+\s*(\;\$format_begin\s*=\s*([0-9]+).*$)/ ) |
|---|
| 93 | { |
|---|
| 94 | $format_begin = int($2); |
|---|
| 95 | $o .= $processed; |
|---|
| 96 | } |
|---|
| 97 | elsif( /^\s*\d+\s*(\;\$format\{\s*\"([^\"]+)\"\s*\}\s*=\s*([0-9]+).*$)/ ) |
|---|
| 98 | { |
|---|
| 99 | $format{$2} = int($3); |
|---|
| 100 | $o .= $processed; |
|---|
| 101 | } |
|---|
| 102 | elsif( /^\s*\d+\s*(\;\$normal_base\s*=\s*0x([0-9a-fA-F]+).*$)/ ) |
|---|
| 103 | { |
|---|
| 104 | $normal_base = hex($2); |
|---|
| 105 | $o .= $processed; |
|---|
| 106 | } |
|---|
| 107 | elsif( /^\s*\d+\s*(\;\$normal_base\s*=\s*([0-9]+).*$)/ ) |
|---|
| 108 | { |
|---|
| 109 | $normal_base = int($2); |
|---|
| 110 | $o .= $processed; |
|---|
| 111 | } |
|---|
| 112 | |
|---|
| 113 | push( @lines, $o ); |
|---|
| 114 | } |
|---|
| 115 | |
|---|
| 116 | # |
|---|
| 117 | # On the second pass, loop through lines of the listing, looking for 'db' lines |
|---|
| 118 | # (and dealing with continuations) and compressing each line as it is encountered. |
|---|
| 119 | # |
|---|
| 120 | for( $l = 0; $l < $#lines; $l++ ) |
|---|
| 121 | { |
|---|
| 122 | $_ = $lines[$l]; |
|---|
| 123 | |
|---|
| 124 | # |
|---|
| 125 | # The <number> indicates a line from an include file, do not include in the output |
|---|
| 126 | # |
|---|
| 127 | if( /^\s*\d+\s*\<\d\>/ ) |
|---|
| 128 | { |
|---|
| 129 | } |
|---|
| 130 | |
|---|
| 131 | # |
|---|
| 132 | # a 'db' line, with or without a label |
|---|
| 133 | # |
|---|
| 134 | elsif( /^\s*\d+\s[0-9A-F]+\s([0-9A-F]+)(-?)\s+([a-z0-9_]+:)?(\s+)(db\s+)(.*)/i ) |
|---|
| 135 | { |
|---|
| 136 | $bytes = $1; |
|---|
| 137 | $continuation = $2; |
|---|
| 138 | $label = $3; |
|---|
| 139 | $spacing = $4; |
|---|
| 140 | $db = $5; |
|---|
| 141 | $string = $6; |
|---|
| 142 | |
|---|
| 143 | print $label.$spacing."; ".$db.$string."\n"; |
|---|
| 144 | |
|---|
| 145 | if( $continuation eq "-" ) |
|---|
| 146 | { |
|---|
| 147 | do |
|---|
| 148 | { |
|---|
| 149 | $_ = $lines[++$l]; |
|---|
| 150 | /^\s*\d+\s[0-9A-F]+\s([0-9A-F]+)(\-?)/i || die "parse error on continuation: '".$_."'"; |
|---|
| 151 | $bytes .= $1; |
|---|
| 152 | $continuation = $2; |
|---|
| 153 | } |
|---|
| 154 | while( $continuation eq "-" ); |
|---|
| 155 | } |
|---|
| 156 | |
|---|
| 157 | &processString( $bytes, $label.$spacing, $db ); |
|---|
| 158 | } |
|---|
| 159 | |
|---|
| 160 | # |
|---|
| 161 | # a ';%%;' prefix line, copy to output without the prefix |
|---|
| 162 | # |
|---|
| 163 | elsif( /^\s*\d+\s*;%%;\s*(.*)$/ ) |
|---|
| 164 | { |
|---|
| 165 | print $1."\n"; |
|---|
| 166 | } |
|---|
| 167 | |
|---|
| 168 | # |
|---|
| 169 | # everything else, copy to the output as is |
|---|
| 170 | # |
|---|
| 171 | elsif( /^\s*\d+\s*(.*)$/ ) |
|---|
| 172 | { |
|---|
| 173 | print $1."\n"; |
|---|
| 174 | } |
|---|
| 175 | } |
|---|
| 176 | |
|---|
| 177 | print ";;; end of input stream\n\n"; |
|---|
| 178 | |
|---|
| 179 | #-------------------------------------------------------------------------------- |
|---|
| 180 | # |
|---|
| 181 | # Output constants and the TranslatesAndFormats table |
|---|
| 182 | # |
|---|
| 183 | |
|---|
| 184 | print "StringsCompressed_NormalBase equ ".$normal_base."\n\n"; |
|---|
| 185 | |
|---|
| 186 | print "StringsCompressed_FormatsBegin equ ".$format_begin."\n\n"; |
|---|
| 187 | |
|---|
| 188 | print "StringsCompressed_TranslatesAndFormats: \n"; |
|---|
| 189 | |
|---|
| 190 | foreach $f (keys(%translate)) |
|---|
| 191 | { |
|---|
| 192 | $translate_index[$translate{$f}] = $f; |
|---|
| 193 | $used{$f} || die "translate $f unused\n"; |
|---|
| 194 | $translate{$f} <= 31 || die $translate{$f}.": translate codes must be below 32"; |
|---|
| 195 | } |
|---|
| 196 | |
|---|
| 197 | for( $g = 0; $translate_index[$g]; $g++ ) |
|---|
| 198 | { |
|---|
| 199 | print " db ".$translate_index[$g]." ; ".$g."\n"; |
|---|
| 200 | } |
|---|
| 201 | |
|---|
| 202 | foreach $f (keys(%format)) |
|---|
| 203 | { |
|---|
| 204 | $n = $f; |
|---|
| 205 | $n =~ s/\-/_/g; |
|---|
| 206 | $format_index[$format{$f}] = "DisplayFormatCompressed_Format_".$n; |
|---|
| 207 | $used{$f} || die "format $f unused\n"; |
|---|
| 208 | $format{$f} <= 31 || die $format{$f}.": format codes must be below 32"; |
|---|
| 209 | } |
|---|
| 210 | |
|---|
| 211 | for( $t = $format_begin; $format_index[$t]; $t++ ) |
|---|
| 212 | { |
|---|
| 213 | print " db (DisplayFormatCompressed_BaseFormatOffset - ".$format_index[$t].") ; ".$t."\n"; |
|---|
| 214 | } |
|---|
| 215 | |
|---|
| 216 | print "\n"; |
|---|
| 217 | |
|---|
| 218 | # |
|---|
| 219 | # Ensure that branch targets are within reach |
|---|
| 220 | # |
|---|
| 221 | print "%ifndef CHECK_FOR_UNUSED_ENTRYPOINTS\n"; |
|---|
| 222 | for( $t = $format_begin; $format_index[$t]; $t++ ) |
|---|
| 223 | { |
|---|
| 224 | print "%if DisplayFormatCompressed_BaseFormatOffset < $format_index[$t] || DisplayFormatCompressed_BaseFormatOffset - $format_index[$t] > 255\n"; |
|---|
| 225 | print "%error \"".$format_index[$t]." is out of range of DisplayFormatCompressed_BaseFormatOffset\"\n"; |
|---|
| 226 | print "%endif\n"; |
|---|
| 227 | } |
|---|
| 228 | print "%endif\n"; |
|---|
| 229 | |
|---|
| 230 | #-------------------------------------------------------------------------------- |
|---|
| 231 | # |
|---|
| 232 | # Output usage statistics |
|---|
| 233 | # |
|---|
| 234 | |
|---|
| 235 | print "\n;; translated usage stats\n"; |
|---|
| 236 | foreach $f (keys(%translate)) |
|---|
| 237 | { |
|---|
| 238 | print ";; ".$f.":".$used{$f}."\n"; |
|---|
| 239 | $translate_count++; |
|---|
| 240 | } |
|---|
| 241 | print ";; total translated: ".$translate_count."\n"; |
|---|
| 242 | |
|---|
| 243 | print "\n;; format usage stats\n"; |
|---|
| 244 | $format_count = 0; |
|---|
| 245 | foreach $f (keys(%format)) |
|---|
| 246 | { |
|---|
| 247 | print ";; ".$f.":".$used{$f}."\n"; |
|---|
| 248 | $format_count++; |
|---|
| 249 | } |
|---|
| 250 | print ";; total format: ".$format_count."\n"; |
|---|
| 251 | |
|---|
| 252 | print "\n;; alphabet usage stats\n"; |
|---|
| 253 | |
|---|
| 254 | $used_count = 0; |
|---|
| 255 | for( $t = $normal_base; $t < $normal_base + 0x40; $t++ ) |
|---|
| 256 | { |
|---|
| 257 | print ";; ".$t.",".chr($t).":".$used{$t}."\n"; |
|---|
| 258 | if( $used{$t} ) |
|---|
| 259 | { |
|---|
| 260 | $used_count++; |
|---|
| 261 | } |
|---|
| 262 | } |
|---|
| 263 | print ";; alphabet used count: ".$used_count."\n"; |
|---|
| 264 | |
|---|
| 265 | #-------------------------------------------------------------------------------- |
|---|
| 266 | # |
|---|
| 267 | # processString does the real compression work... |
|---|
| 268 | # |
|---|
| 269 | |
|---|
| 270 | sub processString |
|---|
| 271 | { |
|---|
| 272 | $chars = $_[0]; |
|---|
| 273 | $label = $_[1]; |
|---|
| 274 | $db = $_[2]; |
|---|
| 275 | |
|---|
| 276 | $label =~ s/[a-z0-9_:]/ /ig; # replace with spaces for proper output spacing |
|---|
| 277 | |
|---|
| 278 | # |
|---|
| 279 | # Copy numeric bytes out of hexadecimal pairs in the listing |
|---|
| 280 | # |
|---|
| 281 | $#v = 0; |
|---|
| 282 | |
|---|
| 283 | $orig = ""; |
|---|
| 284 | for( $g = 0; $g < length($chars); $g += 2 ) |
|---|
| 285 | { |
|---|
| 286 | $i = $g/2; |
|---|
| 287 | $v[$i] = hex(substr($chars,$g,2)); |
|---|
| 288 | $orig .= sprintf( ($v[$i] > 0x9f ? ", %03xh" : ", %02xh"), $v[$i] ); |
|---|
| 289 | } |
|---|
| 290 | $v[length($chars)/2] = 0xff; # guard byte to avoid thinking going past the end of |
|---|
| 291 | # the string is a null |
|---|
| 292 | |
|---|
| 293 | $output = ""; |
|---|
| 294 | # |
|---|
| 295 | # Loop through bytes... |
|---|
| 296 | # looking ahead as needed for possible space and null optimizations, compiling formats |
|---|
| 297 | # |
|---|
| 298 | for( $g = 0; $g < $#v-1; $g++ ) # -1 for the guard byte |
|---|
| 299 | { |
|---|
| 300 | # |
|---|
| 301 | # Special translation of LF,CR to a format |
|---|
| 302 | # |
|---|
| 303 | if( $v[$g] == 10 && $v[$g+1] == 13 ) |
|---|
| 304 | { |
|---|
| 305 | $g++; |
|---|
| 306 | $post = $code_translate; |
|---|
| 307 | $code = $format{"nl"}; |
|---|
| 308 | $used{"nl"}++; |
|---|
| 309 | } |
|---|
| 310 | |
|---|
| 311 | # |
|---|
| 312 | # Format operators |
|---|
| 313 | # |
|---|
| 314 | elsif( $v[$g] == 0x25 ) # "%" |
|---|
| 315 | { |
|---|
| 316 | $fo = ""; |
|---|
| 317 | $g++; |
|---|
| 318 | if( $v[$g] >= ord("0") && $v[$g] <= ord("9") ) |
|---|
| 319 | { |
|---|
| 320 | $fo = $fo.chr($v[$g]); |
|---|
| 321 | $g++; |
|---|
| 322 | } |
|---|
| 323 | if( $v[$g] == ord("-") ) |
|---|
| 324 | { |
|---|
| 325 | $fo = $fo.chr($v[$g]); |
|---|
| 326 | $g++; |
|---|
| 327 | } |
|---|
| 328 | $fo = $fo.chr($v[$g]); |
|---|
| 329 | |
|---|
| 330 | $format{$fo} || die "unknown format operator: '".$fo."'\n"; |
|---|
| 331 | |
|---|
| 332 | $code = $format{$fo}; |
|---|
| 333 | $post = $code_translate; |
|---|
| 334 | $used{$fo}++; |
|---|
| 335 | } |
|---|
| 336 | |
|---|
| 337 | # |
|---|
| 338 | # Translated characters |
|---|
| 339 | # |
|---|
| 340 | elsif( $v[$g] == 32 || $translate{$v[$g]} ) |
|---|
| 341 | { |
|---|
| 342 | $post = $code_translate; |
|---|
| 343 | $code = $translate{$v[$g]}; |
|---|
| 344 | $used{$v[$g]}++; |
|---|
| 345 | } |
|---|
| 346 | |
|---|
| 347 | # |
|---|
| 348 | # "normal" characters (alphabet, and ASCII characters around the alphabet) |
|---|
| 349 | # |
|---|
| 350 | elsif( $v[$g] >= $normal_base && $v[$g] < ($normal_base+0x40) ) |
|---|
| 351 | { |
|---|
| 352 | $used{$v[$g]}++; |
|---|
| 353 | |
|---|
| 354 | $post = $code_normal; |
|---|
| 355 | $code = $v[$g] - $normal_base; |
|---|
| 356 | } |
|---|
| 357 | |
|---|
| 358 | # |
|---|
| 359 | # Not found |
|---|
| 360 | # |
|---|
| 361 | else |
|---|
| 362 | { |
|---|
| 363 | die $v[$g].": no translation or format, and out of normal range - may need to be added\n"; |
|---|
| 364 | } |
|---|
| 365 | |
|---|
| 366 | if( $post == $code_translate ) |
|---|
| 367 | { |
|---|
| 368 | # |
|---|
| 369 | # NULL optimization (space optimization not possible on translate/format) |
|---|
| 370 | # |
|---|
| 371 | if( $v[$g+1] == 0 ) |
|---|
| 372 | { |
|---|
| 373 | $g++; |
|---|
| 374 | $post = $post | $code_translate_null; |
|---|
| 375 | } |
|---|
| 376 | else |
|---|
| 377 | { |
|---|
| 378 | $post = $post | $code_translate_normal; |
|---|
| 379 | } |
|---|
| 380 | } |
|---|
| 381 | else # $post == $code_normal |
|---|
| 382 | { |
|---|
| 383 | # |
|---|
| 384 | # Space optimization |
|---|
| 385 | # |
|---|
| 386 | if( $v[$g+1] == 0x20 && $v[$g+2] != 0 ) |
|---|
| 387 | { |
|---|
| 388 | # can't take this optimization if the next byte is a null, |
|---|
| 389 | # since we can't have both a postfix space and null |
|---|
| 390 | $g++; |
|---|
| 391 | $post = $code_space; |
|---|
| 392 | } |
|---|
| 393 | |
|---|
| 394 | # |
|---|
| 395 | # NULL optimization |
|---|
| 396 | # |
|---|
| 397 | elsif( $v[$g+1] == 0 ) |
|---|
| 398 | { |
|---|
| 399 | $g++; |
|---|
| 400 | $post = $code_null; |
|---|
| 401 | } |
|---|
| 402 | } |
|---|
| 403 | |
|---|
| 404 | $code = $code | $post; |
|---|
| 405 | $output .= sprintf( ($code > 0x9f ? ", %03xh" : ", %02xh"), $code ); |
|---|
| 406 | } |
|---|
| 407 | |
|---|
| 408 | print $label."; ".$db.substr($orig,2)." ; uncompressed\n"; |
|---|
| 409 | print $label." ".$db.substr($output,2); |
|---|
| 410 | for( $t = length($output); $t < length($orig); $t++ ) |
|---|
| 411 | { |
|---|
| 412 | print " "; |
|---|
| 413 | } |
|---|
| 414 | print " ; compressed\n\n"; |
|---|
| 415 | } |
|---|
| 416 | |
|---|
