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 | print "%ifdef STRINGSCOMPRESSED_STRINGS\n\n"; |
---|
69 | |
---|
70 | # |
---|
71 | # On a first pass, look for our table directives. $translate{...}, $format{...}, etc. |
---|
72 | # are expected in the input stream. |
---|
73 | # |
---|
74 | $processed = " [StringsCompress Processed]"; |
---|
75 | while(<>) |
---|
76 | { |
---|
77 | chop; |
---|
78 | $o = $_; |
---|
79 | |
---|
80 | # |
---|
81 | # Table entries for this script |
---|
82 | # |
---|
83 | if( /^\s*\d+\s*(\;\$translate\{\s*ord\(\s*'(.)'\s*\)\s*\}\s*=\s*([0-9]+).*$)/ ) |
---|
84 | { |
---|
85 | $translate{ord($2)} = int($3); |
---|
86 | $o .= $processed; |
---|
87 | } |
---|
88 | elsif( /^\s*\d+\s*(\;\$translate\{\s*([0-9]+)\s*\}\s*=\s*([0-9]+).*$)/ ) |
---|
89 | { |
---|
90 | $translate{int($2)} = int($3); |
---|
91 | $o .= $processed; |
---|
92 | } |
---|
93 | elsif( /^\s*\d+\s*(\;\$format_begin\s*=\s*([0-9]+).*$)/ ) |
---|
94 | { |
---|
95 | $format_begin = int($2); |
---|
96 | $o .= $processed; |
---|
97 | } |
---|
98 | elsif( /^\s*\d+\s*(\;\$format\{\s*\"([^\"]+)\"\s*\}\s*=\s*([0-9]+).*$)/ ) |
---|
99 | { |
---|
100 | $format{$2} = int($3); |
---|
101 | $o .= $processed; |
---|
102 | } |
---|
103 | elsif( /^\s*\d+\s*(\;\$normal_base\s*=\s*0x([0-9a-fA-F]+).*$)/ ) |
---|
104 | { |
---|
105 | $normal_base = hex($2); |
---|
106 | $o .= $processed; |
---|
107 | } |
---|
108 | elsif( /^\s*\d+\s*(\;\$normal_base\s*=\s*([0-9]+).*$)/ ) |
---|
109 | { |
---|
110 | $normal_base = int($2); |
---|
111 | $o .= $processed; |
---|
112 | } |
---|
113 | |
---|
114 | push( @lines, $o ); |
---|
115 | } |
---|
116 | |
---|
117 | # |
---|
118 | # On the second pass, loop through lines of the listing, looking for 'db' lines |
---|
119 | # (and dealing with continuations) and compressing each line as it is encountered. |
---|
120 | # |
---|
121 | for( $l = 0; $l < $#lines; $l++ ) |
---|
122 | { |
---|
123 | $_ = $lines[$l]; |
---|
124 | |
---|
125 | # |
---|
126 | # The <number> indicates a line from an include file, do not include in the output |
---|
127 | # |
---|
128 | if( /^\s*\d+\s*\<\d\>/ ) |
---|
129 | { |
---|
130 | } |
---|
131 | |
---|
132 | # |
---|
133 | # a 'db' line, with or without a label |
---|
134 | # |
---|
135 | elsif( /^\s*\d+\s[0-9A-F]+\s([0-9A-F]+)(-?)\s+([a-z0-9_]+:)?(\s+)(db\s+)(.*)/i ) |
---|
136 | { |
---|
137 | $bytes = $1; |
---|
138 | $continuation = $2; |
---|
139 | $label = $3; |
---|
140 | $spacing = $4; |
---|
141 | $db = $5; |
---|
142 | $string = $6; |
---|
143 | |
---|
144 | print $label.$spacing."; ".$db.$string."\n"; |
---|
145 | |
---|
146 | if( $continuation eq "-" ) |
---|
147 | { |
---|
148 | do |
---|
149 | { |
---|
150 | $_ = $lines[++$l]; |
---|
151 | /^\s*\d+\s[0-9A-F]+\s([0-9A-F]+)(\-?)/i || die "parse error on continuation: '".$_."'"; |
---|
152 | $bytes .= $1; |
---|
153 | $continuation = $2; |
---|
154 | } |
---|
155 | while( $continuation eq "-" ); |
---|
156 | } |
---|
157 | |
---|
158 | &processString( $bytes, $label.$spacing, $db ); |
---|
159 | } |
---|
160 | |
---|
161 | # |
---|
162 | # a ';%%;' prefix line, copy to output without the prefix |
---|
163 | # |
---|
164 | elsif( /^\s*\d+\s*;%%;\s*(.*)$/ ) |
---|
165 | { |
---|
166 | print $1."\n"; |
---|
167 | } |
---|
168 | |
---|
169 | # |
---|
170 | # everything else, copy to the output as is |
---|
171 | # |
---|
172 | elsif( /^\s*\d+\s*(.*)$/ ) |
---|
173 | { |
---|
174 | print $1."\n"; |
---|
175 | } |
---|
176 | } |
---|
177 | |
---|
178 | print ";;; end of input stream\n\n"; |
---|
179 | |
---|
180 | #-------------------------------------------------------------------------------- |
---|
181 | # |
---|
182 | # Output constants and the TranslatesAndFormats table |
---|
183 | # |
---|
184 | |
---|
185 | print "%endif ; STRINGSCOMPRESSED_STRINGS\n\n"; |
---|
186 | print "%ifdef STRINGSCOMPRESSED_TABLES\n\n"; |
---|
187 | |
---|
188 | print "StringsCompressed_NormalBase equ ".$normal_base."\n\n"; |
---|
189 | |
---|
190 | print "StringsCompressed_FormatsBegin equ ".$format_begin."\n\n"; |
---|
191 | |
---|
192 | print "StringsCompressed_TranslatesAndFormats: \n"; |
---|
193 | |
---|
194 | foreach $f (keys(%translate)) |
---|
195 | { |
---|
196 | $translate_index[$translate{$f}] = $f; |
---|
197 | $used{$f} || die "translate $f unused\n"; |
---|
198 | $translate{$f} <= 31 || die $translate{$f}.": translate codes must be below 32"; |
---|
199 | } |
---|
200 | |
---|
201 | for( $g = 0; $translate_index[$g]; $g++ ) |
---|
202 | { |
---|
203 | print " db ".$translate_index[$g]." ; ".$g."\n"; |
---|
204 | } |
---|
205 | |
---|
206 | foreach $f (keys(%format)) |
---|
207 | { |
---|
208 | $n = $f; |
---|
209 | $n =~ s/\-/_/g; |
---|
210 | $format_index[$format{$f}] = "DisplayFormatCompressed_Format_".$n; |
---|
211 | $used{$f} || die "format $f unused\n"; |
---|
212 | $format{$f} <= 31 || die $format{$f}.": format codes must be below 32"; |
---|
213 | } |
---|
214 | |
---|
215 | for( $t = $format_begin; $format_index[$t]; $t++ ) |
---|
216 | { |
---|
217 | print " db (DisplayFormatCompressed_BaseFormatOffset - ".$format_index[$t].") ; ".$t."\n"; |
---|
218 | } |
---|
219 | |
---|
220 | print "\n"; |
---|
221 | |
---|
222 | # |
---|
223 | # Ensure that branch targets are within reach |
---|
224 | # |
---|
225 | print "%ifndef CHECK_FOR_UNUSED_ENTRYPOINTS\n"; |
---|
226 | for( $t = $format_begin; $format_index[$t]; $t++ ) |
---|
227 | { |
---|
228 | print "%if DisplayFormatCompressed_BaseFormatOffset < $format_index[$t] || DisplayFormatCompressed_BaseFormatOffset - $format_index[$t] > 255\n"; |
---|
229 | print "%error \"".$format_index[$t]." is out of range of DisplayFormatCompressed_BaseFormatOffset\"\n"; |
---|
230 | print "%endif\n"; |
---|
231 | } |
---|
232 | print "%endif\n"; |
---|
233 | |
---|
234 | #-------------------------------------------------------------------------------- |
---|
235 | # |
---|
236 | # Output usage statistics |
---|
237 | # |
---|
238 | |
---|
239 | print "\n;; translated usage stats\n"; |
---|
240 | foreach $f (keys(%translate)) |
---|
241 | { |
---|
242 | print ";; ".$f.":".$used{$f}."\n"; |
---|
243 | $translate_count++; |
---|
244 | } |
---|
245 | print ";; total translated: ".$translate_count."\n"; |
---|
246 | |
---|
247 | print "\n;; format usage stats\n"; |
---|
248 | $format_count = 0; |
---|
249 | foreach $f (keys(%format)) |
---|
250 | { |
---|
251 | print ";; ".$f.":".$used{$f}."\n"; |
---|
252 | $format_count++; |
---|
253 | } |
---|
254 | print ";; total format: ".$format_count."\n"; |
---|
255 | |
---|
256 | print "\n;; alphabet usage stats\n"; |
---|
257 | |
---|
258 | $used_count = 0; |
---|
259 | for( $t = $normal_base; $t < $normal_base + 0x40; $t++ ) |
---|
260 | { |
---|
261 | print ";; ".$t.",".chr($t).":".$used{$t}."\n"; |
---|
262 | if( $used{$t} ) |
---|
263 | { |
---|
264 | $used_count++; |
---|
265 | } |
---|
266 | } |
---|
267 | print ";; alphabet used count: ".$used_count."\n"; |
---|
268 | |
---|
269 | print "%endif ; STRINGSCOMPRESSED_TABLES\n\n"; |
---|
270 | |
---|
271 | #-------------------------------------------------------------------------------- |
---|
272 | # |
---|
273 | # processString does the real compression work... |
---|
274 | # |
---|
275 | |
---|
276 | sub processString |
---|
277 | { |
---|
278 | $chars = $_[0]; |
---|
279 | $label = $_[1]; |
---|
280 | $db = $_[2]; |
---|
281 | |
---|
282 | $label =~ s/[a-z0-9_:]/ /ig; # replace with spaces for proper output spacing |
---|
283 | |
---|
284 | # |
---|
285 | # Copy numeric bytes out of hexadecimal pairs in the listing |
---|
286 | # |
---|
287 | $#v = 0; |
---|
288 | |
---|
289 | $orig = ""; |
---|
290 | for( $g = 0; $g < length($chars); $g += 2 ) |
---|
291 | { |
---|
292 | $i = $g/2; |
---|
293 | $v[$i] = hex(substr($chars,$g,2)); |
---|
294 | $orig .= sprintf( ($v[$i] > 0x9f ? ", %03xh" : ", %02xh"), $v[$i] ); |
---|
295 | } |
---|
296 | $v[length($chars)/2] = 0xff; # guard byte to avoid thinking going past the end of |
---|
297 | # the string is a null |
---|
298 | |
---|
299 | $output = ""; |
---|
300 | # |
---|
301 | # Loop through bytes... |
---|
302 | # looking ahead as needed for possible space and null optimizations, compiling formats |
---|
303 | # |
---|
304 | for( $g = 0; $g < $#v-1; $g++ ) # -1 for the guard byte |
---|
305 | { |
---|
306 | # |
---|
307 | # Special translation of LF,CR to a format |
---|
308 | # |
---|
309 | if( $v[$g] == 10 && $v[$g+1] == 13 ) |
---|
310 | { |
---|
311 | $g++; |
---|
312 | $post = $code_translate; |
---|
313 | $code = $format{"nl"}; |
---|
314 | $used{"nl"}++; |
---|
315 | } |
---|
316 | |
---|
317 | # |
---|
318 | # Format operators |
---|
319 | # |
---|
320 | elsif( $v[$g] == 0x25 ) # "%" |
---|
321 | { |
---|
322 | $fo = ""; |
---|
323 | $g++; |
---|
324 | if( $v[$g] >= ord("0") && $v[$g] <= ord("9") ) |
---|
325 | { |
---|
326 | $fo = $fo.chr($v[$g]); |
---|
327 | $g++; |
---|
328 | } |
---|
329 | if( $v[$g] == ord("-") ) |
---|
330 | { |
---|
331 | $fo = $fo.chr($v[$g]); |
---|
332 | $g++; |
---|
333 | } |
---|
334 | $fo = $fo.chr($v[$g]); |
---|
335 | |
---|
336 | $format{$fo} || die "unknown format operator: '".$fo."'\n"; |
---|
337 | |
---|
338 | $code = $format{$fo}; |
---|
339 | $post = $code_translate; |
---|
340 | $used{$fo}++; |
---|
341 | } |
---|
342 | |
---|
343 | # |
---|
344 | # Translated characters |
---|
345 | # |
---|
346 | elsif( $v[$g] == 32 || $translate{$v[$g]} ) |
---|
347 | { |
---|
348 | $post = $code_translate; |
---|
349 | $code = $translate{$v[$g]}; |
---|
350 | $used{$v[$g]}++; |
---|
351 | } |
---|
352 | |
---|
353 | # |
---|
354 | # "normal" characters (alphabet, and ASCII characters around the alphabet) |
---|
355 | # |
---|
356 | elsif( $v[$g] >= $normal_base && $v[$g] < ($normal_base+0x40) ) |
---|
357 | { |
---|
358 | $used{$v[$g]}++; |
---|
359 | |
---|
360 | $post = $code_normal; |
---|
361 | $code = $v[$g] - $normal_base; |
---|
362 | } |
---|
363 | |
---|
364 | # |
---|
365 | # Not found |
---|
366 | # |
---|
367 | else |
---|
368 | { |
---|
369 | die $v[$g].": no translation or format, and out of normal range - may need to be added\n"; |
---|
370 | } |
---|
371 | |
---|
372 | if( $post == $code_translate ) |
---|
373 | { |
---|
374 | # |
---|
375 | # NULL optimization (space optimization not possible on translate/format) |
---|
376 | # |
---|
377 | if( $v[$g+1] == 0 ) |
---|
378 | { |
---|
379 | $g++; |
---|
380 | $post = $post | $code_translate_null; |
---|
381 | } |
---|
382 | else |
---|
383 | { |
---|
384 | $post = $post | $code_translate_normal; |
---|
385 | } |
---|
386 | } |
---|
387 | else # $post == $code_normal |
---|
388 | { |
---|
389 | # |
---|
390 | # Space optimization |
---|
391 | # |
---|
392 | if( $v[$g+1] == 0x20 && $v[$g+2] != 0 ) |
---|
393 | { |
---|
394 | # can't take this optimization if the next byte is a null, |
---|
395 | # since we can't have both a postfix space and null |
---|
396 | $g++; |
---|
397 | $post = $code_space; |
---|
398 | } |
---|
399 | |
---|
400 | # |
---|
401 | # NULL optimization |
---|
402 | # |
---|
403 | elsif( $v[$g+1] == 0 ) |
---|
404 | { |
---|
405 | $g++; |
---|
406 | $post = $code_null; |
---|
407 | } |
---|
408 | } |
---|
409 | |
---|
410 | $code = $code | $post; |
---|
411 | $output .= sprintf( ($code > 0x9f ? ", %03xh" : ", %02xh"), $code ); |
---|
412 | } |
---|
413 | |
---|
414 | print $label."; ".$db.substr($orig,2)." ; uncompressed\n"; |
---|
415 | print $label." ".$db.substr($output,2); |
---|
416 | for( $t = length($output); $t < length($orig); $t++ ) |
---|
417 | { |
---|
418 | print " "; |
---|
419 | } |
---|
420 | print " ; compressed\n\n"; |
---|
421 | } |
---|
422 | |
---|