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 and Format 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 | # 4. Null characters (marking the end of strings) |
---|
32 | # 5. The special string LF,CR |
---|
33 | # |
---|
34 | # If a character or format read at the input cannot be found in one of the above categories, |
---|
35 | # it must be added here before this script will accept it (and DisplayFormatCompressed can |
---|
36 | # display it). |
---|
37 | # |
---|
38 | # Note that these tables are not present in DisplayFormatCompressed, and do not need to |
---|
39 | # updated there. Needed information is put in the compression output that it reads. |
---|
40 | # |
---|
41 | $translate{ord(' ')} = 0; |
---|
42 | $translate{172} = 1; # ONE_QUARTER |
---|
43 | $translate{171} = 2; # ONE_HALF |
---|
44 | $translate{179} = 3; # SINGLE_VERTICAL |
---|
45 | $translate{175} = 4; # ANGLE_QUOTE_RIGHT |
---|
46 | $translate{ord('!')} = 5; |
---|
47 | $translate{ord('"')} = 6; |
---|
48 | $translate{ord(',')} = 7; |
---|
49 | $translate{ord('-')} = 8; |
---|
50 | $translate{ord('.')} = 9; |
---|
51 | $translate{ord('/')} = 10; |
---|
52 | $translate{ord('1')} = 11; |
---|
53 | $translate{ord('2')} = 12; |
---|
54 | $translate{ord('3')} = 13; |
---|
55 | $translate{ord('4')} = 14; |
---|
56 | $translate{ord('5')} = 15; |
---|
57 | $translate{ord('6')} = 16; |
---|
58 | $translate{ord('8')} = 17; |
---|
59 | $translate{200} = 18; # DOUBLE_BOTTOM_LEFT_CORNER |
---|
60 | $translate{181} = 19; # DOUBLE_LEFT_HORIZONTAL_TO_SINGLE_VERTICAL |
---|
61 | |
---|
62 | # |
---|
63 | # Formats begin immediately after the last Translated character (they are in the same table) |
---|
64 | # |
---|
65 | $format_begin = 20; |
---|
66 | |
---|
67 | $format{"s"} = 20; # n/a |
---|
68 | $format{"c"} = 21; # n/a |
---|
69 | $format{"2-I"} = 22; # must be even |
---|
70 | $format{"u"} = 23; # must be odd |
---|
71 | $format{"5-u"} = 24; # must be even |
---|
72 | $format{"x"} = 25; # must be odd |
---|
73 | $format{"5-x"} = 26; # must be even |
---|
74 | $format{"nl"} = 27; # n/a |
---|
75 | $format{"2-u"} = 28; # must be even |
---|
76 | $format{"A"} = 29; # n/a |
---|
77 | |
---|
78 | # NOTE: The last $format cannot exceed 31 (stored in a 5-bit quantity). |
---|
79 | |
---|
80 | # |
---|
81 | # Starting point for the "normal" range, typically around 0x40 to cover upper and lower case |
---|
82 | # letters. If lower case 'z' is not used, 0x3a can be a good choice as it adds ':' to the |
---|
83 | # front end. |
---|
84 | # |
---|
85 | $normal_base = 0x3a; |
---|
86 | |
---|
87 | # |
---|
88 | # High order code bits, determining which type of character we have (translated or not) and |
---|
89 | # if a space or null should come after this character. |
---|
90 | # |
---|
91 | $code_space = 0xc0; |
---|
92 | $code_null = 0x80; |
---|
93 | $code_normal = 0x40; |
---|
94 | $code_translate = 0x00; |
---|
95 | |
---|
96 | # |
---|
97 | # Bit used if it is a translated byte |
---|
98 | # |
---|
99 | $code_translate_null = 0x00; |
---|
100 | $code_translate_normal = 0x20; |
---|
101 | |
---|
102 | print ";;;======================================================================\n"; |
---|
103 | print ";;;\n"; |
---|
104 | print ";;; This file is generated by StringsCompress.pl from source in Strings.asm\n"; |
---|
105 | print ";;; DO NOT EDIT DIRECTLY - See the maekfile for how to rebuild this file.\n"; |
---|
106 | print ";;; This file only needs to be rebuilt if Strings.asm is changed.\n"; |
---|
107 | print ";;;\n"; |
---|
108 | print ";;;======================================================================\n\n"; |
---|
109 | |
---|
110 | # |
---|
111 | # Loop through lines of the listing, looking for 'db' lines (and dealing with continuations) |
---|
112 | # and compressing each line as it is encountered. |
---|
113 | # |
---|
114 | while(<>) |
---|
115 | { |
---|
116 | # |
---|
117 | # The <number> indicates a line from an include file, do not include in the output |
---|
118 | # |
---|
119 | if( /^\s*\d+\s*\<\d\>/ ) |
---|
120 | { |
---|
121 | } |
---|
122 | |
---|
123 | # |
---|
124 | # a 'db' line, with or without a label |
---|
125 | # |
---|
126 | elsif( /^\s*\d+\s[0-9A-F]+\s([0-9A-F]+)(-?)\s+([a-z0-9_]+:)?(\s+)(db\s+)(.*)/i ) |
---|
127 | { |
---|
128 | $bytes = $1; |
---|
129 | $continuation = $2; |
---|
130 | $label = $3; |
---|
131 | $spacing = $4; |
---|
132 | $db = $5; |
---|
133 | $string = $6; |
---|
134 | |
---|
135 | print $label.$spacing."; ".$db.$string."\n"; |
---|
136 | |
---|
137 | if( $continuation eq "-" ) |
---|
138 | { |
---|
139 | do |
---|
140 | { |
---|
141 | $_ = <>; |
---|
142 | /^\s*\d+\s[0-9A-F]+\s([0-9A-F]+)(\-?)/i || die "parse error on continuation"; |
---|
143 | $bytes .= $1; |
---|
144 | $continuation = $2; |
---|
145 | } |
---|
146 | while( $continuation eq "-" ); |
---|
147 | } |
---|
148 | |
---|
149 | &processString( $bytes, $label.$spacing, $db ); |
---|
150 | } |
---|
151 | |
---|
152 | # |
---|
153 | # everything else, copy to the output as is |
---|
154 | # |
---|
155 | elsif( /^\s*\d+\s*(.*)$/ ) |
---|
156 | { |
---|
157 | print $1."\n"; |
---|
158 | } |
---|
159 | } |
---|
160 | |
---|
161 | print ";;; end of strings.asm\n\n"; |
---|
162 | |
---|
163 | #-------------------------------------------------------------------------------- |
---|
164 | # |
---|
165 | # Output constants and the TranslatesAndFormats table |
---|
166 | # |
---|
167 | |
---|
168 | print "StringsCompressed_NormalBase equ ".$normal_base."\n\n"; |
---|
169 | |
---|
170 | print "StringsCompressed_FormatsBegin equ ".$format_begin."\n\n"; |
---|
171 | |
---|
172 | print "StringsCompressed_TranslatesAndFormats: \n"; |
---|
173 | |
---|
174 | foreach $f (keys(%translate)) |
---|
175 | { |
---|
176 | $translate_index[$translate{$f}] = $f; |
---|
177 | $used{$f} || die "translate $f unused\n"; |
---|
178 | $translate{$f} <= 31 || die $translate{$f}.": translate codes must be below 32"; |
---|
179 | } |
---|
180 | |
---|
181 | for( $g = 0; $translate_index[$g]; $g++ ) |
---|
182 | { |
---|
183 | print " db ".$translate_index[$g]." ; ".$g."\n"; |
---|
184 | } |
---|
185 | |
---|
186 | foreach $f (keys(%format)) |
---|
187 | { |
---|
188 | $n = $f; |
---|
189 | $n =~ s/\-/_/g; |
---|
190 | $format_index[$format{$f}] = "DisplayFormatCompressed_Format_".$n; |
---|
191 | $used{$f} || die "format $f unused\n"; |
---|
192 | $format{$f} <= 31 || die $format{$f}.": format codes must be below 32"; |
---|
193 | } |
---|
194 | |
---|
195 | for( $t = $format_begin; $format_index[$t]; $t++ ) |
---|
196 | { |
---|
197 | print " db (DisplayFormatCompressed_BaseFormatOffset - ".$format_index[$t].") ; ".$t."\n"; |
---|
198 | } |
---|
199 | |
---|
200 | print "\n"; |
---|
201 | |
---|
202 | # |
---|
203 | # Ensure that branch targets are within reach |
---|
204 | # |
---|
205 | print "%ifndef CHECK_FOR_UNUSED_ENTRYPOINTS\n"; |
---|
206 | for( $t = $format_begin; $format_index[$t]; $t++ ) |
---|
207 | { |
---|
208 | print "%if DisplayFormatCompressed_BaseFormatOffset < $format_index[$t] || DisplayFormatCompressed_BaseFormatOffset - $format_index[$t] > 255\n"; |
---|
209 | print "%error \"".$format_index[$t]." is out of range of DisplayFormatCompressed_BaseFormatOffset\"\n"; |
---|
210 | print "%endif\n"; |
---|
211 | } |
---|
212 | print "%endif\n"; |
---|
213 | |
---|
214 | #-------------------------------------------------------------------------------- |
---|
215 | # |
---|
216 | # Output usage statistics |
---|
217 | # |
---|
218 | |
---|
219 | print "\n;; translated usage stats\n"; |
---|
220 | foreach $f (keys(%special)) |
---|
221 | { |
---|
222 | print ";; ".$f.":".$used{$f}."\n"; |
---|
223 | $translate_count++; |
---|
224 | } |
---|
225 | print ";; total translated: ".$translate_count."\n"; |
---|
226 | |
---|
227 | print "\n;; format usage stats\n"; |
---|
228 | $format_count = 0; |
---|
229 | foreach $f (keys(%format)) |
---|
230 | { |
---|
231 | print ";; ".$f.":".$used{$f}."\n"; |
---|
232 | $format_count++; |
---|
233 | } |
---|
234 | print ";; total format: ".$format_count."\n"; |
---|
235 | |
---|
236 | print "\n;; alphabet usage stats\n"; |
---|
237 | |
---|
238 | $used_count = 0; |
---|
239 | for( $t = $normal_base; $t < $normal_base + 0x40; $t++ ) |
---|
240 | { |
---|
241 | print ";; ".$t.",".chr($t).":".$used{$t}."\n"; |
---|
242 | if( $used{$t} ) |
---|
243 | { |
---|
244 | $used_count++; |
---|
245 | } |
---|
246 | } |
---|
247 | print ";; alphabet used count: ".$used_count."\n"; |
---|
248 | |
---|
249 | #-------------------------------------------------------------------------------- |
---|
250 | # |
---|
251 | # processString does the real compression work... |
---|
252 | # |
---|
253 | |
---|
254 | sub processString |
---|
255 | { |
---|
256 | $chars = $_[0]; |
---|
257 | $label = $_[1]; |
---|
258 | $db = $_[2]; |
---|
259 | |
---|
260 | $label =~ s/[a-z0-9_:]/ /ig; # replace with spaces for proper output spacing |
---|
261 | |
---|
262 | # |
---|
263 | # Copy numeric bytes out of hexadecimal pairs in the listing |
---|
264 | # |
---|
265 | $#v = 0; |
---|
266 | |
---|
267 | $orig = ""; |
---|
268 | for( $g = 0; $g < length($chars); $g += 2 ) |
---|
269 | { |
---|
270 | $i = $g/2; |
---|
271 | $v[$i] = hex(substr($chars,$g,2)); |
---|
272 | $orig .= sprintf( ($v[$i] > 0x9f ? ", %03xh" : ", %02xh"), $v[$i] ); |
---|
273 | } |
---|
274 | $v[length($chars)/2] = 0xff; # guard byte to avoid thinking going past the end of |
---|
275 | # the string is a null |
---|
276 | |
---|
277 | $output = ""; |
---|
278 | # |
---|
279 | # Loop through bytes... |
---|
280 | # looking ahead as needed for possible space and null optimizations, compiling formats |
---|
281 | # |
---|
282 | for( $g = 0; $g < $#v-1; $g++ ) # -1 for the guard byte |
---|
283 | { |
---|
284 | # |
---|
285 | # Special translation of LF,CR to a format |
---|
286 | # |
---|
287 | if( $v[$g] == 10 && $v[$g+1] == 13 ) |
---|
288 | { |
---|
289 | $g++; |
---|
290 | $post = $code_translate; |
---|
291 | $code = $format{"nl"}; |
---|
292 | $used{"nl"}++; |
---|
293 | } |
---|
294 | |
---|
295 | # |
---|
296 | # Format operators |
---|
297 | # |
---|
298 | elsif( $v[$g] == 0x25 ) # "%" |
---|
299 | { |
---|
300 | $fo = ""; |
---|
301 | $g++; |
---|
302 | if( $v[$g] >= ord("0") && $v[$g] <= ord("9") ) |
---|
303 | { |
---|
304 | $fo = $fo.chr($v[$g]); |
---|
305 | $g++; |
---|
306 | } |
---|
307 | if( $v[$g] == ord("-") ) |
---|
308 | { |
---|
309 | $fo = $fo.chr($v[$g]); |
---|
310 | $g++; |
---|
311 | } |
---|
312 | $fo = $fo.chr($v[$g]); |
---|
313 | |
---|
314 | $format{$fo} || die "unknown format operator: '".$fo."'\n"; |
---|
315 | |
---|
316 | $code = $format{$fo}; |
---|
317 | $post = $code_translate; |
---|
318 | $used{$fo}++; |
---|
319 | } |
---|
320 | |
---|
321 | # |
---|
322 | # Translated characters |
---|
323 | # |
---|
324 | elsif( $v[$g] == 32 || $translate{$v[$g]} ) |
---|
325 | { |
---|
326 | $post = $code_translate; |
---|
327 | $code = $translate{$v[$g]}; |
---|
328 | $used{$v[$g]}++; |
---|
329 | } |
---|
330 | |
---|
331 | # |
---|
332 | # "normal" characters (alphabet, and ASCII characters around the alphabet) |
---|
333 | # |
---|
334 | elsif( $v[$g] >= $normal_base && $v[$g] < ($normal_base+0x40) ) |
---|
335 | { |
---|
336 | $used{$v[$g]}++; |
---|
337 | |
---|
338 | $post = $code_normal; |
---|
339 | $code = $v[$g] - $normal_base; |
---|
340 | } |
---|
341 | |
---|
342 | # |
---|
343 | # Not found |
---|
344 | # |
---|
345 | else |
---|
346 | { |
---|
347 | die $v[$g].": no translation or format, and out of normal range - may need to be added\n"; |
---|
348 | } |
---|
349 | |
---|
350 | if( $post == $code_translate ) |
---|
351 | { |
---|
352 | # |
---|
353 | # NULL optimization (space optimization not possible on translate/format) |
---|
354 | # |
---|
355 | if( $v[$g+1] == 0 ) |
---|
356 | { |
---|
357 | $g++; |
---|
358 | $post = $post | $code_translate_null; |
---|
359 | } |
---|
360 | else |
---|
361 | { |
---|
362 | $post = $post | $code_translate_normal; |
---|
363 | } |
---|
364 | } |
---|
365 | else # $post == $code_normal |
---|
366 | { |
---|
367 | # |
---|
368 | # Space optimization |
---|
369 | # |
---|
370 | if( $v[$g+1] == 0x20 && $v[$g+2] != 0 ) |
---|
371 | { |
---|
372 | # can't take this optimization if the next byte is a null, |
---|
373 | # since we can't have both a postfix space and null |
---|
374 | $g++; |
---|
375 | $post = $code_space; |
---|
376 | } |
---|
377 | |
---|
378 | # |
---|
379 | # NULL optimization |
---|
380 | # |
---|
381 | elsif( $v[$g+1] == 0 ) |
---|
382 | { |
---|
383 | $g++; |
---|
384 | $post = $code_null; |
---|
385 | } |
---|
386 | } |
---|
387 | |
---|
388 | $code = $code | $post; |
---|
389 | $output .= sprintf( ($code > 0x9f ? ", %03xh" : ", %02xh"), $code ); |
---|
390 | } |
---|
391 | |
---|
392 | print $label."; ".$db.substr($orig,2)." ; uncompressed\n"; |
---|
393 | print $label." ".$db.substr($output,2); |
---|
394 | for( $t = length($output); $t < length($orig); $t++ ) |
---|
395 | { |
---|
396 | print " "; |
---|
397 | } |
---|
398 | print " ; compressed\n\n"; |
---|
399 | } |
---|
400 | |
---|