Commit | Line | Data |
---|---|---|
1adfc4f4 AT |
1 | /* |
2 | * (c) 2019 Aaron Taylor <ataylor at subgeniuskitty dot com> | |
3 | * All rights reserved. | |
4 | */ | |
5 | ||
6 | #include <stdio.h> | |
7 | #include <stdlib.h> | |
8 | #include <unistd.h> | |
9 | #include <string.h> | |
10 | #include <errno.h> | |
11 | #include <stdint.h> | |
12 | #include <sys/select.h> | |
13 | #include <getopt.h> | |
14 | ||
15 | #define VERSION 1 | |
16 | ||
17 | #define STACKSIZE 1024 | |
18 | #define HEAPSIZE 1024 | |
19 | #define RETURNSTACKSIZE 1024 | |
20 | ||
21 | void | |
22 | print_usage(char ** argv) | |
23 | { | |
bf43fa3f | 24 | printf( "VVhitespace Interpreter v%d (www.subgeniuskitty.com)\n" |
1adfc4f4 AT |
25 | "Usage: %s -i <file>\n" |
26 | " -h Help (prints this message)\n" | |
bf43fa3f | 27 | " -i <file> Specify a VVhitespace source file to interpret.\n" |
1adfc4f4 AT |
28 | , VERSION, argv[0] |
29 | ); | |
30 | } | |
31 | ||
32 | int | |
33 | stdin_empty(void) | |
34 | { | |
35 | fd_set read_fds; | |
36 | FD_ZERO(&read_fds); | |
37 | FD_SET(STDIN_FILENO, &read_fds); | |
38 | ||
39 | struct timeval timeout; | |
40 | timeout.tv_sec = 0; | |
41 | timeout.tv_usec = 0; | |
42 | ||
43 | int retval = select(1, &read_fds, NULL, NULL, &timeout); | |
44 | /* retval could be -1. Ignoring that for now. */ | |
45 | if (retval > 0) return 0; | |
46 | return 1; | |
47 | } | |
48 | ||
49 | void | |
50 | ws_die(size_t * pc, char * msg) | |
51 | { | |
52 | printf("\n"); | |
53 | printf("SIM_ERROR @ PC %lu: %s\n", *pc, msg); | |
54 | fflush(stdout); | |
55 | exit(EXIT_FAILURE); | |
56 | } | |
57 | ||
58 | void | |
59 | stack_push(int32_t ** sp, int32_t word) | |
60 | { | |
61 | *((*sp)++) = word; | |
62 | } | |
63 | ||
64 | int32_t | |
65 | stack_pop(int32_t ** sp) | |
66 | { | |
67 | return *(--(*sp)); | |
68 | } | |
69 | ||
70 | int32_t | |
71 | stack_peek(int32_t ** sp, size_t offset) | |
72 | /* offset=0 peeks TOS, offset=1 peeks NOS, etc. */ | |
73 | { | |
74 | return *((*sp)-offset-1); | |
75 | } | |
76 | ||
77 | uint8_t | |
78 | next_code_byte(uint8_t * code, size_t * pc) | |
79 | { | |
80 | return code[(*pc)++]; | |
81 | } | |
82 | ||
9686c901 AT |
83 | /* |
84 | * In addition to returning the parsed label, this function advances the PC to | |
85 | * the next instruction. | |
86 | */ | |
1adfc4f4 AT |
87 | uint16_t |
88 | parse_label(uint8_t * code, size_t * pc) | |
89 | { | |
90 | uint16_t label = 0; | |
91 | uint8_t c; | |
92 | while ((c = code[(*pc)++]) != '\n') { | |
93 | label = label << 1; | |
94 | if (c == ' ') label++; | |
95 | } | |
9686c901 AT |
96 | // TODO: Where should I handle attempts to access an unitialized label? |
97 | // For now, leave it undefined in a nasal demon sense. | |
1adfc4f4 AT |
98 | return label; |
99 | } | |
100 | ||
bf43fa3f AT |
101 | void |
102 | populate_labels(uint32_t * labels, uint8_t * code, size_t code_size) | |
103 | { | |
9686c901 AT |
104 | size_t cp = 0; |
105 | while (cp <= code_size) { | |
106 | if (code[cp] == '\v') { | |
107 | uint16_t temp_label = parse_label(code, &cp); | |
108 | labels[temp_label] = cp; | |
109 | } | |
110 | cp++; | |
111 | } | |
bf43fa3f AT |
112 | } |
113 | ||
1adfc4f4 AT |
114 | void |
115 | process_imp_stack(uint8_t * code, size_t * pc, int32_t ** sp) | |
116 | { | |
117 | switch (next_code_byte(code,pc)) { | |
118 | case ' ': | |
119 | /* Push number onto TOS. */ | |
120 | { | |
121 | /* First, pick off the sign */ | |
122 | int32_t sign = 0; | |
123 | switch (next_code_byte(code,pc)) { | |
124 | case ' ' : sign = 1; break; | |
125 | case '\t': sign = -1; break; | |
bf43fa3f | 126 | default : ws_die(pc, "expected sign"); break; |
1adfc4f4 AT |
127 | } |
128 | ||
129 | /* Now, construct the number and push to TOS. */ | |
130 | /* I'm assuming the numbers are read MSb first. */ | |
131 | int32_t temp, number = 0; | |
132 | while ((temp = next_code_byte(code,pc)) != '\n') { | |
bf43fa3f | 133 | if (temp == '\v') ws_die(pc, "non-binary digit in number"); |
1adfc4f4 AT |
134 | number <<= 1; |
135 | if (temp == '\t') number++; | |
136 | } | |
137 | stack_push(sp, number*sign); | |
138 | } | |
139 | break; | |
140 | case '\n': | |
141 | /* Stack sub-command */ | |
142 | { | |
143 | switch (next_code_byte(code,pc)) { | |
144 | /* Duplicate the TOS. */ | |
145 | case ' ': | |
146 | stack_push(sp, stack_peek(sp,0)); | |
147 | break; | |
148 | /* Swap TOS and NOS. */ | |
149 | case '\t': | |
150 | { | |
151 | int32_t t1 = stack_pop(sp); | |
152 | int32_t t2 = stack_pop(sp); | |
153 | stack_push(sp, t1); | |
154 | stack_push(sp, t2); | |
155 | } | |
156 | break; | |
157 | /* Discard TOS. */ | |
158 | case '\n': | |
159 | stack_pop(sp); | |
160 | break; | |
bf43fa3f AT |
161 | default: |
162 | ws_die(pc, "malformed stack IMP"); | |
163 | break; | |
1adfc4f4 AT |
164 | } |
165 | } | |
166 | break; | |
bf43fa3f | 167 | default: ws_die(pc, "malformed stack IMP"); break; |
1adfc4f4 AT |
168 | } |
169 | } | |
170 | ||
171 | void | |
172 | process_imp_arithmetic(uint8_t * code, size_t * pc, int32_t ** sp) | |
173 | { | |
174 | int32_t temp; | |
175 | switch (next_code_byte(code,pc)) { | |
176 | case ' ': | |
177 | { | |
178 | switch (next_code_byte(code,pc)) { | |
179 | case ' ': | |
180 | /* Addition */ | |
181 | stack_push(sp, stack_pop(sp)+stack_pop(sp)); | |
182 | break; | |
183 | case '\t': | |
184 | /* Subtraction */ | |
185 | temp = stack_pop(sp); | |
186 | stack_push(sp, stack_pop(sp)-temp); | |
187 | break; | |
188 | case '\n': | |
189 | /* Multiplication */ | |
190 | stack_push(sp, stack_pop(sp)*stack_pop(sp)); | |
191 | break; | |
bf43fa3f AT |
192 | default: |
193 | ws_die(pc, "malformed arithmetic IMP"); | |
194 | break; | |
1adfc4f4 AT |
195 | } |
196 | } | |
197 | break; | |
198 | case '\t': | |
199 | { | |
200 | switch (next_code_byte(code,pc)) { | |
201 | case ' ': | |
202 | /* Division */ | |
203 | temp = stack_pop(sp); | |
204 | stack_push(sp, stack_pop(sp)/temp); | |
205 | break; | |
206 | case '\t': | |
207 | /* Modulo */ | |
208 | temp = stack_pop(sp); | |
209 | stack_push(sp, stack_pop(sp)%temp); | |
210 | break; | |
bf43fa3f | 211 | default: ws_die(pc, "malformed arithmetic IMP"); break; |
1adfc4f4 AT |
212 | } |
213 | } | |
214 | break; | |
bf43fa3f | 215 | default: ws_die(pc, "malformed arithmetic IMP"); break; |
1adfc4f4 AT |
216 | } |
217 | } | |
218 | ||
219 | void | |
220 | process_imp_flowcontrol(uint8_t * code, size_t * pc, int32_t ** sp, uint32_t * labels, | |
221 | uint32_t ** rsp) | |
222 | { | |
223 | switch (next_code_byte(code,pc)) { | |
224 | case '\n': | |
225 | /* Technically another LF is required but we ignore it. */ | |
226 | printf("\n"); | |
227 | fflush(stdout); | |
228 | exit(EXIT_SUCCESS); | |
229 | case ' ': | |
230 | { | |
231 | switch (next_code_byte(code,pc)) { | |
232 | case ' ': | |
233 | /* Mark a location in the program. */ | |
9686c901 AT |
234 | if (next_code_byte(code,pc) != '\v') ws_die(pc,"expected vtab, " |
235 | "perhaps a whitespace program, rather than vvhitespace?"); | |
236 | /* Jump to next instruction since labels were parsed during startup. */ | |
237 | parse_label( code, pc); | |
1adfc4f4 AT |
238 | break; |
239 | case '\t': | |
240 | /* Call a subroutine. */ | |
241 | *((*rsp)++) = *pc; | |
242 | *pc = labels[parse_label(code, pc)]; | |
243 | break; | |
244 | case '\n': | |
245 | /* Jump unconditionally to a label. */ | |
246 | *pc = labels[parse_label(code, pc)]; | |
247 | break; | |
bf43fa3f AT |
248 | default: |
249 | ws_die(pc, "malformed flow control IMP"); | |
250 | break; | |
1adfc4f4 AT |
251 | } |
252 | } | |
253 | break; | |
254 | case '\t': | |
255 | { | |
256 | switch (next_code_byte(code,pc)) { | |
257 | case ' ': | |
258 | /* Jump to a label if TOS == 0 */ | |
259 | if (stack_peek(sp,0) == 0) *pc = labels[parse_label(code, pc)]; | |
260 | break; | |
261 | case '\t': | |
262 | /* Jump to a label if TOS < 0. */ | |
263 | if (stack_peek(sp,0) < 0) *pc = labels[parse_label(code, pc)]; | |
264 | break; | |
265 | case '\n': | |
266 | /* Return from subroutine. */ | |
267 | *pc = *(--(*rsp)); | |
268 | break; | |
bf43fa3f AT |
269 | default: |
270 | ws_die(pc, "malformed flow control IMP"); | |
271 | break; | |
1adfc4f4 AT |
272 | } |
273 | } | |
274 | break; | |
bf43fa3f AT |
275 | default: |
276 | ws_die(pc, "malformed flow control IMP"); | |
277 | break; | |
1adfc4f4 AT |
278 | } |
279 | } | |
280 | ||
281 | void | |
282 | process_imp_heap(uint8_t * code, size_t * pc, int32_t ** sp, int32_t ** hp) | |
283 | { | |
284 | switch (next_code_byte(code,pc)) { | |
285 | case ' ' : /* Store to heap */ *(*hp + *((*sp)-1)) = **sp; *sp -= 2; break; | |
286 | case '\t': /* Retrieve from heap */ **sp = *(*hp + **sp); break; | |
9686c901 | 287 | default : ws_die(pc, "malformed heap IMP"); break; |
1adfc4f4 AT |
288 | } |
289 | } | |
290 | ||
291 | void | |
292 | process_imp_io(uint8_t * code, size_t * pc, int32_t ** sp, int32_t ** hp) | |
293 | { | |
294 | switch (next_code_byte(code,pc)) { | |
295 | case ' ': | |
296 | /* Output */ | |
297 | { | |
298 | switch (next_code_byte(code,pc)) { | |
299 | case ' ' : /* Output character from TOS */ printf("%c", stack_pop(sp)); break; | |
300 | case '\t': /* Output number from TOS */ printf("%d", stack_pop(sp)); break; | |
9686c901 | 301 | default : ws_die(pc, "malformed output IMP"); break; |
1adfc4f4 AT |
302 | } |
303 | fflush(stdout); | |
304 | } | |
305 | break; | |
306 | case '\t': | |
307 | /* Input */ | |
308 | { | |
309 | while (stdin_empty()) continue; | |
310 | char c = getchar(); | |
311 | switch (next_code_byte(code,pc)) { | |
312 | case '\t': /* Input digit */ c -= '0'; /* fallthrough */ | |
313 | case ' ' : /* Input character */ *(*hp + *((*sp)--)) = c; break; | |
9686c901 | 314 | default : ws_die(pc, "malformed input IMP"); break; |
1adfc4f4 AT |
315 | } |
316 | } | |
317 | break; | |
bf43fa3f | 318 | default: ws_die(pc, "malformed i/o IMP"); break; |
1adfc4f4 AT |
319 | } |
320 | } | |
321 | ||
1adfc4f4 AT |
322 | int |
323 | main(int argc, char ** argv) | |
324 | { | |
325 | /* | |
326 | * Process command line arguments | |
327 | */ | |
328 | int c; | |
329 | FILE * input = NULL; | |
330 | while ((c = getopt(argc,argv,"i:h")) != -1) { | |
331 | switch (c) { | |
332 | case 'i': | |
333 | if ((input = fopen(optarg, "r")) == NULL) { | |
334 | fprintf(stderr, "ERROR: %s: %s\n", optarg, strerror(errno)); | |
335 | } | |
336 | break; | |
337 | case 'h': | |
338 | print_usage(argv); | |
339 | exit(EXIT_SUCCESS); | |
340 | break; | |
341 | default: | |
342 | break; | |
343 | } | |
344 | } | |
345 | if (input == NULL) { | |
bf43fa3f | 346 | fprintf(stderr, "ERROR: Must specify a VVhitespace source file with -f flag.\n"); |
1adfc4f4 AT |
347 | print_usage(argv); |
348 | exit(EXIT_FAILURE); | |
349 | } | |
350 | ||
351 | /* | |
bf43fa3f | 352 | * Read just the VVhitespace source code into memory. |
1adfc4f4 AT |
353 | * We will use the array indices as addresses for the virtual PC when jumping to labels. |
354 | */ | |
355 | size_t ws_code_size = 0; | |
356 | uint8_t temp_byte; | |
357 | while (fread(&temp_byte, 1, 1, input)) { | |
bf43fa3f AT |
358 | if (temp_byte == ' ' || temp_byte == '\t' || temp_byte == '\n' || temp_byte == '\v') { |
359 | ws_code_size++; | |
360 | } | |
1adfc4f4 AT |
361 | } |
362 | rewind(input); | |
363 | uint8_t * ws_code_space = malloc(ws_code_size); | |
364 | ws_code_size = 0; | |
365 | while (fread(&temp_byte, 1, 1, input)) { | |
bf43fa3f AT |
366 | if (temp_byte == ' ' || temp_byte == '\t' || temp_byte == '\n' || temp_byte == '\v') { |
367 | ws_code_space[ws_code_size++] = temp_byte; | |
368 | } | |
1adfc4f4 AT |
369 | } |
370 | fclose(input); | |
371 | ||
372 | /* | |
373 | * Setup a stack and heap. | |
374 | * Assume a 32-bit word size. | |
375 | */ | |
bf43fa3f | 376 | // TODO: Make everything 64-bit. |
1adfc4f4 AT |
377 | int32_t * hp = malloc(HEAPSIZE*4); |
378 | int32_t * sp = malloc(STACKSIZE*4); | |
379 | ||
380 | /* | |
381 | * Setup the return stack and the label array. | |
382 | */ | |
383 | uint32_t * rsp = malloc(RETURNSTACKSIZE*4); | |
bf43fa3f AT |
384 | uint32_t labels[65536] = {0}; |
385 | populate_labels(labels, ws_code_space, ws_code_size); | |
1adfc4f4 AT |
386 | |
387 | /* | |
388 | * Main Loop | |
389 | */ | |
390 | ||
391 | size_t pc = 0; /* Virtual program counter. Operates in the ws_code_space[] address space. */ | |
392 | while (1) { | |
393 | if (pc >= ws_code_size) { | |
bf43fa3f AT |
394 | fprintf(stderr, "SIM_ERROR: PC Overrun\n Requested PC: %lu\n Max Address: %lu\n", |
395 | pc, ws_code_size-1); | |
1adfc4f4 AT |
396 | exit(EXIT_FAILURE); |
397 | } | |
9686c901 AT |
398 | // TODO: Have the SIGTERM signal handler and normal term point return the value |
399 | // on TOS so I can do rudimentary automated tests. | |
1adfc4f4 AT |
400 | |
401 | /* Decode the IMPs */ | |
402 | switch (ws_code_space[pc++]) { | |
403 | case ' ': | |
404 | /* Stack Manipulation */ | |
405 | process_imp_stack(ws_code_space, &pc, &sp); | |
406 | break; | |
407 | case '\n': | |
408 | /* Flow Control */ | |
409 | process_imp_flowcontrol(ws_code_space, &pc, &sp, labels, &rsp); | |
410 | break; | |
411 | case '\t': | |
412 | /* Arithmetic, Heap Access, or I/O */ | |
413 | { | |
414 | switch (ws_code_space[pc++]) { | |
415 | case ' ': | |
416 | /* Arithmetic */ | |
417 | process_imp_arithmetic(ws_code_space, &pc, &sp); | |
418 | break; | |
419 | case '\t': | |
420 | /* Heap Access */ | |
421 | process_imp_heap(ws_code_space, &pc, &sp, &hp); | |
422 | break; | |
423 | case '\n': | |
424 | /* I/O */ | |
425 | process_imp_io(ws_code_space, &pc, &sp, &hp); | |
426 | break; | |
427 | } | |
428 | } | |
429 | break; | |
9686c901 | 430 | default: ws_die(&pc, "unexpected VTab"); break; |
1adfc4f4 AT |
431 | } |
432 | } | |
433 | ||
434 | printf("\n"); | |
435 | printf("Program executed.\n"); | |
436 | ||
437 | exit(EXIT_SUCCESS); | |
438 | } |