1 | #include "header.h" |
2 | |
3 | #define CREATE_SIZE 1 |
4 | |
5 | extern symbol * create_s(void) { |
6 | symbol * p; |
7 | void * mem = malloc(HEAD + (CREATE_SIZE + 1) * sizeof(symbol)); |
8 | if (mem == NULL) return NULL; |
9 | p = (symbol *) (HEAD + (char *) mem); |
10 | CAPACITY(p) = CREATE_SIZE; |
11 | SET_SIZE(p, 0); |
12 | return p; |
13 | } |
14 | |
15 | extern void lose_s(symbol * p) { |
16 | if (p == NULL) return; |
17 | free((char *) p - HEAD); |
18 | } |
19 | |
20 | /* |
21 | new_p = skip_utf8(p, c, lb, l, n); skips n characters forwards from p + c |
22 | if n +ve, or n characters backwards from p + c - 1 if n -ve. new_p is the new |
23 | position, or -1 on failure. |
24 | |
25 | -- used to implement hop and next in the utf8 case. |
26 | */ |
27 | |
28 | extern int skip_utf8(const symbol * p, int c, int lb, int l, int n) { |
29 | int b; |
30 | if (n >= 0) { |
31 | for (; n > 0; n--) { |
32 | if (c >= l) return -1; |
33 | b = p[c++]; |
34 | if (b >= 0xC0) { /* 1100 0000 */ |
35 | while (c < l) { |
36 | b = p[c]; |
37 | if (b >= 0xC0 || b < 0x80) break; |
38 | /* break unless b is 10------ */ |
39 | c++; |
40 | } |
41 | } |
42 | } |
43 | } else { |
44 | for (; n < 0; n++) { |
45 | if (c <= lb) return -1; |
46 | b = p[--c]; |
47 | if (b >= 0x80) { /* 1000 0000 */ |
48 | while (c > lb) { |
49 | b = p[c]; |
50 | if (b >= 0xC0) break; /* 1100 0000 */ |
51 | c--; |
52 | } |
53 | } |
54 | } |
55 | } |
56 | return c; |
57 | } |
58 | |
59 | /* Code for character groupings: utf8 cases */ |
60 | |
61 | static int get_utf8(const symbol * p, int c, int l, int * slot) { |
62 | int b0, b1; |
63 | if (c >= l) return 0; |
64 | b0 = p[c++]; |
65 | if (b0 < 0xC0 || c == l) { /* 1100 0000 */ |
66 | * slot = b0; return 1; |
67 | } |
68 | b1 = p[c++]; |
69 | if (b0 < 0xE0 || c == l) { /* 1110 0000 */ |
70 | * slot = (b0 & 0x1F) << 6 | (b1 & 0x3F); return 2; |
71 | } |
72 | * slot = (b0 & 0xF) << 12 | (b1 & 0x3F) << 6 | (p[c] & 0x3F); return 3; |
73 | } |
74 | |
75 | static int get_b_utf8(const symbol * p, int c, int lb, int * slot) { |
76 | int b0, b1; |
77 | if (c <= lb) return 0; |
78 | b0 = p[--c]; |
79 | if (b0 < 0x80 || c == lb) { /* 1000 0000 */ |
80 | * slot = b0; return 1; |
81 | } |
82 | b1 = p[--c]; |
83 | if (b1 >= 0xC0 || c == lb) { /* 1100 0000 */ |
84 | * slot = (b1 & 0x1F) << 6 | (b0 & 0x3F); return 2; |
85 | } |
86 | * slot = (p[--c] & 0xF) << 12 | (b1 & 0x3F) << 6 | (b0 & 0x3F); return 3; |
87 | } |
88 | |
89 | extern int in_grouping_U(struct SN_env * z, const unsigned char * s, int min, int max, int repeat) { |
90 | do { |
91 | int ch; |
92 | int w = get_utf8(z->p, z->c, z->l, & ch); |
93 | if (!w) return -1; |
94 | if (ch > max || (ch -= min) < 0 || (s[ch >> 3] & (0X1 << (ch & 0X7))) == 0) |
95 | return w; |
96 | z->c += w; |
97 | } while (repeat); |
98 | return 0; |
99 | } |
100 | |
101 | extern int in_grouping_b_U(struct SN_env * z, const unsigned char * s, int min, int max, int repeat) { |
102 | do { |
103 | int ch; |
104 | int w = get_b_utf8(z->p, z->c, z->lb, & ch); |
105 | if (!w) return -1; |
106 | if (ch > max || (ch -= min) < 0 || (s[ch >> 3] & (0X1 << (ch & 0X7))) == 0) |
107 | return w; |
108 | z->c -= w; |
109 | } while (repeat); |
110 | return 0; |
111 | } |
112 | |
113 | extern int out_grouping_U(struct SN_env * z, const unsigned char * s, int min, int max, int repeat) { |
114 | do { |
115 | int ch; |
116 | int w = get_utf8(z->p, z->c, z->l, & ch); |
117 | if (!w) return -1; |
118 | if (!(ch > max || (ch -= min) < 0 || (s[ch >> 3] & (0X1 << (ch & 0X7))) == 0)) |
119 | return w; |
120 | z->c += w; |
121 | } while (repeat); |
122 | return 0; |
123 | } |
124 | |
125 | extern int out_grouping_b_U(struct SN_env * z, const unsigned char * s, int min, int max, int repeat) { |
126 | do { |
127 | int ch; |
128 | int w = get_b_utf8(z->p, z->c, z->lb, & ch); |
129 | if (!w) return -1; |
130 | if (!(ch > max || (ch -= min) < 0 || (s[ch >> 3] & (0X1 << (ch & 0X7))) == 0)) |
131 | return w; |
132 | z->c -= w; |
133 | } while (repeat); |
134 | return 0; |
135 | } |
136 | |
137 | /* Code for character groupings: non-utf8 cases */ |
138 | |
139 | extern int in_grouping(struct SN_env * z, const unsigned char * s, int min, int max, int repeat) { |
140 | do { |
141 | int ch; |
142 | if (z->c >= z->l) return -1; |
143 | ch = z->p[z->c]; |
144 | if (ch > max || (ch -= min) < 0 || (s[ch >> 3] & (0X1 << (ch & 0X7))) == 0) |
145 | return 1; |
146 | z->c++; |
147 | } while (repeat); |
148 | return 0; |
149 | } |
150 | |
151 | extern int in_grouping_b(struct SN_env * z, const unsigned char * s, int min, int max, int repeat) { |
152 | do { |
153 | int ch; |
154 | if (z->c <= z->lb) return -1; |
155 | ch = z->p[z->c - 1]; |
156 | if (ch > max || (ch -= min) < 0 || (s[ch >> 3] & (0X1 << (ch & 0X7))) == 0) |
157 | return 1; |
158 | z->c--; |
159 | } while (repeat); |
160 | return 0; |
161 | } |
162 | |
163 | extern int out_grouping(struct SN_env * z, const unsigned char * s, int min, int max, int repeat) { |
164 | do { |
165 | int ch; |
166 | if (z->c >= z->l) return -1; |
167 | ch = z->p[z->c]; |
168 | if (!(ch > max || (ch -= min) < 0 || (s[ch >> 3] & (0X1 << (ch & 0X7))) == 0)) |
169 | return 1; |
170 | z->c++; |
171 | } while (repeat); |
172 | return 0; |
173 | } |
174 | |
175 | extern int out_grouping_b(struct SN_env * z, const unsigned char * s, int min, int max, int repeat) { |
176 | do { |
177 | int ch; |
178 | if (z->c <= z->lb) return -1; |
179 | ch = z->p[z->c - 1]; |
180 | if (!(ch > max || (ch -= min) < 0 || (s[ch >> 3] & (0X1 << (ch & 0X7))) == 0)) |
181 | return 1; |
182 | z->c--; |
183 | } while (repeat); |
184 | return 0; |
185 | } |
186 | |
187 | extern int eq_s(struct SN_env * z, int s_size, const symbol * s) { |
188 | if (z->l - z->c < s_size || memcmp(z->p + z->c, s, s_size * sizeof(symbol)) != 0) return 0; |
189 | z->c += s_size; return 1; |
190 | } |
191 | |
192 | extern int eq_s_b(struct SN_env * z, int s_size, const symbol * s) { |
193 | if (z->c - z->lb < s_size || memcmp(z->p + z->c - s_size, s, s_size * sizeof(symbol)) != 0) return 0; |
194 | z->c -= s_size; return 1; |
195 | } |
196 | |
197 | extern int eq_v(struct SN_env * z, const symbol * p) { |
198 | return eq_s(z, SIZE(p), p); |
199 | } |
200 | |
201 | extern int eq_v_b(struct SN_env * z, const symbol * p) { |
202 | return eq_s_b(z, SIZE(p), p); |
203 | } |
204 | |
205 | extern int find_among(struct SN_env * z, const struct among * v, int v_size) { |
206 | |
207 | int i = 0; |
208 | int j = v_size; |
209 | |
210 | int c = z->c; int l = z->l; |
211 | symbol * q = z->p + c; |
212 | |
213 | const struct among * w; |
214 | |
215 | int common_i = 0; |
216 | int common_j = 0; |
217 | |
218 | int first_key_inspected = 0; |
219 | |
220 | while(1) { |
221 | int k = i + ((j - i) >> 1); |
222 | int diff = 0; |
223 | int common = common_i < common_j ? common_i : common_j; /* smaller */ |
224 | w = v + k; |
225 | { |
226 | int i2; for (i2 = common; i2 < w->s_size; i2++) { |
227 | if (c + common == l) { diff = -1; break; } |
228 | diff = q[common] - w->s[i2]; |
229 | if (diff != 0) break; |
230 | common++; |
231 | } |
232 | } |
233 | if (diff < 0) { j = k; common_j = common; } |
234 | else { i = k; common_i = common; } |
235 | if (j - i <= 1) { |
236 | if (i > 0) break; /* v->s has been inspected */ |
237 | if (j == i) break; /* only one item in v */ |
238 | |
239 | /* - but now we need to go round once more to get |
240 | v->s inspected. This looks messy, but is actually |
241 | the optimal approach. */ |
242 | |
243 | if (first_key_inspected) break; |
244 | first_key_inspected = 1; |
245 | } |
246 | } |
247 | while(1) { |
248 | w = v + i; |
249 | if (common_i >= w->s_size) { |
250 | z->c = c + w->s_size; |
251 | if (w->function == 0) return w->result; |
252 | { |
253 | int res = w->function(z); |
254 | z->c = c + w->s_size; |
255 | if (res) return w->result; |
256 | } |
257 | } |
258 | i = w->substring_i; |
259 | if (i < 0) return 0; |
260 | } |
261 | } |
262 | |
263 | /* find_among_b is for backwards processing. Same comments apply */ |
264 | |
265 | extern int find_among_b(struct SN_env * z, const struct among * v, int v_size) { |
266 | |
267 | int i = 0; |
268 | int j = v_size; |
269 | |
270 | int c = z->c; int lb = z->lb; |
271 | symbol * q = z->p + c - 1; |
272 | |
273 | const struct among * w; |
274 | |
275 | int common_i = 0; |
276 | int common_j = 0; |
277 | |
278 | int first_key_inspected = 0; |
279 | |
280 | while(1) { |
281 | int k = i + ((j - i) >> 1); |
282 | int diff = 0; |
283 | int common = common_i < common_j ? common_i : common_j; |
284 | w = v + k; |
285 | { |
286 | int i2; for (i2 = w->s_size - 1 - common; i2 >= 0; i2--) { |
287 | if (c - common == lb) { diff = -1; break; } |
288 | diff = q[- common] - w->s[i2]; |
289 | if (diff != 0) break; |
290 | common++; |
291 | } |
292 | } |
293 | if (diff < 0) { j = k; common_j = common; } |
294 | else { i = k; common_i = common; } |
295 | if (j - i <= 1) { |
296 | if (i > 0) break; |
297 | if (j == i) break; |
298 | if (first_key_inspected) break; |
299 | first_key_inspected = 1; |
300 | } |
301 | } |
302 | while(1) { |
303 | w = v + i; |
304 | if (common_i >= w->s_size) { |
305 | z->c = c - w->s_size; |
306 | if (w->function == 0) return w->result; |
307 | { |
308 | int res = w->function(z); |
309 | z->c = c - w->s_size; |
310 | if (res) return w->result; |
311 | } |
312 | } |
313 | i = w->substring_i; |
314 | if (i < 0) return 0; |
315 | } |
316 | } |
317 | |
318 | |
319 | /* Increase the size of the buffer pointed to by p to at least n symbols. |
320 | * If insufficient memory, returns NULL and frees the old buffer. |
321 | */ |
322 | static symbol * increase_size(symbol * p, int n) { |
323 | symbol * q; |
324 | int new_size = n + 20; |
325 | void * mem = realloc((char *) p - HEAD, |
326 | HEAD + (new_size + 1) * sizeof(symbol)); |
327 | if (mem == NULL) { |
328 | lose_s(p); |
329 | return NULL; |
330 | } |
331 | q = (symbol *) (HEAD + (char *)mem); |
332 | CAPACITY(q) = new_size; |
333 | return q; |
334 | } |
335 | |
336 | /* to replace symbols between c_bra and c_ket in z->p by the |
337 | s_size symbols at s. |
338 | Returns 0 on success, -1 on error. |
339 | Also, frees z->p (and sets it to NULL) on error. |
340 | */ |
341 | extern int replace_s(struct SN_env * z, int c_bra, int c_ket, int s_size, const symbol * s, int * adjptr) |
342 | { |
343 | int adjustment; |
344 | int len; |
345 | if (z->p == NULL) { |
346 | z->p = create_s(); |
347 | if (z->p == NULL) return -1; |
348 | } |
349 | adjustment = s_size - (c_ket - c_bra); |
350 | len = SIZE(z->p); |
351 | if (adjustment != 0) { |
352 | if (adjustment + len > CAPACITY(z->p)) { |
353 | z->p = increase_size(z->p, adjustment + len); |
354 | if (z->p == NULL) return -1; |
355 | } |
356 | memmove(z->p + c_ket + adjustment, |
357 | z->p + c_ket, |
358 | (len - c_ket) * sizeof(symbol)); |
359 | SET_SIZE(z->p, adjustment + len); |
360 | z->l += adjustment; |
361 | if (z->c >= c_ket) |
362 | z->c += adjustment; |
363 | else |
364 | if (z->c > c_bra) |
365 | z->c = c_bra; |
366 | } |
367 | if (s_size) memmove(z->p + c_bra, s, s_size * sizeof(symbol)); |
368 | if (adjptr != NULL) |
369 | *adjptr = adjustment; |
370 | return 0; |
371 | } |
372 | |
373 | static int slice_check(struct SN_env * z) { |
374 | |
375 | if (z->bra < 0 || |
376 | z->bra > z->ket || |
377 | z->ket > z->l || |
378 | z->p == NULL || |
379 | z->l > SIZE(z->p)) /* this line could be removed */ |
380 | { |
381 | #if 0 |
382 | fprintf(stderr, "faulty slice operation:\n" ); |
383 | debug(z, -1, 0); |
384 | #endif |
385 | return -1; |
386 | } |
387 | return 0; |
388 | } |
389 | |
390 | extern int slice_from_s(struct SN_env * z, int s_size, const symbol * s) { |
391 | if (slice_check(z)) return -1; |
392 | return replace_s(z, z->bra, z->ket, s_size, s, NULL); |
393 | } |
394 | |
395 | extern int slice_from_v(struct SN_env * z, const symbol * p) { |
396 | return slice_from_s(z, SIZE(p), p); |
397 | } |
398 | |
399 | extern int slice_del(struct SN_env * z) { |
400 | return slice_from_s(z, 0, 0); |
401 | } |
402 | |
403 | extern int insert_s(struct SN_env * z, int bra, int ket, int s_size, const symbol * s) { |
404 | int adjustment; |
405 | if (replace_s(z, bra, ket, s_size, s, &adjustment)) |
406 | return -1; |
407 | if (bra <= z->bra) z->bra += adjustment; |
408 | if (bra <= z->ket) z->ket += adjustment; |
409 | return 0; |
410 | } |
411 | |
412 | extern int insert_v(struct SN_env * z, int bra, int ket, const symbol * p) { |
413 | return insert_s(z, bra, ket, SIZE(p), p); |
414 | } |
415 | |
416 | extern symbol * slice_to(struct SN_env * z, symbol * p) { |
417 | if (slice_check(z)) { |
418 | lose_s(p); |
419 | return NULL; |
420 | } |
421 | { |
422 | int len = z->ket - z->bra; |
423 | if (CAPACITY(p) < len) { |
424 | p = increase_size(p, len); |
425 | if (p == NULL) |
426 | return NULL; |
427 | } |
428 | memmove(p, z->p + z->bra, len * sizeof(symbol)); |
429 | SET_SIZE(p, len); |
430 | } |
431 | return p; |
432 | } |
433 | |
434 | extern symbol * assign_to(struct SN_env * z, symbol * p) { |
435 | int len = z->l; |
436 | if (CAPACITY(p) < len) { |
437 | p = increase_size(p, len); |
438 | if (p == NULL) |
439 | return NULL; |
440 | } |
441 | memmove(p, z->p, len * sizeof(symbol)); |
442 | SET_SIZE(p, len); |
443 | return p; |
444 | } |
445 | |
446 | extern int len_utf8(const symbol * p) { |
447 | int size = SIZE(p); |
448 | int len = 0; |
449 | while (size--) { |
450 | symbol b = *p++; |
451 | if (b >= 0xC0 || b < 0x80) ++len; |
452 | } |
453 | return len; |
454 | } |
455 | |
456 | #if 0 |
457 | extern void debug(struct SN_env * z, int number, int line_count) { |
458 | int i; |
459 | int limit = SIZE(z->p); |
460 | /*if (number >= 0) printf("%3d (line %4d): '", number, line_count);*/ |
461 | if (number >= 0) printf("%3d (line %4d): [%d]'" , number, line_count,limit); |
462 | for (i = 0; i <= limit; i++) { |
463 | if (z->lb == i) printf("{" ); |
464 | if (z->bra == i) printf("[" ); |
465 | if (z->c == i) printf("|" ); |
466 | if (z->ket == i) printf("]" ); |
467 | if (z->l == i) printf("}" ); |
468 | if (i < limit) |
469 | { int ch = z->p[i]; |
470 | if (ch == 0) ch = '#'; |
471 | printf("%c" , ch); |
472 | } |
473 | } |
474 | printf("'\n" ); |
475 | } |
476 | #endif |
477 | |