Annotation of parser3/src/lib/cord/cordbscs.c, revision 1.1.2.8
1.1.2.1 paf 1: /*
2: * Copyright (c) 1993-1994 by Xerox Corporation. All rights reserved.
3: *
4: * THIS MATERIAL IS PROVIDED AS IS, WITH ABSOLUTELY NO WARRANTY EXPRESSED
5: * OR IMPLIED. ANY USE IS AT YOUR OWN RISK.
6: *
7: * Permission is hereby granted to use or copy this program
8: * for any purpose, provided the above notices are retained on all copies.
9: * Permission to modify the code and to distribute modified code is granted,
10: * provided the above notices are retained, and a notice that the code was
11: * modified is included with the above copyright notice.
12: *
13: * Author: Hans-J. Boehm (boehm@parc.xerox.com)
14: */
15: /* Boehm, October 3, 1994 5:19 pm PDT */
16: # include "gc.h"
17: # include "cord.h"
18: # include <stdlib.h>
19: # include <stdio.h>
20: # include <string.h>
21:
22: /* An implementation of the cord primitives. These are the only */
23: /* Functions that understand the representation. We perform only */
24: /* minimal checks on arguments to these functions. Out of bounds */
25: /* arguments to the iteration functions may result in client functions */
26: /* invoked on garbage data. In most cases, client functions should be */
27: /* programmed defensively enough that this does not result in memory */
28: /* smashes. */
29:
30: typedef void (* oom_fn)(void);
31:
32: oom_fn CORD_oom_fn = (oom_fn) 0;
33:
34: # define OUT_OF_MEMORY { if (CORD_oom_fn != (oom_fn) 0) (*CORD_oom_fn)(); \
35: ABORT("Out of memory\n"); }
36: # define ABORT(msg) { fprintf(stderr, "%s\n", msg); abort(); }
37:
38: typedef unsigned long word;
39:
40: typedef union {
41: struct Concatenation {
42: char null;
43: char header;
44: char depth; /* concatenation nesting depth. */
45: unsigned char left_len;
46: /* Length of left child if it is sufficiently */
47: /* short; 0 otherwise. */
48: # define MAX_LEFT_LEN 255
49: word len;
50: CORD left; /* length(left) > 0 */
51: CORD right; /* length(right) > 0 */
52: } concatenation;
53: struct Function {
54: char null;
55: char header;
56: char depth; /* always 0 */
57: char left_len; /* always 0 */
58: word len;
59: CORD_fn fn;
60: void * client_data;
61: } function;
62: struct Generic {
63: char null;
64: char header;
65: char depth;
66: char left_len;
67: word len;
68: } generic;
69: char string[1];
70: } CordRep;
71:
72: # define CONCAT_HDR 1
73:
74: # define FN_HDR 4
75: # define SUBSTR_HDR 6
76: /* Substring nodes are a special case of function nodes. */
77: /* The client_data field is known to point to a substr_args */
78: /* structure, and the function is either CORD_apply_access_fn */
79: /* or CORD_index_access_fn. */
80:
81: /* The following may be applied only to function and concatenation nodes: */
82: #define IS_CONCATENATION(s) (((CordRep *)s)->generic.header == CONCAT_HDR)
83:
84: #define IS_FUNCTION(s) ((((CordRep *)s)->generic.header & FN_HDR) != 0)
85:
86: #define IS_SUBSTR(s) (((CordRep *)s)->generic.header == SUBSTR_HDR)
87:
88: #define LEN(s) (((CordRep *)s) -> generic.len)
89: #define DEPTH(s) (((CordRep *)s) -> generic.depth)
90: #define GEN_LEN(s) (CORD_IS_STRING(s) ? strlen(s) : LEN(s))
91:
92: #define LEFT_LEN(c) ((c) -> left_len != 0? \
93: (c) -> left_len \
94: : (CORD_IS_STRING((c) -> left) ? \
95: (c) -> len - GEN_LEN((c) -> right) \
96: : LEN((c) -> left)))
97:
98: #define SHORT_LIMIT (sizeof(CordRep) - 1)
99: /* Cords shorter than this are C strings */
100:
101:
102: /* Dump the internal representation of x to stdout, with initial */
103: /* indentation level n. */
104: void CORD_dump_inner(CORD x, unsigned n)
105: {
106: register size_t i;
107:
108: for (i = 0; i < (size_t)n; i++) {
109: fputs(" ", stdout);
110: }
111: if (x == 0) {
112: fputs("NIL\n", stdout);
113: } else if (CORD_IS_STRING(x)) {
114: for (i = 0; i <= SHORT_LIMIT; i++) {
115: if (x[i] == '\0') break;
116: putchar(x[i]);
117: }
118: if (x[i] != '\0') fputs("...", stdout);
119: putchar('\n');
120: } else if (IS_CONCATENATION(x)) {
121: register struct Concatenation * conc =
122: &(((CordRep *)x) -> concatenation);
123: printf("Concatenation: %p (len: %d, depth: %d)\n",
124: x, (int)(conc -> len), (int)(conc -> depth));
125: CORD_dump_inner(conc -> left, n+1);
126: CORD_dump_inner(conc -> right, n+1);
127: } else /* function */{
128: register struct Function * func =
129: &(((CordRep *)x) -> function);
130: if (IS_SUBSTR(x)) printf("(Substring) ");
131: printf("Function: %p (len: %d): ", x, (int)(func -> len));
132: for (i = 0; i < 20 && i < func -> len; i++) {
133: putchar((*(func -> fn))(i, func -> client_data));
134: }
135: if (i < func -> len) fputs("...", stdout);
136: putchar('\n');
137: }
138: }
139:
140: /* Dump the internal representation of x to stdout */
141: void CORD_dump(CORD x)
142: {
143: CORD_dump_inner(x, 0);
144: fflush(stdout);
145: }
146:
1.1.2.2 paf 147: CORD CORD_cat_char_star(CORD x, const char* y, size_t leny)
1.1.2.1 paf 148: {
149: register size_t result_len;
150: register size_t lenx;
151: register int depth;
1.1.2.3 paf 152:
1.1.2.1 paf 153: if (x == CORD_EMPTY) return(y);
1.1.2.6 paf 154: if (leny == 0) leny=strlen(y); // PAF
155: if (leny == 0) ABORT("CORD_cat_char_star(,y,) is empty"); // PAF
1.1.2.5 paf 156:
1.1.2.1 paf 157: if (CORD_IS_STRING(x)) {
158: lenx = strlen(x);
159: result_len = lenx + leny;
160: if (result_len <= SHORT_LIMIT) {
161: register char * result = GC_MALLOC_ATOMIC(result_len+1);
1.1.2.7 paf 162:
1.1.2.1 paf 163: if (result == 0) OUT_OF_MEMORY;
164: memcpy(result, x, lenx);
165: memcpy(result + lenx, y, leny);
166: result[result_len] = '\0';
167: return((CORD) result);
168: } else {
169: depth = 1;
170: }
171: } else {
172: register CORD right;
173: register CORD left;
174: register char * new_right;
175: register size_t right_len;
176:
177: lenx = LEN(x);
178:
179: if (leny <= SHORT_LIMIT/2
180: && IS_CONCATENATION(x)
181: && CORD_IS_STRING(right = ((CordRep *)x) -> concatenation.right)) {
182: /* Merge y into right part of x. */
183: if (!CORD_IS_STRING(left = ((CordRep *)x) -> concatenation.left)) {
184: right_len = lenx - LEN(left);
185: } else if (((CordRep *)x) -> concatenation.left_len != 0) {
186: right_len = lenx - ((CordRep *)x) -> concatenation.left_len;
187: } else {
188: right_len = strlen(right);
189: }
190: result_len = right_len + leny; /* length of new_right */
191: if (result_len <= SHORT_LIMIT) {
192: new_right = GC_MALLOC_ATOMIC(result_len + 1);
193: memcpy(new_right, right, right_len);
194: memcpy(new_right + right_len, y, leny);
195: new_right[result_len] = '\0';
196: y = new_right;
197: leny = result_len;
198: x = left;
199: lenx -= right_len;
200: /* Now fall through to concatenate the two pieces: */
201: }
202: if (CORD_IS_STRING(x)) {
203: depth = 1;
204: } else {
205: depth = DEPTH(x) + 1;
206: }
207: } else {
208: depth = DEPTH(x) + 1;
209: }
210: result_len = lenx + leny;
211: }
212: {
213: /* The general case; lenx, result_len is known: */
214: register struct Concatenation * result;
215:
216: result = GC_NEW(struct Concatenation);
217: if (result == 0) OUT_OF_MEMORY;
218: result->header = CONCAT_HDR;
219: result->depth = depth;
220: if (lenx <= MAX_LEFT_LEN) result->left_len = lenx;
221: result->len = result_len;
222: result->left = x;
223: result->right = y;
224: if (depth > MAX_DEPTH) {
225: return(CORD_balance((CORD)result));
226: } else {
227: return((CORD) result);
228: }
229: }
230: }
231:
232:
233: CORD CORD_cat(CORD x, CORD y)
234: {
235: register size_t result_len;
236: register int depth;
237: register size_t lenx;
238:
239: if (x == CORD_EMPTY) return(y);
240: if (y == CORD_EMPTY) return(x);
241: if (CORD_IS_STRING(y)) {
242: return(CORD_cat_char_star(x, y, strlen(y)));
243: } else if (CORD_IS_STRING(x)) {
244: lenx = strlen(x);
245: depth = DEPTH(y) + 1;
246: } else {
247: register int depthy = DEPTH(y);
248:
249: lenx = LEN(x);
250: depth = DEPTH(x) + 1;
251: if (depthy >= depth) depth = depthy + 1;
252: }
253: result_len = lenx + LEN(y);
254: {
255: register struct Concatenation * result;
256:
257: result = GC_NEW(struct Concatenation);
258: if (result == 0) OUT_OF_MEMORY;
259: result->header = CONCAT_HDR;
260: result->depth = depth;
1.1.2.8 ! paf 261: // printf("depth=%d\n", depth);
1.1.2.1 paf 262: if (lenx <= MAX_LEFT_LEN) result->left_len = lenx;
263: result->len = result_len;
264: result->left = x;
265: result->right = y;
1.1.2.8 ! paf 266: // PAF@design.ru bug fix:
! 267: if (depth > MAX_DEPTH) {
! 268: return(CORD_balance((CORD)result));
! 269: } else {
! 270: return((CORD) result);
! 271: }
1.1.2.1 paf 272: }
273: }
274:
275:
276:
277: CORD CORD_from_fn(CORD_fn fn, void * client_data, size_t len)
278: {
279: if (len <= 0) return(0);
280: if (len <= SHORT_LIMIT) {
281: register char * result;
282: register size_t i;
283: char buf[SHORT_LIMIT+1];
284: register char c;
285:
286: for (i = 0; i < len; i++) {
287: c = (*fn)(i, client_data);
288: if (c == '\0') goto gen_case;
289: buf[i] = c;
290: }
291: buf[i] = '\0';
292: result = GC_MALLOC_ATOMIC(len+1);
293: if (result == 0) OUT_OF_MEMORY;
294: strcpy(result, buf);
295: result[len] = '\0';
296: return((CORD) result);
297: }
298: gen_case:
299: {
300: register struct Function * result;
301:
302: result = GC_NEW(struct Function);
303: if (result == 0) OUT_OF_MEMORY;
304: result->header = FN_HDR;
305: /* depth is already 0 */
306: result->len = len;
307: result->fn = fn;
308: result->client_data = client_data;
309: return((CORD) result);
310: }
311: }
312:
313: size_t CORD_len(CORD x)
314: {
315: if (x == 0) {
316: return(0);
317: } else {
318: return(GEN_LEN(x));
319: }
320: }
321:
322: struct substr_args {
323: CordRep * sa_cord;
324: size_t sa_index;
325: };
326:
327: char CORD_index_access_fn(size_t i, void * client_data)
328: {
329: register struct substr_args *descr = (struct substr_args *)client_data;
330:
331: return(((char *)(descr->sa_cord))[i + descr->sa_index]);
332: }
333:
334: char CORD_apply_access_fn(size_t i, void * client_data)
335: {
336: register struct substr_args *descr = (struct substr_args *)client_data;
337: register struct Function * fn_cord = &(descr->sa_cord->function);
338:
339: return((*(fn_cord->fn))(i + descr->sa_index, fn_cord->client_data));
340: }
341:
342: /* A version of CORD_substr that simply returns a function node, thus */
343: /* postponing its work. The fourth argument is a function that may */
344: /* be used for efficient access to the ith character. */
345: /* Assumes i >= 0 and i + n < length(x). */
346: CORD CORD_substr_closure(CORD x, size_t i, size_t n, CORD_fn f)
347: {
348: register struct substr_args * sa = GC_NEW(struct substr_args);
349: CORD result;
350:
351: if (sa == 0) OUT_OF_MEMORY;
352: sa->sa_cord = (CordRep *)x;
353: sa->sa_index = i;
354: result = CORD_from_fn(f, (void *)sa, n);
355: ((CordRep *)result) -> function.header = SUBSTR_HDR;
356: return (result);
357: }
358:
359: # define SUBSTR_LIMIT (10 * SHORT_LIMIT)
360: /* Substrings of function nodes and flat strings shorter than */
361: /* this are flat strings. Othewise we use a functional */
362: /* representation, which is significantly slower to access. */
363:
364: /* A version of CORD_substr that assumes i >= 0, n > 0, and i + n < length(x).*/
365: CORD CORD_substr_checked(CORD x, size_t i, size_t n)
366: {
367: if (CORD_IS_STRING(x)) {
368: if (n > SUBSTR_LIMIT) {
369: return(CORD_substr_closure(x, i, n, CORD_index_access_fn));
370: } else {
371: register char * result = GC_MALLOC_ATOMIC(n+1);
372:
373: if (result == 0) OUT_OF_MEMORY;
374: strncpy(result, x+i, n);
375: result[n] = '\0';
376: return(result);
377: }
378: } else if (IS_CONCATENATION(x)) {
379: register struct Concatenation * conc
380: = &(((CordRep *)x) -> concatenation);
381: register size_t left_len;
382: register size_t right_len;
383:
384: left_len = LEFT_LEN(conc);
385: right_len = conc -> len - left_len;
386: if (i >= left_len) {
387: if (n == right_len) return(conc -> right);
388: return(CORD_substr_checked(conc -> right, i - left_len, n));
389: } else if (i+n <= left_len) {
390: if (n == left_len) return(conc -> left);
391: return(CORD_substr_checked(conc -> left, i, n));
392: } else {
393: /* Need at least one character from each side. */
394: register CORD left_part;
395: register CORD right_part;
396: register size_t left_part_len = left_len - i;
397:
398: if (i == 0) {
399: left_part = conc -> left;
400: } else {
401: left_part = CORD_substr_checked(conc -> left, i, left_part_len);
402: }
403: if (i + n == right_len + left_len) {
404: right_part = conc -> right;
405: } else {
406: right_part = CORD_substr_checked(conc -> right, 0,
407: n - left_part_len);
408: }
409: return(CORD_cat(left_part, right_part));
410: }
411: } else /* function */ {
412: if (n > SUBSTR_LIMIT) {
413: if (IS_SUBSTR(x)) {
414: /* Avoid nesting substring nodes. */
415: register struct Function * f = &(((CordRep *)x) -> function);
416: register struct substr_args *descr =
417: (struct substr_args *)(f -> client_data);
418:
419: return(CORD_substr_closure((CORD)descr->sa_cord,
420: i + descr->sa_index,
421: n, f -> fn));
422: } else {
423: return(CORD_substr_closure(x, i, n, CORD_apply_access_fn));
424: }
425: } else {
426: char * result;
427: register struct Function * f = &(((CordRep *)x) -> function);
428: char buf[SUBSTR_LIMIT+1];
429: register char * p = buf;
430: register char c;
431: register int j;
432: register int lim = i + n;
433:
434: for (j = i; j < lim; j++) {
435: c = (*(f -> fn))(j, f -> client_data);
436: if (c == '\0') {
437: return(CORD_substr_closure(x, i, n, CORD_apply_access_fn));
438: }
439: *p++ = c;
440: }
441: *p = '\0';
442: result = GC_MALLOC_ATOMIC(n+1);
443: if (result == 0) OUT_OF_MEMORY;
444: strcpy(result, buf);
445: return(result);
446: }
447: }
448: }
449:
450: CORD CORD_substr(CORD x, size_t i, size_t n)
451: {
452: register size_t len = CORD_len(x);
453:
454: if (i >= len || n <= 0) return(0);
455: /* n < 0 is impossible in a correct C implementation, but */
456: /* quite possible under SunOS 4.X. */
457: if (i + n > len) n = len - i;
458: # ifndef __STDC__
459: if (i < 0) ABORT("CORD_substr: second arg. negative");
460: /* Possible only if both client and C implementation are buggy. */
461: /* But empirically this happens frequently. */
462: # endif
463: return(CORD_substr_checked(x, i, n));
464: }
465:
466: /* See cord.h for definition. We assume i is in range. */
467: int CORD_iter5(CORD x, size_t i, CORD_iter_fn f1,
468: CORD_batched_iter_fn f2, void * client_data)
469: {
470: if (x == 0) return(0);
471: if (CORD_IS_STRING(x)) {
1.1.2.2 paf 472: register const char* p = x+i;
1.1.2.1 paf 473:
474: if (*p == '\0') ABORT("2nd arg to CORD_iter5 too big");
475: if (f2 != CORD_NO_FN) {
476: return((*f2)(p, client_data));
477: } else {
478: while (*p) {
479: if ((*f1)(*p, client_data)) return(1);
480: p++;
481: }
482: return(0);
483: }
484: } else if (IS_CONCATENATION(x)) {
485: register struct Concatenation * conc
486: = &(((CordRep *)x) -> concatenation);
487:
488:
489: if (i > 0) {
490: register size_t left_len = LEFT_LEN(conc);
491:
492: if (i >= left_len) {
493: return(CORD_iter5(conc -> right, i - left_len, f1, f2,
494: client_data));
495: }
496: }
497: if (CORD_iter5(conc -> left, i, f1, f2, client_data)) {
498: return(1);
499: }
500: return(CORD_iter5(conc -> right, 0, f1, f2, client_data));
501: } else /* function */ {
502: register struct Function * f = &(((CordRep *)x) -> function);
503: register size_t j;
504: register size_t lim = f -> len;
505:
506: for (j = i; j < lim; j++) {
507: if ((*f1)((*(f -> fn))(j, f -> client_data), client_data)) {
508: return(1);
509: }
510: }
511: return(0);
512: }
513: }
514:
515: #undef CORD_iter
516: int CORD_iter(CORD x, CORD_iter_fn f1, void * client_data)
517: {
518: return(CORD_iter5(x, 0, f1, CORD_NO_FN, client_data));
519: }
520:
521: int CORD_riter4(CORD x, size_t i, CORD_iter_fn f1, void * client_data)
522: {
523: if (x == 0) return(0);
524: if (CORD_IS_STRING(x)) {
1.1.2.2 paf 525: register const char* p = x + i;
1.1.2.1 paf 526: register char c;
527:
528: for(;;) {
529: c = *p;
530: if (c == '\0') ABORT("2nd arg to CORD_riter4 too big");
531: if ((*f1)(c, client_data)) return(1);
532: if (p == x) break;
533: p--;
534: }
535: return(0);
536: } else if (IS_CONCATENATION(x)) {
537: register struct Concatenation * conc
538: = &(((CordRep *)x) -> concatenation);
539: register CORD left_part = conc -> left;
540: register size_t left_len;
541:
542: left_len = LEFT_LEN(conc);
543: if (i >= left_len) {
544: if (CORD_riter4(conc -> right, i - left_len, f1, client_data)) {
545: return(1);
546: }
547: return(CORD_riter4(left_part, left_len - 1, f1, client_data));
548: } else {
549: return(CORD_riter4(left_part, i, f1, client_data));
550: }
551: } else /* function */ {
552: register struct Function * f = &(((CordRep *)x) -> function);
553: register size_t j;
554:
555: for (j = i; ; j--) {
556: if ((*f1)((*(f -> fn))(j, f -> client_data), client_data)) {
557: return(1);
558: }
559: if (j == 0) return(0);
560: }
561: }
562: }
563:
564: int CORD_riter(CORD x, CORD_iter_fn f1, void * client_data)
565: {
566: return(CORD_riter4(x, CORD_len(x) - 1, f1, client_data));
567: }
568:
569: /*
570: * The following functions are concerned with balancing cords.
571: * Strategy:
572: * Scan the cord from left to right, keeping the cord scanned so far
573: * as a forest of balanced trees of exponentialy decreasing length.
574: * When a new subtree needs to be added to the forest, we concatenate all
575: * shorter ones to the new tree in the appropriate order, and then insert
576: * the result into the forest.
577: * Crucial invariants:
578: * 1. The concatenation of the forest (in decreasing order) with the
579: * unscanned part of the rope is equal to the rope being balanced.
580: * 2. All trees in the forest are balanced.
581: * 3. forest[i] has depth at most i.
582: */
583:
584: typedef struct {
585: CORD c;
586: size_t len; /* Actual length of c */
587: } ForestElement;
588:
589: static size_t min_len [ MAX_DEPTH ];
590:
591: static int min_len_init = 0;
592:
593: int CORD_max_len;
594:
595: typedef ForestElement Forest [ MAX_DEPTH ];
596: /* forest[i].len >= fib(i+1) */
597: /* The string is the concatenation */
598: /* of the forest in order of DECREASING */
599: /* indices. */
600:
601: void CORD_init_min_len()
602: {
603: register int i;
604: register size_t last, previous, current;
605:
606: min_len[0] = previous = 1;
607: min_len[1] = last = 2;
608: for (i = 2; i < MAX_DEPTH; i++) {
609: current = last + previous;
610: if (current < last) /* overflow */ current = last;
611: min_len[i] = current;
612: previous = last;
613: last = current;
614: }
615: CORD_max_len = last - 1;
616: min_len_init = 1;
617: }
618:
619:
620: void CORD_init_forest(ForestElement * forest, size_t max_len)
621: {
622: register int i;
623:
624: for (i = 0; i < MAX_DEPTH; i++) {
625: forest[i].c = 0;
626: if (min_len[i] > max_len) return;
627: }
628: ABORT("Cord too long");
629: }
630:
631: /* Add a leaf to the appropriate level in the forest, cleaning */
632: /* out lower levels as necessary. */
633: /* Also works if x is a balanced tree of concatenations; however */
634: /* in this case an extra concatenation node may be inserted above x; */
635: /* This node should not be counted in the statement of the invariants. */
636: void CORD_add_forest(ForestElement * forest, CORD x, size_t len)
637: {
638: register int i = 0;
639: register CORD sum = CORD_EMPTY;
640: register size_t sum_len = 0;
641:
642: while (len > min_len[i + 1]) {
643: if (forest[i].c != 0) {
644: sum = CORD_cat(forest[i].c, sum);
645: sum_len += forest[i].len;
646: forest[i].c = 0;
647: }
648: i++;
649: }
650: /* Sum has depth at most 1 greter than what would be required */
651: /* for balance. */
652: sum = CORD_cat(sum, x);
653: sum_len += len;
654: /* If x was a leaf, then sum is now balanced. To see this */
655: /* consider the two cases in which forest[i-1] either is or is */
656: /* not empty. */
657: while (sum_len >= min_len[i]) {
658: if (forest[i].c != 0) {
659: sum = CORD_cat(forest[i].c, sum);
660: sum_len += forest[i].len;
661: /* This is again balanced, since sum was balanced, and has */
662: /* allowable depth that differs from i by at most 1. */
663: forest[i].c = 0;
664: }
665: i++;
666: }
667: i--;
668: forest[i].c = sum;
669: forest[i].len = sum_len;
670: }
671:
672: CORD CORD_concat_forest(ForestElement * forest, size_t expected_len)
673: {
674: register int i = 0;
675: CORD sum = 0;
676: size_t sum_len = 0;
677:
678: while (sum_len != expected_len) {
679: if (forest[i].c != 0) {
680: sum = CORD_cat(forest[i].c, sum);
681: sum_len += forest[i].len;
682: }
683: i++;
684: }
685: return(sum);
686: }
687:
688: /* Insert the frontier of x into forest. Balanced subtrees are */
689: /* treated as leaves. This potentially adds one to the depth */
690: /* of the final tree. */
691: void CORD_balance_insert(CORD x, size_t len, ForestElement * forest)
692: {
693: register int depth;
694:
695: if (CORD_IS_STRING(x)) {
696: CORD_add_forest(forest, x, len);
697: } else if (IS_CONCATENATION(x)
698: && ((depth = DEPTH(x)) >= MAX_DEPTH
699: || len < min_len[depth])) {
700: register struct Concatenation * conc
701: = &(((CordRep *)x) -> concatenation);
702: size_t left_len = LEFT_LEN(conc);
703:
704: CORD_balance_insert(conc -> left, left_len, forest);
705: CORD_balance_insert(conc -> right, len - left_len, forest);
706: } else /* function or balanced */ {
707: CORD_add_forest(forest, x, len);
708: }
709: }
710:
711:
712: CORD CORD_balance(CORD x)
713: {
714: Forest forest;
715: register size_t len;
716:
717: if (x == 0) return(0);
718: if (CORD_IS_STRING(x)) return(x);
719: if (!min_len_init) CORD_init_min_len();
720: len = LEN(x);
721: CORD_init_forest(forest, len);
722: CORD_balance_insert(x, len, forest);
723: return(CORD_concat_forest(forest, len));
724: }
725:
726:
727: /* Position primitives */
728:
729: /* Private routines to deal with the hard cases only: */
730:
731: /* P contains a prefix of the path to cur_pos. Extend it to a full */
732: /* path and set up leaf info. */
733: /* Return 0 if past the end of cord, 1 o.w. */
734: void CORD__extend_path(register CORD_pos p)
735: {
736: register struct CORD_pe * current_pe = &(p[0].path[p[0].path_len]);
737: register CORD top = current_pe -> pe_cord;
738: register size_t pos = p[0].cur_pos;
739: register size_t top_pos = current_pe -> pe_start_pos;
740: register size_t top_len = GEN_LEN(top);
741:
742: /* Fill in the rest of the path. */
743: while(!CORD_IS_STRING(top) && IS_CONCATENATION(top)) {
744: register struct Concatenation * conc =
745: &(((CordRep *)top) -> concatenation);
746: register size_t left_len;
747:
748: left_len = LEFT_LEN(conc);
749: current_pe++;
750: if (pos >= top_pos + left_len) {
751: current_pe -> pe_cord = top = conc -> right;
752: current_pe -> pe_start_pos = top_pos = top_pos + left_len;
753: top_len -= left_len;
754: } else {
755: current_pe -> pe_cord = top = conc -> left;
756: current_pe -> pe_start_pos = top_pos;
757: top_len = left_len;
758: }
759: p[0].path_len++;
760: }
761: /* Fill in leaf description for fast access. */
762: if (CORD_IS_STRING(top)) {
763: p[0].cur_leaf = top;
764: p[0].cur_start = top_pos;
765: p[0].cur_end = top_pos + top_len;
766: } else {
767: p[0].cur_end = 0;
768: }
769: if (pos >= top_pos + top_len) p[0].path_len = CORD_POS_INVALID;
770: }
771:
772: char CORD__pos_fetch(register CORD_pos p)
773: {
774: /* Leaf is a function node */
775: struct CORD_pe * pe = &((p)[0].path[(p)[0].path_len]);
776: CORD leaf = pe -> pe_cord;
777: register struct Function * f = &(((CordRep *)leaf) -> function);
778:
779: if (!IS_FUNCTION(leaf)) ABORT("CORD_pos_fetch: bad leaf");
780: return ((*(f -> fn))(p[0].cur_pos - pe -> pe_start_pos, f -> client_data));
781: }
782:
783: void CORD__next(register CORD_pos p)
784: {
785: register size_t cur_pos = p[0].cur_pos + 1;
786: register struct CORD_pe * current_pe = &((p)[0].path[(p)[0].path_len]);
787: register CORD leaf = current_pe -> pe_cord;
788:
789: /* Leaf is not a string or we're at end of leaf */
790: p[0].cur_pos = cur_pos;
791: if (!CORD_IS_STRING(leaf)) {
792: /* Function leaf */
793: register struct Function * f = &(((CordRep *)leaf) -> function);
794: register size_t start_pos = current_pe -> pe_start_pos;
795: register size_t end_pos = start_pos + f -> len;
796:
797: if (cur_pos < end_pos) {
798: /* Fill cache and return. */
799: register size_t i;
800: register size_t limit = cur_pos + FUNCTION_BUF_SZ;
801: register CORD_fn fn = f -> fn;
802: register void * client_data = f -> client_data;
803:
804: if (limit > end_pos) {
805: limit = end_pos;
806: }
807: for (i = cur_pos; i < limit; i++) {
808: p[0].function_buf[i - cur_pos] =
809: (*fn)(i - start_pos, client_data);
810: }
811: p[0].cur_start = cur_pos;
812: p[0].cur_leaf = p[0].function_buf;
813: p[0].cur_end = limit;
814: return;
815: }
816: }
817: /* End of leaf */
818: /* Pop the stack until we find two concatenation nodes with the */
819: /* same start position: this implies we were in left part. */
820: {
821: while (p[0].path_len > 0
822: && current_pe[0].pe_start_pos != current_pe[-1].pe_start_pos) {
823: p[0].path_len--;
824: current_pe--;
825: }
826: if (p[0].path_len == 0) {
827: p[0].path_len = CORD_POS_INVALID;
828: return;
829: }
830: }
831: p[0].path_len--;
832: CORD__extend_path(p);
833: }
834:
835: void CORD__prev(register CORD_pos p)
836: {
837: register struct CORD_pe * pe = &(p[0].path[p[0].path_len]);
838:
839: if (p[0].cur_pos == 0) {
840: p[0].path_len = CORD_POS_INVALID;
841: return;
842: }
843: p[0].cur_pos--;
844: if (p[0].cur_pos >= pe -> pe_start_pos) return;
845:
846: /* Beginning of leaf */
847:
848: /* Pop the stack until we find two concatenation nodes with the */
849: /* different start position: this implies we were in right part. */
850: {
851: register struct CORD_pe * current_pe = &((p)[0].path[(p)[0].path_len]);
852:
853: while (p[0].path_len > 0
854: && current_pe[0].pe_start_pos == current_pe[-1].pe_start_pos) {
855: p[0].path_len--;
856: current_pe--;
857: }
858: }
859: p[0].path_len--;
860: CORD__extend_path(p);
861: }
862:
863: #undef CORD_pos_fetch
864: #undef CORD_next
865: #undef CORD_prev
866: #undef CORD_pos_to_index
867: #undef CORD_pos_to_cord
868: #undef CORD_pos_valid
869:
870: char CORD_pos_fetch(register CORD_pos p)
871: {
872: if (p[0].cur_start <= p[0].cur_pos && p[0].cur_pos < p[0].cur_end) {
873: return(p[0].cur_leaf[p[0].cur_pos - p[0].cur_start]);
874: } else {
875: return(CORD__pos_fetch(p));
876: }
877: }
878:
879: void CORD_next(CORD_pos p)
880: {
881: if (p[0].cur_pos < p[0].cur_end - 1) {
882: p[0].cur_pos++;
883: } else {
884: CORD__next(p);
885: }
886: }
887:
888: void CORD_prev(CORD_pos p)
889: {
890: if (p[0].cur_end != 0 && p[0].cur_pos > p[0].cur_start) {
891: p[0].cur_pos--;
892: } else {
893: CORD__prev(p);
894: }
895: }
896:
897: size_t CORD_pos_to_index(CORD_pos p)
898: {
899: return(p[0].cur_pos);
900: }
901:
902: CORD CORD_pos_to_cord(CORD_pos p)
903: {
904: return(p[0].path[0].pe_cord);
905: }
906:
907: int CORD_pos_valid(CORD_pos p)
908: {
909: return(p[0].path_len != CORD_POS_INVALID);
910: }
911:
912: void CORD_set_pos(CORD_pos p, CORD x, size_t i)
913: {
914: if (x == CORD_EMPTY) {
915: p[0].path_len = CORD_POS_INVALID;
916: return;
917: }
918: p[0].path[0].pe_cord = x;
919: p[0].path[0].pe_start_pos = 0;
920: p[0].path_len = 0;
921: p[0].cur_pos = i;
922: CORD__extend_path(p);
923: }
E-mail:
![[BACK]](/cvsweb/icons/back.gif){kind=icon}
Return to cordbscs.c CVS log ![[TXT]](/cvsweb/icons/text.gif){kind=icon}
![[DIR]](/cvsweb/icons/dir.gif){kind=icon}
Up to [parser3project] / parser3 / src / lib / cord