1/*
2 * LibXDiff by Davide Libenzi ( File Differential Library )
3 * Copyright (C) 2003 Davide Libenzi
4 *
5 * This library is free software; you can redistribute it and/or
6 * modify it under the terms of the GNU Lesser General Public
7 * License as published by the Free Software Foundation; either
8 * version 2.1 of the License, or (at your option) any later version.
9 *
10 * This library is distributed in the hope that it will be useful,
11 * but WITHOUT ANY WARRANTY; without even the implied warranty of
12 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
13 * Lesser General Public License for more details.
14 *
15 * You should have received a copy of the GNU Lesser General Public
16 * License along with this library; if not, write to the Free Software
17 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
18 *
19 * Davide Libenzi <davidel@xmailserver.org>
20 *
21 */
22
23#include "xinclude.h"
24
25
26
27
28long xdl_bogosqrt(long n) {
29 long i;
30
31 /*
32 * Classical integer square root approximation using shifts.
33 */
34 for (i = 1; n > 0; n >>= 2)
35 i <<= 1;
36
37 return i;
38}
39
40
41int xdl_emit_diffrec(char const *rec, long size, char const *pre, long psize,
42 xdemitcb_t *ecb) {
43 int i = 2;
44 mmbuffer_t mb[3];
45
46 mb[0].ptr = (char *) pre;
47 mb[0].size = psize;
48 mb[1].ptr = (char *) rec;
49 mb[1].size = size;
50 if (size > 0 && rec[size - 1] != '\n') {
51 mb[2].ptr = (char *) "\n\\ No newline at end of file\n";
52 mb[2].size = strlen(mb[2].ptr);
53 i++;
54 }
55 if (ecb->outf(ecb->priv, mb, i) < 0) {
56
57 return -1;
58 }
59
60 return 0;
61}
62
63void *xdl_mmfile_first(mmfile_t *mmf, long *size)
64{
65 *size = mmf->size;
66 return mmf->ptr;
67}
68
69
70long xdl_mmfile_size(mmfile_t *mmf)
71{
72 return mmf->size;
73}
74
75
76int xdl_cha_init(chastore_t *cha, long isize, long icount) {
77
78 cha->head = cha->tail = NULL;
79 cha->isize = isize;
80 cha->nsize = icount * isize;
81 cha->ancur = cha->sncur = NULL;
82 cha->scurr = 0;
83
84 return 0;
85}
86
87
88void xdl_cha_free(chastore_t *cha) {
89 chanode_t *cur, *tmp;
90
91 for (cur = cha->head; (tmp = cur) != NULL;) {
92 cur = cur->next;
93 xdl_free(tmp);
94 }
95}
96
97
98void *xdl_cha_alloc(chastore_t *cha) {
99 chanode_t *ancur;
100 void *data;
101
102 if (!(ancur = cha->ancur) || ancur->icurr == cha->nsize) {
103 if (!(ancur = (chanode_t *) xdl_malloc(sizeof(chanode_t) + cha->nsize))) {
104
105 return NULL;
106 }
107 ancur->icurr = 0;
108 ancur->next = NULL;
109 if (cha->tail)
110 cha->tail->next = ancur;
111 if (!cha->head)
112 cha->head = ancur;
113 cha->tail = ancur;
114 cha->ancur = ancur;
115 }
116
117 data = (char *) ancur + sizeof(chanode_t) + ancur->icurr;
118 ancur->icurr += cha->isize;
119
120 return data;
121}
122
123
124void *xdl_cha_first(chastore_t *cha) {
125 chanode_t *sncur;
126
127 if (!(cha->sncur = sncur = cha->head))
128 return NULL;
129
130 cha->scurr = 0;
131
132 return (char *) sncur + sizeof(chanode_t) + cha->scurr;
133}
134
135
136void *xdl_cha_next(chastore_t *cha) {
137 chanode_t *sncur;
138
139 if (!(sncur = cha->sncur))
140 return NULL;
141 cha->scurr += cha->isize;
142 if (cha->scurr == sncur->icurr) {
143 if (!(sncur = cha->sncur = sncur->next))
144 return NULL;
145 cha->scurr = 0;
146 }
147
148 return (char *) sncur + sizeof(chanode_t) + cha->scurr;
149}
150
151
152long xdl_guess_lines(mmfile_t *mf, long sample) {
153 long nl = 0, size, tsize = 0;
154 char const *data, *cur, *top;
155
156 if ((cur = data = xdl_mmfile_first(mf, &size)) != NULL) {
157 for (top = data + size; nl < sample && cur < top; ) {
158 nl++;
159 if (!(cur = memchr(cur, '\n', top - cur)))
160 cur = top;
161 else
162 cur++;
163 }
164 tsize += (long) (cur - data);
165 }
166
167 if (nl && tsize)
168 nl = xdl_mmfile_size(mf) / (tsize / nl);
169
170 return nl + 1;
171}
172
173int xdl_recmatch(const char *l1, long s1, const char *l2, long s2, long flags)
174{
175 int i1, i2;
176
177 if (s1 == s2 && !memcmp(l1, l2, s1))
178 return 1;
179 if (!(flags & XDF_WHITESPACE_FLAGS))
180 return 0;
181
182 i1 = 0;
183 i2 = 0;
184
185 /*
186 * -w matches everything that matches with -b, and -b in turn
187 * matches everything that matches with --ignore-space-at-eol.
188 *
189 * Each flavor of ignoring needs different logic to skip whitespaces
190 * while we have both sides to compare.
191 */
192 if (flags & XDF_IGNORE_WHITESPACE) {
193 goto skip_ws;
194 while (i1 < s1 && i2 < s2) {
195 if (l1[i1++] != l2[i2++])
196 return 0;
197 skip_ws:
198 while (i1 < s1 && XDL_ISSPACE(l1[i1]))
199 i1++;
200 while (i2 < s2 && XDL_ISSPACE(l2[i2]))
201 i2++;
202 }
203 } else if (flags & XDF_IGNORE_WHITESPACE_CHANGE) {
204 while (i1 < s1 && i2 < s2) {
205 if (XDL_ISSPACE(l1[i1]) && XDL_ISSPACE(l2[i2])) {
206 /* Skip matching spaces and try again */
207 while (i1 < s1 && XDL_ISSPACE(l1[i1]))
208 i1++;
209 while (i2 < s2 && XDL_ISSPACE(l2[i2]))
210 i2++;
211 continue;
212 }
213 if (l1[i1++] != l2[i2++])
214 return 0;
215 }
216 } else if (flags & XDF_IGNORE_WHITESPACE_AT_EOL) {
217 while (i1 < s1 && i2 < s2 && l1[i1++] == l2[i2++])
218 ; /* keep going */
219 }
220
221 /*
222 * After running out of one side, the remaining side must have
223 * nothing but whitespace for the lines to match. Note that
224 * ignore-whitespace-at-eol case may break out of the loop
225 * while there still are characters remaining on both lines.
226 */
227 if (i1 < s1) {
228 while (i1 < s1 && XDL_ISSPACE(l1[i1]))
229 i1++;
230 if (s1 != i1)
231 return 0;
232 }
233 if (i2 < s2) {
234 while (i2 < s2 && XDL_ISSPACE(l2[i2]))
235 i2++;
236 return (s2 == i2);
237 }
238 return 1;
239}
240
241static unsigned long xdl_hash_record_with_whitespace(char const **data,
242 char const *top, long flags) {
243 unsigned long ha = 5381;
244 char const *ptr = *data;
245
246 for (; ptr < top && *ptr != '\n'; ptr++) {
247 if (XDL_ISSPACE(*ptr)) {
248 const char *ptr2 = ptr;
249 int at_eol;
250 while (ptr + 1 < top && XDL_ISSPACE(ptr[1])
251 && ptr[1] != '\n')
252 ptr++;
253 at_eol = (top <= ptr + 1 || ptr[1] == '\n');
254 if (flags & XDF_IGNORE_WHITESPACE)
255 ; /* already handled */
256 else if (flags & XDF_IGNORE_WHITESPACE_CHANGE
257 && !at_eol) {
258 ha += (ha << 5);
259 ha ^= (unsigned long) ' ';
260 }
261 else if (flags & XDF_IGNORE_WHITESPACE_AT_EOL
262 && !at_eol) {
263 while (ptr2 != ptr + 1) {
264 ha += (ha << 5);
265 ha ^= (unsigned long) *ptr2;
266 ptr2++;
267 }
268 }
269 continue;
270 }
271 ha += (ha << 5);
272 ha ^= (unsigned long) *ptr;
273 }
274 *data = ptr < top ? ptr + 1: ptr;
275
276 return ha;
277}
278
279
280unsigned long xdl_hash_record(char const **data, char const *top, long flags) {
281 unsigned long ha = 5381;
282 char const *ptr = *data;
283
284 if (flags & XDF_WHITESPACE_FLAGS)
285 return xdl_hash_record_with_whitespace(data, top, flags);
286
287 for (; ptr < top && *ptr != '\n'; ptr++) {
288 ha += (ha << 5);
289 ha ^= (unsigned long) *ptr;
290 }
291 *data = ptr < top ? ptr + 1: ptr;
292
293 return ha;
294}
295
296
297unsigned int xdl_hashbits(unsigned int size) {
298 unsigned int val = 1, bits = 0;
299
300 for (; val < size && bits < CHAR_BIT * sizeof(unsigned int); val <<= 1, bits++);
301 return bits ? bits: 1;
302}
303
304
305int xdl_num_out(char *out, long val) {
306 char *ptr, *str = out;
307 char buf[32];
308
309 ptr = buf + sizeof(buf) - 1;
310 *ptr = '\0';
311 if (val < 0) {
312 *--ptr = '-';
313 val = -val;
314 }
315 for (; val && ptr > buf; val /= 10)
316 *--ptr = "0123456789"[val % 10];
317 if (*ptr)
318 for (; *ptr; ptr++, str++)
319 *str = *ptr;
320 else
321 *str++ = '0';
322 *str = '\0';
323
324 return str - out;
325}
326
327
328long xdl_atol(char const *str, char const **next) {
329 long val, base;
330 char const *top;
331
332 for (top = str; XDL_ISDIGIT(*top); top++);
333 if (next)
334 *next = top;
335 for (val = 0, base = 1, top--; top >= str; top--, base *= 10)
336 val += base * (long)(*top - '0');
337 return val;
338}
339
340
341int xdl_emit_hunk_hdr(long s1, long c1, long s2, long c2,
342 const char *func, long funclen, xdemitcb_t *ecb) {
343 int nb = 0;
344 mmbuffer_t mb;
345 char buf[128];
346
347 memcpy(buf, "@@ -", 4);
348 nb += 4;
349
350 nb += xdl_num_out(buf + nb, c1 ? s1: s1 - 1);
351
352 if (c1 != 1) {
353 memcpy(buf + nb, ",", 1);
354 nb += 1;
355
356 nb += xdl_num_out(buf + nb, c1);
357 }
358
359 memcpy(buf + nb, " +", 2);
360 nb += 2;
361
362 nb += xdl_num_out(buf + nb, c2 ? s2: s2 - 1);
363
364 if (c2 != 1) {
365 memcpy(buf + nb, ",", 1);
366 nb += 1;
367
368 nb += xdl_num_out(buf + nb, c2);
369 }
370
371 memcpy(buf + nb, " @@", 3);
372 nb += 3;
373 if (func && funclen) {
374 buf[nb++] = ' ';
375 if (funclen > sizeof(buf) - nb - 1)
376 funclen = sizeof(buf) - nb - 1;
377 memcpy(buf + nb, func, funclen);
378 nb += funclen;
379 }
380 buf[nb++] = '\n';
381
382 mb.ptr = buf;
383 mb.size = nb;
384 if (ecb->outf(ecb->priv, &mb, 1) < 0)
385 return -1;
386
387 return 0;
388}
389
390int xdl_fall_back_diff(xdfenv_t *diff_env, xpparam_t const *xpp,
391 int line1, int count1, int line2, int count2)
392{
393 /*
394 * This probably does not work outside Git, since
395 * we have a very simple mmfile structure.
396 *
397 * Note: ideally, we would reuse the prepared environment, but
398 * the libxdiff interface does not (yet) allow for diffing only
399 * ranges of lines instead of the whole files.
400 */
401 mmfile_t subfile1, subfile2;
402 xdfenv_t env;
403
404 subfile1.ptr = (char *)diff_env->xdf1.recs[line1 - 1]->ptr;
405 subfile1.size = diff_env->xdf1.recs[line1 + count1 - 2]->ptr +
406 diff_env->xdf1.recs[line1 + count1 - 2]->size - subfile1.ptr;
407 subfile2.ptr = (char *)diff_env->xdf2.recs[line2 - 1]->ptr;
408 subfile2.size = diff_env->xdf2.recs[line2 + count2 - 2]->ptr +
409 diff_env->xdf2.recs[line2 + count2 - 2]->size - subfile2.ptr;
410 if (xdl_do_diff(&subfile1, &subfile2, xpp, &env) < 0)
411 return -1;
412
413 memcpy(diff_env->xdf1.rchg + line1 - 1, env.xdf1.rchg, count1);
414 memcpy(diff_env->xdf2.rchg + line2 - 1, env.xdf2.rchg, count2);
415
416 xdl_free_env(&env);
417
418 return 0;
419}