xdiff/xprepare: use a smaller sample size for histogram diff
[git.git] / xdiff / xutils.c
bloba45e89bbed5ba582f31aff62b13fd6ddb27b1a85
1 /*
2 * LibXDiff by Davide Libenzi ( File Differential Library )
3 * Copyright (C) 2003 Davide Libenzi
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.
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.
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
19 * Davide Libenzi <davidel@xmailserver.org>
23 #include "xinclude.h"
28 long xdl_bogosqrt(long n) {
29 long i;
32 * Classical integer square root approximation using shifts.
34 for (i = 1; n > 0; n >>= 2)
35 i <<= 1;
37 return i;
41 int 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];
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++;
55 if (ecb->outf(ecb->priv, mb, i) < 0) {
57 return -1;
60 return 0;
63 void *xdl_mmfile_first(mmfile_t *mmf, long *size)
65 *size = mmf->size;
66 return mmf->ptr;
70 void *xdl_mmfile_next(mmfile_t *mmf, long *size)
72 return NULL;
76 long xdl_mmfile_size(mmfile_t *mmf)
78 return mmf->size;
82 int xdl_cha_init(chastore_t *cha, long isize, long icount) {
84 cha->head = cha->tail = NULL;
85 cha->isize = isize;
86 cha->nsize = icount * isize;
87 cha->ancur = cha->sncur = NULL;
88 cha->scurr = 0;
90 return 0;
94 void xdl_cha_free(chastore_t *cha) {
95 chanode_t *cur, *tmp;
97 for (cur = cha->head; (tmp = cur) != NULL;) {
98 cur = cur->next;
99 xdl_free(tmp);
104 void *xdl_cha_alloc(chastore_t *cha) {
105 chanode_t *ancur;
106 void *data;
108 if (!(ancur = cha->ancur) || ancur->icurr == cha->nsize) {
109 if (!(ancur = (chanode_t *) xdl_malloc(sizeof(chanode_t) + cha->nsize))) {
111 return NULL;
113 ancur->icurr = 0;
114 ancur->next = NULL;
115 if (cha->tail)
116 cha->tail->next = ancur;
117 if (!cha->head)
118 cha->head = ancur;
119 cha->tail = ancur;
120 cha->ancur = ancur;
123 data = (char *) ancur + sizeof(chanode_t) + ancur->icurr;
124 ancur->icurr += cha->isize;
126 return data;
130 void *xdl_cha_first(chastore_t *cha) {
131 chanode_t *sncur;
133 if (!(cha->sncur = sncur = cha->head))
134 return NULL;
136 cha->scurr = 0;
138 return (char *) sncur + sizeof(chanode_t) + cha->scurr;
142 void *xdl_cha_next(chastore_t *cha) {
143 chanode_t *sncur;
145 if (!(sncur = cha->sncur))
146 return NULL;
147 cha->scurr += cha->isize;
148 if (cha->scurr == sncur->icurr) {
149 if (!(sncur = cha->sncur = sncur->next))
150 return NULL;
151 cha->scurr = 0;
154 return (char *) sncur + sizeof(chanode_t) + cha->scurr;
158 long xdl_guess_lines(mmfile_t *mf, long sample) {
159 long nl = 0, size, tsize = 0;
160 char const *data, *cur, *top;
162 if ((cur = data = xdl_mmfile_first(mf, &size)) != NULL) {
163 for (top = data + size; nl < sample;) {
164 if (cur >= top) {
165 tsize += (long) (cur - data);
166 if (!(cur = data = xdl_mmfile_next(mf, &size)))
167 break;
168 top = data + size;
170 nl++;
171 if (!(cur = memchr(cur, '\n', top - cur)))
172 cur = top;
173 else
174 cur++;
176 tsize += (long) (cur - data);
179 if (nl && tsize)
180 nl = xdl_mmfile_size(mf) / (tsize / nl);
182 return nl + 1;
185 int xdl_recmatch(const char *l1, long s1, const char *l2, long s2, long flags)
187 int i1, i2;
189 if (s1 == s2 && !memcmp(l1, l2, s1))
190 return 1;
191 if (!(flags & XDF_WHITESPACE_FLAGS))
192 return 0;
194 i1 = 0;
195 i2 = 0;
198 * -w matches everything that matches with -b, and -b in turn
199 * matches everything that matches with --ignore-space-at-eol.
201 * Each flavor of ignoring needs different logic to skip whitespaces
202 * while we have both sides to compare.
204 if (flags & XDF_IGNORE_WHITESPACE) {
205 goto skip_ws;
206 while (i1 < s1 && i2 < s2) {
207 if (l1[i1++] != l2[i2++])
208 return 0;
209 skip_ws:
210 while (i1 < s1 && XDL_ISSPACE(l1[i1]))
211 i1++;
212 while (i2 < s2 && XDL_ISSPACE(l2[i2]))
213 i2++;
215 } else if (flags & XDF_IGNORE_WHITESPACE_CHANGE) {
216 while (i1 < s1 && i2 < s2) {
217 if (XDL_ISSPACE(l1[i1]) && XDL_ISSPACE(l2[i2])) {
218 /* Skip matching spaces and try again */
219 while (i1 < s1 && XDL_ISSPACE(l1[i1]))
220 i1++;
221 while (i2 < s2 && XDL_ISSPACE(l2[i2]))
222 i2++;
223 continue;
225 if (l1[i1++] != l2[i2++])
226 return 0;
228 } else if (flags & XDF_IGNORE_WHITESPACE_AT_EOL) {
229 while (i1 < s1 && i2 < s2 && l1[i1++] == l2[i2++])
230 ; /* keep going */
234 * After running out of one side, the remaining side must have
235 * nothing but whitespace for the lines to match. Note that
236 * ignore-whitespace-at-eol case may break out of the loop
237 * while there still are characters remaining on both lines.
239 if (i1 < s1) {
240 while (i1 < s1 && XDL_ISSPACE(l1[i1]))
241 i1++;
242 if (s1 != i1)
243 return 0;
245 if (i2 < s2) {
246 while (i2 < s2 && XDL_ISSPACE(l2[i2]))
247 i2++;
248 return (s2 == i2);
250 return 1;
253 static unsigned long xdl_hash_record_with_whitespace(char const **data,
254 char const *top, long flags) {
255 unsigned long ha = 5381;
256 char const *ptr = *data;
258 for (; ptr < top && *ptr != '\n'; ptr++) {
259 if (XDL_ISSPACE(*ptr)) {
260 const char *ptr2 = ptr;
261 int at_eol;
262 while (ptr + 1 < top && XDL_ISSPACE(ptr[1])
263 && ptr[1] != '\n')
264 ptr++;
265 at_eol = (top <= ptr + 1 || ptr[1] == '\n');
266 if (flags & XDF_IGNORE_WHITESPACE)
267 ; /* already handled */
268 else if (flags & XDF_IGNORE_WHITESPACE_CHANGE
269 && !at_eol) {
270 ha += (ha << 5);
271 ha ^= (unsigned long) ' ';
273 else if (flags & XDF_IGNORE_WHITESPACE_AT_EOL
274 && !at_eol) {
275 while (ptr2 != ptr + 1) {
276 ha += (ha << 5);
277 ha ^= (unsigned long) *ptr2;
278 ptr2++;
281 continue;
283 ha += (ha << 5);
284 ha ^= (unsigned long) *ptr;
286 *data = ptr < top ? ptr + 1: ptr;
288 return ha;
292 unsigned long xdl_hash_record(char const **data, char const *top, long flags) {
293 unsigned long ha = 5381;
294 char const *ptr = *data;
296 if (flags & XDF_WHITESPACE_FLAGS)
297 return xdl_hash_record_with_whitespace(data, top, flags);
299 for (; ptr < top && *ptr != '\n'; ptr++) {
300 ha += (ha << 5);
301 ha ^= (unsigned long) *ptr;
303 *data = ptr < top ? ptr + 1: ptr;
305 return ha;
309 unsigned int xdl_hashbits(unsigned int size) {
310 unsigned int val = 1, bits = 0;
312 for (; val < size && bits < CHAR_BIT * sizeof(unsigned int); val <<= 1, bits++);
313 return bits ? bits: 1;
317 int xdl_num_out(char *out, long val) {
318 char *ptr, *str = out;
319 char buf[32];
321 ptr = buf + sizeof(buf) - 1;
322 *ptr = '\0';
323 if (val < 0) {
324 *--ptr = '-';
325 val = -val;
327 for (; val && ptr > buf; val /= 10)
328 *--ptr = "0123456789"[val % 10];
329 if (*ptr)
330 for (; *ptr; ptr++, str++)
331 *str = *ptr;
332 else
333 *str++ = '0';
334 *str = '\0';
336 return str - out;
340 long xdl_atol(char const *str, char const **next) {
341 long val, base;
342 char const *top;
344 for (top = str; XDL_ISDIGIT(*top); top++);
345 if (next)
346 *next = top;
347 for (val = 0, base = 1, top--; top >= str; top--, base *= 10)
348 val += base * (long)(*top - '0');
349 return val;
353 int xdl_emit_hunk_hdr(long s1, long c1, long s2, long c2,
354 const char *func, long funclen, xdemitcb_t *ecb) {
355 int nb = 0;
356 mmbuffer_t mb;
357 char buf[128];
359 memcpy(buf, "@@ -", 4);
360 nb += 4;
362 nb += xdl_num_out(buf + nb, c1 ? s1: s1 - 1);
364 if (c1 != 1) {
365 memcpy(buf + nb, ",", 1);
366 nb += 1;
368 nb += xdl_num_out(buf + nb, c1);
371 memcpy(buf + nb, " +", 2);
372 nb += 2;
374 nb += xdl_num_out(buf + nb, c2 ? s2: s2 - 1);
376 if (c2 != 1) {
377 memcpy(buf + nb, ",", 1);
378 nb += 1;
380 nb += xdl_num_out(buf + nb, c2);
383 memcpy(buf + nb, " @@", 3);
384 nb += 3;
385 if (func && funclen) {
386 buf[nb++] = ' ';
387 if (funclen > sizeof(buf) - nb - 1)
388 funclen = sizeof(buf) - nb - 1;
389 memcpy(buf + nb, func, funclen);
390 nb += funclen;
392 buf[nb++] = '\n';
394 mb.ptr = buf;
395 mb.size = nb;
396 if (ecb->outf(ecb->priv, &mb, 1) < 0)
397 return -1;
399 return 0;
402 int xdl_fall_back_diff(xdfenv_t *diff_env, xpparam_t const *xpp,
403 int line1, int count1, int line2, int count2)
406 * This probably does not work outside Git, since
407 * we have a very simple mmfile structure.
409 * Note: ideally, we would reuse the prepared environment, but
410 * the libxdiff interface does not (yet) allow for diffing only
411 * ranges of lines instead of the whole files.
413 mmfile_t subfile1, subfile2;
414 xdfenv_t env;
416 subfile1.ptr = (char *)diff_env->xdf1.recs[line1 - 1]->ptr;
417 subfile1.size = diff_env->xdf1.recs[line1 + count1 - 2]->ptr +
418 diff_env->xdf1.recs[line1 + count1 - 2]->size - subfile1.ptr;
419 subfile2.ptr = (char *)diff_env->xdf2.recs[line2 - 1]->ptr;
420 subfile2.size = diff_env->xdf2.recs[line2 + count2 - 2]->ptr +
421 diff_env->xdf2.recs[line2 + count2 - 2]->size - subfile2.ptr;
422 if (xdl_do_diff(&subfile1, &subfile2, xpp, &env) < 0)
423 return -1;
425 memcpy(diff_env->xdf1.rchg + line1 - 1, env.xdf1.rchg, count1);
426 memcpy(diff_env->xdf2.rchg + line2 - 1, env.xdf2.rchg, count2);
428 xdl_free_env(&env);
430 return 0;