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, see
17  *  <http://www.gnu.org/licenses/>.
18  *
19  *  Davide Libenzi <davidel@xmailserver.org>
20  *
21  */
22 
23 #include "xinclude.h"
24 
25 
26 
27 
xdl_bogosqrt(long n)28 long 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 
xdl_emit_diffrec(char const * rec,long size,char const * pre,long psize,xdemitcb_t * ecb)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];
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 
xdl_mmfile_first(mmfile_t * mmf,long * size)63 void *xdl_mmfile_first(mmfile_t *mmf, long *size)
64 {
65 	*size = mmf->size;
66 	return mmf->ptr;
67 }
68 
69 
xdl_mmfile_size(mmfile_t * mmf)70 long xdl_mmfile_size(mmfile_t *mmf)
71 {
72 	return mmf->size;
73 }
74 
75 
xdl_cha_init(chastore_t * cha,long isize,long icount)76 int 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 
xdl_cha_free(chastore_t * cha)88 void 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 
xdl_cha_alloc(chastore_t * cha)98 void *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 
xdl_guess_lines(mmfile_t * mf,long sample)123 long xdl_guess_lines(mmfile_t *mf, long sample) {
124 	long nl = 0, size, tsize = 0;
125 	char const *data, *cur, *top;
126 
127 	if ((cur = data = xdl_mmfile_first(mf, &size)) != NULL) {
128 		for (top = data + size; nl < sample && cur < top; ) {
129 			nl++;
130 			if (!(cur = memchr(cur, '\n', top - cur)))
131 				cur = top;
132 			else
133 				cur++;
134 		}
135 		tsize += (long) (cur - data);
136 	}
137 
138 	if (nl && tsize)
139 		nl = xdl_mmfile_size(mf) / (tsize / nl);
140 
141 	return nl + 1;
142 }
143 
xdl_blankline(const char * line,long size,long flags)144 int xdl_blankline(const char *line, long size, long flags)
145 {
146 	long i;
147 
148 	if (!(flags & XDF_WHITESPACE_FLAGS))
149 		return (size <= 1);
150 
151 	for (i = 0; i < size && XDL_ISSPACE(line[i]); i++)
152 		;
153 
154 	return (i == size);
155 }
156 
157 /*
158  * Have we eaten everything on the line, except for an optional
159  * CR at the very end?
160  */
ends_with_optional_cr(const char * l,long s,long i)161 static int ends_with_optional_cr(const char *l, long s, long i)
162 {
163 	int complete = s && l[s-1] == '\n';
164 
165 	if (complete)
166 		s--;
167 	if (s == i)
168 		return 1;
169 	/* do not ignore CR at the end of an incomplete line */
170 	if (complete && s == i + 1 && l[i] == '\r')
171 		return 1;
172 	return 0;
173 }
174 
xdl_recmatch(const char * l1,long s1,const char * l2,long s2,long flags)175 int xdl_recmatch(const char *l1, long s1, const char *l2, long s2, long flags)
176 {
177 	int i1, i2;
178 
179 	if (s1 == s2 && !memcmp(l1, l2, s1))
180 		return 1;
181 	if (!(flags & XDF_WHITESPACE_FLAGS))
182 		return 0;
183 
184 	i1 = 0;
185 	i2 = 0;
186 
187 	/*
188 	 * -w matches everything that matches with -b, and -b in turn
189 	 * matches everything that matches with --ignore-space-at-eol,
190 	 * which in turn matches everything that matches with --ignore-cr-at-eol.
191 	 *
192 	 * Each flavor of ignoring needs different logic to skip whitespaces
193 	 * while we have both sides to compare.
194 	 */
195 	if (flags & XDF_IGNORE_WHITESPACE) {
196 		goto skip_ws;
197 		while (i1 < s1 && i2 < s2) {
198 			if (l1[i1++] != l2[i2++])
199 				return 0;
200 		skip_ws:
201 			while (i1 < s1 && XDL_ISSPACE(l1[i1]))
202 				i1++;
203 			while (i2 < s2 && XDL_ISSPACE(l2[i2]))
204 				i2++;
205 		}
206 	} else if (flags & XDF_IGNORE_WHITESPACE_CHANGE) {
207 		while (i1 < s1 && i2 < s2) {
208 			if (XDL_ISSPACE(l1[i1]) && XDL_ISSPACE(l2[i2])) {
209 				/* Skip matching spaces and try again */
210 				while (i1 < s1 && XDL_ISSPACE(l1[i1]))
211 					i1++;
212 				while (i2 < s2 && XDL_ISSPACE(l2[i2]))
213 					i2++;
214 				continue;
215 			}
216 			if (l1[i1++] != l2[i2++])
217 				return 0;
218 		}
219 	} else if (flags & XDF_IGNORE_WHITESPACE_AT_EOL) {
220 		while (i1 < s1 && i2 < s2 && l1[i1] == l2[i2]) {
221 			i1++;
222 			i2++;
223 		}
224 	} else if (flags & XDF_IGNORE_CR_AT_EOL) {
225 		/* Find the first difference and see how the line ends */
226 		while (i1 < s1 && i2 < s2 && l1[i1] == l2[i2]) {
227 			i1++;
228 			i2++;
229 		}
230 		return (ends_with_optional_cr(l1, s1, i1) &&
231 			ends_with_optional_cr(l2, s2, i2));
232 	}
233 
234 	/*
235 	 * After running out of one side, the remaining side must have
236 	 * nothing but whitespace for the lines to match.  Note that
237 	 * ignore-whitespace-at-eol case may break out of the loop
238 	 * while there still are characters remaining on both lines.
239 	 */
240 	if (i1 < s1) {
241 		while (i1 < s1 && XDL_ISSPACE(l1[i1]))
242 			i1++;
243 		if (s1 != i1)
244 			return 0;
245 	}
246 	if (i2 < s2) {
247 		while (i2 < s2 && XDL_ISSPACE(l2[i2]))
248 			i2++;
249 		return (s2 == i2);
250 	}
251 	return 1;
252 }
253 
xdl_hash_record_with_whitespace(char const ** data,char const * top,long flags)254 static unsigned long xdl_hash_record_with_whitespace(char const **data,
255 		char const *top, long flags) {
256 	unsigned long ha = 5381;
257 	char const *ptr = *data;
258 	int cr_at_eol_only = (flags & XDF_WHITESPACE_FLAGS) == XDF_IGNORE_CR_AT_EOL;
259 
260 	for (; ptr < top && *ptr != '\n'; ptr++) {
261 		if (cr_at_eol_only) {
262 			/* do not ignore CR at the end of an incomplete line */
263 			if (*ptr == '\r' &&
264 			    (ptr + 1 < top && ptr[1] == '\n'))
265 				continue;
266 		}
267 		else if (XDL_ISSPACE(*ptr)) {
268 			const char *ptr2 = ptr;
269 			int at_eol;
270 			while (ptr + 1 < top && XDL_ISSPACE(ptr[1])
271 					&& ptr[1] != '\n')
272 				ptr++;
273 			at_eol = (top <= ptr + 1 || ptr[1] == '\n');
274 			if (flags & XDF_IGNORE_WHITESPACE)
275 				; /* already handled */
276 			else if (flags & XDF_IGNORE_WHITESPACE_CHANGE
277 				 && !at_eol) {
278 				ha += (ha << 5);
279 				ha ^= (unsigned long) ' ';
280 			}
281 			else if (flags & XDF_IGNORE_WHITESPACE_AT_EOL
282 				 && !at_eol) {
283 				while (ptr2 != ptr + 1) {
284 					ha += (ha << 5);
285 					ha ^= (unsigned long) *ptr2;
286 					ptr2++;
287 				}
288 			}
289 			continue;
290 		}
291 		ha += (ha << 5);
292 		ha ^= (unsigned long) *ptr;
293 	}
294 	*data = ptr < top ? ptr + 1: ptr;
295 
296 	return ha;
297 }
298 
xdl_hash_record(char const ** data,char const * top,long flags)299 unsigned long xdl_hash_record(char const **data, char const *top, long flags) {
300 	unsigned long ha = 5381;
301 	char const *ptr = *data;
302 
303 	if (flags & XDF_WHITESPACE_FLAGS)
304 		return xdl_hash_record_with_whitespace(data, top, flags);
305 
306 	for (; ptr < top && *ptr != '\n'; ptr++) {
307 		ha += (ha << 5);
308 		ha ^= (unsigned long) *ptr;
309 	}
310 	*data = ptr < top ? ptr + 1: ptr;
311 
312 	return ha;
313 }
314 
xdl_hashbits(unsigned int size)315 unsigned int xdl_hashbits(unsigned int size) {
316 	unsigned int val = 1, bits = 0;
317 
318 	for (; val < size && bits < CHAR_BIT * sizeof(unsigned int); val <<= 1, bits++);
319 	return bits ? bits: 1;
320 }
321 
322 
xdl_num_out(char * out,long val)323 int xdl_num_out(char *out, long val) {
324 	char *ptr, *str = out;
325 	char buf[32];
326 
327 	ptr = buf + sizeof(buf) - 1;
328 	*ptr = '\0';
329 	if (val < 0) {
330 		*--ptr = '-';
331 		val = -val;
332 	}
333 	for (; val && ptr > buf; val /= 10)
334 		*--ptr = "0123456789"[val % 10];
335 	if (*ptr)
336 		for (; *ptr; ptr++, str++)
337 			*str = *ptr;
338 	else
339 		*str++ = '0';
340 	*str = '\0';
341 
342 	return str - out;
343 }
344 
xdl_emit_hunk_hdr(long s1,long c1,long s2,long c2,const char * func,long funclen,xdemitcb_t * ecb)345 int xdl_emit_hunk_hdr(long s1, long c1, long s2, long c2,
346 		      const char *func, long funclen, xdemitcb_t *ecb) {
347 	int nb = 0;
348 	mmbuffer_t mb;
349 	char buf[128];
350 
351 	memcpy(buf, "@@ -", 4);
352 	nb += 4;
353 
354 	nb += xdl_num_out(buf + nb, c1 ? s1: s1 - 1);
355 
356 	if (c1 != 1) {
357 		memcpy(buf + nb, ",", 1);
358 		nb += 1;
359 
360 		nb += xdl_num_out(buf + nb, c1);
361 	}
362 
363 	memcpy(buf + nb, " +", 2);
364 	nb += 2;
365 
366 	nb += xdl_num_out(buf + nb, c2 ? s2: s2 - 1);
367 
368 	if (c2 != 1) {
369 		memcpy(buf + nb, ",", 1);
370 		nb += 1;
371 
372 		nb += xdl_num_out(buf + nb, c2);
373 	}
374 
375 	memcpy(buf + nb, " @@", 3);
376 	nb += 3;
377 	if (func && funclen) {
378 		buf[nb++] = ' ';
379 		if (funclen > (long)(sizeof(buf) - nb - 1))
380 			funclen = sizeof(buf) - nb - 1;
381 		memcpy(buf + nb, func, funclen);
382 		nb += funclen;
383 	}
384 	buf[nb++] = '\n';
385 
386 	mb.ptr = buf;
387 	mb.size = nb;
388 	if (ecb->outf(ecb->priv, &mb, 1) < 0)
389 		return -1;
390 
391 	return 0;
392 }
393 
xdl_fall_back_diff(xdfenv_t * diff_env,xpparam_t const * xpp,int line1,int count1,int line2,int count2)394 int xdl_fall_back_diff(xdfenv_t *diff_env, xpparam_t const *xpp,
395 		int line1, int count1, int line2, int count2)
396 {
397 	/*
398 	 * This probably does not work outside Git, since
399 	 * we have a very simple mmfile structure.
400 	 *
401 	 * Note: ideally, we would reuse the prepared environment, but
402 	 * the libxdiff interface does not (yet) allow for diffing only
403 	 * ranges of lines instead of the whole files.
404 	 */
405 	mmfile_t subfile1, subfile2;
406 	xdfenv_t env;
407 
408 	subfile1.ptr = (char *)diff_env->xdf1.recs[line1 - 1]->ptr;
409 	subfile1.size = diff_env->xdf1.recs[line1 + count1 - 2]->ptr +
410 		diff_env->xdf1.recs[line1 + count1 - 2]->size - subfile1.ptr;
411 	subfile2.ptr = (char *)diff_env->xdf2.recs[line2 - 1]->ptr;
412 	subfile2.size = diff_env->xdf2.recs[line2 + count2 - 2]->ptr +
413 		diff_env->xdf2.recs[line2 + count2 - 2]->size - subfile2.ptr;
414 	if (xdl_do_diff(&subfile1, &subfile2, xpp, &env) < 0)
415 		return -1;
416 
417 	memcpy(diff_env->xdf1.rchg + line1 - 1, env.xdf1.rchg, count1);
418 	memcpy(diff_env->xdf2.rchg + line2 - 1, env.xdf2.rchg, count2);
419 
420 	xdl_free_env(&env);
421 
422 	return 0;
423 }
424