]>
Commit | Line | Data |
---|---|---|
3443546f LT |
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 | #define XDL_MAX_COST_MIN 256 | |
28 | #define XDL_HEUR_MIN_COST 256 | |
f630cfda | 29 | #define XDL_LINE_MAX (long)((1UL << (CHAR_BIT * sizeof(long) - 1)) - 1) |
3443546f LT |
30 | #define XDL_SNAKE_CNT 20 |
31 | #define XDL_K_HEUR 4 | |
32 | ||
33 | ||
34 | ||
35 | typedef struct s_xdpsplit { | |
36 | long i1, i2; | |
37 | int min_lo, min_hi; | |
38 | } xdpsplit_t; | |
39 | ||
40 | ||
41 | ||
42 | ||
43 | static long xdl_split(unsigned long const *ha1, long off1, long lim1, | |
44 | unsigned long const *ha2, long off2, long lim2, | |
45 | long *kvdf, long *kvdb, int need_min, xdpsplit_t *spl, | |
46 | xdalgoenv_t *xenv); | |
47 | static xdchange_t *xdl_add_change(xdchange_t *xscr, long i1, long i2, long chg1, long chg2); | |
295ba2fb | 48 | |
3443546f LT |
49 | |
50 | ||
51 | ||
52 | ||
53 | /* | |
54 | * See "An O(ND) Difference Algorithm and its Variations", by Eugene Myers. | |
55 | * Basically considers a "box" (off1, off2, lim1, lim2) and scan from both | |
56 | * the forward diagonal starting from (off1, off2) and the backward diagonal | |
57 | * starting from (lim1, lim2). If the K values on the same diagonal crosses | |
58 | * returns the furthest point of reach. We might end up having to expensive | |
59 | * cases using this algorithm is full, so a little bit of heuristic is needed | |
60 | * to cut the search and to return a suboptimal point. | |
61 | */ | |
62 | static long xdl_split(unsigned long const *ha1, long off1, long lim1, | |
63 | unsigned long const *ha2, long off2, long lim2, | |
64 | long *kvdf, long *kvdb, int need_min, xdpsplit_t *spl, | |
65 | xdalgoenv_t *xenv) { | |
66 | long dmin = off1 - lim2, dmax = lim1 - off2; | |
67 | long fmid = off1 - off2, bmid = lim1 - lim2; | |
68 | long odd = (fmid - bmid) & 1; | |
69 | long fmin = fmid, fmax = fmid; | |
70 | long bmin = bmid, bmax = bmid; | |
71 | long ec, d, i1, i2, prev1, best, dd, v, k; | |
72 | ||
73 | /* | |
74 | * Set initial diagonal values for both forward and backward path. | |
75 | */ | |
76 | kvdf[fmid] = off1; | |
77 | kvdb[bmid] = lim1; | |
78 | ||
79 | for (ec = 1;; ec++) { | |
80 | int got_snake = 0; | |
81 | ||
82 | /* | |
83 | * We need to extent the diagonal "domain" by one. If the next | |
84 | * values exits the box boundaries we need to change it in the | |
85 | * opposite direction because (max - min) must be a power of two. | |
82e5a82f | 86 | * Also we initialize the external K value to -1 so that we can |
3443546f LT |
87 | * avoid extra conditions check inside the core loop. |
88 | */ | |
89 | if (fmin > dmin) | |
90 | kvdf[--fmin - 1] = -1; | |
91 | else | |
92 | ++fmin; | |
93 | if (fmax < dmax) | |
94 | kvdf[++fmax + 1] = -1; | |
95 | else | |
96 | --fmax; | |
97 | ||
98 | for (d = fmax; d >= fmin; d -= 2) { | |
99 | if (kvdf[d - 1] >= kvdf[d + 1]) | |
100 | i1 = kvdf[d - 1] + 1; | |
101 | else | |
102 | i1 = kvdf[d + 1]; | |
103 | prev1 = i1; | |
104 | i2 = i1 - d; | |
105 | for (; i1 < lim1 && i2 < lim2 && ha1[i1] == ha2[i2]; i1++, i2++); | |
106 | if (i1 - prev1 > xenv->snake_cnt) | |
107 | got_snake = 1; | |
108 | kvdf[d] = i1; | |
109 | if (odd && bmin <= d && d <= bmax && kvdb[d] <= i1) { | |
110 | spl->i1 = i1; | |
111 | spl->i2 = i2; | |
112 | spl->min_lo = spl->min_hi = 1; | |
113 | return ec; | |
114 | } | |
115 | } | |
116 | ||
117 | /* | |
118 | * We need to extent the diagonal "domain" by one. If the next | |
119 | * values exits the box boundaries we need to change it in the | |
120 | * opposite direction because (max - min) must be a power of two. | |
82e5a82f | 121 | * Also we initialize the external K value to -1 so that we can |
3443546f LT |
122 | * avoid extra conditions check inside the core loop. |
123 | */ | |
124 | if (bmin > dmin) | |
125 | kvdb[--bmin - 1] = XDL_LINE_MAX; | |
126 | else | |
127 | ++bmin; | |
128 | if (bmax < dmax) | |
129 | kvdb[++bmax + 1] = XDL_LINE_MAX; | |
130 | else | |
131 | --bmax; | |
132 | ||
133 | for (d = bmax; d >= bmin; d -= 2) { | |
134 | if (kvdb[d - 1] < kvdb[d + 1]) | |
135 | i1 = kvdb[d - 1]; | |
136 | else | |
137 | i1 = kvdb[d + 1] - 1; | |
138 | prev1 = i1; | |
139 | i2 = i1 - d; | |
140 | for (; i1 > off1 && i2 > off2 && ha1[i1 - 1] == ha2[i2 - 1]; i1--, i2--); | |
141 | if (prev1 - i1 > xenv->snake_cnt) | |
142 | got_snake = 1; | |
143 | kvdb[d] = i1; | |
144 | if (!odd && fmin <= d && d <= fmax && i1 <= kvdf[d]) { | |
145 | spl->i1 = i1; | |
146 | spl->i2 = i2; | |
147 | spl->min_lo = spl->min_hi = 1; | |
148 | return ec; | |
149 | } | |
150 | } | |
151 | ||
152 | if (need_min) | |
153 | continue; | |
154 | ||
155 | /* | |
156 | * If the edit cost is above the heuristic trigger and if | |
157 | * we got a good snake, we sample current diagonals to see | |
158 | * if some of the, have reached an "interesting" path. Our | |
159 | * measure is a function of the distance from the diagonal | |
160 | * corner (i1 + i2) penalized with the distance from the | |
161 | * mid diagonal itself. If this value is above the current | |
162 | * edit cost times a magic factor (XDL_K_HEUR) we consider | |
163 | * it interesting. | |
164 | */ | |
165 | if (got_snake && ec > xenv->heur_min) { | |
166 | for (best = 0, d = fmax; d >= fmin; d -= 2) { | |
167 | dd = d > fmid ? d - fmid: fmid - d; | |
168 | i1 = kvdf[d]; | |
169 | i2 = i1 - d; | |
170 | v = (i1 - off1) + (i2 - off2) - dd; | |
171 | ||
172 | if (v > XDL_K_HEUR * ec && v > best && | |
173 | off1 + xenv->snake_cnt <= i1 && i1 < lim1 && | |
174 | off2 + xenv->snake_cnt <= i2 && i2 < lim2) { | |
175 | for (k = 1; ha1[i1 - k] == ha2[i2 - k]; k++) | |
176 | if (k == xenv->snake_cnt) { | |
177 | best = v; | |
178 | spl->i1 = i1; | |
179 | spl->i2 = i2; | |
180 | break; | |
181 | } | |
182 | } | |
183 | } | |
184 | if (best > 0) { | |
185 | spl->min_lo = 1; | |
186 | spl->min_hi = 0; | |
187 | return ec; | |
188 | } | |
189 | ||
190 | for (best = 0, d = bmax; d >= bmin; d -= 2) { | |
191 | dd = d > bmid ? d - bmid: bmid - d; | |
192 | i1 = kvdb[d]; | |
193 | i2 = i1 - d; | |
194 | v = (lim1 - i1) + (lim2 - i2) - dd; | |
195 | ||
196 | if (v > XDL_K_HEUR * ec && v > best && | |
197 | off1 < i1 && i1 <= lim1 - xenv->snake_cnt && | |
198 | off2 < i2 && i2 <= lim2 - xenv->snake_cnt) { | |
199 | for (k = 0; ha1[i1 + k] == ha2[i2 + k]; k++) | |
200 | if (k == xenv->snake_cnt - 1) { | |
201 | best = v; | |
202 | spl->i1 = i1; | |
203 | spl->i2 = i2; | |
204 | break; | |
205 | } | |
206 | } | |
207 | } | |
208 | if (best > 0) { | |
209 | spl->min_lo = 0; | |
210 | spl->min_hi = 1; | |
211 | return ec; | |
212 | } | |
213 | } | |
214 | ||
215 | /* | |
216 | * Enough is enough. We spent too much time here and now we collect | |
217 | * the furthest reaching path using the (i1 + i2) measure. | |
218 | */ | |
219 | if (ec >= xenv->mxcost) { | |
220 | long fbest, fbest1, bbest, bbest1; | |
221 | ||
0ed49a3e | 222 | fbest = fbest1 = -1; |
3443546f LT |
223 | for (d = fmax; d >= fmin; d -= 2) { |
224 | i1 = XDL_MIN(kvdf[d], lim1); | |
225 | i2 = i1 - d; | |
226 | if (lim2 < i2) | |
227 | i1 = lim2 + d, i2 = lim2; | |
228 | if (fbest < i1 + i2) { | |
229 | fbest = i1 + i2; | |
230 | fbest1 = i1; | |
231 | } | |
232 | } | |
233 | ||
0ed49a3e | 234 | bbest = bbest1 = XDL_LINE_MAX; |
3443546f LT |
235 | for (d = bmax; d >= bmin; d -= 2) { |
236 | i1 = XDL_MAX(off1, kvdb[d]); | |
237 | i2 = i1 - d; | |
238 | if (i2 < off2) | |
239 | i1 = off2 + d, i2 = off2; | |
240 | if (i1 + i2 < bbest) { | |
241 | bbest = i1 + i2; | |
242 | bbest1 = i1; | |
243 | } | |
244 | } | |
245 | ||
246 | if ((lim1 + lim2) - bbest < fbest - (off1 + off2)) { | |
247 | spl->i1 = fbest1; | |
248 | spl->i2 = fbest - fbest1; | |
249 | spl->min_lo = 1; | |
250 | spl->min_hi = 0; | |
251 | } else { | |
252 | spl->i1 = bbest1; | |
253 | spl->i2 = bbest - bbest1; | |
254 | spl->min_lo = 0; | |
255 | spl->min_hi = 1; | |
256 | } | |
257 | return ec; | |
258 | } | |
259 | } | |
3443546f LT |
260 | } |
261 | ||
262 | ||
263 | /* | |
264 | * Rule: "Divide et Impera". Recursively split the box in sub-boxes by calling | |
265 | * the box splitting function. Note that the real job (marking changed lines) | |
266 | * is done in the two boundary reaching checks. | |
267 | */ | |
268 | int xdl_recs_cmp(diffdata_t *dd1, long off1, long lim1, | |
269 | diffdata_t *dd2, long off2, long lim2, | |
270 | long *kvdf, long *kvdb, int need_min, xdalgoenv_t *xenv) { | |
271 | unsigned long const *ha1 = dd1->ha, *ha2 = dd2->ha; | |
272 | ||
273 | /* | |
274 | * Shrink the box by walking through each diagonal snake (SW and NE). | |
275 | */ | |
276 | for (; off1 < lim1 && off2 < lim2 && ha1[off1] == ha2[off2]; off1++, off2++); | |
277 | for (; off1 < lim1 && off2 < lim2 && ha1[lim1 - 1] == ha2[lim2 - 1]; lim1--, lim2--); | |
278 | ||
279 | /* | |
280 | * If one dimension is empty, then all records on the other one must | |
281 | * be obviously changed. | |
282 | */ | |
283 | if (off1 == lim1) { | |
284 | char *rchg2 = dd2->rchg; | |
285 | long *rindex2 = dd2->rindex; | |
286 | ||
287 | for (; off2 < lim2; off2++) | |
288 | rchg2[rindex2[off2]] = 1; | |
289 | } else if (off2 == lim2) { | |
290 | char *rchg1 = dd1->rchg; | |
291 | long *rindex1 = dd1->rindex; | |
292 | ||
293 | for (; off1 < lim1; off1++) | |
294 | rchg1[rindex1[off1]] = 1; | |
295 | } else { | |
3443546f | 296 | xdpsplit_t spl; |
0ed49a3e | 297 | spl.i1 = spl.i2 = 0; |
3443546f LT |
298 | |
299 | /* | |
300 | * Divide ... | |
301 | */ | |
8e24cbae BK |
302 | if (xdl_split(ha1, off1, lim1, ha2, off2, lim2, kvdf, kvdb, |
303 | need_min, &spl, xenv) < 0) { | |
3443546f LT |
304 | |
305 | return -1; | |
306 | } | |
307 | ||
308 | /* | |
309 | * ... et Impera. | |
310 | */ | |
311 | if (xdl_recs_cmp(dd1, off1, spl.i1, dd2, off2, spl.i2, | |
312 | kvdf, kvdb, spl.min_lo, xenv) < 0 || | |
313 | xdl_recs_cmp(dd1, spl.i1, lim1, dd2, spl.i2, lim2, | |
314 | kvdf, kvdb, spl.min_hi, xenv) < 0) { | |
315 | ||
316 | return -1; | |
317 | } | |
318 | } | |
319 | ||
320 | return 0; | |
321 | } | |
322 | ||
323 | ||
324 | int xdl_do_diff(mmfile_t *mf1, mmfile_t *mf2, xpparam_t const *xpp, | |
325 | xdfenv_t *xe) { | |
326 | long ndiags; | |
327 | long *kvd, *kvdf, *kvdb; | |
328 | xdalgoenv_t xenv; | |
329 | diffdata_t dd1, dd2; | |
330 | ||
307ab20b | 331 | if (XDF_DIFF_ALG(xpp->flags) == XDF_PATIENCE_DIFF) |
92b7de93 JS |
332 | return xdl_do_patience_diff(mf1, mf2, xpp, xe); |
333 | ||
307ab20b | 334 | if (XDF_DIFF_ALG(xpp->flags) == XDF_HISTOGRAM_DIFF) |
8c912eea TRC |
335 | return xdl_do_histogram_diff(mf1, mf2, xpp, xe); |
336 | ||
3443546f LT |
337 | if (xdl_prepare_env(mf1, mf2, xpp, xe) < 0) { |
338 | ||
339 | return -1; | |
340 | } | |
341 | ||
342 | /* | |
343 | * Allocate and setup K vectors to be used by the differential algorithm. | |
344 | * One is to store the forward path and one to store the backward path. | |
345 | */ | |
346 | ndiags = xe->xdf1.nreff + xe->xdf2.nreff + 3; | |
347 | if (!(kvd = (long *) xdl_malloc((2 * ndiags + 2) * sizeof(long)))) { | |
348 | ||
349 | xdl_free_env(xe); | |
350 | return -1; | |
351 | } | |
352 | kvdf = kvd; | |
353 | kvdb = kvdf + ndiags; | |
354 | kvdf += xe->xdf2.nreff + 1; | |
355 | kvdb += xe->xdf2.nreff + 1; | |
356 | ||
ca557aff | 357 | xenv.mxcost = xdl_bogosqrt(ndiags); |
3443546f LT |
358 | if (xenv.mxcost < XDL_MAX_COST_MIN) |
359 | xenv.mxcost = XDL_MAX_COST_MIN; | |
360 | xenv.snake_cnt = XDL_SNAKE_CNT; | |
361 | xenv.heur_min = XDL_HEUR_MIN_COST; | |
362 | ||
363 | dd1.nrec = xe->xdf1.nreff; | |
364 | dd1.ha = xe->xdf1.ha; | |
365 | dd1.rchg = xe->xdf1.rchg; | |
366 | dd1.rindex = xe->xdf1.rindex; | |
367 | dd2.nrec = xe->xdf2.nreff; | |
368 | dd2.ha = xe->xdf2.ha; | |
369 | dd2.rchg = xe->xdf2.rchg; | |
370 | dd2.rindex = xe->xdf2.rindex; | |
371 | ||
372 | if (xdl_recs_cmp(&dd1, 0, dd1.nrec, &dd2, 0, dd2.nrec, | |
373 | kvdf, kvdb, (xpp->flags & XDF_NEED_MINIMAL) != 0, &xenv) < 0) { | |
374 | ||
375 | xdl_free(kvd); | |
376 | xdl_free_env(xe); | |
377 | return -1; | |
378 | } | |
379 | ||
380 | xdl_free(kvd); | |
381 | ||
382 | return 0; | |
383 | } | |
384 | ||
385 | ||
386 | static xdchange_t *xdl_add_change(xdchange_t *xscr, long i1, long i2, long chg1, long chg2) { | |
387 | xdchange_t *xch; | |
388 | ||
389 | if (!(xch = (xdchange_t *) xdl_malloc(sizeof(xdchange_t)))) | |
390 | return NULL; | |
391 | ||
392 | xch->next = xscr; | |
393 | xch->i1 = i1; | |
394 | xch->i2 = i2; | |
395 | xch->chg1 = chg1; | |
396 | xch->chg2 = chg2; | |
36617af7 | 397 | xch->ignore = 0; |
3443546f LT |
398 | |
399 | return xch; | |
400 | } | |
401 | ||
402 | ||
d634d61e SB |
403 | static int is_blank_line(xrecord_t **recs, long ix, long flags) |
404 | { | |
405 | return xdl_blankline(recs[ix]->ptr, recs[ix]->size, flags); | |
406 | } | |
407 | ||
92e5b62f JK |
408 | static int recs_match(xrecord_t **recs, long ixs, long ix, long flags) |
409 | { | |
410 | return (recs[ixs]->ha == recs[ix]->ha && | |
411 | xdl_recmatch(recs[ixs]->ptr, recs[ixs]->size, | |
412 | recs[ix]->ptr, recs[ix]->size, | |
413 | flags)); | |
414 | } | |
415 | ||
857b933e | 416 | int xdl_change_compact(xdfile_t *xdf, xdfile_t *xdfo, long flags) { |
295ba2fb DL |
417 | long ix, ixo, ixs, ixref, grpsiz, nrec = xdf->nrec; |
418 | char *rchg = xdf->rchg, *rchgo = xdfo->rchg; | |
d634d61e | 419 | unsigned int blank_lines; |
295ba2fb DL |
420 | xrecord_t **recs = xdf->recs; |
421 | ||
422 | /* | |
423 | * This is the same of what GNU diff does. Move back and forward | |
424 | * change groups for a consistent and pretty diff output. This also | |
82e5a82f | 425 | * helps in finding joinable change groups and reduce the diff size. |
295ba2fb DL |
426 | */ |
427 | for (ix = ixo = 0;;) { | |
428 | /* | |
429 | * Find the first changed line in the to-be-compacted file. | |
430 | * We need to keep track of both indexes, so if we find a | |
431 | * changed lines group on the other file, while scanning the | |
432 | * to-be-compacted file, we need to skip it properly. Note | |
433 | * that loops that are testing for changed lines on rchg* do | |
434 | * not need index bounding since the array is prepared with | |
435 | * a zero at position -1 and N. | |
436 | */ | |
437 | for (; ix < nrec && !rchg[ix]; ix++) | |
438 | while (rchgo[ixo++]); | |
439 | if (ix == nrec) | |
440 | break; | |
441 | ||
442 | /* | |
443 | * Record the start of a changed-group in the to-be-compacted file | |
444 | * and find the end of it, on both to-be-compacted and other file | |
445 | * indexes (ix and ixo). | |
446 | */ | |
447 | ixs = ix; | |
448 | for (ix++; rchg[ix]; ix++); | |
449 | for (; rchgo[ixo]; ixo++); | |
450 | ||
451 | do { | |
452 | grpsiz = ix - ixs; | |
d634d61e | 453 | blank_lines = 0; |
295ba2fb DL |
454 | |
455 | /* | |
456 | * If the line before the current change group, is equal to | |
457 | * the last line of the current change group, shift backward | |
458 | * the group. | |
459 | */ | |
92e5b62f | 460 | while (ixs > 0 && recs_match(recs, ixs - 1, ix - 1, flags)) { |
295ba2fb DL |
461 | rchg[--ixs] = 1; |
462 | rchg[--ix] = 0; | |
463 | ||
464 | /* | |
465 | * This change might have joined two change groups, | |
466 | * so we try to take this scenario in account by moving | |
467 | * the start index accordingly (and so the other-file | |
468 | * end-of-group index). | |
469 | */ | |
470 | for (; rchg[ixs - 1]; ixs--); | |
471 | while (rchgo[--ixo]); | |
472 | } | |
473 | ||
474 | /* | |
475 | * Record the end-of-group position in case we are matched | |
476 | * with a group of changes in the other file (that is, the | |
3ea3c215 | 477 | * change record before the end-of-group index in the other |
295ba2fb DL |
478 | * file is set). |
479 | */ | |
480 | ixref = rchgo[ixo - 1] ? ix: nrec; | |
481 | ||
482 | /* | |
483 | * If the first line of the current change group, is equal to | |
484 | * the line next of the current change group, shift forward | |
485 | * the group. | |
486 | */ | |
92e5b62f | 487 | while (ix < nrec && recs_match(recs, ixs, ix, flags)) { |
d634d61e SB |
488 | blank_lines += is_blank_line(recs, ix, flags); |
489 | ||
295ba2fb DL |
490 | rchg[ixs++] = 0; |
491 | rchg[ix++] = 1; | |
492 | ||
493 | /* | |
494 | * This change might have joined two change groups, | |
495 | * so we try to take this scenario in account by moving | |
496 | * the start index accordingly (and so the other-file | |
497 | * end-of-group index). Keep tracking the reference | |
498 | * index in case we are shifting together with a | |
499 | * corresponding group of changes in the other file. | |
500 | */ | |
501 | for (; rchg[ix]; ix++); | |
502 | while (rchgo[++ixo]) | |
503 | ixref = ix; | |
504 | } | |
505 | } while (grpsiz != ix - ixs); | |
506 | ||
507 | /* | |
508 | * Try to move back the possibly merged group of changes, to match | |
41ccfdd9 | 509 | * the recorded position in the other file. |
295ba2fb DL |
510 | */ |
511 | while (ixref < ix) { | |
512 | rchg[--ixs] = 1; | |
513 | rchg[--ix] = 0; | |
514 | while (rchgo[--ixo]); | |
515 | } | |
d634d61e SB |
516 | |
517 | /* | |
518 | * If a group can be moved back and forth, see if there is a | |
519 | * blank line in the moving space. If there is a blank line, | |
520 | * make sure the last blank line is the end of the group. | |
521 | * | |
522 | * As we already shifted the group forward as far as possible | |
523 | * in the earlier loop, we need to shift it back only if at all. | |
524 | */ | |
525 | if ((flags & XDF_COMPACTION_HEURISTIC) && blank_lines) { | |
526 | while (ixs > 0 && | |
527 | !is_blank_line(recs, ix - 1, flags) && | |
528 | recs_match(recs, ixs - 1, ix - 1, flags)) { | |
529 | rchg[--ixs] = 1; | |
530 | rchg[--ix] = 0; | |
531 | } | |
532 | } | |
295ba2fb DL |
533 | } |
534 | ||
535 | return 0; | |
536 | } | |
537 | ||
538 | ||
3443546f LT |
539 | int xdl_build_script(xdfenv_t *xe, xdchange_t **xscr) { |
540 | xdchange_t *cscr = NULL, *xch; | |
541 | char *rchg1 = xe->xdf1.rchg, *rchg2 = xe->xdf2.rchg; | |
542 | long i1, i2, l1, l2; | |
543 | ||
544 | /* | |
545 | * Trivial. Collects "groups" of changes and creates an edit script. | |
546 | */ | |
547 | for (i1 = xe->xdf1.nrec, i2 = xe->xdf2.nrec; i1 >= 0 || i2 >= 0; i1--, i2--) | |
548 | if (rchg1[i1 - 1] || rchg2[i2 - 1]) { | |
549 | for (l1 = i1; rchg1[i1 - 1]; i1--); | |
550 | for (l2 = i2; rchg2[i2 - 1]; i2--); | |
551 | ||
552 | if (!(xch = xdl_add_change(cscr, i1, i2, l1 - i1, l2 - i2))) { | |
553 | xdl_free_script(cscr); | |
554 | return -1; | |
555 | } | |
556 | cscr = xch; | |
557 | } | |
558 | ||
559 | *xscr = cscr; | |
560 | ||
561 | return 0; | |
562 | } | |
563 | ||
564 | ||
565 | void xdl_free_script(xdchange_t *xscr) { | |
566 | xdchange_t *xch; | |
567 | ||
568 | while ((xch = xscr) != NULL) { | |
569 | xscr = xscr->next; | |
570 | xdl_free(xch); | |
571 | } | |
572 | } | |
573 | ||
467d348c RS |
574 | static int xdl_call_hunk_func(xdfenv_t *xe, xdchange_t *xscr, xdemitcb_t *ecb, |
575 | xdemitconf_t const *xecfg) | |
576 | { | |
577 | xdchange_t *xch, *xche; | |
578 | ||
579 | for (xch = xscr; xch; xch = xche->next) { | |
36617af7 AP |
580 | xche = xdl_get_hunk(&xch, xecfg); |
581 | if (!xch) | |
582 | break; | |
467d348c RS |
583 | if (xecfg->hunk_func(xch->i1, xche->i1 + xche->chg1 - xch->i1, |
584 | xch->i2, xche->i2 + xche->chg2 - xch->i2, | |
585 | ecb->priv) < 0) | |
586 | return -1; | |
587 | } | |
588 | return 0; | |
589 | } | |
3443546f | 590 | |
36617af7 AP |
591 | static void xdl_mark_ignorable(xdchange_t *xscr, xdfenv_t *xe, long flags) |
592 | { | |
593 | xdchange_t *xch; | |
594 | ||
595 | for (xch = xscr; xch; xch = xch->next) { | |
596 | int ignore = 1; | |
597 | xrecord_t **rec; | |
598 | long i; | |
599 | ||
600 | rec = &xe->xdf1.recs[xch->i1]; | |
601 | for (i = 0; i < xch->chg1 && ignore; i++) | |
602 | ignore = xdl_blankline(rec[i]->ptr, rec[i]->size, flags); | |
603 | ||
604 | rec = &xe->xdf2.recs[xch->i2]; | |
605 | for (i = 0; i < xch->chg2 && ignore; i++) | |
606 | ignore = xdl_blankline(rec[i]->ptr, rec[i]->size, flags); | |
607 | ||
608 | xch->ignore = ignore; | |
609 | } | |
610 | } | |
611 | ||
3443546f LT |
612 | int xdl_diff(mmfile_t *mf1, mmfile_t *mf2, xpparam_t const *xpp, |
613 | xdemitconf_t const *xecfg, xdemitcb_t *ecb) { | |
614 | xdchange_t *xscr; | |
615 | xdfenv_t xe; | |
3319e606 | 616 | emit_func_t ef = xecfg->hunk_func ? xdl_call_hunk_func : xdl_emit_diff; |
467d348c | 617 | |
3443546f LT |
618 | if (xdl_do_diff(mf1, mf2, xpp, &xe) < 0) { |
619 | ||
620 | return -1; | |
621 | } | |
0d21efa5 JS |
622 | if (xdl_change_compact(&xe.xdf1, &xe.xdf2, xpp->flags) < 0 || |
623 | xdl_change_compact(&xe.xdf2, &xe.xdf1, xpp->flags) < 0 || | |
295ba2fb | 624 | xdl_build_script(&xe, &xscr) < 0) { |
3443546f LT |
625 | |
626 | xdl_free_env(&xe); | |
627 | return -1; | |
628 | } | |
3443546f | 629 | if (xscr) { |
36617af7 AP |
630 | if (xpp->flags & XDF_IGNORE_BLANK_LINES) |
631 | xdl_mark_ignorable(xscr, &xe, xpp->flags); | |
632 | ||
ef2e62fe | 633 | if (ef(&xe, xscr, ecb, xecfg) < 0) { |
3443546f LT |
634 | |
635 | xdl_free_script(xscr); | |
636 | xdl_free_env(&xe); | |
637 | return -1; | |
638 | } | |
3443546f LT |
639 | xdl_free_script(xscr); |
640 | } | |
3443546f LT |
641 | xdl_free_env(&xe); |
642 | ||
643 | return 0; | |
644 | } |