1 /*
2  * Copyright (c) 2020 Neels Hofmeyr <neels@hofmeyr.de>
3  * Copyright (c) 2020 Stefan Sperling <stsp@openbsd.org>
4  *
5  * Permission to use, copy, modify, and distribute this software for any
6  * purpose with or without fee is hereby granted, provided that the above
7  * copyright notice and this permission notice appear in all copies.
8  *
9  * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
10  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
11  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
12  * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
13  * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
14  * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
15  * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
16  */
17 
18 #include <sys/mman.h>
19 #include <sys/stat.h>
20 #include <sys/queue.h>
21 
22 #include <errno.h>
23 #include <stdio.h>
24 #include <stdlib.h>
25 #include <string.h>
26 
27 #include "got_compat.h"
28 
29 #include "got_object.h"
30 #include "got_opentemp.h"
31 #include "got_error.h"
32 #include "got_diff.h"
33 
34 #include "got_lib_diff.h"
35 
36 const struct diff_algo_config myers_then_patience;
37 const struct diff_algo_config myers_then_myers_divide;
38 const struct diff_algo_config patience;
39 const struct diff_algo_config myers_divide;
40 
41 const struct diff_algo_config myers_then_patience = (struct diff_algo_config){
42 	.impl = diff_algo_myers,
43 	.permitted_state_size = 1024 * 1024 * sizeof(int),
44 	.fallback_algo = &patience,
45 };
46 
47 const struct diff_algo_config myers_then_myers_divide =
48 	(struct diff_algo_config){
49 	.impl = diff_algo_myers,
50 	.permitted_state_size = 1024 * 1024 * sizeof(int),
51 	.fallback_algo = &myers_divide,
52 };
53 
54 const struct diff_algo_config patience = (struct diff_algo_config){
55 	.impl = diff_algo_patience,
56 	/* After subdivision, do Patience again: */
57 	.inner_algo = &patience,
58 	/* If subdivision failed, do Myers Divide et Impera: */
59 	.fallback_algo = &myers_then_myers_divide,
60 };
61 
62 const struct diff_algo_config myers_divide = (struct diff_algo_config){
63 	.impl = diff_algo_myers_divide,
64 	/* When division succeeded, start from the top: */
65 	.inner_algo = &myers_then_myers_divide,
66 	/* (fallback_algo = NULL implies diff_algo_none). */
67 };
68 
69 /* If the state for a forward-Myers is small enough, use Myers, otherwise first
70  * do a Myers-divide. */
71 const struct diff_config diff_config_myers_then_myers_divide = {
72 	.atomize_func = diff_atomize_text_by_line,
73 	.algo = &myers_then_myers_divide,
74 };
75 
76 /* If the state for a forward-Myers is small enough, use Myers, otherwise first
77  * do a Patience. */
78 const struct diff_config diff_config_myers_then_patience = {
79 	.atomize_func = diff_atomize_text_by_line,
80 	.algo = &myers_then_patience,
81 };
82 
83 /* Directly force Patience as a first divider of the source file. */
84 const struct diff_config diff_config_patience = {
85 	.atomize_func = diff_atomize_text_by_line,
86 	.algo = &patience,
87 };
88 
89 /* Directly force Patience as a first divider of the source file. */
90 const struct diff_config diff_config_no_algo = {
91 	.atomize_func = diff_atomize_text_by_line,
92 };
93 
94 const struct got_error *
95 got_diffreg_close(char *p1, size_t size1, char *p2, size_t size2)
96 {
97 	const struct got_error *err = NULL;
98 
99 	if (p1 && munmap(p1, size1) == -1 && err == NULL)
100 		err = got_error_from_errno("munmap");
101 	if (p2 && munmap(p2, size2) == -1 && err == NULL)
102 		err = got_error_from_errno("munmap");
103 	return err;
104 }
105 
106 const struct got_error *
107 got_diff_get_config(struct diff_config **cfg,
108     enum got_diff_algorithm algorithm,
109     diff_atomize_func_t atomize_func, void *atomize_func_data)
110 {
111 	*cfg = calloc(1, sizeof(**cfg));
112 	if (*cfg == NULL)
113 		return got_error_from_errno("calloc");
114 
115 	switch (algorithm) {
116 	case GOT_DIFF_ALGORITHM_PATIENCE:
117 		(*cfg)->algo = &patience;
118 		break;
119 	case GOT_DIFF_ALGORITHM_MYERS:
120 		(*cfg)->algo = &myers_then_myers_divide;
121 		break;
122 	default:
123 		return got_error_msg(GOT_ERR_NOT_IMPL, "bad diff algorithm");
124 	}
125 
126 	if (atomize_func) {
127 		(*cfg)->atomize_func = atomize_func;
128 		(*cfg)->atomize_func_data = atomize_func_data;
129 	} else
130 		(*cfg)->atomize_func = diff_atomize_text_by_line;
131 
132 	(*cfg)->max_recursion_depth = 0; /* use default recursion depth */
133 
134 	return NULL;
135 }
136 
137 const struct got_error *
138 got_diff_prepare_file(FILE *f, char **p, size_t *size,
139     struct diff_data *diff_data, const struct diff_config *cfg,
140     int ignore_whitespace, int force_text_diff)
141 {
142 	const struct got_error *err = NULL;
143 	struct stat st;
144 	int diff_flags = 0, rc;
145 
146 	*size = 0;
147 
148 	diff_flags |= DIFF_FLAG_SHOW_PROTOTYPES;
149 	if (ignore_whitespace)
150 		diff_flags |= DIFF_FLAG_IGNORE_WHITESPACE;
151 	if (force_text_diff)
152 		diff_flags |= DIFF_FLAG_FORCE_TEXT_DATA;
153 
154 	if (fstat(fileno(f), &st) == -1) {
155 		err = got_error_from_errno("fstat");
156 		goto done;
157 	}
158 #ifndef GOT_DIFF_NO_MMAP
159 	*p = mmap(NULL, st.st_size, PROT_READ, MAP_PRIVATE,
160 	    fileno(f), 0);
161 	if (*p == MAP_FAILED)
162 #endif
163 		*p = NULL; /* fall back on file I/O */
164 
165 	rc = diff_atomize_file(diff_data, cfg, f, *p, st.st_size, diff_flags);
166 	if (rc) {
167 		err = got_error_set_errno(rc, "diff_atomize_file");
168 		goto done;
169 	}
170 done:
171 	if (err)
172 		diff_data_free(diff_data);
173 	else
174 		*size = st.st_size;
175 	return err;
176 }
177 
178 const struct got_error *
179 got_diffreg(struct got_diffreg_result **diffreg_result, FILE *f1, FILE *f2,
180     enum got_diff_algorithm algorithm, int ignore_whitespace,
181     int force_text_diff)
182 {
183 	const struct got_error *err = NULL;
184 	struct diff_config *cfg = NULL;
185 	char *p1 = NULL, *p2 = NULL;
186 	size_t size1, size2;
187 	struct diff_data d_left, d_right;
188 	struct diff_data *left, *right;
189 	struct diff_result *diff_result;
190 
191 	if (diffreg_result) {
192 		*diffreg_result = calloc(1, sizeof(**diffreg_result));
193 		if (*diffreg_result == NULL)
194 			return got_error_from_errno("calloc");
195 		left = &(*diffreg_result)->left;
196 		right = &(*diffreg_result)->right;
197 	} else {
198 		memset(&d_left, 0, sizeof(d_left));
199 		memset(&d_right, 0, sizeof(d_right));
200 		left = &d_left;
201 		right = &d_right;
202 	}
203 
204 	err = got_diff_get_config(&cfg, algorithm, NULL, NULL);
205 	if (err)
206 		goto done;
207 
208 	err = got_diff_prepare_file(f1, &p1, &size1, left, cfg,
209 	    ignore_whitespace, force_text_diff);
210 	if (err)
211 		goto done;
212 
213 	err = got_diff_prepare_file(f2, &p2, &size2, right, cfg,
214 	    ignore_whitespace, force_text_diff);
215 	if (err)
216 		goto done;
217 
218 	diff_result = diff_main(cfg, left, right);
219 	if (diff_result == NULL) {
220 		err = got_error_set_errno(ENOMEM, "malloc");
221 		goto done;
222 	}
223 	if (diff_result->rc != DIFF_RC_OK) {
224 		err = got_error_set_errno(diff_result->rc, "diff");
225 		goto done;
226 	}
227 
228 	if (diffreg_result) {
229 		(*diffreg_result)->result = diff_result;
230 		(*diffreg_result)->map1 = p1;
231 		(*diffreg_result)->size1 = size1;
232 		(*diffreg_result)->map2 = p2;
233 		(*diffreg_result)->size2 = size2;
234 	}
235 done:
236 	free(cfg);
237 	if (diffreg_result == NULL) {
238 		diff_data_free(left);
239 		diff_data_free(right);
240 	}
241 	if (err) {
242 		got_diffreg_close(p1, size1, p2, size2);
243 		if (diffreg_result) {
244 			diff_data_free(left);
245 			diff_data_free(right);
246 			free(*diffreg_result);
247 			*diffreg_result = NULL;
248 		}
249 	}
250 
251 	return err;
252 }
253 
254 const struct got_error *
255 got_diffreg_output(struct got_diff_line **lines, size_t *nlines,
256     struct got_diffreg_result *diff_result, int f1_exists, int f2_exists,
257     const char *path1, const char *path2,
258     enum got_diff_output_format output_format, int context_lines, FILE *outfile)
259 {
260 	struct diff_input_info info = {
261 		.left_path = path1,
262 		.right_path = path2,
263 		.flags = 0,
264 	};
265 	int rc;
266 	struct diff_output_info *output_info;
267 
268 	if (!f1_exists)
269 		info.flags |= DIFF_INPUT_LEFT_NONEXISTENT;
270 	if (!f2_exists)
271 		info.flags |= DIFF_INPUT_RIGHT_NONEXISTENT;
272 
273 	switch (output_format) {
274 	case GOT_DIFF_OUTPUT_UNIDIFF:
275 		rc = diff_output_unidiff(
276 		    lines ? &output_info : NULL, outfile, &info,
277 		    diff_result->result, context_lines);
278 		if (rc != DIFF_RC_OK)
279 			return got_error_set_errno(rc, "diff_output_unidiff");
280 		break;
281 	case GOT_DIFF_OUTPUT_PLAIN:
282 		rc = diff_output_plain(lines ? &output_info : NULL,
283 		    outfile, &info, diff_result->result, 1);
284 		if (rc != DIFF_RC_OK)
285 			return got_error_set_errno(rc, "diff_output_edscript");
286 		break;
287 
288 	}
289 
290 	if (lines && *lines) {
291 		if (output_info->line_offsets.len > 0) {
292 			struct got_diff_line *p;
293 			off_t prev_offset = 0, *o;
294 			uint8_t *o2;
295 			int i, len;
296 			if (*nlines > 0) {
297 				prev_offset = (*lines)[*nlines - 1].offset;
298 				/*
299 				 * First line offset is always zero. Skip it
300 				 * when appending to a pre-populated array.
301 				 */
302 				o = &output_info->line_offsets.head[1];
303 				o2 = &output_info->line_types.head[1];
304 				len = output_info->line_offsets.len - 1;
305 			} else {
306 				o = &output_info->line_offsets.head[0];
307 				o2 = &output_info->line_types.head[0];
308 				len = output_info->line_offsets.len;
309 			}
310 			p = reallocarray(*lines, *nlines + len, sizeof(**lines));
311 			if (p == NULL)
312 				return got_error_from_errno("calloc");
313 			for (i = 0; i < len; i++) {
314 				p[*nlines + i].offset = o[i] + prev_offset;
315 				p[*nlines + i].type = o2[i];
316 			}
317 			*lines = p;
318 			*nlines += len;
319 		}
320 		diff_output_info_free(output_info);
321 	}
322 
323 	return NULL;
324 }
325 
326 const struct got_error *
327 got_diffreg_result_free(struct got_diffreg_result *diffreg_result)
328 {
329 	const struct got_error *err;
330 
331 	diff_result_free(diffreg_result->result);
332 	diff_data_free(&diffreg_result->left);
333 	diff_data_free(&diffreg_result->right);
334 	err = got_diffreg_close(diffreg_result->map1, diffreg_result->size1,
335 	    diffreg_result->map2, diffreg_result->size2);
336 	free(diffreg_result);
337 	return err;
338 }
339 
340 const struct got_error *
341 got_diffreg_result_free_left(struct got_diffreg_result *diffreg_result)
342 {
343 	diff_data_free(&diffreg_result->left);
344 	memset(&diffreg_result->left, 0, sizeof(diffreg_result->left));
345 	return got_diffreg_close(diffreg_result->map1, diffreg_result->size1,
346 	    NULL, 0);
347 }
348 
349 const struct got_error *
350 got_diffreg_result_free_right(struct got_diffreg_result *diffreg_result)
351 {
352 	diff_data_free(&diffreg_result->right);
353 	memset(&diffreg_result->right, 0, sizeof(diffreg_result->right));
354 	return got_diffreg_close(NULL, 0, diffreg_result->map2,
355 	    diffreg_result->size2);
356 }