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_object.h"
28 #include "got_opentemp.h"
29 #include "got_error.h"
30 
31 #include "got_lib_diff.h"
32 
33 const struct diff_algo_config myers_then_patience;
34 const struct diff_algo_config myers_then_myers_divide;
35 const struct diff_algo_config patience;
36 const struct diff_algo_config myers_divide;
37 
38 const struct diff_algo_config myers_then_patience = (struct diff_algo_config){
39 	.impl = diff_algo_myers,
40 	.permitted_state_size = 1024 * 1024 * sizeof(int),
41 	.fallback_algo = &patience,
42 };
43 
44 const struct diff_algo_config myers_then_myers_divide =
45 	(struct diff_algo_config){
46 	.impl = diff_algo_myers,
47 	.permitted_state_size = 1024 * 1024 * sizeof(int),
48 	.fallback_algo = &myers_divide,
49 };
50 
51 const struct diff_algo_config patience = (struct diff_algo_config){
52 	.impl = diff_algo_patience,
53 	/* After subdivision, do Patience again: */
54 	.inner_algo = &patience,
55 	/* If subdivision failed, do Myers Divide et Impera: */
56 	.fallback_algo = &myers_then_myers_divide,
57 };
58 
59 const struct diff_algo_config myers_divide = (struct diff_algo_config){
60 	.impl = diff_algo_myers_divide,
61 	/* When division succeeded, start from the top: */
62 	.inner_algo = &myers_then_myers_divide,
63 	/* (fallback_algo = NULL implies diff_algo_none). */
64 };
65 
66 /* If the state for a forward-Myers is small enough, use Myers, otherwise first
67  * do a Myers-divide. */
68 const struct diff_config diff_config_myers_then_myers_divide = {
69 	.atomize_func = diff_atomize_text_by_line,
70 	.algo = &myers_then_myers_divide,
71 };
72 
73 /* If the state for a forward-Myers is small enough, use Myers, otherwise first
74  * do a Patience. */
75 const struct diff_config diff_config_myers_then_patience = {
76 	.atomize_func = diff_atomize_text_by_line,
77 	.algo = &myers_then_patience,
78 };
79 
80 /* Directly force Patience as a first divider of the source file. */
81 const struct diff_config diff_config_patience = {
82 	.atomize_func = diff_atomize_text_by_line,
83 	.algo = &patience,
84 };
85 
86 /* Directly force Patience as a first divider of the source file. */
87 const struct diff_config diff_config_no_algo = {
88 	.atomize_func = diff_atomize_text_by_line,
89 };
90 
91 const struct got_error *
92 got_diffreg_close(FILE *f1, char *p1, size_t size1,
93     FILE *f2, char *p2, size_t size2)
94 {
95 	const struct got_error *err = NULL;
96 
97 	if (p1 && munmap(p1, size1) == -1 && err == NULL)
98 		err = got_error_from_errno("munmap");
99 	if (p2 && munmap(p2, size2) == -1 && err == NULL)
100 		err = got_error_from_errno("munmap");
101 	if (f1 && fclose(f1) != 0 && err == NULL)
102 		err = got_error_from_errno("fclose");
103 	if (f2 && fclose(f2) != 0 && err == NULL)
104 		err = got_error_from_errno("fclose");
105 	return err;
106 }
107 
108 const struct diff_config *
109 got_diff_get_config(enum got_diff_algorithm algorithm)
110 {
111 	switch (algorithm) {
112 	case GOT_DIFF_ALGORITHM_PATIENCE:
113 		return &diff_config_patience;
114 	case GOT_DIFF_ALGORITHM_MYERS:
115 		return &diff_config_myers_then_myers_divide;
116 	}
117 	return NULL; /* should not happen */
118 }
119 
120 const struct got_error *
121 got_diff_prepare_file(FILE *f, char **p, size_t *size,
122     struct diff_data *diff_data, const struct diff_config *cfg,
123     int ignore_whitespace)
124 {
125 	const struct got_error *err = NULL;
126 	struct stat st;
127 	int diff_flags = 0, rc;
128 
129 	*size = 0;
130 
131 	diff_flags |= DIFF_FLAG_SHOW_PROTOTYPES;
132 	if (ignore_whitespace)
133 		diff_flags |= DIFF_FLAG_IGNORE_WHITESPACE;
134 
135 	if (fstat(fileno(f), &st) == -1) {
136 		err = got_error_from_errno("fstat");
137 		goto done;
138 	}
139 #ifndef GOT_DIFF_NO_MMAP
140 	*p = mmap(NULL, st.st_size, PROT_READ, MAP_PRIVATE,
141 	    fileno(f), 0);
142 	if (*p == MAP_FAILED)
143 #endif
144 		*p = NULL; /* fall back on file I/O */
145 
146 	rc = diff_atomize_file(diff_data, cfg, f, *p, st.st_size, diff_flags);
147 	if (rc) {
148 		err = got_error_set_errno(rc, "diff_atomize_file");
149 		goto done;
150 	}
151 done:
152 	if (err)
153 		diff_data_free(diff_data);
154 	else
155 		*size = st.st_size;
156 	return err;
157 }
158 
159 const struct got_error *
160 got_diffreg(struct got_diffreg_result **diffreg_result, FILE *f1, FILE *f2,
161     enum got_diff_algorithm algorithm, int ignore_whitespace)
162 {
163 	const struct got_error *err = NULL;
164 	const struct diff_config *cfg;
165 	char *p1 = NULL, *p2 = NULL;
166 	int f1_created = 0, f2_created = 0;
167 	size_t size1, size2;
168 	struct diff_data d_left, d_right;
169 	struct diff_data *left, *right;
170 	struct diff_result *diff_result;
171 
172 	if (diffreg_result) {
173 		*diffreg_result = calloc(1, sizeof(**diffreg_result));
174 		if (*diffreg_result == NULL)
175 			return got_error_from_errno("calloc");
176 		left = &(*diffreg_result)->left;
177 		right = &(*diffreg_result)->right;
178 	} else {
179 		memset(&d_left, 0, sizeof(d_left));
180 		memset(&d_right, 0, sizeof(d_right));
181 		left = &d_left;
182 		right = &d_right;
183 	}
184 
185 	cfg = got_diff_get_config(algorithm);
186 	if (cfg == NULL) {
187 		err = got_error(GOT_ERR_NOT_IMPL);
188 		goto done;
189 	}
190 
191 	if (f1 == NULL) {
192 		f1_created = 1;
193 		f1 = got_opentemp();
194 		if (f1 == NULL) {
195 			err = got_error_from_errno("got_opentemp");
196 			goto done;
197 		}
198 	}
199 	if (f2 == NULL) {
200 		f2_created = 1;
201 		f2 = got_opentemp();
202 		if (f2 == NULL) {
203 			err = got_error_from_errno("got_opentemp");
204 			goto done;
205 		}
206 	}
207 
208 	err = got_diff_prepare_file(f1, &p1, &size1, left, cfg,
209 	    ignore_whitespace);
210 	if (err)
211 		goto done;
212 
213 	err = got_diff_prepare_file(f2, &p2, &size2, right, cfg,
214 	    ignore_whitespace);
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 		if (f1_created)
231 			(*diffreg_result)->f1 = f1;
232 		(*diffreg_result)->map1 = p1;
233 		(*diffreg_result)->size1 = size1;
234 		if (f2_created)
235 			(*diffreg_result)->f2 = f2;
236 		(*diffreg_result)->map2 = p2;
237 		(*diffreg_result)->size2 = size2;
238 	}
239 done:
240 	if (diffreg_result == NULL) {
241 		diff_data_free(left);
242 		diff_data_free(right);
243 	}
244 	if (err) {
245 		got_diffreg_close(f1_created ? f1 : NULL, p1, size1,
246 		    f2_created ? f2 : NULL, p2, size2);
247 		if (diffreg_result) {
248 			diff_data_free(left);
249 			diff_data_free(right);
250 			free(*diffreg_result);
251 			*diffreg_result = NULL;
252 		}
253 	}
254 	
255 	return err;
256 }
257 
258 const struct got_error *
259 got_diffreg_output(off_t **line_offsets, size_t *nlines,
260     struct got_diffreg_result *diff_result, FILE *f1, FILE *f2,
261     const char *path1, const char *path2,
262     enum got_diff_output_format output_format, int context_lines, FILE *outfile)
263 {
264 	struct diff_input_info info = {
265 		.left_path = path1,
266 		.right_path = path2,
267 	};
268 	int rc;
269 	struct diff_output_info *output_info;
270 
271 	switch (output_format) {
272 	case GOT_DIFF_OUTPUT_UNIDIFF:
273 		rc = diff_output_unidiff(
274 		    line_offsets ? &output_info : NULL, outfile, &info,
275 		    diff_result->result, context_lines);
276 		if (rc != DIFF_RC_OK)
277 			return got_error_set_errno(rc, "diff_output_unidiff");
278 		break;
279 	case GOT_DIFF_OUTPUT_EDSCRIPT:
280 		rc = diff_output_edscript(line_offsets ? &output_info : NULL,
281 		    outfile, &info, diff_result->result);
282 		if (rc != DIFF_RC_OK)
283 			return got_error_set_errno(rc, "diff_output_edscript");
284 		break;
285 
286 	}
287 
288 	if (line_offsets && *line_offsets) {
289 		if (output_info->line_offsets.len > 0) {
290 			off_t prev_offset = 0, *p, *o;
291 			int i, len;
292 			if (*nlines > 0) {
293 				prev_offset = (*line_offsets)[*nlines - 1];
294 				/*
295 				 * First line offset is always zero. Skip it
296 				 * when appending to a pre-populated array.
297 				 */
298 				o = &output_info->line_offsets.head[1];
299 				len = output_info->line_offsets.len - 1;
300 			} else {
301 				o = &output_info->line_offsets.head[0];
302 				len = output_info->line_offsets.len;
303 			}
304 			p = reallocarray(*line_offsets, *nlines + len,
305 			    sizeof(off_t));
306 			if (p == NULL)
307 				return got_error_from_errno("calloc");
308 			for (i = 0; i < len; i++)
309 				p[*nlines + i] = o[i] + prev_offset;
310 			*line_offsets = p;
311 			*nlines += len;
312 		}
313 		diff_output_info_free(output_info);
314 	}
315 
316 	return NULL;
317 }
318 
319 const struct got_error *
320 got_diffreg_result_free(struct got_diffreg_result *diffreg_result)
321 {
322 	const struct got_error *err;
323 
324 	diff_result_free(diffreg_result->result);
325 	diff_data_free(&diffreg_result->left);
326 	diff_data_free(&diffreg_result->right);
327 	err = got_diffreg_close(diffreg_result->f1, diffreg_result->map1,
328 	    diffreg_result->size1, diffreg_result->f2,
329 	    diffreg_result->map2, diffreg_result->size2);
330 	free(diffreg_result);
331 	return err;
332 }
333 
334 const struct got_error *
335 got_diffreg_result_free_left(struct got_diffreg_result *diffreg_result)
336 {
337 	diff_data_free(&diffreg_result->left);
338 	memset(&diffreg_result->left, 0, sizeof(diffreg_result->left));
339 	return got_diffreg_close(diffreg_result->f1, diffreg_result->map1,
340 	    diffreg_result->size1, NULL, NULL, 0);
341 }
342 
343 const struct got_error *
344 got_diffreg_result_free_right(struct got_diffreg_result *diffreg_result)
345 {
346 	diff_data_free(&diffreg_result->right);
347 	memset(&diffreg_result->right, 0, sizeof(diffreg_result->right));
348 	return got_diffreg_close(NULL, NULL, 0, diffreg_result->f2,
349 	    diffreg_result->map2, diffreg_result->size2);
350 }