1 /* Common parts for printing diff output */
2 /*
3  * Copyright (c) 2020 Neels Hofmeyr <neels@hofmeyr.de>
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 <ctype.h>
19 #include <errno.h>
20 #include <stdbool.h>
21 #include <stdint.h>
22 #include <stdio.h>
23 #include <stdlib.h>
24 #include <string.h>
25 #include <unistd.h>
26 
27 #include <arraylist.h>
28 #include <diff_main.h>
29 #include <diff_output.h>
30 
31 #include "diff_internal.h"
32 
33 static int
34 get_atom_byte(int *ch, struct diff_atom *atom, off_t off)
35 {
36 	off_t cur;
37 
38 	if (atom->at != NULL) {
39 		*ch = atom->at[off];
40 		return 0;
41 	}
42 
43 	cur = ftello(atom->root->f);
44 	if (cur == -1)
45 		return errno;
46 
47 	if (cur != atom->pos + off &&
48 	    fseeko(atom->root->f, atom->pos + off, SEEK_SET) == -1)
49 		return errno;
50 
51 	*ch = fgetc(atom->root->f);
52 	if (*ch == EOF && ferror(atom->root->f))
53 		return errno;
54 
55 	return 0;
56 }
57 
58 #define DIFF_OUTPUT_BUF_SIZE	512
59 
60 int
61 diff_output_lines(struct diff_output_info *outinfo, FILE *dest,
62 		  const char *prefix, struct diff_atom *start_atom,
63 		  unsigned int count)
64 {
65 	struct diff_atom *atom;
66 	off_t outoff = 0, *offp;
67 	uint8_t *typep;
68 	int rc;
69 
70 	if (outinfo && outinfo->line_offsets.len > 0) {
71 		unsigned int idx = outinfo->line_offsets.len - 1;
72 		outoff = outinfo->line_offsets.head[idx];
73 	}
74 
75 	foreach_diff_atom(atom, start_atom, count) {
76 		off_t outlen = 0;
77 		int i, ch, nbuf = 0;
78 		unsigned int len = atom->len;
79 		unsigned char buf[DIFF_OUTPUT_BUF_SIZE + 1 /* '\n' */];
80 		size_t n;
81 
82 		n = strlcpy(buf, prefix, sizeof(buf));
83 		if (n >= DIFF_OUTPUT_BUF_SIZE) /* leave room for '\n' */
84 			return ENOBUFS;
85 		nbuf += n;
86 
87 		if (len) {
88 			rc = get_atom_byte(&ch, atom, len - 1);
89 			if (rc)
90 				return rc;
91 			if (ch == '\n')
92 				len--;
93 		}
94 
95 		for (i = 0; i < len; i++) {
96 			rc = get_atom_byte(&ch, atom, i);
97 			if (rc)
98 				return rc;
99 			if (nbuf >= DIFF_OUTPUT_BUF_SIZE) {
100 				rc = fwrite(buf, 1, nbuf, dest);
101 				if (rc != nbuf)
102 					return errno;
103 				outlen += rc;
104 				nbuf = 0;
105 			}
106 			buf[nbuf++] = ch;
107 		}
108 		buf[nbuf++] = '\n';
109 		rc = fwrite(buf, 1, nbuf, dest);
110 		if (rc != nbuf)
111 			return errno;
112 		outlen += rc;
113 		if (outinfo) {
114 			ARRAYLIST_ADD(offp, outinfo->line_offsets);
115 			if (offp == NULL)
116 				return ENOMEM;
117 			outoff += outlen;
118 			*offp = outoff;
119 			ARRAYLIST_ADD(typep, outinfo->line_types);
120 			if (typep == NULL)
121 				return ENOMEM;
122 			*typep = *prefix == ' ' ? DIFF_LINE_CONTEXT :
123 			    *prefix == '-' ? DIFF_LINE_MINUS :
124 			    *prefix == '+' ? DIFF_LINE_PLUS : DIFF_LINE_NONE;
125 		}
126 	}
127 
128 	return DIFF_RC_OK;
129 }
130 
131 int
132 diff_output_chunk_left_version(struct diff_output_info **output_info,
133 			       FILE *dest,
134 			       const struct diff_input_info *info,
135 			       const struct diff_result *result,
136 			       const struct diff_chunk_context *cc)
137 {
138 	int rc, c_idx;
139 	struct diff_output_info *outinfo = NULL;
140 
141 	if (diff_range_empty(&cc->left))
142 		return DIFF_RC_OK;
143 
144 	if (output_info) {
145 		*output_info = diff_output_info_alloc();
146 		if (*output_info == NULL)
147 			return ENOMEM;
148 		outinfo = *output_info;
149 	}
150 
151 	/* Write out all chunks on the left side. */
152 	for (c_idx = cc->chunk.start; c_idx < cc->chunk.end; c_idx++) {
153 		const struct diff_chunk *c = &result->chunks.head[c_idx];
154 
155 		if (c->left_count) {
156 			rc = diff_output_lines(outinfo, dest, "",
157 			    c->left_start, c->left_count);
158 			if (rc)
159 				return rc;
160 		}
161 	}
162 
163 	return DIFF_RC_OK;
164 }
165 
166 int
167 diff_output_chunk_right_version(struct diff_output_info **output_info,
168 				FILE *dest,
169 				const struct diff_input_info *info,
170 				const struct diff_result *result,
171 				const struct diff_chunk_context *cc)
172 {
173 	int rc, c_idx;
174 	struct diff_output_info *outinfo = NULL;
175 
176 	if (diff_range_empty(&cc->right))
177 		return DIFF_RC_OK;
178 
179 	if (output_info) {
180 		*output_info = diff_output_info_alloc();
181 		if (*output_info == NULL)
182 			return ENOMEM;
183 		outinfo = *output_info;
184 	}
185 
186 	/* Write out all chunks on the right side. */
187 	for (c_idx = cc->chunk.start; c_idx < cc->chunk.end; c_idx++) {
188 		const struct diff_chunk *c = &result->chunks.head[c_idx];
189 
190 		if (c->right_count) {
191 			rc = diff_output_lines(outinfo, dest, "", c->right_start,
192 			    c->right_count);
193 			if (rc)
194 				return rc;
195 		}
196 	}
197 
198 	return DIFF_RC_OK;
199 }
200 
201 int
202 diff_output_trailing_newline_msg(struct diff_output_info *outinfo, FILE *dest,
203 				 const struct diff_chunk *c)
204 {
205 	enum diff_chunk_type chunk_type = diff_chunk_type(c);
206 	struct diff_atom *atom, *start_atom;
207 	unsigned int atom_count;
208 	int rc, ch;
209 	off_t outoff = 0, *offp;
210 	uint8_t *typep;
211 
212 
213 	if (chunk_type == CHUNK_MINUS || chunk_type == CHUNK_SAME) {
214 		start_atom = c->left_start;
215 		atom_count = c->left_count;
216 	} else if (chunk_type == CHUNK_PLUS) {
217 		start_atom = c->right_start;
218 		atom_count = c->right_count;
219 	} else
220 		return EINVAL;
221 
222 	/* Locate the last atom. */
223 	if (atom_count == 0)
224 		return EINVAL;
225 	atom = &start_atom[atom_count - 1];
226 
227 	rc = get_atom_byte(&ch, atom, atom->len - 1);
228 	if (rc != DIFF_RC_OK)
229 		return rc;
230 
231 	if (ch != '\n') {
232 		if (outinfo && outinfo->line_offsets.len > 0) {
233 			unsigned int idx = outinfo->line_offsets.len - 1;
234 			outoff = outinfo->line_offsets.head[idx];
235 		}
236 		rc = fprintf(dest, "\\ No newline at end of file\n");
237 		if (rc < 0)
238 			return errno;
239 		if (outinfo) {
240 			ARRAYLIST_ADD(offp, outinfo->line_offsets);
241 			if (offp == NULL)
242 				return ENOMEM;
243 			outoff += rc;
244 			*offp = outoff;
245 			ARRAYLIST_ADD(typep, outinfo->line_types);
246 			if (typep == NULL)
247 				return ENOMEM;
248 			*typep = DIFF_LINE_NONE;
249 		}
250 	}
251 
252 	return DIFF_RC_OK;
253 }
254 
255 static bool
256 is_function_prototype(unsigned char ch)
257 {
258 	return (isalpha((unsigned char)ch) || ch == '_' || ch == '$');
259 }
260 
261 #define begins_with(s, pre) (strncmp(s, pre, sizeof(pre)-1) == 0)
262 
263 int
264 diff_output_match_function_prototype(char *prototype, size_t prototype_size,
265     int *last_prototype_idx, const struct diff_result *result,
266     const struct diff_chunk_context *cc)
267 {
268 	struct diff_atom *start_atom, *atom;
269 	const struct diff_data *data;
270 	unsigned char buf[DIFF_FUNCTION_CONTEXT_SIZE];
271 	const char *state = NULL;
272 	int rc, i, ch;
273 
274 	if (result->left->atoms.len > 0 && cc->left.start > 0) {
275 		data = result->left;
276 		start_atom = &data->atoms.head[cc->left.start - 1];
277 	} else
278 		return DIFF_RC_OK;
279 
280 	diff_data_foreach_atom_backwards_from(start_atom, atom, data) {
281 		int atom_idx = diff_atom_root_idx(data, atom);
282 		if (atom_idx < *last_prototype_idx)
283 			break;
284 		rc = get_atom_byte(&ch, atom, 0);
285 		if (rc)
286 			return rc;
287 		buf[0] = (unsigned char)ch;
288 		if (!is_function_prototype(buf[0]))
289 			continue;
290 		for (i = 1; i < atom->len && i < sizeof(buf) - 1; i++) {
291 			rc = get_atom_byte(&ch, atom, i);
292 			if (rc)
293 				return rc;
294 			if (ch == '\n')
295 				break;
296 			buf[i] = (unsigned char)ch;
297 		}
298 		buf[i] = '\0';
299 		if (begins_with(buf, "private:")) {
300 			if (!state)
301 				state = " (private)";
302 		} else if (begins_with(buf, "protected:")) {
303 			if (!state)
304 				state = " (protected)";
305 		} else if (begins_with(buf, "public:")) {
306 			if (!state)
307 				state = " (public)";
308 		} else {
309 			if (state)  /* don't care about truncation */
310 				strlcat(buf, state, sizeof(buf));
311 			strlcpy(prototype, buf, prototype_size);
312 			break;
313 		}
314 	}
315 
316 	*last_prototype_idx = diff_atom_root_idx(data, start_atom);
317 	return DIFF_RC_OK;
318 }
319 
320 struct diff_output_info *
321 diff_output_info_alloc(void)
322 {
323 	struct diff_output_info *output_info;
324 	off_t *offp;
325 	uint8_t *typep;
326 
327 	output_info = malloc(sizeof(*output_info));
328 	if (output_info != NULL) {
329 		ARRAYLIST_INIT(output_info->line_offsets, 128);
330 		ARRAYLIST_ADD(offp, output_info->line_offsets);
331 		if (offp == NULL) {
332 			diff_output_info_free(output_info);
333 			return NULL;
334 		}
335 		*offp = 0;
336 		ARRAYLIST_INIT(output_info->line_types, 128);
337 		ARRAYLIST_ADD(typep, output_info->line_types);
338 		if (typep == NULL) {
339 			diff_output_info_free(output_info);
340 			return NULL;
341 		}
342 		*typep = DIFF_LINE_NONE;
343 	}
344 	return output_info;
345 }
346 
347 void
348 diff_output_info_free(struct diff_output_info *output_info)
349 {
350 	ARRAYLIST_FREE(output_info->line_offsets);
351 	ARRAYLIST_FREE(output_info->line_types);
352 	free(output_info);
353 }
354 
355 const char *
356 diff_output_get_label_left(const struct diff_input_info *info)
357 {
358 	if (info->flags & DIFF_INPUT_LEFT_NONEXISTENT)
359 		return "/dev/null";
360 
361 	return info->left_path ? info->left_path : "a";
362 }
363 
364 const char *
365 diff_output_get_label_right(const struct diff_input_info *info)
366 {
367 	if (info->flags & DIFF_INPUT_RIGHT_NONEXISTENT)
368 		return "/dev/null";
369 
370 	return info->right_path ? info->right_path : "b";
371 }