1 /* Generic infrastructure to implement various diff algorithms. */
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 #pragma once
19 
20 #include <stdint.h>
21 #include <stdlib.h>
22 #include <stdbool.h>
23 #include <string.h>
24 #include <errno.h>
25 
26 #include <diff/arraylist.h>
27 
28 #ifndef MAX
29 #define MAX(A,B) ((A)>(B)?(A):(B))
30 #endif
31 #ifndef MIN
32 #define MIN(A,B) ((A)<(B)?(A):(B))
33 #endif
34 
35 struct range {
36 	int start;
37 	int end;
38 };
39 
40 static inline bool
41 range_empty(const struct range *r)
42 {
43 	return r->start == r->end;
44 }
45 
46 static inline bool
47 ranges_touch(const struct range *a, const struct range *b)
48 {
49 	return (a->end >= b->start) && (a->start <= b->end);
50 }
51 
52 static inline void
53 ranges_merge(struct range *a, const struct range *b)
54 {
55 	*a = (struct range){
56 		.start = MIN(a->start, b->start),
57 		.end = MAX(a->end, b->end),
58 	};
59 }
60 
61 static inline int
62 range_len(const struct range *r)
63 {
64 	if (!r)
65 		return 0;
66 	return r->end - r->start;
67 }
68 
69 /* List of all possible return codes of a diff invocation. */
70 enum diff_rc {
71 	DIFF_RC_USE_DIFF_ALGO_FALLBACK = -1,
72 	DIFF_RC_OK = 0,
73 	DIFF_RC_ENOTSUP = ENOTSUP,
74 	DIFF_RC_ENOMEM = ENOMEM,
75 	DIFF_RC_EINVAL = EINVAL,
76 };
77 
78 struct diff_atom {
79 	const uint8_t *at;
80 	size_t len;
81 
82 	/* This hash is just a very cheap speed up for finding *mismatching* atoms. When hashes match, we still need to
83 	 * compare entire atoms to find out whether they are indeed identical or not. */
84 	unsigned int hash;
85 
86 	/* State for the Patience Diff algorithm */
87 	/* TODO: keep a separate array for the patience state */
88 	struct {
89 		bool unique_here;
90 		bool unique_in_both;
91 		struct diff_atom *pos_in_other;
92 		struct diff_atom *prev_stack;
93 		struct range identical_lines;
94 	} patience;
95 };
96 
97 static inline bool
98 diff_atom_same(const struct diff_atom *left, const struct diff_atom *right)
99 {
100 	return left->hash == right->hash
101 		&& left->len == right->len
102 		&& memcmp(left->at, right->at, left->len) == 0;
103 }
104 
105 /* For each file, there is a "root" struct diff_data referencing the entire file, which the atoms are parsed from. In
106  * recursion of diff algorithm, there may be "child" struct diff_data only referencing a subsection of the file,
107  * re-using the atoms parsing. For "root" structs, atoms_allocated will be nonzero, indicating that the array of atoms
108  * is owned by that struct. For "child" structs, atoms_allocated == 0, to indicate that the struct is referencing a
109  * subset of atoms. */
110 struct diff_data {
111 	const uint8_t *data;
112 	size_t len;
113 	ARRAYLIST(struct diff_atom) atoms;
114 	struct diff_data *root;
115 };
116 
117 void diff_data_free(struct diff_data *diff_data);
118 
119 /* The atom's index in the entire file. For atoms divided by lines of text, this yields the line number (starting with
120  * 0). Also works for diff_data that reference only a subsection of a file, always reflecting the global position in
121  * the file (and not the relative position within the subsection). */
122 #define diff_atom_root_idx(DIFF_DATA, ATOM) \
123 	((ATOM) && ((ATOM) >= (DIFF_DATA)->root->atoms.head) \
124 	 ? (unsigned int)((ATOM) - ((DIFF_DATA)->root->atoms.head)) \
125 	 : (DIFF_DATA)->root->atoms.len)
126 
127 /* The atom's index within DIFF_DATA. For atoms divided by lines of text, this yields the line number (starting with
128  * 0). */
129 #define diff_atom_idx(DIFF_DATA, ATOM) \
130 	((ATOM) && ((ATOM) >= (DIFF_DATA)->atoms.head) \
131 	 ? (unsigned int)((ATOM) - ((DIFF_DATA)->atoms.head)) \
132 	 : (DIFF_DATA)->atoms.len)
133 
134 #define foreach_diff_atom(ATOM, FIRST_ATOM, COUNT) \
135 	for ((ATOM) = (FIRST_ATOM); \
136 	     (ATOM) && ((ATOM) >= (FIRST_ATOM)) && ((ATOM) - (FIRST_ATOM) < (COUNT)); \
137 	     (ATOM)++)
138 
139 #define diff_data_foreach_atom(ATOM, DIFF_DATA) \
140 	foreach_diff_atom(ATOM, (DIFF_DATA)->atoms.head, (DIFF_DATA)->atoms.len)
141 
142 #define diff_data_foreach_atom_from(FROM, ATOM, DIFF_DATA) \
143 	for ((ATOM) = (FROM); \
144 	     (ATOM) && ((ATOM) >= (DIFF_DATA)->atoms.head) && ((ATOM) - (DIFF_DATA)->atoms.head < (DIFF_DATA)->atoms.len); \
145 	     (ATOM)++)
146 
147 /* A diff chunk represents a set of atoms on the left and/or a set of atoms on the right.
148  *
149  * If solved == false:
150  * The diff algorithm has divided the source file, and this is a chunk that the inner_algo should run on next.
151  * The lines on the left should be diffed against the lines on the right.
152  * (If there are no left lines or no right lines, it implies solved == true, because there is nothing to diff.)
153  *
154  * If solved == true:
155  * If there are only left atoms, it is a chunk removing atoms from the left ("a minus chunk").
156  * If there are only right atoms, it is a chunk adding atoms from the right ("a plus chunk").
157  * If there are both left and right lines, it is a chunk of equal content on both sides,
158  * and left_count == right_count:
159  *
160  * - foo  }
161  * - bar  }-- diff_chunk{ left_start = &left.atoms.head[0], left_count = 3,
162  * - baz  }               right_start = NULL, right_count = 0 }
163  *   moo    }
164  *   goo    }-- diff_chunk{ left_start = &left.atoms.head[3], left_count = 3,
165  *   zoo    }               right_start = &right.atoms.head[0], right_count = 3 }
166  *  +loo      }
167  *  +roo      }-- diff_chunk{ left_start = NULL, left_count = 0,
168  *  +too      }               right_start = &right.atoms.head[3], right_count = 3 }
169  *
170  */
171 struct diff_chunk {
172 	bool solved;
173 	struct diff_atom *left_start;
174 	unsigned int left_count;
175 	struct diff_atom *right_start;
176 	unsigned int right_count;
177 };
178 
179 typedef ARRAYLIST(struct diff_chunk) diff_chunk_arraylist_t;
180 #define DIFF_RESULT_ALLOC_BLOCKSIZE 128
181 
182 struct diff_result {
183 	enum diff_rc rc;
184 	struct diff_data left;
185 	struct diff_data right;
186 	diff_chunk_arraylist_t chunks;
187 };
188 
189 struct diff_state {
190 	/* The final result passed to the original diff caller. */
191 	struct diff_result *result;
192 
193 	/* The root diff_data is in result->left,right, these are (possibly) subsections of the root data. */
194 	struct diff_data left;
195 	struct diff_data right;
196 
197 	unsigned int recursion_depth_left;
198 
199 	/* Remaining chunks from one diff algorithm pass, if any solved == false chunks came up. */
200 	diff_chunk_arraylist_t temp_result;
201 };
202 
203 struct diff_chunk *diff_state_add_chunk(struct diff_state *state, bool solved,
204 					struct diff_atom *left_start, unsigned int left_count,
205 					struct diff_atom *right_start, unsigned int right_count);
206 
207 /* Signature of a utility function to divide both source files into diff atoms.
208  * It is possible that a (future) algorithm requires both source files to decide on atom split points, hence this gets
209  * both left and right to atomize at the same time.
210  * An example is diff_atomize_text_by_line() in diff_atomize_text.c.
211  *
212  * func_data: context pointer (free to be used by implementation).
213  * left: struct diff_data with left->data and left->len already set up, and left->atoms to be created.
214  * right: struct diff_data with right->data and right->len already set up, and right->atoms to be created.
215  */
216 typedef enum diff_rc (*diff_atomize_func_t)(void *func_data, struct diff_data *left, struct diff_data *right);
217 
218 extern enum diff_rc diff_atomize_text_by_line(void *func_data, struct diff_data *left, struct diff_data *right);
219 
220 struct diff_algo_config;
221 typedef enum diff_rc (*diff_algo_impl_t)(const struct diff_algo_config *algo_config, struct diff_state *state);
222 
223 /* Form a result with all left-side removed and all right-side added, i.e. no actual diff algorithm involved. */
224 enum diff_rc diff_algo_none(const struct diff_algo_config *algo_config, struct diff_state *state);
225 
226 /* Myers Diff tracing from the start all the way through to the end, requiring quadratic amounts of memory. This can
227  * fail if the required space surpasses algo_config->permitted_state_size. */
228 extern enum diff_rc diff_algo_myers(const struct diff_algo_config *algo_config, struct diff_state *state);
229 
230 /* Myers "Divide et Impera": tracing forwards from the start and backwards from the end to find a midpoint that divides
231  * the problem into smaller chunks. Requires only linear amounts of memory. */
232 extern enum diff_rc diff_algo_myers_divide(const struct diff_algo_config *algo_config, struct diff_state *state);
233 
234 /* Patience Diff algorithm, which divides a larger diff into smaller chunks. For very specific scenarios, it may lead to
235  * a complete diff result by itself, but needs a fallback algo to solve chunks that don't have common-unique atoms. */
236 extern enum diff_rc diff_algo_patience(const struct diff_algo_config *algo_config, struct diff_state *state);
237 
238 /* Diff algorithms to use, possibly nested. For example:
239  *
240  * struct diff_algo_config myers, patience, myers_divide;
241  *
242  * myers = (struct diff_algo_config){
243  *         .impl = diff_algo_myers,
244  *         .permitted_state_size = 32 * 1024 * 1024,
245  *         .fallback_algo = &patience, // when too large, do diff_algo_patience
246  * };
247  *
248  * patience = (struct diff_algo_config){
249  *         .impl = diff_algo_patience,
250  *         .inner_algo = &patience,        // After subdivision, do Patience again.
251  *         .fallback_algo = &myers_divide, // If subdivision failed, do Myers Divide et Impera.
252  * };
253  *
254  * myers_divide = (struct diff_algo_config){
255  *         .impl = diff_algo_myers_divide,
256  *         .inner_algo = &myers,           // When division succeeded, start from the top.
257  *                                         // (fallback_algo = NULL implies diff_algo_none).
258  * };
259  *
260  * struct diff_config config = {
261  *         .algo = &myers,
262  *         ...
263  * };
264  * diff_main(&config, ...);
265  */
266 struct diff_algo_config {
267 	diff_algo_impl_t impl;
268 
269 	/* Fail this algo if it would use more than this amount of memory, and instead use fallback_algo
270 	 * (diff_algo_myers). permitted_state_size == 0 means no limitation. */
271 	size_t permitted_state_size;
272 
273 	/* For algorithms that divide into smaller chunks, use this algorithm to solve the divided chunks. */
274 	const struct diff_algo_config *inner_algo;
275 
276 	/* If the algorithm fails (e.g. diff_algo_myers_if_small needs too large state, or diff_algo_patience can't find
277 	 * any common-unique atoms), then use this algorithm instead. */
278 	const struct diff_algo_config *fallback_algo;
279 };
280 
281 struct diff_config {
282 	diff_atomize_func_t atomize_func;
283 	void *atomize_func_data;
284 
285 	const struct diff_algo_config *algo;
286 
287 	/* How deep to step into subdivisions of a source file, a paranoia / safety measure to guard against infinite
288 	 * loops through diff algorithms. When the maximum recursion is reached, employ diff_algo_none (i.e. remove all
289 	 * left atoms and add all right atoms). */
290 	unsigned int max_recursion_depth;
291 };
292 
293 struct diff_result *diff_main(const struct diff_config *config,
294 			      const uint8_t *left_data, size_t left_len,
295 			      const uint8_t *right_data, size_t right_len);
296 void diff_result_free(struct diff_result *result);