1 /* Produce a unidiff output from a diff_result. */
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 <errno.h>
19 #include <inttypes.h>
20 #include <stdbool.h>
21 #include <stdio.h>
22 #include <stdlib.h>
23 
24 #include <diff/arraylist.h>
25 #include <diff/diff_main.h>
26 #include <diff/diff_output.h>
27 
28 #include "debug.h"
29 
30 enum chunk_type {
31 	CHUNK_EMPTY,
32 	CHUNK_PLUS,
33 	CHUNK_MINUS,
34 	CHUNK_SAME,
35 	CHUNK_WEIRD,
36 };
37 
38 static inline enum chunk_type
39 chunk_type(const struct diff_chunk *chunk)
40 {
41 	if (!chunk->left_count && !chunk->right_count)
42 		return CHUNK_EMPTY;
43 	if (!chunk->solved)
44 		return CHUNK_WEIRD;
45 	if (!chunk->right_count)
46 		return CHUNK_MINUS;
47 	if (!chunk->left_count)
48 		return CHUNK_PLUS;
49 	if (chunk->left_count != chunk->right_count)
50 		return CHUNK_WEIRD;
51 	return CHUNK_SAME;
52 }
53 
54 struct chunk_context {
55 	struct diff_range chunk;
56 	struct diff_range left, right;
57 };
58 
59 static bool
60 chunk_context_empty(const struct chunk_context *cc)
61 {
62 	return diff_range_empty(&cc->chunk);
63 }
64 
65 static void
66 chunk_context_get(struct chunk_context *cc, const struct diff_result *r,
67 		  int chunk_idx, int context_lines)
68 {
69 	const struct diff_chunk *c = &r->chunks.head[chunk_idx];
70 	int left_start = diff_atom_root_idx(&r->left, c->left_start);
71 	int left_end = MIN(r->left.atoms.len,
72 			   left_start + c->left_count + context_lines);
73 	int right_start = diff_atom_root_idx(&r->right, c->right_start);
74 	int right_end = MIN(r->right.atoms.len,
75 			    right_start + c->right_count + context_lines);
76 
77 	left_start = MAX(0, left_start - context_lines);
78 	right_start = MAX(0, right_start - context_lines);
79 
80 	*cc = (struct chunk_context){
81 		.chunk = {
82 			.start = chunk_idx,
83 			.end = chunk_idx + 1,
84 		},
85 		.left = {
86 			.start = left_start,
87 			.end = left_end,
88 		},
89 		.right = {
90 			.start = right_start,
91 			.end = right_end,
92 		},
93 	};
94 }
95 
96 static bool
97 chunk_contexts_touch(const struct chunk_context *cc,
98 		     const struct chunk_context *other)
99 {
100 	return diff_ranges_touch(&cc->chunk, &other->chunk)
101 		|| diff_ranges_touch(&cc->left, &other->left)
102 		|| diff_ranges_touch(&cc->right, &other->right);
103 }
104 
105 static void
106 chunk_contexts_merge(struct chunk_context *cc,
107 		     const struct chunk_context *other)
108 {
109 	diff_ranges_merge(&cc->chunk, &other->chunk);
110 	diff_ranges_merge(&cc->left, &other->left);
111 	diff_ranges_merge(&cc->right, &other->right);
112 }
113 
114 static void
115 diff_output_unidiff_chunk(FILE *dest, bool *header_printed,
116 			  const struct diff_input_info *info,
117 			  const struct diff_result *result,
118 			  const struct chunk_context *cc)
119 {
120 	if (diff_range_empty(&cc->left) && diff_range_empty(&cc->right))
121 		return;
122 
123 	if (!(*header_printed)) {
124 		fprintf(dest, "--- %s\n+++ %s\n",
125 			info->left_path ? : "a",
126 			info->right_path ? : "b");
127 		*header_printed = true;
128 	}
129 
130 	fprintf(dest, "@@ -%d,%d +%d,%d @@\n",
131 		cc->left.start + 1, cc->left.end - cc->left.start,
132 		cc->right.start + 1, cc->right.end - cc->right.start);
133 
134 	/* Got the absolute line numbers where to start printing, and the index
135 	 * of the interesting (non-context) chunk.
136 	 * To print context lines above the interesting chunk, nipping on the
137 	 * previous chunk index may be necessary.
138 	 * It is guaranteed to be only context lines where left == right, so it
139 	 * suffices to look on the left. */
140 	const struct diff_chunk *first_chunk;
141 	int chunk_start_line;
142 	first_chunk = &result->chunks.head[cc->chunk.start];
143 	chunk_start_line = diff_atom_root_idx(&result->left,
144 					      first_chunk->left_start);
145 	if (cc->left.start < chunk_start_line)
146 		diff_output_lines(dest, " ",
147 				  &result->left.atoms.head[cc->left.start],
148 				  chunk_start_line - cc->left.start);
149 
150 	/* Now write out all the joined chunks and contexts between them */
151 	int c_idx;
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 && c->right_count)
156 			diff_output_lines(dest,
157 					  c->solved ? " " : "?",
158 					  c->left_start, c->left_count);
159 		else if (c->left_count && !c->right_count)
160 			diff_output_lines(dest,
161 					  c->solved ? "-" : "?",
162 					  c->left_start, c->left_count);
163 		else if (c->right_count && !c->left_count)
164 			diff_output_lines(dest,
165 					  c->solved ? "+" : "?",
166 					  c->right_start, c->right_count);
167 	}
168 
169 	/* Trailing context? */
170 	const struct diff_chunk *last_chunk;
171 	int chunk_end_line;
172 	last_chunk = &result->chunks.head[cc->chunk.end - 1];
173 	chunk_end_line = diff_atom_root_idx(&result->left,
174 					    last_chunk->left_start
175 					    + last_chunk->left_count);
176 	if (cc->left.end > chunk_end_line)
177 		diff_output_lines(dest, " ",
178 				  &result->left.atoms.head[chunk_end_line],
179 				  cc->left.end - chunk_end_line);
180 }
181 
182 int
183 diff_output_unidiff(FILE *dest, const struct diff_input_info *info,
184 		    const struct diff_result *result,
185 		    unsigned int context_lines)
186 {
187 	if (!result)
188 		return EINVAL;
189 	if (result->rc != DIFF_RC_OK)
190 		return result->rc;
191 
192 	struct chunk_context cc = {};
193 	bool header_printed = false;
194 
195 	int i;
196 	for (i = 0; i < result->chunks.len; i++) {
197 		struct diff_chunk *c = &result->chunks.head[i];
198 		enum chunk_type t = chunk_type(c);
199 		struct chunk_context next;
200 
201 		if (t != CHUNK_MINUS && t != CHUNK_PLUS)
202 			continue;
203 
204 		if (chunk_context_empty(&cc)) {
205 			/* These are the first lines being printed.
206 			 * Note down the start point, any number of subsequent
207 			 * chunks may be joined up to this unidiff chunk by
208 			 * context lines or by being directly adjacent. */
209 			chunk_context_get(&cc, result, i, context_lines);
210 			debug("new chunk to be printed:"
211 			      " chunk %d-%d left %d-%d right %d-%d\n",
212 			      cc.chunk.start, cc.chunk.end,
213 			      cc.left.start, cc.left.end,
214 			      cc.right.start, cc.right.end);
215 			continue;
216 		}
217 
218 		/* There already is a previous chunk noted down for being
219 		 * printed. Does it join up with this one? */
220 		chunk_context_get(&next, result, i, context_lines);
221 		debug("new chunk to be printed:"
222 		      " chunk %d-%d left %d-%d right %d-%d\n",
223 		      next.chunk.start, next.chunk.end,
224 		      next.left.start, next.left.end,
225 		      next.right.start, next.right.end);
226 
227 		if (chunk_contexts_touch(&cc, &next)) {
228 			/* This next context touches or overlaps the previous
229 			 * one, join. */
230 			chunk_contexts_merge(&cc, &next);
231 			debug("new chunk to be printed touches previous chunk,"
232 			      " now: left %d-%d right %d-%d\n",
233 			      cc.left.start, cc.left.end,
234 			      cc.right.start, cc.right.end);
235 			continue;
236 		}
237 
238 		/* No touching, so the previous context is complete with a gap
239 		 * between it and this next one. Print the previous one and
240 		 * start fresh here. */
241 		debug("new chunk to be printed does not touch previous chunk;"
242 		      " print left %d-%d right %d-%d\n",
243 		      cc.left.start, cc.left.end, cc.right.start, cc.right.end);
244 		diff_output_unidiff_chunk(dest, &header_printed, info, result,
245 					  &cc);
246 		cc = next;
247 		debug("new unprinted chunk is left %d-%d right %d-%d\n",
248 		      cc.left.start, cc.left.end, cc.right.start, cc.right.end);
249 	}
250 
251 	if (!chunk_context_empty(&cc))
252 		diff_output_unidiff_chunk(dest, &header_printed, info, result,
253 					  &cc);
254 	return DIFF_RC_OK;
255 }