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 <diff/diff_output.h>
19 
20 #include "debug.h"
21 
22 enum chunk_type {
23 	CHUNK_EMPTY,
24 	CHUNK_PLUS,
25 	CHUNK_MINUS,
26 	CHUNK_SAME,
27 	CHUNK_WEIRD,
28 };
29 
30 static inline enum chunk_type
31 chunk_type(const struct diff_chunk *chunk)
32 {
33 	if (!chunk->left_count && !chunk->right_count)
34 		return CHUNK_EMPTY;
35 	if (!chunk->solved)
36 		return CHUNK_WEIRD;
37 	if (!chunk->right_count)
38 		return CHUNK_MINUS;
39 	if (!chunk->left_count)
40 		return CHUNK_PLUS;
41 	if (chunk->left_count != chunk->right_count)
42 		return CHUNK_WEIRD;
43 	return CHUNK_SAME;
44 }
45 
46 struct chunk_context {
47 	struct range chunk;
48 	struct range left, right;
49 };
50 
51 static bool
52 chunk_context_empty(const struct chunk_context *cc)
53 {
54 	return range_empty(&cc->chunk);
55 }
56 
57 static void
58 chunk_context_get(struct chunk_context *cc, const struct diff_result *r, int chunk_idx,
59 		  int context_lines)
60 {
61 	const struct diff_chunk *c = &r->chunks.head[chunk_idx];
62 	int left_start = diff_atom_root_idx(&r->left, c->left_start);
63 	int right_start = diff_atom_root_idx(&r->right, c->right_start);
64 
65 	*cc = (struct chunk_context){
66 		.chunk = {
67 			.start = chunk_idx,
68 			.end = chunk_idx + 1,
69 		},
70 		.left = {
71 			.start = MAX(0, left_start - context_lines),
72 			.end = MIN(r->left.atoms.len, left_start + c->left_count + context_lines),
73 		},
74 		.right = {
75 			.start = MAX(0, right_start - context_lines),
76 			.end = MIN(r->right.atoms.len, right_start + c->right_count + context_lines),
77 		},
78 	};
79 }
80 
81 static bool
82 chunk_contexts_touch(const struct chunk_context *cc, const struct chunk_context *other)
83 {
84 	return ranges_touch(&cc->chunk, &other->chunk)
85 		|| ranges_touch(&cc->left, &other->left)
86 		|| ranges_touch(&cc->right, &other->right);
87 }
88 
89 static void
90 chunk_contexts_merge(struct chunk_context *cc, const struct chunk_context *other)
91 {
92 	ranges_merge(&cc->chunk, &other->chunk);
93 	ranges_merge(&cc->left, &other->left);
94 	ranges_merge(&cc->right, &other->right);
95 }
96 
97 static void
98 diff_output_unidiff_chunk(FILE *dest, bool *header_printed, const struct diff_input_info *info,
99 			  const struct diff_result *result, const struct chunk_context *cc)
100 {
101 	if (range_empty(&cc->left) && range_empty(&cc->right))
102 		return;
103 
104 	if (!(*header_printed)) {
105 		fprintf(dest, "--- %s\n+++ %s\n",
106 			info->left_path ? : "a",
107 			info->right_path ? : "b");
108 		*header_printed = true;
109 	}
110 
111 	fprintf(dest, "@@ -%d,%d +%d,%d @@\n",
112 		cc->left.start + 1, cc->left.end - cc->left.start,
113 		cc->right.start + 1, cc->right.end - cc->right.start);
114 
115 	/* Got the absolute line numbers where to start printing, and the index of the interesting (non-context) chunk.
116 	 * To print context lines above the interesting chunk, nipping on the previous chunk index may be necessary.
117 	 * It is guaranteed to be only context lines where left == right, so it suffices to look on the left. */
118 	const struct diff_chunk *first_chunk = &result->chunks.head[cc->chunk.start];
119 	int chunk_start_line = diff_atom_root_idx(&result->left, first_chunk->left_start);
120 	if (cc->left.start < chunk_start_line)
121 		diff_output_lines(dest, " ", &result->left.atoms.head[cc->left.start],
122 				  chunk_start_line - cc->left.start);
123 
124 	/* Now write out all the joined chunks and contexts between them */
125 	int c_idx;
126 	for (c_idx = cc->chunk.start; c_idx < cc->chunk.end; c_idx++) {
127 		const struct diff_chunk *c = &result->chunks.head[c_idx];
128 
129 		if (c->left_count && c->right_count)
130 			diff_output_lines(dest, c->solved ? " " : "?", c->left_start, c->left_count);
131 		else if (c->left_count && !c->right_count)
132 			diff_output_lines(dest, c->solved ? "-" : "?", c->left_start, c->left_count);
133 		else if (c->right_count && !c->left_count)
134 			diff_output_lines(dest, c->solved ? "+" : "?", c->right_start, c->right_count);
135 	}
136 
137 	/* Trailing context? */
138 	const struct diff_chunk *last_chunk = &result->chunks.head[cc->chunk.end - 1];
139 	int chunk_end_line = diff_atom_root_idx(&result->left, last_chunk->left_start + last_chunk->left_count);
140 	if (cc->left.end > chunk_end_line)
141 		diff_output_lines(dest, " ", &result->left.atoms.head[chunk_end_line],
142 				  cc->left.end - chunk_end_line);
143 }
144 
145 enum diff_rc
146 diff_output_unidiff(FILE *dest, const struct diff_input_info *info,
147 		    const struct diff_result *result, unsigned int context_lines)
148 {
149 	if (!result)
150 		return DIFF_RC_EINVAL;
151 	if (result->rc != DIFF_RC_OK)
152 		return result->rc;
153 
154 	struct chunk_context cc = {};
155 	bool header_printed = false;
156 
157 	int i;
158 	for (i = 0; i < result->chunks.len; i++) {
159 		struct diff_chunk *c = &result->chunks.head[i];
160 		enum chunk_type t = chunk_type(c);
161 		struct chunk_context next;
162 
163 		if (t != CHUNK_MINUS && t != CHUNK_PLUS)
164 			continue;
165 
166 		if (chunk_context_empty(&cc)) {
167 			/* These are the first lines being printed.
168 			 * Note down the start point, any number of subsequent chunks may be joined up to this
169 			 * unidiff chunk by context lines or by being directly adjacent. */
170 			chunk_context_get(&cc, result, i, context_lines);
171 			debug("new chunk to be printed: chunk %d-%d left %d-%d right %d-%d\n",
172 			      cc.chunk.start, cc.chunk.end,
173 			      cc.left.start, cc.left.end,
174 			      cc.right.start, cc.right.end);
175 			continue;
176 		}
177 
178 		/* There already is a previous chunk noted down for being printed.
179 		 * Does it join up with this one? */
180 		chunk_context_get(&next, result, i, context_lines);
181 		debug("new chunk to be printed: chunk %d-%d left %d-%d right %d-%d\n",
182 		      next.chunk.start, next.chunk.end,
183 		      next.left.start, next.left.end,
184 		      next.right.start, next.right.end);
185 
186 		if (chunk_contexts_touch(&cc, &next)) {
187 			/* This next context touches or overlaps the previous one, join. */
188 			chunk_contexts_merge(&cc, &next);
189 			debug("new chunk to be printed touches previous chunk, now: left %d-%d right %d-%d\n",
190 			      cc.left.start, cc.left.end,
191 			      cc.right.start, cc.right.end);
192 			continue;
193 		}
194 
195 		/* No touching, so the previous context is complete with a gap between it and
196 		 * this next one. Print the previous one and start fresh here. */
197 		debug("new chunk to be printed does not touch previous chunk; print left %d-%d right %d-%d\n",
198 		      cc.left.start, cc.left.end, cc.right.start, cc.right.end);
199 		diff_output_unidiff_chunk(dest, &header_printed, info, result, &cc);
200 		cc = next;
201 		debug("new unprinted chunk is left %d-%d right %d-%d\n",
202 		      cc.left.start, cc.left.end, cc.right.start, cc.right.end);
203 	}
204 
205 	if (!chunk_context_empty(&cc))
206 		diff_output_unidiff_chunk(dest, &header_printed, info, result, &cc);
207 	return DIFF_RC_OK;
208 }