1 /*
2  * Copyright (c) 2018 Stefan Sperling <stsp@openbsd.org>
3  *
4  * Permission to use, copy, modify, and distribute this software for any
5  * purpose with or without fee is hereby granted, provided that the above
6  * copyright notice and this permission notice appear in all copies.
7  *
8  * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
9  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
10  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
11  * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
12  * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
13  * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
14  * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
15  */
16 
17 enum got_diff_algorithm {
18 	GOT_DIFF_ALGORITHM_MYERS,
19 	GOT_DIFF_ALGORITHM_PATIENCE,
20 };
21 
22 /*
23  * Compute the differences between two blobs and write unified diff text
24  * to the provided output file. Two open temporary files must be provided
25  * for internal use; these files can be obtained from got_opentemp() and
26  * must be closed by the caller.
27  * If one of the blobs being diffed does not exist, all corresponding
28  * blob object arguments should be set to NULL.
29  * Two const char * diff header labels may be provided which will be used
30  * to identify each blob in the diff output.
31  * If a label is NULL, use the blob's SHA1 checksum instead.
32  * The number of context lines to show in the diff must be specified as well.
33  * Whitespace differences may optionally be ignored.
34  * If not NULL, the two initial output arguments will be populated with an
35  * array of line offsets for, and the number of lines in, the unidiff text.
36  */
37 const struct got_error *got_diff_blob(off_t **, size_t *,
38     struct got_blob_object *, struct got_blob_object *, FILE *, FILE *,
39     const char *, const char *, enum got_diff_algorithm, int, int, int,
40     FILE *);
41 
42 /*
43  * Compute the differences between a blob and a file and write unified diff
44  * text to the provided output file. The blob object, its content, and its
45  * size must be provided. The file's size must be provided, as well as a
46  * const char * diff header label which identifies the file.
47  * An optional const char * diff header label for the blob may be provided, too.
48  * The number of context lines to show in the diff must be specified as well.
49  * Whitespace differences may optionally be ignored.
50  */
51 const struct got_error *got_diff_blob_file(struct got_blob_object *, FILE *,
52     off_t, const char *, FILE *, int, size_t, const char *,
53     enum got_diff_algorithm, int, int, int, FILE *);
54 
55 /*
56  * A callback function invoked to handle the differences between two blobs
57  * when diffing trees with got_diff_tree(). This callback receives two blobs,
58  * their respective IDs, and two corresponding paths within the diffed trees.
59  * The first blob contains content from the old side of the diff, and
60  * the second blob contains content on the new side of the diff.
61  * Two open temporary files must be provided for internal use; these files
62  * can be obtained from got_opentemp() and must be closed by the caller.
63  * The blob object argument for either blob may be NULL to indicate
64  * that no content is present on its respective side of the diff.
65  * File modes from relevant tree objects which contain the blobs may
66  * also be passed. These will be zero if not available.
67  */
68 typedef const struct got_error *(*got_diff_blob_cb)(void *,
69     struct got_blob_object *, struct got_blob_object *, FILE *, FILE *,
70     struct got_object_id *, struct got_object_id *,
71     const char *, const char *, mode_t, mode_t, struct got_repository *);
72 
73 /*
74  * A pre-defined implementation of got_diff_blob_cb() which appends unidiff
75  * output to a file. The caller must allocate and fill in the argument
76  * structure.
77  */
78 struct got_diff_blob_output_unidiff_arg {
79 	FILE *outfile;		/* Unidiff text will be written here. */
80 	int diff_context;	/* Sets the number of context lines. */
81 	int ignore_whitespace;	/* Ignore whitespace differences. */
82 	int force_text_diff;	/* Assume text even if binary data detected. */
83 	enum got_diff_algorithm diff_algo; /* Diffing algorithm to use. */
84 
85 	/*
86 	 * The number of lines contained in produced unidiff text output,
87 	 * and an array of byte offsets to each line. May be initialized to
88 	 * zero and NULL to ignore line offsets. If not NULL, then the line
89 	 * offsets array will be populated. Optionally, the array can be
90 	 * pre-populated with line offsets, with nlines > 0 indicating
91 	 * the length of the pre-populated array. This is useful if the
92 	 * output file already contains some lines of text.
93 	 * The array will be grown as needed to accomodate additional line
94 	 * offsets, and the last offset found in a pre-populated array will
95 	 * be added to all subsequent offsets.
96 	 */
97 	size_t nlines;
98 	off_t *line_offsets;	/* Dispose of with free(3) when done. */
99 };
100 const struct got_error *got_diff_blob_output_unidiff(void *,
101     struct got_blob_object *, struct got_blob_object *, FILE *, FILE *,
102     struct got_object_id *, struct got_object_id *,
103     const char *, const char *, mode_t, mode_t, struct got_repository *);
104 
105 /*
106  * Compute the differences between two trees and invoke the provided
107  * got_diff_blob_cb() callback when content differs.
108  * Diffing of blob content can be suppressed by passing zero for the
109  * 'diff_content' parameter. The callback will then only receive blob
110  * object IDs and diff labels, but NULL pointers instead of blob objects.
111  * If 'diff_content' is set, two open temporary FILEs and two open
112  * temporary file descriptors must be provided for internal use; these
113  * files can be obtained from got_opentemp() and got_opentempfd(),
114  * and must be closed by the caller. Otherwise the files can be NULL.
115  * The set of arguments relating to either tree may be NULL to indicate
116  * that no content is present on its respective side of the diff.
117  */
118 const struct got_error *got_diff_tree(struct got_tree_object *,
119     struct got_tree_object *, FILE *, FILE *, int, int,
120     const char *, const char *,
121     struct got_repository *, got_diff_blob_cb cb, void *cb_arg, int);
122 
123 /*
124  * A pre-defined implementation of got_diff_blob_cb() which collects a list
125  * of file paths that differ between two trees.
126  * The caller must allocate and initialize a got_pathlist_head * argument.
127  * Data pointers of entries added to the path list will point to a struct
128  * got_diff_changed_path object.
129  * The caller is expected to free both the path and data pointers of all
130  * entries on the path list.
131  */
132 struct got_diff_changed_path {
133 	/*
134 	 * The modification status of this path. It can be GOT_STATUS_ADD,
135 	 * GOT_STATUS_DELETE, GOT_STATUS_MODIFY, or GOT_STATUS_MODE_CHANGE.
136 	 */
137 	int status;
138 };
139 const struct got_error *got_diff_tree_collect_changed_paths(void *,
140     struct got_blob_object *, struct got_blob_object *, FILE *, FILE *,
141     struct got_object_id *, struct got_object_id *,
142     const char *, const char *, mode_t, mode_t, struct got_repository *);
143 
144 /*
145  * Diff two objects, assuming both objects are blobs. Two const char * diff
146  * header labels may be provided which will be used to identify each blob in
147  * the diff output. If a label is NULL, use the blob's SHA1 checksum instead.
148  * Two open temporary files and two temporary file descriptors must be
149  * provided for internal use; these files can be obtained from
150  * got_opentemp() and got_opentempfd(), and must be closed by the caller.
151  * The set of arguments relating to either blob may be NULL/-1 to indicate
152  * that no content is present on its respective side of the diff.
153  * The number of context lines to show in the diff must be specified as well.
154  * Write unified diff text to the provided output FILE.
155  * If not NULL, the two initial output arguments will be populated with an
156  * array of line offsets for, and the number of lines in, the unidiff text.
157  */
158 const struct got_error *got_diff_objects_as_blobs(off_t **, size_t *,
159     FILE *, FILE *, int, int, struct got_object_id *, struct got_object_id *,
160     const char *, const char *, enum got_diff_algorithm, int, int, int,
161     struct got_repository *, FILE *);
162 
163 struct got_pathlist_head;
164 
165 /*
166  * Diff two objects, assuming both objects are trees. Two const char * diff
167  * header labels may be provided which will be used to identify each blob in
168  * the trees. If a label is NULL, use the blob's SHA1 checksum instead.
169  * The number of context lines to show in diffs must be specified.
170  * Two open temporary files and two temporary file descriptors must be
171  * provided for internal use; these files can be obtained from
172  * got_opentemp() and got_opentempfd(), and must be closed by the caller.
173  * If 'diff_content' is not set, the files may be NULL / -1.
174  * The set of arguments relating to either tree may be NULL to indicate
175  * that no content is present on its respective side of the diff.
176  * Write unified diff text to the provided output FILE.
177  * If not NULL, the two initial output arguments will be populated with an
178  * array of line offsets for, and the number of lines in, the unidiff text.
179  */
180 const struct got_error *got_diff_objects_as_trees(off_t **, size_t *,
181     FILE *, FILE *, int, int, struct got_object_id *, struct got_object_id *,
182     struct got_pathlist_head *, const char *, const char *,
183     enum got_diff_algorithm, int, int, int, struct got_repository *, FILE *);
184 
185 /*
186  * Diff two objects, assuming both objects are commits.
187  * The number of context lines to show in diffs must be specified.
188  * Two open temporary files and two temporary file descriptors must be
189  * provided for internal use; these files can be obtained from
190  * got_opentemp() and got_opentempfd(), and must be closed by the caller.
191  * The set of arguments relating to either commit may be NULL to indicate
192  * that no content is present on its respective side of the diff.
193  * Write unified diff text to the provided output FILE.
194  * If not NULL, the two initial output arguments will be populated with an
195  * array of line offsets for, and the number of lines in, the unidiff text.
196  */
197 const struct got_error *got_diff_objects_as_commits(off_t **, size_t *,
198     FILE *, FILE *, int, int, struct got_object_id *, struct got_object_id *,
199     struct got_pathlist_head *, enum got_diff_algorithm, int, int, int,
200     struct got_repository *, FILE *);
201 
202 #define GOT_DIFF_MAX_CONTEXT	64