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  * List of all line types in a diff (including '{got,tog} log' lines).
24  * XXX GOT_DIFF_LINE_HUNK to GOT_DIFF_LINE_NONE inclusive must map to the
25  * DIFF_LINE_* macro counterparts defined in lib/diff_output.h (i.e., 60-64).
26  */
27 enum got_diff_line_type {
28 	GOT_DIFF_LINE_LOGMSG,
29 	GOT_DIFF_LINE_AUTHOR,
30 	GOT_DIFF_LINE_DATE,
31 	GOT_DIFF_LINE_CHANGES,
32 	GOT_DIFF_LINE_META,
33 	GOT_DIFF_LINE_BLOB_MIN,
34 	GOT_DIFF_LINE_BLOB_PLUS,
35 	GOT_DIFF_LINE_HUNK = 60,
36 	GOT_DIFF_LINE_MINUS,
37 	GOT_DIFF_LINE_PLUS,
38 	GOT_DIFF_LINE_CONTEXT,
39 	GOT_DIFF_LINE_NONE
40 };
41 
42 struct got_diff_line {
43 	off_t	offset;
44 	uint8_t	type;
45 };
46 
47 /*
48  * Compute the differences between two blobs and write unified diff text
49  * to the provided output file. Two open temporary files must be provided
50  * for internal use; these files can be obtained from got_opentemp() and
51  * must be closed by the caller.
52  * If one of the blobs being diffed does not exist, all corresponding
53  * blob object arguments should be set to NULL.
54  * Two const char * diff header labels may be provided which will be used
55  * to identify each blob in the diff output.
56  * If a label is NULL, use the blob's SHA1 checksum instead.
57  * The number of context lines to show in the diff must be specified as well.
58  * Whitespace differences may optionally be ignored.
59  * If not NULL, the two initial output arguments will be populated with an
60  * array of line offsets for, and the number of lines in, the unidiff text.
61  */
62 const struct got_error *got_diff_blob(struct got_diff_line **, size_t *,
63     struct got_blob_object *, struct got_blob_object *, FILE *, FILE *,
64     const char *, const char *, enum got_diff_algorithm, int, int, int,
65     FILE *);
66 
67 /*
68  * Compute the differences between a blob and a file and write unified diff
69  * text to the provided output file. The blob object, its content, and its
70  * size must be provided. The file's size must be provided, as well as a
71  * const char * diff header label which identifies the file.
72  * An optional const char * diff header label for the blob may be provided, too.
73  * The number of context lines to show in the diff must be specified as well.
74  * Whitespace differences may optionally be ignored.
75  */
76 const struct got_error *got_diff_blob_file(struct got_blob_object *, FILE *,
77     off_t, const char *, FILE *, int, struct stat *, const char *,
78     enum got_diff_algorithm, int, int, int, FILE *);
79 
80 /*
81  * A callback function invoked to handle the differences between two blobs
82  * when diffing trees with got_diff_tree(). This callback receives two blobs,
83  * their respective IDs, and two corresponding paths within the diffed trees.
84  * The first blob contains content from the old side of the diff, and
85  * the second blob contains content on the new side of the diff.
86  * Two open temporary files must be provided for internal use; these files
87  * can be obtained from got_opentemp() and must be closed by the caller.
88  * The blob object argument for either blob may be NULL to indicate
89  * that no content is present on its respective side of the diff.
90  * File modes from relevant tree objects which contain the blobs may
91  * also be passed. These will be zero if not available.
92  */
93 typedef const struct got_error *(*got_diff_blob_cb)(void *,
94     struct got_blob_object *, struct got_blob_object *, FILE *, FILE *,
95     struct got_object_id *, struct got_object_id *,
96     const char *, const char *, mode_t, mode_t, struct got_repository *);
97 
98 /*
99  * A pre-defined implementation of got_diff_blob_cb() which appends unidiff
100  * output to a file. The caller must allocate and fill in the argument
101  * structure.
102  */
103 struct got_diff_blob_output_unidiff_arg {
104 	FILE *outfile;		/* Unidiff text will be written here. */
105 	int diff_context;	/* Sets the number of context lines. */
106 	int ignore_whitespace;	/* Ignore whitespace differences. */
107 	int force_text_diff;	/* Assume text even if binary data detected. */
108 	enum got_diff_algorithm diff_algo; /* Diffing algorithm to use. */
109 
110 	/*
111 	 * The number of lines contained in produced unidiff text output,
112 	 * and an array of got_diff_lines with byte offset and line type to
113 	 * each line. May be initialized to zero and NULL to ignore line
114 	 * metadata. If not NULL, then the array of line offsets and types will
115 	 * be populated. Optionally, the array can be pre-populated with line
116 	 * offsets and types, with nlines > 0 indicating the length of the
117 	 * pre-populated array.  This is useful if the output file already
118 	 * contains some lines of text.  The array will be grown as needed to
119 	 * accomodate additional offsets and types, and the last offset found
120 	 * in a pre-populated array will be added to all subsequent offsets.
121 	 */
122 	size_t nlines;
123 	struct got_diff_line *lines; /* Dispose of with free(3) when done. */
124 };
125 const struct got_error *got_diff_blob_output_unidiff(void *,
126     struct got_blob_object *, struct got_blob_object *, FILE *, FILE *,
127     struct got_object_id *, struct got_object_id *,
128     const char *, const char *, mode_t, mode_t, struct got_repository *);
129 
130 /*
131  * Compute the differences between two trees and invoke the provided
132  * got_diff_blob_cb() callback when content differs.
133  * Diffing of blob content can be suppressed by passing zero for the
134  * 'diff_content' parameter. The callback will then only receive blob
135  * object IDs and diff labels, but NULL pointers instead of blob objects.
136  * If 'diff_content' is set, two open temporary FILEs and two open
137  * temporary file descriptors must be provided for internal use; these
138  * files can be obtained from got_opentemp() and got_opentempfd(),
139  * and must be closed by the caller. Otherwise the files can be NULL.
140  * The set of arguments relating to either tree may be NULL to indicate
141  * that no content is present on its respective side of the diff.
142  */
143 const struct got_error *got_diff_tree(struct got_tree_object *,
144     struct got_tree_object *, FILE *, FILE *, int, int,
145     const char *, const char *,
146     struct got_repository *, got_diff_blob_cb cb, void *cb_arg, int);
147 
148 /*
149  * Pre-defined implementations of got_diff_blob_cb(): the first of which
150  * collects a list of file paths that differ between two trees; the second
151  * also computes a diffstat of added/removed lines for each collected path
152  * and requires passing an initialized got_diffstat_cb_arg argument.
153  * The caller must allocate and initialize a got_pathlist_head * argument.
154  * Data pointers of entries added to the path list will point to a struct
155  * got_diff_changed_path object.
156  * The caller is expected to free both the path and data pointers of all
157  * entries on the path list.
158  */
159 struct got_diff_changed_path {
160 	uint32_t add;	/* number of lines added */
161 	uint32_t rm;	/* number of lines removed */
162 	/*
163 	 * The modification status of this path. It can be GOT_STATUS_ADD,
164 	 * GOT_STATUS_DELETE, GOT_STATUS_MODIFY, or GOT_STATUS_MODE_CHANGE.
165 	 */
166 	int status;
167 };
168 const struct got_error *got_diff_tree_collect_changed_paths(void *,
169     struct got_blob_object *, struct got_blob_object *, FILE *, FILE *,
170     struct got_object_id *, struct got_object_id *,
171     const char *, const char *, mode_t, mode_t, struct got_repository *);
172 
173 struct got_diffstat_cb_arg {
174 	size_t max_path_len;
175 	uint32_t ins;
176 	uint32_t del;
177 	int add_cols;
178 	int rm_cols;
179 	int nfiles;
180 	struct got_pathlist_head *paths;
181 	int ignore_ws;
182 	int force_text;
183 	enum got_diff_algorithm diff_algo;
184 };
185 const struct got_error *got_diff_tree_compute_diffstat(void *,
186     struct got_blob_object *, struct got_blob_object *, FILE *, FILE *,
187     struct got_object_id *, struct got_object_id *, const char *, const char *,
188     mode_t, mode_t, struct got_repository *);
189 
190 /*
191  * Diff two objects, assuming both objects are blobs. Two const char * diff
192  * header labels may be provided which will be used to identify each blob in
193  * the diff output. If a label is NULL, use the blob's SHA1 checksum instead.
194  * Two open temporary files and two temporary file descriptors must be
195  * provided for internal use; these files can be obtained from
196  * got_opentemp() and got_opentempfd(), and must be closed by the caller.
197  * The set of arguments relating to either blob may be NULL/-1 to indicate
198  * that no content is present on its respective side of the diff.
199  * The number of context lines to show in the diff must be specified as well.
200  * Write unified diff text to the provided output FILE.
201  * If not NULL, the two initial output arguments will be populated with an
202  * array of line offsets for, and the number of lines in, the unidiff text.
203  */
204 const struct got_error *got_diff_objects_as_blobs(struct got_diff_line **,
205     size_t *, FILE *, FILE *, int, int, struct got_object_id *,
206     struct got_object_id *, const char *, const char *, enum got_diff_algorithm,
207     int, int, int, struct got_repository *, FILE *);
208 
209 struct got_pathlist_head;
210 
211 /*
212  * Diff two objects, assuming both objects are trees. Two const char * diff
213  * header labels may be provided which will be used to identify each blob in
214  * the trees. If a label is NULL, use the blob's SHA1 checksum instead.
215  * The number of context lines to show in diffs must be specified.
216  * Two open temporary files and two temporary file descriptors must be
217  * provided for internal use; these files can be obtained from
218  * got_opentemp() and got_opentempfd(), and must be closed by the caller.
219  * If 'diff_content' is not set, the files may be NULL / -1.
220  * The set of arguments relating to either tree may be NULL to indicate
221  * that no content is present on its respective side of the diff.
222  * Write unified diff text to the provided output FILE.
223  * If not NULL, the two initial output arguments will be populated with an
224  * array of line offsets for, and the number of lines in, the unidiff text.
225  */
226 const struct got_error *got_diff_objects_as_trees(struct got_diff_line **,
227     size_t *, FILE *, FILE *, int, int, struct got_object_id *,
228     struct got_object_id *, struct got_pathlist_head *, const char *,
229     const char *, enum got_diff_algorithm, int, int, int,
230     struct got_repository *, FILE *);
231 
232 /*
233  * Diff two objects, assuming both objects are commits.
234  * The number of context lines to show in diffs must be specified.
235  * Two open temporary files and two temporary file descriptors must be
236  * provided for internal use; these files can be obtained from
237  * got_opentemp() and got_opentempfd(), and must be closed by the caller.
238  * The set of arguments relating to either commit may be NULL to indicate
239  * that no content is present on its respective side of the diff.
240  * Write unified diff text to the provided output FILE.
241  * If not NULL, the two initial output arguments will be populated with an
242  * array of line offsets for, and the number of lines in, the unidiff text.
243  */
244 const struct got_error *got_diff_objects_as_commits(struct got_diff_line **,
245     size_t *, FILE *, FILE *, int, int, struct got_object_id *,
246     struct got_object_id *, struct got_pathlist_head *, enum got_diff_algorithm,
247     int, int, int, struct got_repository *, FILE *);
248 
249 #define GOT_DIFF_MAX_CONTEXT	64