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 struct got_diffstat_cb_arg;
48 
49 /*
50  * Compute the differences between two blobs and write unified diff text
51  * to the provided output file. Two open temporary files must be provided
52  * for internal use; these files can be obtained from got_opentemp() and
53  * must be closed by the caller.
54  * If one of the blobs being diffed does not exist, all corresponding
55  * blob object arguments should be set to NULL.
56  * Two const char * diff header labels may be provided which will be used
57  * to identify each blob in the diff output.
58  * If a label is NULL, use the blob's SHA1 checksum instead.
59  * The number of context lines to show in the diff must be specified as well.
60  * Whitespace differences may optionally be ignored.
61  * If not NULL, the two initial output arguments will be populated with an
62  * array of line offsets for, and the number of lines in, the unidiff text.
63  */
64 const struct got_error *got_diff_blob(struct got_diff_line **, size_t *,
65     struct got_blob_object *, struct got_blob_object *, FILE *, FILE *,
66     const char *, const char *, enum got_diff_algorithm, int, int, int,
67     struct got_diffstat_cb_arg *, FILE *);
68 
69 /*
70  * Compute the differences between a blob and a file and write unified diff
71  * text to the provided output file. The blob object, its content, and its
72  * size must be provided. The file's size must be provided, as well as a
73  * const char * diff header label which identifies the file.
74  * An optional const char * diff header label for the blob may be provided, too.
75  * The number of context lines to show in the diff must be specified as well.
76  * Whitespace differences may optionally be ignored.
77  */
78 const struct got_error *got_diff_blob_file(struct got_blob_object *, FILE *,
79     off_t, const char *, FILE *, int, struct stat *, const char *,
80     enum got_diff_algorithm, int, int, int, struct got_diffstat_cb_arg *,
81     FILE *);
82 
83 /*
84  * A callback function invoked to handle the differences between two blobs
85  * when diffing trees with got_diff_tree(). This callback receives two blobs,
86  * their respective IDs, and two corresponding paths within the diffed trees.
87  * The first blob contains content from the old side of the diff, and
88  * the second blob contains content on the new side of the diff.
89  * Two open temporary files must be provided for internal use; these files
90  * can be obtained from got_opentemp() and must be closed by the caller.
91  * The blob object argument for either blob may be NULL to indicate
92  * that no content is present on its respective side of the diff.
93  * File modes from relevant tree objects which contain the blobs may
94  * also be passed. These will be zero if not available.
95  */
96 typedef const struct got_error *(*got_diff_blob_cb)(void *,
97     struct got_blob_object *, struct got_blob_object *, FILE *, FILE *,
98     struct got_object_id *, struct got_object_id *,
99     const char *, const char *, mode_t, mode_t, struct got_repository *);
100 
101 /*
102  * A pre-defined implementation of got_diff_blob_cb() which appends unidiff
103  * output to a file. The caller must allocate and fill in the argument
104  * structure.
105  */
106 struct got_diff_blob_output_unidiff_arg {
107 	FILE *outfile;		/* Unidiff text will be written here. */
108 	int diff_context;	/* Sets the number of context lines. */
109 	int ignore_whitespace;	/* Ignore whitespace differences. */
110 	int force_text_diff;	/* Assume text even if binary data detected. */
111 	struct got_diffstat_cb_arg *diffstat; /* Compute diffstat of changes */
112 	enum got_diff_algorithm diff_algo; /* Diffing algorithm to use. */
113 
114 	/*
115 	 * The number of lines contained in produced unidiff text output,
116 	 * and an array of got_diff_lines with byte offset and line type to
117 	 * each line. May be initialized to zero and NULL to ignore line
118 	 * metadata. If not NULL, then the array of line offsets and types will
119 	 * be populated. Optionally, the array can be pre-populated with line
120 	 * offsets and types, with nlines > 0 indicating the length of the
121 	 * pre-populated array.  This is useful if the output file already
122 	 * contains some lines of text.  The array will be grown as needed to
123 	 * accomodate additional offsets and types, and the last offset found
124 	 * in a pre-populated array will be added to all subsequent offsets.
125 	 */
126 	size_t nlines;
127 	struct got_diff_line *lines; /* Dispose of with free(3) when done. */
128 };
129 const struct got_error *got_diff_blob_output_unidiff(void *,
130     struct got_blob_object *, struct got_blob_object *, FILE *, FILE *,
131     struct got_object_id *, struct got_object_id *,
132     const char *, const char *, mode_t, mode_t, struct got_repository *);
133 
134 /*
135  * Compute the differences between two trees and invoke the provided
136  * got_diff_blob_cb() callback when content differs.
137  * Diffing of blob content can be suppressed by passing zero for the
138  * 'diff_content' parameter. The callback will then only receive blob
139  * object IDs and diff labels, but NULL pointers instead of blob objects.
140  * If 'diff_content' is set, two open temporary FILEs and two open
141  * temporary file descriptors must be provided for internal use; these
142  * files can be obtained from got_opentemp() and got_opentempfd(),
143  * and must be closed by the caller. Otherwise the files can be NULL.
144  * The set of arguments relating to either tree may be NULL to indicate
145  * that no content is present on its respective side of the diff.
146  */
147 const struct got_error *got_diff_tree(struct got_tree_object *,
148     struct got_tree_object *, FILE *, FILE *, int, int,
149     const char *, const char *,
150     struct got_repository *, got_diff_blob_cb cb, void *cb_arg, int);
151 
152 /*
153  * Pre-defined implementations of got_diff_blob_cb(): the first of which
154  * collects a list of file paths that differ between two trees; the second
155  * also computes a diffstat of added/removed lines for each collected path
156  * and requires passing an initialized got_diffstat_cb_arg argument.
157  * The caller must allocate and initialize a got_pathlist_head * argument.
158  * Data pointers of entries added to the path list will point to a struct
159  * got_diff_changed_path object.
160  * The caller is expected to free both the path and data pointers of all
161  * entries on the path list.
162  */
163 struct got_diff_changed_path {
164 	uint32_t add;	/* number of lines added */
165 	uint32_t rm;	/* number of lines removed */
166 	/*
167 	 * The modification status of this path. It can be GOT_STATUS_ADD,
168 	 * GOT_STATUS_DELETE, GOT_STATUS_MODIFY, or GOT_STATUS_MODE_CHANGE.
169 	 */
170 	int status;
171 };
172 const struct got_error *got_diff_tree_collect_changed_paths(void *,
173     struct got_blob_object *, struct got_blob_object *, FILE *, FILE *,
174     struct got_object_id *, struct got_object_id *,
175     const char *, const char *, mode_t, mode_t, struct got_repository *);
176 
177 struct got_diffstat_cb_arg {
178 	size_t max_path_len;
179 	uint32_t ins;
180 	uint32_t del;
181 	int add_cols;
182 	int rm_cols;
183 	int nfiles;
184 	struct got_pathlist_head *paths;
185 	int ignore_ws;
186 	int force_text;
187 	enum got_diff_algorithm diff_algo;
188 };
189 const struct got_error *got_diff_tree_compute_diffstat(void *,
190     struct got_blob_object *, struct got_blob_object *, FILE *, FILE *,
191     struct got_object_id *, struct got_object_id *, const char *, const char *,
192     mode_t, mode_t, struct got_repository *);
193 
194 /*
195  * Diff two objects, assuming both objects are blobs. Two const char * diff
196  * header labels may be provided which will be used to identify each blob in
197  * the diff output. If a label is NULL, use the blob's SHA1 checksum instead.
198  * Two open temporary files and two temporary file descriptors must be
199  * provided for internal use; these files can be obtained from
200  * got_opentemp() and got_opentempfd(), and must be closed by the caller.
201  * The set of arguments relating to either blob may be NULL/-1 to indicate
202  * that no content is present on its respective side of the diff.
203  * The number of context lines to show in the diff must be specified as well.
204  * Write unified diff text to the provided output FILE.
205  * If not NULL, the two initial output arguments will be populated with an
206  * array of line offsets for, and the number of lines in, the unidiff text.
207  */
208 const struct got_error *got_diff_objects_as_blobs(struct got_diff_line **,
209     size_t *, FILE *, FILE *, int, int, struct got_object_id *,
210     struct got_object_id *, const char *, const char *, enum got_diff_algorithm,
211     int, int, int, struct got_diffstat_cb_arg *, struct got_repository *,
212     FILE *);
213 
214 struct got_pathlist_head;
215 
216 /*
217  * Diff two objects, assuming both objects are trees. Two const char * diff
218  * header labels may be provided which will be used to identify each blob in
219  * the trees. If a label is NULL, use the blob's SHA1 checksum instead.
220  * The number of context lines to show in diffs must be specified.
221  * Two open temporary files and two temporary file descriptors must be
222  * provided for internal use; these files can be obtained from
223  * got_opentemp() and got_opentempfd(), and must be closed by the caller.
224  * If 'diff_content' is not set, the files may be NULL / -1.
225  * The set of arguments relating to either tree may be NULL to indicate
226  * that no content is present on its respective side of the diff.
227  * Write unified diff text to the provided output FILE.
228  * If not NULL, the two initial output arguments will be populated with an
229  * array of line offsets for, and the number of lines in, the unidiff text.
230  */
231 const struct got_error *got_diff_objects_as_trees(struct got_diff_line **,
232     size_t *, FILE *, FILE *, int, int, struct got_object_id *,
233     struct got_object_id *, struct got_pathlist_head *, const char *,
234     const char *, enum got_diff_algorithm, int, int, int,
235     struct got_diffstat_cb_arg *, struct got_repository *, FILE *);
236 
237 /*
238  * Diff two objects, assuming both objects are commits.
239  * The number of context lines to show in diffs must be specified.
240  * Two open temporary files and two temporary file descriptors must be
241  * provided for internal use; these files can be obtained from
242  * got_opentemp() and got_opentempfd(), and must be closed by the caller.
243  * The set of arguments relating to either commit may be NULL to indicate
244  * that no content is present on its respective side of the diff.
245  * Write unified diff text to the provided output FILE.
246  * If not NULL, the two initial output arguments will be populated with an
247  * array of line offsets for, and the number of lines in, the unidiff text.
248  */
249 const struct got_error *got_diff_objects_as_commits(struct got_diff_line **,
250     size_t *, FILE *, FILE *, int, int, struct got_object_id *,
251     struct got_object_id *, struct got_pathlist_head *, enum got_diff_algorithm,
252     int, int, int, struct got_diffstat_cb_arg *, struct got_repository *,
253     FILE *);
254 
255 #define GOT_DIFF_MAX_CONTEXT	64