include/got_diff.h

   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 hash 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  * If not NULL, the two initial output arguments will be populated with an
  78  * array of line metadata for, and the number of lines in, the unidiff text.
  79  */
  80 const struct got_error *got_diff_blob_file(struct got_diff_line **, size_t *,
  81     struct got_blob_object *, FILE *, off_t, const char *, FILE *, int,
  82     struct stat *, const char *, enum got_diff_algorithm, int, int, int,
  83     struct got_diffstat_cb_arg *, FILE *);
  84
  85 /*
  86  * A callback function invoked to handle the differences between two blobs
  87  * when diffing trees with got_diff_tree(). This callback receives two blobs,
  88  * their respective IDs, and two corresponding paths within the diffed trees.
  89  * The first blob contains content from the old side of the diff, and
  90  * the second blob contains content on the new side of the diff.
  91  * Two open temporary files must be provided for internal use; these files
  92  * can be obtained from got_opentemp() and must be closed by the caller.
  93  * The blob object argument for either blob may be NULL to indicate
  94  * that no content is present on its respective side of the diff.
  95  * File modes from relevant tree objects which contain the blobs may
  96  * also be passed. These will be zero if not available.
  97  */
  98 typedef const struct got_error *(*got_diff_blob_cb)(void *,
  99     struct got_blob_object *, struct got_blob_object *, FILE *, FILE *,
 100     struct got_object_id *, struct got_object_id *,
 101     const char *, const char *, mode_t, mode_t, struct got_repository *);
 102
 103 /*
 104  * A pre-defined implementation of got_diff_blob_cb() which appends unidiff
 105  * output to a file. The caller must allocate and fill in the argument
 106  * structure.
 107  */
 108 struct got_diff_blob_output_unidiff_arg {
 109         FILE *outfile;          /* Unidiff text will be written here. */
 110         int diff_context;       /* Sets the number of context lines. */
 111         int ignore_whitespace;  /* Ignore whitespace differences. */
 112         int force_text_diff;    /* Assume text even if binary data detected. */
 113         struct got_diffstat_cb_arg *diffstat; /* Compute diffstat of changes */
 114         enum got_diff_algorithm diff_algo; /* Diffing algorithm to use. */
 115
 116         /*
 117          * The number of lines contained in produced unidiff text output,
 118          * and an array of got_diff_lines with byte offset and line type to
 119          * each line. May be initialized to zero and NULL to ignore line
 120          * metadata. If not NULL, then the array of line offsets and types will
 121          * be populated. Optionally, the array can be pre-populated with line
 122          * offsets and types, with nlines > 0 indicating the length of the
 123          * pre-populated array.  This is useful if the output file already
 124          * contains some lines of text.  The array will be grown as needed to
 125          * accommodate additional offsets and types, and the last offset found
 126          * in a pre-populated array will be added to all subsequent offsets.
 127          */
 128         size_t nlines;
 129         struct got_diff_line *lines; /* Dispose of with free(3) when done. */
 130 };
 131 const struct got_error *got_diff_blob_output_unidiff(void *,
 132     struct got_blob_object *, struct got_blob_object *, FILE *, FILE *,
 133     struct got_object_id *, struct got_object_id *,
 134     const char *, const char *, mode_t, mode_t, struct got_repository *);
 135
 136 /*
 137  * Compute the differences between two trees and invoke the provided
 138  * got_diff_blob_cb() callback when content differs.
 139  * Diffing of blob content can be suppressed by passing zero for the
 140  * 'diff_content' parameter. The callback will then only receive blob
 141  * object IDs and diff labels, but NULL pointers instead of blob objects.
 142  * If 'diff_content' is set, two open temporary FILEs and two open
 143  * temporary file descriptors must be provided for internal use; these
 144  * files can be obtained from got_opentemp() and got_opentempfd(),
 145  * and must be closed by the caller. Otherwise the files can be NULL.
 146  * The set of arguments relating to either tree may be NULL to indicate
 147  * that no content is present on its respective side of the diff.
 148  */
 149 const struct got_error *got_diff_tree(struct got_tree_object *,
 150     struct got_tree_object *, FILE *, FILE *, int, int,
 151     const char *, const char *,
 152     struct got_repository *, got_diff_blob_cb cb, void *cb_arg, int);
 153
 154 /*
 155  * Pre-defined implementations of got_diff_blob_cb(): the first of which
 156  * collects a list of file paths that differ between two trees; the second
 157  * also computes a diffstat of added/removed lines for each collected path
 158  * and requires passing an initialized got_diffstat_cb_arg argument.
 159  * The caller must allocate and initialize a got_pathlist_head * argument.
 160  * Data pointers of entries added to the path list will point to a struct
 161  * got_diff_changed_path object.
 162  * The caller is expected to free both the path and data pointers of all
 163  * entries on the path list.
 164  */
 165 struct got_diff_changed_path {
 166         uint32_t add;   /* number of lines added */
 167         uint32_t rm;    /* number of lines removed */
 168         /*
 169          * The modification status of this path. It can be GOT_STATUS_ADD,
 170          * GOT_STATUS_DELETE, GOT_STATUS_MODIFY, or GOT_STATUS_MODE_CHANGE.
 171          */
 172         int status;
 173 };
 174 const struct got_error *got_diff_tree_collect_changed_paths(void *,
 175     struct got_blob_object *, struct got_blob_object *, FILE *, FILE *,
 176     struct got_object_id *, struct got_object_id *,
 177     const char *, const char *, mode_t, mode_t, struct got_repository *);
 178
 179 struct got_diffstat_cb_arg {
 180         size_t max_path_len;
 181         uint32_t ins;
 182         uint32_t del;
 183         int add_cols;
 184         int rm_cols;
 185         int nfiles;
 186         struct got_pathlist_head *paths;
 187         int ignore_ws;
 188         int force_text;
 189         enum got_diff_algorithm diff_algo;
 190 };
 191 const struct got_error *got_diff_tree_compute_diffstat(void *,
 192     struct got_blob_object *, struct got_blob_object *, FILE *, FILE *,
 193     struct got_object_id *, struct got_object_id *, const char *, const char *,
 194     mode_t, mode_t, struct got_repository *);
 195
 196 /*
 197  * Diff two objects, assuming both objects are blobs. Two const char * diff
 198  * header labels may be provided which will be used to identify each blob in
 199  * the diff output. If a label is NULL, use the blob's hash instead.
 200  * Two open temporary files and two temporary file descriptors must be
 201  * provided for internal use; these files can be obtained from
 202  * got_opentemp() and got_opentempfd(), and must be closed by the caller.
 203  * The set of arguments relating to either blob may be NULL/-1 to indicate
 204  * that no content is present on its respective side of the diff.
 205  * The number of context lines to show in the diff must be specified as well.
 206  * Write unified diff text to the provided output FILE.
 207  * If not NULL, the two initial output arguments will be populated with an
 208  * array of line offsets for, and the number of lines in, the unidiff text.
 209  */
 210 const struct got_error *got_diff_objects_as_blobs(struct got_diff_line **,
 211     size_t *, FILE *, FILE *, int, int, struct got_object_id *,
 212     struct got_object_id *, const char *, const char *, enum got_diff_algorithm,
 213     int, int, int, struct got_diffstat_cb_arg *, struct got_repository *,
 214     FILE *);
 215
 216 struct got_pathlist_head;
 217
 218 /*
 219  * Diff two objects, assuming both objects are trees. Two const char * diff
 220  * header labels may be provided which will be used to identify each blob in
 221  * the trees. If a label is NULL, use the blob's hash instead.
 222  * The number of context lines to show in diffs must be specified.
 223  * Two open temporary files and two temporary file descriptors must be
 224  * provided for internal use; these files can be obtained from
 225  * got_opentemp() and got_opentempfd(), and must be closed by the caller.
 226  * If 'diff_content' is not set, the files may be NULL / -1.
 227  * The set of arguments relating to either tree may be NULL to indicate
 228  * that no content is present on its respective side of the diff.
 229  * Write unified diff text to the provided output FILE.
 230  * If not NULL, the two initial output arguments will be populated with an
 231  * array of line offsets for, and the number of lines in, the unidiff text.
 232  */
 233 const struct got_error *got_diff_objects_as_trees(struct got_diff_line **,
 234     size_t *, FILE *, FILE *, int, int, struct got_object_id *,
 235     struct got_object_id *, struct got_pathlist_head *, const char *,
 236     const char *, enum got_diff_algorithm, int, int, int,
 237     struct got_diffstat_cb_arg *, struct got_repository *, FILE *);
 238
 239 /*
 240  * Diff two objects, assuming both objects are commits.
 241  * The number of context lines to show in diffs must be specified.
 242  * Two open temporary files and two temporary file descriptors must be
 243  * provided for internal use; these files can be obtained from
 244  * got_opentemp() and got_opentempfd(), and must be closed by the caller.
 245  * The set of arguments relating to either commit may be NULL to indicate
 246  * that no content is present on its respective side of the diff.
 247  * Write unified diff text to the provided output FILE.
 248  * If not NULL, the two initial output arguments will be populated with an
 249  * array of line offsets for, and the number of lines in, the unidiff text.
 250  */
 251 const struct got_error *got_diff_objects_as_commits(struct got_diff_line **,
 252     size_t *, FILE *, FILE *, int, int, struct got_object_id *,
 253     struct got_object_id *, struct got_pathlist_head *, enum got_diff_algorithm,
 254     int, int, int, struct got_diffstat_cb_arg *, struct got_repository *,
 255     FILE *);
 256
 257 #define GOT_DIFF_MAX_CONTEXT    64