builtin/shortlog.c

   1 #define USE_THE_REPOSITORY_VARIABLE
   2 #include "builtin.h"
   3 #include "config.h"
   4 #include "commit.h"
   5 #include "diff.h"
   6 #include "environment.h"
   7 #include "gettext.h"
   8 #include "string-list.h"
   9 #include "revision.h"
  10 #include "utf8.h"
  11 #include "mailmap.h"
  12 #include "setup.h"
  13 #include "shortlog.h"
  14 #include "parse-options.h"
  15 #include "trailer.h"
  16 #include "strmap.h"
  17
  18 static char const * const shortlog_usage[] = {
  19         N_("git shortlog [<options>] [<revision-range>] [[--] <path>...]"),
  20         N_("git log --pretty=short | git shortlog [<options>]"),
  21         NULL
  22 };
  23
  24 /*
  25  * The util field of our string_list_items will contain one of two things:
  26  *
  27  *   - if --summary is not in use, it will point to a string list of the
  28  *     oneline subjects assigned to this author
  29  *
  30  *   - if --summary is in use, we don't need that list; we only need to know
  31  *     its size. So we abuse the pointer slot to store our integer counter.
  32  *
  33  *  This macro accesses the latter.
  34  */
  35 #define UTIL_TO_INT(x) ((intptr_t)(x)->util)
  36
  37 static int compare_by_counter(const void *a1, const void *a2)
  38 {
  39         const struct string_list_item *i1 = a1, *i2 = a2;
  40         return UTIL_TO_INT(i2) - UTIL_TO_INT(i1);
  41 }
  42
  43 static int compare_by_list(const void *a1, const void *a2)
  44 {
  45         const struct string_list_item *i1 = a1, *i2 = a2;
  46         const struct string_list *l1 = i1->util, *l2 = i2->util;
  47
  48         if (l1->nr < l2->nr)
  49                 return 1;
  50         else if (l1->nr == l2->nr)
  51                 return 0;
  52         else
  53                 return -1;
  54 }
  55
  56 static void insert_one_record(struct shortlog *log,
  57                               const char *ident,
  58                               const char *oneline)
  59 {
  60         struct string_list_item *item;
  61
  62         item = string_list_insert(&log->list, ident);
  63
  64         if (log->summary)
  65                 item->util = (void *)(UTIL_TO_INT(item) + 1);
  66         else {
  67                 char *buffer;
  68                 struct strbuf subject = STRBUF_INIT;
  69                 const char *eol;
  70
  71                 /* Skip any leading whitespace, including any blank lines. */
  72                 while (*oneline && isspace(*oneline))
  73                         oneline++;
  74                 eol = strchr(oneline, '\n');
  75                 if (!eol)
  76                         eol = oneline + strlen(oneline);
  77                 if (starts_with(oneline, "[PATCH")) {
  78                         char *eob = strchr(oneline, ']');
  79                         if (eob && (!eol || eob < eol))
  80                                 oneline = eob + 1;
  81                 }
  82                 while (*oneline && isspace(*oneline) && *oneline != '\n')
  83                         oneline++;
  84                 format_subject(&subject, oneline, " ");
  85                 buffer = strbuf_detach(&subject, NULL);
  86
  87                 if (!item->util) {
  88                         item->util = xmalloc(sizeof(struct string_list));
  89                         string_list_init_nodup(item->util);
  90                 }
  91                 string_list_append(item->util, buffer);
  92         }
  93 }
  94
  95 static int parse_ident(struct shortlog *log,
  96                        struct strbuf *out, const char *in)
  97 {
  98         const char *mailbuf, *namebuf;
  99         size_t namelen, maillen;
 100         struct ident_split ident;
 101
 102         if (split_ident_line(&ident, in, strlen(in)))
 103                 return -1;
 104
 105         namebuf = ident.name_begin;
 106         mailbuf = ident.mail_begin;
 107         namelen = ident.name_end - ident.name_begin;
 108         maillen = ident.mail_end - ident.mail_begin;
 109
 110         map_user(&log->mailmap, &mailbuf, &maillen, &namebuf, &namelen);
 111         strbuf_add(out, namebuf, namelen);
 112         if (log->email)
 113                 strbuf_addf(out, " <%.*s>", (int)maillen, mailbuf);
 114
 115         return 0;
 116 }
 117
 118 static void read_from_stdin(struct shortlog *log)
 119 {
 120         struct strbuf ident = STRBUF_INIT;
 121         struct strbuf mapped_ident = STRBUF_INIT;
 122         struct strbuf oneline = STRBUF_INIT;
 123         static const char *author_match[2] = { "Author: ", "author " };
 124         static const char *committer_match[2] = { "Commit: ", "committer " };
 125         const char **match;
 126
 127         if (HAS_MULTI_BITS(log->groups))
 128                 die(_("using multiple --group options with stdin is not supported"));
 129
 130         switch (log->groups) {
 131         case SHORTLOG_GROUP_AUTHOR:
 132                 match = author_match;
 133                 break;
 134         case SHORTLOG_GROUP_COMMITTER:
 135                 match = committer_match;
 136                 break;
 137         case SHORTLOG_GROUP_TRAILER:
 138                 die(_("using %s with stdin is not supported"), "--group=trailer");
 139         case SHORTLOG_GROUP_FORMAT:
 140                 die(_("using %s with stdin is not supported"), "--group=format");
 141         default:
 142                 BUG("unhandled shortlog group");
 143         }
 144
 145         while (strbuf_getline_lf(&ident, stdin) != EOF) {
 146                 const char *v;
 147                 if (!skip_prefix(ident.buf, match[0], &v) &&
 148                     !skip_prefix(ident.buf, match[1], &v))
 149                         continue;
 150                 while (strbuf_getline_lf(&oneline, stdin) != EOF &&
 151                        oneline.len)
 152                         ; /* discard headers */
 153                 while (strbuf_getline_lf(&oneline, stdin) != EOF &&
 154                        !oneline.len)
 155                         ; /* discard blanks */
 156
 157                 strbuf_reset(&mapped_ident);
 158                 if (parse_ident(log, &mapped_ident, v) < 0)
 159                         continue;
 160
 161                 insert_one_record(log, mapped_ident.buf, oneline.buf);
 162         }
 163         strbuf_release(&ident);
 164         strbuf_release(&mapped_ident);
 165         strbuf_release(&oneline);
 166 }
 167
 168 static void insert_records_from_trailers(struct shortlog *log,
 169                                          struct strset *dups,
 170                                          struct commit *commit,
 171                                          struct pretty_print_context *ctx,
 172                                          const char *oneline)
 173 {
 174         struct trailer_iterator iter;
 175         const char *commit_buffer, *body;
 176         struct strbuf ident = STRBUF_INIT;
 177
 178         if (!log->trailers.nr)
 179                 return;
 180
 181         /*
 182          * Using repo_format_commit_message("%B") would be simpler here, but
 183          * this saves us copying the message.
 184          */
 185         commit_buffer = repo_logmsg_reencode(the_repository, commit, NULL,
 186                                              ctx->output_encoding);
 187         body = strstr(commit_buffer, "\n\n");
 188         if (!body)
 189                 return;
 190
 191         trailer_iterator_init(&iter, body);
 192         while (trailer_iterator_advance(&iter)) {
 193                 const char *value = iter.val.buf;
 194
 195                 if (!string_list_has_string(&log->trailers, iter.key.buf))
 196                         continue;
 197
 198                 strbuf_reset(&ident);
 199                 if (!parse_ident(log, &ident, value))
 200                         value = ident.buf;
 201
 202                 if (!strset_add(dups, value))
 203                         continue;
 204                 insert_one_record(log, value, oneline);
 205         }
 206         trailer_iterator_release(&iter);
 207
 208         strbuf_release(&ident);
 209         repo_unuse_commit_buffer(the_repository, commit, commit_buffer);
 210 }
 211
 212 static int shortlog_needs_dedup(const struct shortlog *log)
 213 {
 214         return HAS_MULTI_BITS(log->groups) || log->format.nr > 1 || log->trailers.nr;
 215 }
 216
 217 static void insert_records_from_format(struct shortlog *log,
 218                                        struct strset *dups,
 219                                        struct commit *commit,
 220                                        struct pretty_print_context *ctx,
 221                                        const char *oneline)
 222 {
 223         struct strbuf buf = STRBUF_INIT;
 224         struct string_list_item *item;
 225
 226         for_each_string_list_item(item, &log->format) {
 227                 strbuf_reset(&buf);
 228
 229                 repo_format_commit_message(the_repository, commit,
 230                                            item->string, &buf, ctx);
 231
 232                 if (!shortlog_needs_dedup(log) || strset_add(dups, buf.buf))
 233                         insert_one_record(log, buf.buf, oneline);
 234         }
 235
 236         strbuf_release(&buf);
 237 }
 238
 239 void shortlog_add_commit(struct shortlog *log, struct commit *commit)
 240 {
 241         struct strbuf oneline = STRBUF_INIT;
 242         struct strset dups = STRSET_INIT;
 243         struct pretty_print_context ctx = {0};
 244         const char *oneline_str;
 245
 246         ctx.fmt = CMIT_FMT_USERFORMAT;
 247         ctx.abbrev = log->abbrev;
 248         ctx.date_mode = log->date_mode;
 249         ctx.output_encoding = get_log_output_encoding();
 250
 251         if (!log->summary) {
 252                 if (log->user_format)
 253                         pretty_print_commit(&ctx, commit, &oneline);
 254                 else
 255                         repo_format_commit_message(the_repository, commit,
 256                                                    "%s", &oneline, &ctx);
 257         }
 258         oneline_str = oneline.len ? oneline.buf : "<none>";
 259
 260         insert_records_from_trailers(log, &dups, commit, &ctx, oneline_str);
 261         insert_records_from_format(log, &dups, commit, &ctx, oneline_str);
 262
 263         strset_clear(&dups);
 264         strbuf_release(&oneline);
 265 }
 266
 267 static void get_from_rev(struct rev_info *rev, struct shortlog *log)
 268 {
 269         struct commit *commit;
 270
 271         if (prepare_revision_walk(rev))
 272                 die(_("revision walk setup failed"));
 273         while ((commit = get_revision(rev)) != NULL)
 274                 shortlog_add_commit(log, commit);
 275 }
 276
 277 static int parse_uint(char const **arg, int comma, int defval)
 278 {
 279         unsigned long ul;
 280         int ret;
 281         char *endp;
 282
 283         ul = strtoul(*arg, &endp, 10);
 284         if (*endp && *endp != comma)
 285                 return -1;
 286         if (ul > INT_MAX)
 287                 return -1;
 288         ret = *arg == endp ? defval : (int)ul;
 289         *arg = *endp ? endp + 1 : endp;
 290         return ret;
 291 }
 292
 293 static const char wrap_arg_usage[] = "-w[<width>[,<indent1>[,<indent2>]]]";
 294 #define DEFAULT_WRAPLEN 76
 295 #define DEFAULT_INDENT1 6
 296 #define DEFAULT_INDENT2 9
 297
 298 static int parse_wrap_args(const struct option *opt, const char *arg, int unset)
 299 {
 300         struct shortlog *log = opt->value;
 301
 302         log->wrap_lines = !unset;
 303         if (unset)
 304                 return 0;
 305         if (!arg) {
 306                 log->wrap = DEFAULT_WRAPLEN;
 307                 log->in1 = DEFAULT_INDENT1;
 308                 log->in2 = DEFAULT_INDENT2;
 309                 return 0;
 310         }
 311
 312         log->wrap = parse_uint(&arg, ',', DEFAULT_WRAPLEN);
 313         log->in1 = parse_uint(&arg, ',', DEFAULT_INDENT1);
 314         log->in2 = parse_uint(&arg, '\0', DEFAULT_INDENT2);
 315         if (log->wrap < 0 || log->in1 < 0 || log->in2 < 0)
 316                 return error(wrap_arg_usage);
 317         if (log->wrap &&
 318             ((log->in1 && log->wrap <= log->in1) ||
 319              (log->in2 && log->wrap <= log->in2)))
 320                 return error(wrap_arg_usage);
 321         return 0;
 322 }
 323
 324 static int parse_group_option(const struct option *opt, const char *arg, int unset)
 325 {
 326         struct shortlog *log = opt->value;
 327         const char *field;
 328
 329         if (unset) {
 330                 log->groups = 0;
 331                 string_list_clear(&log->trailers, 0);
 332                 string_list_clear(&log->format, 0);
 333         } else if (!strcasecmp(arg, "author"))
 334                 log->groups |= SHORTLOG_GROUP_AUTHOR;
 335         else if (!strcasecmp(arg, "committer"))
 336                 log->groups |= SHORTLOG_GROUP_COMMITTER;
 337         else if (skip_prefix(arg, "trailer:", &field)) {
 338                 log->groups |= SHORTLOG_GROUP_TRAILER;
 339                 string_list_append(&log->trailers, field);
 340         } else if (skip_prefix(arg, "format:", &field)) {
 341                 log->groups |= SHORTLOG_GROUP_FORMAT;
 342                 string_list_append(&log->format, field);
 343         } else if (strchr(arg, '%')) {
 344                 log->groups |= SHORTLOG_GROUP_FORMAT;
 345                 string_list_append(&log->format, arg);
 346         } else {
 347                 return error(_("unknown group type: %s"), arg);
 348         }
 349
 350         return 0;
 351 }
 352
 353
 354 void shortlog_init(struct shortlog *log)
 355 {
 356         memset(log, 0, sizeof(*log));
 357
 358         read_mailmap(&log->mailmap);
 359
 360         log->list.strdup_strings = 1;
 361         log->wrap = DEFAULT_WRAPLEN;
 362         log->in1 = DEFAULT_INDENT1;
 363         log->in2 = DEFAULT_INDENT2;
 364         log->trailers.strdup_strings = 1;
 365         log->trailers.cmp = strcasecmp;
 366         log->format.strdup_strings = 1;
 367 }
 368
 369 void shortlog_finish_setup(struct shortlog *log)
 370 {
 371         if (log->groups & SHORTLOG_GROUP_AUTHOR)
 372                 string_list_append(&log->format,
 373                                    log->email ? "%aN <%aE>" : "%aN");
 374         if (log->groups & SHORTLOG_GROUP_COMMITTER)
 375                 string_list_append(&log->format,
 376                                    log->email ? "%cN <%cE>" : "%cN");
 377
 378         string_list_sort(&log->trailers);
 379 }
 380
 381 int cmd_shortlog(int argc,
 382                  const char **argv,
 383                  const char *prefix,
 384                  struct repository *repo UNUSED)
 385 {
 386         struct shortlog log = { STRING_LIST_INIT_NODUP };
 387         struct rev_info rev;
 388         int nongit = !startup_info->have_repository;
 389
 390         const struct option options[] = {
 391                 OPT_BIT('c', "committer", &log.groups,
 392                         N_("group by committer rather than author"),
 393                         SHORTLOG_GROUP_COMMITTER),
 394                 OPT_BOOL('n', "numbered", &log.sort_by_number,
 395                          N_("sort output according to the number of commits per author")),
 396                 OPT_BOOL('s', "summary", &log.summary,
 397                          N_("suppress commit descriptions, only provides commit count")),
 398                 OPT_BOOL('e', "email", &log.email,
 399                          N_("show the email address of each author")),
 400                 OPT_CALLBACK_F('w', NULL, &log, N_("<w>[,<i1>[,<i2>]]"),
 401                         N_("linewrap output"), PARSE_OPT_OPTARG,
 402                         &parse_wrap_args),
 403                 OPT_CALLBACK(0, "group", &log, N_("field"),
 404                         N_("group by field"), parse_group_option),
 405                 OPT_END(),
 406         };
 407
 408         struct parse_opt_ctx_t ctx;
 409
 410         /*
 411          * NEEDSWORK: Later on we'll call parse_revision_opt which relies on
 412          * the hash algorithm being set but since we are operating outside of a
 413          * Git repository we cannot determine one. This is only needed because
 414          * parse_revision_opt expects hexsz for --abbrev which is irrelevant
 415          * for shortlog outside of a git repository. For now explicitly set
 416          * SHA1, but ideally the parsing machinery would be split between
 417          * git/nongit so that we do not have to do this.
 418          */
 419         if (nongit && !the_hash_algo)
 420                 repo_set_hash_algo(the_repository, GIT_HASH_SHA1);
 421
 422         git_config(git_default_config, NULL);
 423         shortlog_init(&log);
 424         repo_init_revisions(the_repository, &rev, prefix);
 425         parse_options_start(&ctx, argc, argv, prefix, options,
 426                             PARSE_OPT_KEEP_DASHDASH | PARSE_OPT_KEEP_ARGV0);
 427
 428         for (;;) {
 429                 switch (parse_options_step(&ctx, options, shortlog_usage)) {
 430                 case PARSE_OPT_NON_OPTION:
 431                 case PARSE_OPT_UNKNOWN:
 432                         break;
 433                 case PARSE_OPT_HELP:
 434                 case PARSE_OPT_ERROR:
 435                 case PARSE_OPT_SUBCOMMAND:
 436                         exit(129);
 437                 case PARSE_OPT_COMPLETE:
 438                         exit(0);
 439                 case PARSE_OPT_DONE:
 440                         goto parse_done;
 441                 }
 442                 parse_revision_opt(&rev, &ctx, options, shortlog_usage);
 443         }
 444 parse_done:
 445         revision_opts_finish(&rev);
 446         argc = parse_options_end(&ctx);
 447
 448         if (nongit && argc > 1) {
 449                 error(_("too many arguments given outside repository"));
 450                 usage_with_options(shortlog_usage, options);
 451         }
 452
 453         if (!nongit && setup_revisions(argc, argv, &rev, NULL) != 1) {
 454                 error(_("unrecognized argument: %s"), argv[1]);
 455                 usage_with_options(shortlog_usage, options);
 456         }
 457
 458         log.user_format = rev.commit_format == CMIT_FMT_USERFORMAT;
 459         log.abbrev = rev.abbrev;
 460         log.file = rev.diffopt.file;
 461         log.date_mode = rev.date_mode;
 462
 463         if (!log.groups)
 464                 log.groups = SHORTLOG_GROUP_AUTHOR;
 465         shortlog_finish_setup(&log);
 466
 467         /* assume HEAD if from a tty */
 468         if (!nongit && !rev.pending.nr && isatty(0))
 469                 add_head_to_pending(&rev);
 470         if (rev.pending.nr == 0) {
 471                 if (isatty(0))
 472                         fprintf(stderr, _("(reading log message from standard input)\n"));
 473                 read_from_stdin(&log);
 474         }
 475         else
 476                 get_from_rev(&rev, &log);
 477
 478         shortlog_output(&log);
 479         release_revisions(&rev);
 480         return 0;
 481 }
 482
 483 static void add_wrapped_shortlog_msg(struct strbuf *sb, const char *s,
 484                                      const struct shortlog *log)
 485 {
 486         strbuf_add_wrapped_text(sb, s, log->in1, log->in2, log->wrap);
 487         strbuf_addch(sb, '\n');
 488 }
 489
 490 void shortlog_output(struct shortlog *log)
 491 {
 492         size_t i, j;
 493         struct strbuf sb = STRBUF_INIT;
 494
 495         if (log->sort_by_number)
 496                 STABLE_QSORT(log->list.items, log->list.nr,
 497                       log->summary ? compare_by_counter : compare_by_list);
 498         for (i = 0; i < log->list.nr; i++) {
 499                 const struct string_list_item *item = &log->list.items[i];
 500                 if (log->summary) {
 501                         fprintf(log->file, "%6d\t%s\n",
 502                                 (int)UTIL_TO_INT(item), item->string);
 503                 } else {
 504                         struct string_list *onelines = item->util;
 505                         fprintf(log->file, "%s (%"PRIuMAX"):\n",
 506                                 item->string, (uintmax_t)onelines->nr);
 507                         for (j = onelines->nr; j >= 1; j--) {
 508                                 const char *msg = onelines->items[j - 1].string;
 509
 510                                 if (log->wrap_lines) {
 511                                         strbuf_reset(&sb);
 512                                         add_wrapped_shortlog_msg(&sb, msg, log);
 513                                         fwrite(sb.buf, sb.len, 1, log->file);
 514                                 }
 515                                 else
 516                                         fprintf(log->file, "      %s\n", msg);
 517                         }
 518                         putc('\n', log->file);
 519                         onelines->strdup_strings = 1;
 520                         string_list_clear(onelines, 0);
 521                         free(onelines);
 522                 }
 523
 524                 log->list.items[i].util = NULL;
 525         }
 526
 527         strbuf_release(&sb);
 528         log->list.strdup_strings = 1;
 529         string_list_clear(&log->list, 1);
 530         clear_mailmap(&log->mailmap);
 531         string_list_clear(&log->format, 0);
 532         string_list_clear(&log->trailers, 0);
 533 }