builtin/shortlog.c

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