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 --group=trailer with stdin is not supported"));
 136         default:
 137                 BUG("unhandled shortlog group");
 138         }
 139
 140         while (strbuf_getline_lf(&ident, stdin) != EOF) {
 141                 const char *v;
 142                 if (!skip_prefix(ident.buf, match[0], &v) &&
 143                     !skip_prefix(ident.buf, match[1], &v))
 144                         continue;
 145                 while (strbuf_getline_lf(&oneline, stdin) != EOF &&
 146                        oneline.len)
 147                         ; /* discard headers */
 148                 while (strbuf_getline_lf(&oneline, stdin) != EOF &&
 149                        !oneline.len)
 150                         ; /* discard blanks */
 151
 152                 strbuf_reset(&mapped_ident);
 153                 if (parse_ident(log, &mapped_ident, v) < 0)
 154                         continue;
 155
 156                 insert_one_record(log, mapped_ident.buf, oneline.buf);
 157         }
 158         strbuf_release(&ident);
 159         strbuf_release(&mapped_ident);
 160         strbuf_release(&oneline);
 161 }
 162
 163 static void insert_records_from_trailers(struct shortlog *log,
 164                                          struct strset *dups,
 165                                          struct commit *commit,
 166                                          struct pretty_print_context *ctx,
 167                                          const char *oneline)
 168 {
 169         struct trailer_iterator iter;
 170         const char *commit_buffer, *body;
 171         struct strbuf ident = STRBUF_INIT;
 172
 173         /*
 174          * Using format_commit_message("%B") would be simpler here, but
 175          * this saves us copying the message.
 176          */
 177         commit_buffer = logmsg_reencode(commit, NULL, ctx->output_encoding);
 178         body = strstr(commit_buffer, "\n\n");
 179         if (!body)
 180                 return;
 181
 182         trailer_iterator_init(&iter, body);
 183         while (trailer_iterator_advance(&iter)) {
 184                 const char *value = iter.val.buf;
 185
 186                 if (!string_list_has_string(&log->trailers, iter.key.buf))
 187                         continue;
 188
 189                 strbuf_reset(&ident);
 190                 if (!parse_ident(log, &ident, value))
 191                         value = ident.buf;
 192
 193                 if (!strset_add(dups, value))
 194                         continue;
 195                 insert_one_record(log, value, oneline);
 196         }
 197         trailer_iterator_release(&iter);
 198
 199         strbuf_release(&ident);
 200         unuse_commit_buffer(commit, commit_buffer);
 201 }
 202
 203 void shortlog_add_commit(struct shortlog *log, struct commit *commit)
 204 {
 205         struct strbuf ident = STRBUF_INIT;
 206         struct strbuf oneline = STRBUF_INIT;
 207         struct strset dups = STRSET_INIT;
 208         struct pretty_print_context ctx = {0};
 209         const char *oneline_str;
 210
 211         ctx.fmt = CMIT_FMT_USERFORMAT;
 212         ctx.abbrev = log->abbrev;
 213         ctx.print_email_subject = 1;
 214         ctx.date_mode.type = DATE_NORMAL;
 215         ctx.output_encoding = get_log_output_encoding();
 216
 217         if (!log->summary) {
 218                 if (log->user_format)
 219                         pretty_print_commit(&ctx, commit, &oneline);
 220                 else
 221                         format_commit_message(commit, "%s", &oneline, &ctx);
 222         }
 223         oneline_str = oneline.len ? oneline.buf : "<none>";
 224
 225         if (log->groups & SHORTLOG_GROUP_AUTHOR) {
 226                 strbuf_reset(&ident);
 227                 format_commit_message(commit,
 228                                       log->email ? "%aN <%aE>" : "%aN",
 229                                       &ident, &ctx);
 230                 if (!HAS_MULTI_BITS(log->groups) ||
 231                     strset_add(&dups, ident.buf))
 232                         insert_one_record(log, ident.buf, oneline_str);
 233         }
 234         if (log->groups & SHORTLOG_GROUP_COMMITTER) {
 235                 strbuf_reset(&ident);
 236                 format_commit_message(commit,
 237                                       log->email ? "%cN <%cE>" : "%cN",
 238                                       &ident, &ctx);
 239                 if (!HAS_MULTI_BITS(log->groups) ||
 240                     strset_add(&dups, ident.buf))
 241                         insert_one_record(log, ident.buf, oneline_str);
 242         }
 243         if (log->groups & SHORTLOG_GROUP_TRAILER) {
 244                 insert_records_from_trailers(log, &dups, commit, &ctx, oneline_str);
 245         }
 246
 247         strset_clear(&dups);
 248         strbuf_release(&ident);
 249         strbuf_release(&oneline);
 250 }
 251
 252 static void get_from_rev(struct rev_info *rev, struct shortlog *log)
 253 {
 254         struct commit *commit;
 255
 256         if (prepare_revision_walk(rev))
 257                 die(_("revision walk setup failed"));
 258         while ((commit = get_revision(rev)) != NULL)
 259                 shortlog_add_commit(log, commit);
 260 }
 261
 262 static int parse_uint(char const **arg, int comma, int defval)
 263 {
 264         unsigned long ul;
 265         int ret;
 266         char *endp;
 267
 268         ul = strtoul(*arg, &endp, 10);
 269         if (*endp && *endp != comma)
 270                 return -1;
 271         if (ul > INT_MAX)
 272                 return -1;
 273         ret = *arg == endp ? defval : (int)ul;
 274         *arg = *endp ? endp + 1 : endp;
 275         return ret;
 276 }
 277
 278 static const char wrap_arg_usage[] = "-w[<width>[,<indent1>[,<indent2>]]]";
 279 #define DEFAULT_WRAPLEN 76
 280 #define DEFAULT_INDENT1 6
 281 #define DEFAULT_INDENT2 9
 282
 283 static int parse_wrap_args(const struct option *opt, const char *arg, int unset)
 284 {
 285         struct shortlog *log = opt->value;
 286
 287         log->wrap_lines = !unset;
 288         if (unset)
 289                 return 0;
 290         if (!arg) {
 291                 log->wrap = DEFAULT_WRAPLEN;
 292                 log->in1 = DEFAULT_INDENT1;
 293                 log->in2 = DEFAULT_INDENT2;
 294                 return 0;
 295         }
 296
 297         log->wrap = parse_uint(&arg, ',', DEFAULT_WRAPLEN);
 298         log->in1 = parse_uint(&arg, ',', DEFAULT_INDENT1);
 299         log->in2 = parse_uint(&arg, '\0', DEFAULT_INDENT2);
 300         if (log->wrap < 0 || log->in1 < 0 || log->in2 < 0)
 301                 return error(wrap_arg_usage);
 302         if (log->wrap &&
 303             ((log->in1 && log->wrap <= log->in1) ||
 304              (log->in2 && log->wrap <= log->in2)))
 305                 return error(wrap_arg_usage);
 306         return 0;
 307 }
 308
 309 static int parse_group_option(const struct option *opt, const char *arg, int unset)
 310 {
 311         struct shortlog *log = opt->value;
 312         const char *field;
 313
 314         if (unset) {
 315                 log->groups = 0;
 316                 string_list_clear(&log->trailers, 0);
 317         } else if (!strcasecmp(arg, "author"))
 318                 log->groups |= SHORTLOG_GROUP_AUTHOR;
 319         else if (!strcasecmp(arg, "committer"))
 320                 log->groups |= SHORTLOG_GROUP_COMMITTER;
 321         else if (skip_prefix(arg, "trailer:", &field)) {
 322                 log->groups |= SHORTLOG_GROUP_TRAILER;
 323                 string_list_append(&log->trailers, field);
 324         } else
 325                 return error(_("unknown group type: %s"), arg);
 326
 327         return 0;
 328 }
 329
 330
 331 void shortlog_init(struct shortlog *log)
 332 {
 333         memset(log, 0, sizeof(*log));
 334
 335         read_mailmap(&log->mailmap);
 336
 337         log->list.strdup_strings = 1;
 338         log->wrap = DEFAULT_WRAPLEN;
 339         log->in1 = DEFAULT_INDENT1;
 340         log->in2 = DEFAULT_INDENT2;
 341         log->trailers.strdup_strings = 1;
 342         log->trailers.cmp = strcasecmp;
 343 }
 344
 345 int cmd_shortlog(int argc, const char **argv, const char *prefix)
 346 {
 347         struct shortlog log = { STRING_LIST_INIT_NODUP };
 348         struct rev_info rev;
 349         int nongit = !startup_info->have_repository;
 350
 351         const struct option options[] = {
 352                 OPT_BIT('c', "committer", &log.groups,
 353                         N_("group by committer rather than author"),
 354                         SHORTLOG_GROUP_COMMITTER),
 355                 OPT_BOOL('n', "numbered", &log.sort_by_number,
 356                          N_("sort output according to the number of commits per author")),
 357                 OPT_BOOL('s', "summary", &log.summary,
 358                          N_("suppress commit descriptions, only provides commit count")),
 359                 OPT_BOOL('e', "email", &log.email,
 360                          N_("show the email address of each author")),
 361                 OPT_CALLBACK_F('w', NULL, &log, N_("<w>[,<i1>[,<i2>]]"),
 362                         N_("linewrap output"), PARSE_OPT_OPTARG,
 363                         &parse_wrap_args),
 364                 OPT_CALLBACK(0, "group", &log, N_("field"),
 365                         N_("group by field"), parse_group_option),
 366                 OPT_END(),
 367         };
 368
 369         struct parse_opt_ctx_t ctx;
 370
 371         git_config(git_default_config, NULL);
 372         shortlog_init(&log);
 373         repo_init_revisions(the_repository, &rev, prefix);
 374         parse_options_start(&ctx, argc, argv, prefix, options,
 375                             PARSE_OPT_KEEP_DASHDASH | PARSE_OPT_KEEP_ARGV0);
 376
 377         for (;;) {
 378                 switch (parse_options_step(&ctx, options, shortlog_usage)) {
 379                 case PARSE_OPT_NON_OPTION:
 380                 case PARSE_OPT_UNKNOWN:
 381                         break;
 382                 case PARSE_OPT_HELP:
 383                 case PARSE_OPT_ERROR:
 384                         exit(129);
 385                 case PARSE_OPT_COMPLETE:
 386                         exit(0);
 387                 case PARSE_OPT_DONE:
 388                         goto parse_done;
 389                 }
 390                 parse_revision_opt(&rev, &ctx, options, shortlog_usage);
 391         }
 392 parse_done:
 393         revision_opts_finish(&rev);
 394         argc = parse_options_end(&ctx);
 395
 396         if (nongit && argc > 1) {
 397                 error(_("too many arguments given outside repository"));
 398                 usage_with_options(shortlog_usage, options);
 399         }
 400
 401         if (setup_revisions(argc, argv, &rev, NULL) != 1) {
 402                 error(_("unrecognized argument: %s"), argv[1]);
 403                 usage_with_options(shortlog_usage, options);
 404         }
 405
 406         log.user_format = rev.commit_format == CMIT_FMT_USERFORMAT;
 407         log.abbrev = rev.abbrev;
 408         log.file = rev.diffopt.file;
 409
 410         if (!log.groups)
 411                 log.groups = SHORTLOG_GROUP_AUTHOR;
 412         string_list_sort(&log.trailers);
 413
 414         /* assume HEAD if from a tty */
 415         if (!nongit && !rev.pending.nr && isatty(0))
 416                 add_head_to_pending(&rev);
 417         if (rev.pending.nr == 0) {
 418                 if (isatty(0))
 419                         fprintf(stderr, _("(reading log message from standard input)\n"));
 420                 read_from_stdin(&log);
 421         }
 422         else
 423                 get_from_rev(&rev, &log);
 424
 425         release_revisions(&rev);
 426
 427         shortlog_output(&log);
 428         if (log.file != stdout)
 429                 fclose(log.file);
 430         return 0;
 431 }
 432
 433 static void add_wrapped_shortlog_msg(struct strbuf *sb, const char *s,
 434                                      const struct shortlog *log)
 435 {
 436         strbuf_add_wrapped_text(sb, s, log->in1, log->in2, log->wrap);
 437         strbuf_addch(sb, '\n');
 438 }
 439
 440 void shortlog_output(struct shortlog *log)
 441 {
 442         size_t i, j;
 443         struct strbuf sb = STRBUF_INIT;
 444
 445         if (log->sort_by_number)
 446                 STABLE_QSORT(log->list.items, log->list.nr,
 447                       log->summary ? compare_by_counter : compare_by_list);
 448         for (i = 0; i < log->list.nr; i++) {
 449                 const struct string_list_item *item = &log->list.items[i];
 450                 if (log->summary) {
 451                         fprintf(log->file, "%6d\t%s\n",
 452                                 (int)UTIL_TO_INT(item), item->string);
 453                 } else {
 454                         struct string_list *onelines = item->util;
 455                         fprintf(log->file, "%s (%"PRIuMAX"):\n",
 456                                 item->string, (uintmax_t)onelines->nr);
 457                         for (j = onelines->nr; j >= 1; j--) {
 458                                 const char *msg = onelines->items[j - 1].string;
 459
 460                                 if (log->wrap_lines) {
 461                                         strbuf_reset(&sb);
 462                                         add_wrapped_shortlog_msg(&sb, msg, log);
 463                                         fwrite(sb.buf, sb.len, 1, log->file);
 464                                 }
 465                                 else
 466                                         fprintf(log->file, "      %s\n", msg);
 467                         }
 468                         putc('\n', log->file);
 469                         onelines->strdup_strings = 1;
 470                         string_list_clear(onelines, 0);
 471                         free(onelines);
 472                 }
 473
 474                 log->list.items[i].util = NULL;
 475         }
 476
 477         strbuf_release(&sb);
 478         log->list.strdup_strings = 1;
 479         string_list_clear(&log->list, 1);
 480         clear_mailmap(&log->mailmap);
 481 }