grep.c

   1 #include "cache.h"
   2 #include "config.h"
   3 #include "grep.h"
   4 #include "object-store.h"
   5 #include "userdiff.h"
   6 #include "xdiff-interface.h"
   7 #include "diff.h"
   8 #include "diffcore.h"
   9 #include "commit.h"
  10 #include "quote.h"
  11 #include "help.h"
  12
  13 static int grep_source_load(struct grep_source *gs);
  14 static int grep_source_is_binary(struct grep_source *gs,
  15                                  struct index_state *istate);
  16
  17 static struct grep_opt grep_defaults;
  18
  19 static const char *color_grep_slots[] = {
  20         [GREP_COLOR_CONTEXT]        = "context",
  21         [GREP_COLOR_FILENAME]       = "filename",
  22         [GREP_COLOR_FUNCTION]       = "function",
  23         [GREP_COLOR_LINENO]         = "lineNumber",
  24         [GREP_COLOR_COLUMNNO]       = "column",
  25         [GREP_COLOR_MATCH_CONTEXT]  = "matchContext",
  26         [GREP_COLOR_MATCH_SELECTED] = "matchSelected",
  27         [GREP_COLOR_SELECTED]       = "selected",
  28         [GREP_COLOR_SEP]            = "separator",
  29 };
  30
  31 static void std_output(struct grep_opt *opt, const void *buf, size_t size)
  32 {
  33         fwrite(buf, size, 1, stdout);
  34 }
  35
  36 static void color_set(char *dst, const char *color_bytes)
  37 {
  38         xsnprintf(dst, COLOR_MAXLEN, "%s", color_bytes);
  39 }
  40
  41 /*
  42  * Initialize the grep_defaults template with hardcoded defaults.
  43  * We could let the compiler do this, but without C99 initializers
  44  * the code gets unwieldy and unreadable, so...
  45  */
  46 void init_grep_defaults(struct repository *repo)
  47 {
  48         struct grep_opt *opt = &grep_defaults;
  49         static int run_once;
  50
  51         if (run_once)
  52                 return;
  53         run_once++;
  54
  55         memset(opt, 0, sizeof(*opt));
  56         opt->repo = repo;
  57         opt->relative = 1;
  58         opt->pathname = 1;
  59         opt->max_depth = -1;
  60         opt->pattern_type_option = GREP_PATTERN_TYPE_UNSPECIFIED;
  61         color_set(opt->colors[GREP_COLOR_CONTEXT], "");
  62         color_set(opt->colors[GREP_COLOR_FILENAME], "");
  63         color_set(opt->colors[GREP_COLOR_FUNCTION], "");
  64         color_set(opt->colors[GREP_COLOR_LINENO], "");
  65         color_set(opt->colors[GREP_COLOR_COLUMNNO], "");
  66         color_set(opt->colors[GREP_COLOR_MATCH_CONTEXT], GIT_COLOR_BOLD_RED);
  67         color_set(opt->colors[GREP_COLOR_MATCH_SELECTED], GIT_COLOR_BOLD_RED);
  68         color_set(opt->colors[GREP_COLOR_SELECTED], "");
  69         color_set(opt->colors[GREP_COLOR_SEP], GIT_COLOR_CYAN);
  70         opt->only_matching = 0;
  71         opt->color = -1;
  72         opt->output = std_output;
  73 }
  74
  75 static int parse_pattern_type_arg(const char *opt, const char *arg)
  76 {
  77         if (!strcmp(arg, "default"))
  78                 return GREP_PATTERN_TYPE_UNSPECIFIED;
  79         else if (!strcmp(arg, "basic"))
  80                 return GREP_PATTERN_TYPE_BRE;
  81         else if (!strcmp(arg, "extended"))
  82                 return GREP_PATTERN_TYPE_ERE;
  83         else if (!strcmp(arg, "fixed"))
  84                 return GREP_PATTERN_TYPE_FIXED;
  85         else if (!strcmp(arg, "perl"))
  86                 return GREP_PATTERN_TYPE_PCRE;
  87         die("bad %s argument: %s", opt, arg);
  88 }
  89
  90 define_list_config_array_extra(color_grep_slots, {"match"});
  91
  92 /*
  93  * Read the configuration file once and store it in
  94  * the grep_defaults template.
  95  */
  96 int grep_config(const char *var, const char *value, void *cb)
  97 {
  98         struct grep_opt *opt = &grep_defaults;
  99         const char *slot;
 100
 101         if (userdiff_config(var, value) < 0)
 102                 return -1;
 103
 104         if (!strcmp(var, "grep.extendedregexp")) {
 105                 opt->extended_regexp_option = git_config_bool(var, value);
 106                 return 0;
 107         }
 108
 109         if (!strcmp(var, "grep.patterntype")) {
 110                 opt->pattern_type_option = parse_pattern_type_arg(var, value);
 111                 return 0;
 112         }
 113
 114         if (!strcmp(var, "grep.linenumber")) {
 115                 opt->linenum = git_config_bool(var, value);
 116                 return 0;
 117         }
 118         if (!strcmp(var, "grep.column")) {
 119                 opt->columnnum = git_config_bool(var, value);
 120                 return 0;
 121         }
 122
 123         if (!strcmp(var, "grep.fullname")) {
 124                 opt->relative = !git_config_bool(var, value);
 125                 return 0;
 126         }
 127
 128         if (!strcmp(var, "color.grep"))
 129                 opt->color = git_config_colorbool(var, value);
 130         if (!strcmp(var, "color.grep.match")) {
 131                 if (grep_config("color.grep.matchcontext", value, cb) < 0)
 132                         return -1;
 133                 if (grep_config("color.grep.matchselected", value, cb) < 0)
 134                         return -1;
 135         } else if (skip_prefix(var, "color.grep.", &slot)) {
 136                 int i = LOOKUP_CONFIG(color_grep_slots, slot);
 137                 char *color;
 138
 139                 if (i < 0)
 140                         return -1;
 141                 color = opt->colors[i];
 142                 if (!value)
 143                         return config_error_nonbool(var);
 144                 return color_parse(value, color);
 145         }
 146         return 0;
 147 }
 148
 149 /*
 150  * Initialize one instance of grep_opt and copy the
 151  * default values from the template we read the configuration
 152  * information in an earlier call to git_config(grep_config).
 153  */
 154 void grep_init(struct grep_opt *opt, struct repository *repo, const char *prefix)
 155 {
 156         struct grep_opt *def = &grep_defaults;
 157         int i;
 158
 159         memset(opt, 0, sizeof(*opt));
 160         opt->repo = repo;
 161         opt->prefix = prefix;
 162         opt->prefix_length = (prefix && *prefix) ? strlen(prefix) : 0;
 163         opt->pattern_tail = &opt->pattern_list;
 164         opt->header_tail = &opt->header_list;
 165
 166         opt->only_matching = def->only_matching;
 167         opt->color = def->color;
 168         opt->extended_regexp_option = def->extended_regexp_option;
 169         opt->pattern_type_option = def->pattern_type_option;
 170         opt->linenum = def->linenum;
 171         opt->columnnum = def->columnnum;
 172         opt->max_depth = def->max_depth;
 173         opt->pathname = def->pathname;
 174         opt->relative = def->relative;
 175         opt->output = def->output;
 176
 177         for (i = 0; i < NR_GREP_COLORS; i++)
 178                 color_set(opt->colors[i], def->colors[i]);
 179 }
 180
 181 static void grep_set_pattern_type_option(enum grep_pattern_type pattern_type, struct grep_opt *opt)
 182 {
 183         /*
 184          * When committing to the pattern type by setting the relevant
 185          * fields in grep_opt it's generally not necessary to zero out
 186          * the fields we're not choosing, since they won't have been
 187          * set by anything. The extended_regexp_option field is the
 188          * only exception to this.
 189          *
 190          * This is because in the process of parsing grep.patternType
 191          * & grep.extendedRegexp we set opt->pattern_type_option and
 192          * opt->extended_regexp_option, respectively. We then
 193          * internally use opt->extended_regexp_option to see if we're
 194          * compiling an ERE. It must be unset if that's not actually
 195          * the case.
 196          */
 197         if (pattern_type != GREP_PATTERN_TYPE_ERE &&
 198             opt->extended_regexp_option)
 199                 opt->extended_regexp_option = 0;
 200
 201         switch (pattern_type) {
 202         case GREP_PATTERN_TYPE_UNSPECIFIED:
 203                 /* fall through */
 204
 205         case GREP_PATTERN_TYPE_BRE:
 206                 break;
 207
 208         case GREP_PATTERN_TYPE_ERE:
 209                 opt->extended_regexp_option = 1;
 210                 break;
 211
 212         case GREP_PATTERN_TYPE_FIXED:
 213                 opt->fixed = 1;
 214                 break;
 215
 216         case GREP_PATTERN_TYPE_PCRE:
 217 #ifdef USE_LIBPCRE2
 218                 opt->pcre2 = 1;
 219 #else
 220                 /*
 221                  * It's important that pcre1 always be assigned to
 222                  * even when there's no USE_LIBPCRE* defined. We still
 223                  * call the PCRE stub function, it just dies with
 224                  * "cannot use Perl-compatible regexes[...]".
 225                  */
 226                 opt->pcre1 = 1;
 227 #endif
 228                 break;
 229         }
 230 }
 231
 232 void grep_commit_pattern_type(enum grep_pattern_type pattern_type, struct grep_opt *opt)
 233 {
 234         if (pattern_type != GREP_PATTERN_TYPE_UNSPECIFIED)
 235                 grep_set_pattern_type_option(pattern_type, opt);
 236         else if (opt->pattern_type_option != GREP_PATTERN_TYPE_UNSPECIFIED)
 237                 grep_set_pattern_type_option(opt->pattern_type_option, opt);
 238         else if (opt->extended_regexp_option)
 239                 /*
 240                  * This branch *must* happen after setting from the
 241                  * opt->pattern_type_option above, we don't want
 242                  * grep.extendedRegexp to override grep.patternType!
 243                  */
 244                 grep_set_pattern_type_option(GREP_PATTERN_TYPE_ERE, opt);
 245 }
 246
 247 static struct grep_pat *create_grep_pat(const char *pat, size_t patlen,
 248                                         const char *origin, int no,
 249                                         enum grep_pat_token t,
 250                                         enum grep_header_field field)
 251 {
 252         struct grep_pat *p = xcalloc(1, sizeof(*p));
 253         p->pattern = xmemdupz(pat, patlen);
 254         p->patternlen = patlen;
 255         p->origin = origin;
 256         p->no = no;
 257         p->token = t;
 258         p->field = field;
 259         return p;
 260 }
 261
 262 static void do_append_grep_pat(struct grep_pat ***tail, struct grep_pat *p)
 263 {
 264         **tail = p;
 265         *tail = &p->next;
 266         p->next = NULL;
 267
 268         switch (p->token) {
 269         case GREP_PATTERN: /* atom */
 270         case GREP_PATTERN_HEAD:
 271         case GREP_PATTERN_BODY:
 272                 for (;;) {
 273                         struct grep_pat *new_pat;
 274                         size_t len = 0;
 275                         char *cp = p->pattern + p->patternlen, *nl = NULL;
 276                         while (++len <= p->patternlen) {
 277                                 if (*(--cp) == '\n') {
 278                                         nl = cp;
 279                                         break;
 280                                 }
 281                         }
 282                         if (!nl)
 283                                 break;
 284                         new_pat = create_grep_pat(nl + 1, len - 1, p->origin,
 285                                                   p->no, p->token, p->field);
 286                         new_pat->next = p->next;
 287                         if (!p->next)
 288                                 *tail = &new_pat->next;
 289                         p->next = new_pat;
 290                         *nl = '\0';
 291                         p->patternlen -= len;
 292                 }
 293                 break;
 294         default:
 295                 break;
 296         }
 297 }
 298
 299 void append_header_grep_pattern(struct grep_opt *opt,
 300                                 enum grep_header_field field, const char *pat)
 301 {
 302         struct grep_pat *p = create_grep_pat(pat, strlen(pat), "header", 0,
 303                                              GREP_PATTERN_HEAD, field);
 304         if (field == GREP_HEADER_REFLOG)
 305                 opt->use_reflog_filter = 1;
 306         do_append_grep_pat(&opt->header_tail, p);
 307 }
 308
 309 void append_grep_pattern(struct grep_opt *opt, const char *pat,
 310                          const char *origin, int no, enum grep_pat_token t)
 311 {
 312         append_grep_pat(opt, pat, strlen(pat), origin, no, t);
 313 }
 314
 315 void append_grep_pat(struct grep_opt *opt, const char *pat, size_t patlen,
 316                      const char *origin, int no, enum grep_pat_token t)
 317 {
 318         struct grep_pat *p = create_grep_pat(pat, patlen, origin, no, t, 0);
 319         do_append_grep_pat(&opt->pattern_tail, p);
 320 }
 321
 322 struct grep_opt *grep_opt_dup(const struct grep_opt *opt)
 323 {
 324         struct grep_pat *pat;
 325         struct grep_opt *ret = xmalloc(sizeof(struct grep_opt));
 326         *ret = *opt;
 327
 328         ret->pattern_list = NULL;
 329         ret->pattern_tail = &ret->pattern_list;
 330
 331         for(pat = opt->pattern_list; pat != NULL; pat = pat->next)
 332         {
 333                 if(pat->token == GREP_PATTERN_HEAD)
 334                         append_header_grep_pattern(ret, pat->field,
 335                                                    pat->pattern);
 336                 else
 337                         append_grep_pat(ret, pat->pattern, pat->patternlen,
 338                                         pat->origin, pat->no, pat->token);
 339         }
 340
 341         return ret;
 342 }
 343
 344 static NORETURN void compile_regexp_failed(const struct grep_pat *p,
 345                 const char *error)
 346 {
 347         char where[1024];
 348
 349         if (p->no)
 350                 xsnprintf(where, sizeof(where), "In '%s' at %d, ", p->origin, p->no);
 351         else if (p->origin)
 352                 xsnprintf(where, sizeof(where), "%s, ", p->origin);
 353         else
 354                 where[0] = 0;
 355
 356         die("%s'%s': %s", where, p->pattern, error);
 357 }
 358
 359 static int is_fixed(const char *s, size_t len)
 360 {
 361         size_t i;
 362
 363         for (i = 0; i < len; i++) {
 364                 if (is_regex_special(s[i]))
 365                         return 0;
 366         }
 367
 368         return 1;
 369 }
 370
 371 #ifdef USE_LIBPCRE1
 372 static void compile_pcre1_regexp(struct grep_pat *p, const struct grep_opt *opt)
 373 {
 374         const char *error;
 375         int erroffset;
 376         int options = PCRE_MULTILINE;
 377
 378         if (opt->ignore_case) {
 379                 if (!opt->ignore_locale && has_non_ascii(p->pattern))
 380                         p->pcre1_tables = pcre_maketables();
 381                 options |= PCRE_CASELESS;
 382         }
 383         if (!opt->ignore_locale && is_utf8_locale() && has_non_ascii(p->pattern))
 384                 options |= PCRE_UTF8;
 385
 386         p->pcre1_regexp = pcre_compile(p->pattern, options, &error, &erroffset,
 387                                       p->pcre1_tables);
 388         if (!p->pcre1_regexp)
 389                 compile_regexp_failed(p, error);
 390
 391         p->pcre1_extra_info = pcre_study(p->pcre1_regexp, GIT_PCRE_STUDY_JIT_COMPILE, &error);
 392         if (!p->pcre1_extra_info && error)
 393                 die("%s", error);
 394
 395 #ifdef GIT_PCRE1_USE_JIT
 396         pcre_config(PCRE_CONFIG_JIT, &p->pcre1_jit_on);
 397 #endif
 398 }
 399
 400 static int pcre1match(struct grep_pat *p, const char *line, const char *eol,
 401                 regmatch_t *match, int eflags)
 402 {
 403         int ovector[30], ret, flags = 0;
 404
 405         if (eflags & REG_NOTBOL)
 406                 flags |= PCRE_NOTBOL;
 407
 408         ret = pcre_exec(p->pcre1_regexp, p->pcre1_extra_info, line,
 409                         eol - line, 0, flags, ovector,
 410                         ARRAY_SIZE(ovector));
 411
 412         if (ret < 0 && ret != PCRE_ERROR_NOMATCH)
 413                 die("pcre_exec failed with error code %d", ret);
 414         if (ret > 0) {
 415                 ret = 0;
 416                 match->rm_so = ovector[0];
 417                 match->rm_eo = ovector[1];
 418         }
 419
 420         return ret;
 421 }
 422
 423 static void free_pcre1_regexp(struct grep_pat *p)
 424 {
 425         pcre_free(p->pcre1_regexp);
 426 #ifdef GIT_PCRE1_USE_JIT
 427         if (p->pcre1_jit_on)
 428                 pcre_free_study(p->pcre1_extra_info);
 429         else
 430 #endif
 431                 pcre_free(p->pcre1_extra_info);
 432         pcre_free((void *)p->pcre1_tables);
 433 }
 434 #else /* !USE_LIBPCRE1 */
 435 static void compile_pcre1_regexp(struct grep_pat *p, const struct grep_opt *opt)
 436 {
 437         die("cannot use Perl-compatible regexes when not compiled with USE_LIBPCRE");
 438 }
 439
 440 static int pcre1match(struct grep_pat *p, const char *line, const char *eol,
 441                 regmatch_t *match, int eflags)
 442 {
 443         return 1;
 444 }
 445
 446 static void free_pcre1_regexp(struct grep_pat *p)
 447 {
 448 }
 449 #endif /* !USE_LIBPCRE1 */
 450
 451 #ifdef USE_LIBPCRE2
 452 static void compile_pcre2_pattern(struct grep_pat *p, const struct grep_opt *opt)
 453 {
 454         int error;
 455         PCRE2_UCHAR errbuf[256];
 456         PCRE2_SIZE erroffset;
 457         int options = PCRE2_MULTILINE;
 458         const uint8_t *character_tables = NULL;
 459         int jitret;
 460         int patinforet;
 461         size_t jitsizearg;
 462
 463         assert(opt->pcre2);
 464
 465         p->pcre2_compile_context = NULL;
 466
 467         if (opt->ignore_case) {
 468                 if (!opt->ignore_locale && has_non_ascii(p->pattern)) {
 469                         character_tables = pcre2_maketables(NULL);
 470                         p->pcre2_compile_context = pcre2_compile_context_create(NULL);
 471                         pcre2_set_character_tables(p->pcre2_compile_context, character_tables);
 472                 }
 473                 options |= PCRE2_CASELESS;
 474         }
 475         if (!opt->ignore_locale && is_utf8_locale() && has_non_ascii(p->pattern))
 476                 options |= PCRE2_UTF;
 477
 478         p->pcre2_pattern = pcre2_compile((PCRE2_SPTR)p->pattern,
 479                                          p->patternlen, options, &error, &erroffset,
 480                                          p->pcre2_compile_context);
 481
 482         if (p->pcre2_pattern) {
 483                 p->pcre2_match_data = pcre2_match_data_create_from_pattern(p->pcre2_pattern, NULL);
 484                 if (!p->pcre2_match_data)
 485                         die("Couldn't allocate PCRE2 match data");
 486         } else {
 487                 pcre2_get_error_message(error, errbuf, sizeof(errbuf));
 488                 compile_regexp_failed(p, (const char *)&errbuf);
 489         }
 490
 491         pcre2_config(PCRE2_CONFIG_JIT, &p->pcre2_jit_on);
 492         if (p->pcre2_jit_on) {
 493                 jitret = pcre2_jit_compile(p->pcre2_pattern, PCRE2_JIT_COMPLETE);
 494                 if (jitret)
 495                         die("Couldn't JIT the PCRE2 pattern '%s', got '%d'\n", p->pattern, jitret);
 496
 497                 /*
 498                  * The pcre2_config(PCRE2_CONFIG_JIT, ...) call just
 499                  * tells us whether the library itself supports JIT,
 500                  * but to see whether we're going to be actually using
 501                  * JIT we need to extract PCRE2_INFO_JITSIZE from the
 502                  * pattern *after* we do pcre2_jit_compile() above.
 503                  *
 504                  * This is because if the pattern contains the
 505                  * (*NO_JIT) verb (see pcre2syntax(3))
 506                  * pcre2_jit_compile() will exit early with 0. If we
 507                  * then proceed to call pcre2_jit_match() further down
 508                  * the line instead of pcre2_match() we'll either
 509                  * segfault (pre PCRE 10.31) or run into a fatal error
 510                  * (post PCRE2 10.31)
 511                  */
 512                 patinforet = pcre2_pattern_info(p->pcre2_pattern, PCRE2_INFO_JITSIZE, &jitsizearg);
 513                 if (patinforet)
 514                         BUG("pcre2_pattern_info() failed: %d", patinforet);
 515                 if (jitsizearg == 0) {
 516                         p->pcre2_jit_on = 0;
 517                         return;
 518                 }
 519         }
 520 }
 521
 522 static int pcre2match(struct grep_pat *p, const char *line, const char *eol,
 523                 regmatch_t *match, int eflags)
 524 {
 525         int ret, flags = 0;
 526         PCRE2_SIZE *ovector;
 527         PCRE2_UCHAR errbuf[256];
 528
 529         if (eflags & REG_NOTBOL)
 530                 flags |= PCRE2_NOTBOL;
 531
 532         if (p->pcre2_jit_on)
 533                 ret = pcre2_jit_match(p->pcre2_pattern, (unsigned char *)line,
 534                                       eol - line, 0, flags, p->pcre2_match_data,
 535                                       NULL);
 536         else
 537                 ret = pcre2_match(p->pcre2_pattern, (unsigned char *)line,
 538                                   eol - line, 0, flags, p->pcre2_match_data,
 539                                   NULL);
 540
 541         if (ret < 0 && ret != PCRE2_ERROR_NOMATCH) {
 542                 pcre2_get_error_message(ret, errbuf, sizeof(errbuf));
 543                 die("%s failed with error code %d: %s",
 544                     (p->pcre2_jit_on ? "pcre2_jit_match" : "pcre2_match"), ret,
 545                     errbuf);
 546         }
 547         if (ret > 0) {
 548                 ovector = pcre2_get_ovector_pointer(p->pcre2_match_data);
 549                 ret = 0;
 550                 match->rm_so = (int)ovector[0];
 551                 match->rm_eo = (int)ovector[1];
 552         }
 553
 554         return ret;
 555 }
 556
 557 static void free_pcre2_pattern(struct grep_pat *p)
 558 {
 559         pcre2_compile_context_free(p->pcre2_compile_context);
 560         pcre2_code_free(p->pcre2_pattern);
 561         pcre2_match_data_free(p->pcre2_match_data);
 562 }
 563 #else /* !USE_LIBPCRE2 */
 564 static void compile_pcre2_pattern(struct grep_pat *p, const struct grep_opt *opt)
 565 {
 566         /*
 567          * Unreachable until USE_LIBPCRE2 becomes synonymous with
 568          * USE_LIBPCRE. See the sibling comment in
 569          * grep_set_pattern_type_option().
 570          */
 571         die("cannot use Perl-compatible regexes when not compiled with USE_LIBPCRE");
 572 }
 573
 574 static int pcre2match(struct grep_pat *p, const char *line, const char *eol,
 575                 regmatch_t *match, int eflags)
 576 {
 577         return 1;
 578 }
 579
 580 static void free_pcre2_pattern(struct grep_pat *p)
 581 {
 582 }
 583
 584 static void compile_fixed_regexp(struct grep_pat *p, struct grep_opt *opt)
 585 {
 586         struct strbuf sb = STRBUF_INIT;
 587         int err;
 588         int regflags = 0;
 589
 590         basic_regex_quote_buf(&sb, p->pattern);
 591         if (opt->ignore_case)
 592                 regflags |= REG_ICASE;
 593         err = regcomp(&p->regexp, sb.buf, regflags);
 594         if (opt->debug)
 595                 fprintf(stderr, "fixed %s\n", sb.buf);
 596         strbuf_release(&sb);
 597         if (err) {
 598                 char errbuf[1024];
 599                 regerror(err, &p->regexp, errbuf, sizeof(errbuf));
 600                 compile_regexp_failed(p, errbuf);
 601         }
 602 }
 603 #endif /* !USE_LIBPCRE2 */
 604
 605 static void compile_regexp(struct grep_pat *p, struct grep_opt *opt)
 606 {
 607         int err;
 608         int regflags = REG_NEWLINE;
 609         int pat_is_fixed;
 610
 611         p->word_regexp = opt->word_regexp;
 612         p->ignore_case = opt->ignore_case;
 613         p->fixed = opt->fixed;
 614
 615         if (memchr(p->pattern, 0, p->patternlen) && !opt->pcre2)
 616                 die(_("given pattern contains NULL byte (via -f <file>). This is only supported with -P under PCRE v2"));
 617
 618         pat_is_fixed = is_fixed(p->pattern, p->patternlen);
 619         if (p->fixed || pat_is_fixed) {
 620 #ifdef USE_LIBPCRE2
 621                 opt->pcre2 = 1;
 622                 if (pat_is_fixed) {
 623                         compile_pcre2_pattern(p, opt);
 624                 } else {
 625                         /*
 626                          * E.g. t7811-grep-open.sh relies on the
 627                          * pattern being restored.
 628                          */
 629                         char *old_pattern = p->pattern;
 630                         size_t old_patternlen = p->patternlen;
 631                         struct strbuf sb = STRBUF_INIT;
 632
 633                         /*
 634                          * There is the PCRE2_LITERAL flag, but it's
 635                          * only in PCRE v2 10.30 and later. Needing to
 636                          * ifdef our way around that and dealing with
 637                          * it + PCRE2_MULTILINE being an error is more
 638                          * complex than just quoting this ourselves.
 639                         */
 640                         strbuf_add(&sb, "\\Q", 2);
 641                         strbuf_add(&sb, p->pattern, p->patternlen);
 642                         strbuf_add(&sb, "\\E", 2);
 643
 644                         p->pattern = sb.buf;
 645                         p->patternlen = sb.len;
 646                         compile_pcre2_pattern(p, opt);
 647                         p->pattern = old_pattern;
 648                         p->patternlen = old_patternlen;
 649                         strbuf_release(&sb);
 650                 }
 651 #else /* !USE_LIBPCRE2 */
 652                 compile_fixed_regexp(p, opt);
 653 #endif /* !USE_LIBPCRE2 */
 654                 return;
 655         }
 656
 657         if (opt->pcre2) {
 658                 compile_pcre2_pattern(p, opt);
 659                 return;
 660         }
 661
 662         if (opt->pcre1) {
 663                 compile_pcre1_regexp(p, opt);
 664                 return;
 665         }
 666
 667         if (p->ignore_case)
 668                 regflags |= REG_ICASE;
 669         if (opt->extended_regexp_option)
 670                 regflags |= REG_EXTENDED;
 671         err = regcomp(&p->regexp, p->pattern, regflags);
 672         if (err) {
 673                 char errbuf[1024];
 674                 regerror(err, &p->regexp, errbuf, 1024);
 675                 compile_regexp_failed(p, errbuf);
 676         }
 677 }
 678
 679 static struct grep_expr *compile_pattern_or(struct grep_pat **);
 680 static struct grep_expr *compile_pattern_atom(struct grep_pat **list)
 681 {
 682         struct grep_pat *p;
 683         struct grep_expr *x;
 684
 685         p = *list;
 686         if (!p)
 687                 return NULL;
 688         switch (p->token) {
 689         case GREP_PATTERN: /* atom */
 690         case GREP_PATTERN_HEAD:
 691         case GREP_PATTERN_BODY:
 692                 x = xcalloc(1, sizeof (struct grep_expr));
 693                 x->node = GREP_NODE_ATOM;
 694                 x->u.atom = p;
 695                 *list = p->next;
 696                 return x;
 697         case GREP_OPEN_PAREN:
 698                 *list = p->next;
 699                 x = compile_pattern_or(list);
 700                 if (!*list || (*list)->token != GREP_CLOSE_PAREN)
 701                         die("unmatched parenthesis");
 702                 *list = (*list)->next;
 703                 return x;
 704         default:
 705                 return NULL;
 706         }
 707 }
 708
 709 static struct grep_expr *compile_pattern_not(struct grep_pat **list)
 710 {
 711         struct grep_pat *p;
 712         struct grep_expr *x;
 713
 714         p = *list;
 715         if (!p)
 716                 return NULL;
 717         switch (p->token) {
 718         case GREP_NOT:
 719                 if (!p->next)
 720                         die("--not not followed by pattern expression");
 721                 *list = p->next;
 722                 x = xcalloc(1, sizeof (struct grep_expr));
 723                 x->node = GREP_NODE_NOT;
 724                 x->u.unary = compile_pattern_not(list);
 725                 if (!x->u.unary)
 726                         die("--not followed by non pattern expression");
 727                 return x;
 728         default:
 729                 return compile_pattern_atom(list);
 730         }
 731 }
 732
 733 static struct grep_expr *compile_pattern_and(struct grep_pat **list)
 734 {
 735         struct grep_pat *p;
 736         struct grep_expr *x, *y, *z;
 737
 738         x = compile_pattern_not(list);
 739         p = *list;
 740         if (p && p->token == GREP_AND) {
 741                 if (!p->next)
 742                         die("--and not followed by pattern expression");
 743                 *list = p->next;
 744                 y = compile_pattern_and(list);
 745                 if (!y)
 746                         die("--and not followed by pattern expression");
 747                 z = xcalloc(1, sizeof (struct grep_expr));
 748                 z->node = GREP_NODE_AND;
 749                 z->u.binary.left = x;
 750                 z->u.binary.right = y;
 751                 return z;
 752         }
 753         return x;
 754 }
 755
 756 static struct grep_expr *compile_pattern_or(struct grep_pat **list)
 757 {
 758         struct grep_pat *p;
 759         struct grep_expr *x, *y, *z;
 760
 761         x = compile_pattern_and(list);
 762         p = *list;
 763         if (x && p && p->token != GREP_CLOSE_PAREN) {
 764                 y = compile_pattern_or(list);
 765                 if (!y)
 766                         die("not a pattern expression %s", p->pattern);
 767                 z = xcalloc(1, sizeof (struct grep_expr));
 768                 z->node = GREP_NODE_OR;
 769                 z->u.binary.left = x;
 770                 z->u.binary.right = y;
 771                 return z;
 772         }
 773         return x;
 774 }
 775
 776 static struct grep_expr *compile_pattern_expr(struct grep_pat **list)
 777 {
 778         return compile_pattern_or(list);
 779 }
 780
 781 static void indent(int in)
 782 {
 783         while (in-- > 0)
 784                 fputc(' ', stderr);
 785 }
 786
 787 static void dump_grep_pat(struct grep_pat *p)
 788 {
 789         switch (p->token) {
 790         case GREP_AND: fprintf(stderr, "*and*"); break;
 791         case GREP_OPEN_PAREN: fprintf(stderr, "*(*"); break;
 792         case GREP_CLOSE_PAREN: fprintf(stderr, "*)*"); break;
 793         case GREP_NOT: fprintf(stderr, "*not*"); break;
 794         case GREP_OR: fprintf(stderr, "*or*"); break;
 795
 796         case GREP_PATTERN: fprintf(stderr, "pattern"); break;
 797         case GREP_PATTERN_HEAD: fprintf(stderr, "pattern_head"); break;
 798         case GREP_PATTERN_BODY: fprintf(stderr, "pattern_body"); break;
 799         }
 800
 801         switch (p->token) {
 802         default: break;
 803         case GREP_PATTERN_HEAD:
 804                 fprintf(stderr, "<head %d>", p->field); break;
 805         case GREP_PATTERN_BODY:
 806                 fprintf(stderr, "<body>"); break;
 807         }
 808         switch (p->token) {
 809         default: break;
 810         case GREP_PATTERN_HEAD:
 811         case GREP_PATTERN_BODY:
 812         case GREP_PATTERN:
 813                 fprintf(stderr, "%.*s", (int)p->patternlen, p->pattern);
 814                 break;
 815         }
 816         fputc('\n', stderr);
 817 }
 818
 819 static void dump_grep_expression_1(struct grep_expr *x, int in)
 820 {
 821         indent(in);
 822         switch (x->node) {
 823         case GREP_NODE_TRUE:
 824                 fprintf(stderr, "true\n");
 825                 break;
 826         case GREP_NODE_ATOM:
 827                 dump_grep_pat(x->u.atom);
 828                 break;
 829         case GREP_NODE_NOT:
 830                 fprintf(stderr, "(not\n");
 831                 dump_grep_expression_1(x->u.unary, in+1);
 832                 indent(in);
 833                 fprintf(stderr, ")\n");
 834                 break;
 835         case GREP_NODE_AND:
 836                 fprintf(stderr, "(and\n");
 837                 dump_grep_expression_1(x->u.binary.left, in+1);
 838                 dump_grep_expression_1(x->u.binary.right, in+1);
 839                 indent(in);
 840                 fprintf(stderr, ")\n");
 841                 break;
 842         case GREP_NODE_OR:
 843                 fprintf(stderr, "(or\n");
 844                 dump_grep_expression_1(x->u.binary.left, in+1);
 845                 dump_grep_expression_1(x->u.binary.right, in+1);
 846                 indent(in);
 847                 fprintf(stderr, ")\n");
 848                 break;
 849         }
 850 }
 851
 852 static void dump_grep_expression(struct grep_opt *opt)
 853 {
 854         struct grep_expr *x = opt->pattern_expression;
 855
 856         if (opt->all_match)
 857                 fprintf(stderr, "[all-match]\n");
 858         dump_grep_expression_1(x, 0);
 859         fflush(NULL);
 860 }
 861
 862 static struct grep_expr *grep_true_expr(void)
 863 {
 864         struct grep_expr *z = xcalloc(1, sizeof(*z));
 865         z->node = GREP_NODE_TRUE;
 866         return z;
 867 }
 868
 869 static struct grep_expr *grep_or_expr(struct grep_expr *left, struct grep_expr *right)
 870 {
 871         struct grep_expr *z = xcalloc(1, sizeof(*z));
 872         z->node = GREP_NODE_OR;
 873         z->u.binary.left = left;
 874         z->u.binary.right = right;
 875         return z;
 876 }
 877
 878 static struct grep_expr *prep_header_patterns(struct grep_opt *opt)
 879 {
 880         struct grep_pat *p;
 881         struct grep_expr *header_expr;
 882         struct grep_expr *(header_group[GREP_HEADER_FIELD_MAX]);
 883         enum grep_header_field fld;
 884
 885         if (!opt->header_list)
 886                 return NULL;
 887
 888         for (p = opt->header_list; p; p = p->next) {
 889                 if (p->token != GREP_PATTERN_HEAD)
 890                         BUG("a non-header pattern in grep header list.");
 891                 if (p->field < GREP_HEADER_FIELD_MIN ||
 892                     GREP_HEADER_FIELD_MAX <= p->field)
 893                         BUG("unknown header field %d", p->field);
 894                 compile_regexp(p, opt);
 895         }
 896
 897         for (fld = 0; fld < GREP_HEADER_FIELD_MAX; fld++)
 898                 header_group[fld] = NULL;
 899
 900         for (p = opt->header_list; p; p = p->next) {
 901                 struct grep_expr *h;
 902                 struct grep_pat *pp = p;
 903
 904                 h = compile_pattern_atom(&pp);
 905                 if (!h || pp != p->next)
 906                         BUG("malformed header expr");
 907                 if (!header_group[p->field]) {
 908                         header_group[p->field] = h;
 909                         continue;
 910                 }
 911                 header_group[p->field] = grep_or_expr(h, header_group[p->field]);
 912         }
 913
 914         header_expr = NULL;
 915
 916         for (fld = 0; fld < GREP_HEADER_FIELD_MAX; fld++) {
 917                 if (!header_group[fld])
 918                         continue;
 919                 if (!header_expr)
 920                         header_expr = grep_true_expr();
 921                 header_expr = grep_or_expr(header_group[fld], header_expr);
 922         }
 923         return header_expr;
 924 }
 925
 926 static struct grep_expr *grep_splice_or(struct grep_expr *x, struct grep_expr *y)
 927 {
 928         struct grep_expr *z = x;
 929
 930         while (x) {
 931                 assert(x->node == GREP_NODE_OR);
 932                 if (x->u.binary.right &&
 933                     x->u.binary.right->node == GREP_NODE_TRUE) {
 934                         x->u.binary.right = y;
 935                         break;
 936                 }
 937                 x = x->u.binary.right;
 938         }
 939         return z;
 940 }
 941
 942 static void compile_grep_patterns_real(struct grep_opt *opt)
 943 {
 944         struct grep_pat *p;
 945         struct grep_expr *header_expr = prep_header_patterns(opt);
 946
 947         for (p = opt->pattern_list; p; p = p->next) {
 948                 switch (p->token) {
 949                 case GREP_PATTERN: /* atom */
 950                 case GREP_PATTERN_HEAD:
 951                 case GREP_PATTERN_BODY:
 952                         compile_regexp(p, opt);
 953                         break;
 954                 default:
 955                         opt->extended = 1;
 956                         break;
 957                 }
 958         }
 959
 960         if (opt->all_match || header_expr)
 961                 opt->extended = 1;
 962         else if (!opt->extended && !opt->debug)
 963                 return;
 964
 965         p = opt->pattern_list;
 966         if (p)
 967                 opt->pattern_expression = compile_pattern_expr(&p);
 968         if (p)
 969                 die("incomplete pattern expression: %s", p->pattern);
 970
 971         if (!header_expr)
 972                 return;
 973
 974         if (!opt->pattern_expression)
 975                 opt->pattern_expression = header_expr;
 976         else if (opt->all_match)
 977                 opt->pattern_expression = grep_splice_or(header_expr,
 978                                                          opt->pattern_expression);
 979         else
 980                 opt->pattern_expression = grep_or_expr(opt->pattern_expression,
 981                                                        header_expr);
 982         opt->all_match = 1;
 983 }
 984
 985 void compile_grep_patterns(struct grep_opt *opt)
 986 {
 987         compile_grep_patterns_real(opt);
 988         if (opt->debug)
 989                 dump_grep_expression(opt);
 990 }
 991
 992 static void free_pattern_expr(struct grep_expr *x)
 993 {
 994         switch (x->node) {
 995         case GREP_NODE_TRUE:
 996         case GREP_NODE_ATOM:
 997                 break;
 998         case GREP_NODE_NOT:
 999                 free_pattern_expr(x->u.unary);
1000                 break;
1001         case GREP_NODE_AND:
1002         case GREP_NODE_OR:
1003                 free_pattern_expr(x->u.binary.left);
1004                 free_pattern_expr(x->u.binary.right);
1005                 break;
1006         }
1007         free(x);
1008 }
1009
1010 void free_grep_patterns(struct grep_opt *opt)
1011 {
1012         struct grep_pat *p, *n;
1013
1014         for (p = opt->pattern_list; p; p = n) {
1015                 n = p->next;
1016                 switch (p->token) {
1017                 case GREP_PATTERN: /* atom */
1018                 case GREP_PATTERN_HEAD:
1019                 case GREP_PATTERN_BODY:
1020                         if (p->pcre1_regexp)
1021                                 free_pcre1_regexp(p);
1022                         else if (p->pcre2_pattern)
1023                                 free_pcre2_pattern(p);
1024                         else
1025                                 regfree(&p->regexp);
1026                         free(p->pattern);
1027                         break;
1028                 default:
1029                         break;
1030                 }
1031                 free(p);
1032         }
1033
1034         if (!opt->extended)
1035                 return;
1036         free_pattern_expr(opt->pattern_expression);
1037 }
1038
1039 static char *end_of_line(char *cp, unsigned long *left)
1040 {
1041         unsigned long l = *left;
1042         while (l && *cp != '\n') {
1043                 l--;
1044                 cp++;
1045         }
1046         *left = l;
1047         return cp;
1048 }
1049
1050 static int word_char(char ch)
1051 {
1052         return isalnum(ch) || ch == '_';
1053 }
1054
1055 static void output_color(struct grep_opt *opt, const void *data, size_t size,
1056                          const char *color)
1057 {
1058         if (want_color(opt->color) && color && color[0]) {
1059                 opt->output(opt, color, strlen(color));
1060                 opt->output(opt, data, size);
1061                 opt->output(opt, GIT_COLOR_RESET, strlen(GIT_COLOR_RESET));
1062         } else
1063                 opt->output(opt, data, size);
1064 }
1065
1066 static void output_sep(struct grep_opt *opt, char sign)
1067 {
1068         if (opt->null_following_name)
1069                 opt->output(opt, "\0", 1);
1070         else
1071                 output_color(opt, &sign, 1, opt->colors[GREP_COLOR_SEP]);
1072 }
1073
1074 static void show_name(struct grep_opt *opt, const char *name)
1075 {
1076         output_color(opt, name, strlen(name), opt->colors[GREP_COLOR_FILENAME]);
1077         opt->output(opt, opt->null_following_name ? "\0" : "\n", 1);
1078 }
1079
1080 static int patmatch(struct grep_pat *p, char *line, char *eol,
1081                     regmatch_t *match, int eflags)
1082 {
1083         int hit;
1084
1085         if (p->pcre1_regexp)
1086                 hit = !pcre1match(p, line, eol, match, eflags);
1087         else if (p->pcre2_pattern)
1088                 hit = !pcre2match(p, line, eol, match, eflags);
1089         else
1090                 hit = !regexec_buf(&p->regexp, line, eol - line, 1, match,
1091                                    eflags);
1092
1093         return hit;
1094 }
1095
1096 static int strip_timestamp(char *bol, char **eol_p)
1097 {
1098         char *eol = *eol_p;
1099         int ch;
1100
1101         while (bol < --eol) {
1102                 if (*eol != '>')
1103                         continue;
1104                 *eol_p = ++eol;
1105                 ch = *eol;
1106                 *eol = '\0';
1107                 return ch;
1108         }
1109         return 0;
1110 }
1111
1112 static struct {
1113         const char *field;
1114         size_t len;
1115 } header_field[] = {
1116         { "author ", 7 },
1117         { "committer ", 10 },
1118         { "reflog ", 7 },
1119 };
1120
1121 static int match_one_pattern(struct grep_pat *p, char *bol, char *eol,
1122                              enum grep_context ctx,
1123                              regmatch_t *pmatch, int eflags)
1124 {
1125         int hit = 0;
1126         int saved_ch = 0;
1127         const char *start = bol;
1128
1129         if ((p->token != GREP_PATTERN) &&
1130             ((p->token == GREP_PATTERN_HEAD) != (ctx == GREP_CONTEXT_HEAD)))
1131                 return 0;
1132
1133         if (p->token == GREP_PATTERN_HEAD) {
1134                 const char *field;
1135                 size_t len;
1136                 assert(p->field < ARRAY_SIZE(header_field));
1137                 field = header_field[p->field].field;
1138                 len = header_field[p->field].len;
1139                 if (strncmp(bol, field, len))
1140                         return 0;
1141                 bol += len;
1142                 switch (p->field) {
1143                 case GREP_HEADER_AUTHOR:
1144                 case GREP_HEADER_COMMITTER:
1145                         saved_ch = strip_timestamp(bol, &eol);
1146                         break;
1147                 default:
1148                         break;
1149                 }
1150         }
1151
1152  again:
1153         hit = patmatch(p, bol, eol, pmatch, eflags);
1154
1155         if (hit && p->word_regexp) {
1156                 if ((pmatch[0].rm_so < 0) ||
1157                     (eol - bol) < pmatch[0].rm_so ||
1158                     (pmatch[0].rm_eo < 0) ||
1159                     (eol - bol) < pmatch[0].rm_eo)
1160                         die("regexp returned nonsense");
1161
1162                 /* Match beginning must be either beginning of the
1163                  * line, or at word boundary (i.e. the last char must
1164                  * not be a word char).  Similarly, match end must be
1165                  * either end of the line, or at word boundary
1166                  * (i.e. the next char must not be a word char).
1167                  */
1168                 if ( ((pmatch[0].rm_so == 0) ||
1169                       !word_char(bol[pmatch[0].rm_so-1])) &&
1170                      ((pmatch[0].rm_eo == (eol-bol)) ||
1171                       !word_char(bol[pmatch[0].rm_eo])) )
1172                         ;
1173                 else
1174                         hit = 0;
1175
1176                 /* Words consist of at least one character. */
1177                 if (pmatch->rm_so == pmatch->rm_eo)
1178                         hit = 0;
1179
1180                 if (!hit && pmatch[0].rm_so + bol + 1 < eol) {
1181                         /* There could be more than one match on the
1182                          * line, and the first match might not be
1183                          * strict word match.  But later ones could be!
1184                          * Forward to the next possible start, i.e. the
1185                          * next position following a non-word char.
1186                          */
1187                         bol = pmatch[0].rm_so + bol + 1;
1188                         while (word_char(bol[-1]) && bol < eol)
1189                                 bol++;
1190                         eflags |= REG_NOTBOL;
1191                         if (bol < eol)
1192                                 goto again;
1193                 }
1194         }
1195         if (p->token == GREP_PATTERN_HEAD && saved_ch)
1196                 *eol = saved_ch;
1197         if (hit) {
1198                 pmatch[0].rm_so += bol - start;
1199                 pmatch[0].rm_eo += bol - start;
1200         }
1201         return hit;
1202 }
1203
1204 static int match_expr_eval(struct grep_opt *opt, struct grep_expr *x, char *bol,
1205                            char *eol, enum grep_context ctx, ssize_t *col,
1206                            ssize_t *icol, int collect_hits)
1207 {
1208         int h = 0;
1209
1210         if (!x)
1211                 die("Not a valid grep expression");
1212         switch (x->node) {
1213         case GREP_NODE_TRUE:
1214                 h = 1;
1215                 break;
1216         case GREP_NODE_ATOM:
1217                 {
1218                         regmatch_t tmp;
1219                         h = match_one_pattern(x->u.atom, bol, eol, ctx,
1220                                               &tmp, 0);
1221                         if (h && (*col < 0 || tmp.rm_so < *col))
1222                                 *col = tmp.rm_so;
1223                 }
1224                 break;
1225         case GREP_NODE_NOT:
1226                 /*
1227                  * Upon visiting a GREP_NODE_NOT, col and icol become swapped.
1228                  */
1229                 h = !match_expr_eval(opt, x->u.unary, bol, eol, ctx, icol, col,
1230                                      0);
1231                 break;
1232         case GREP_NODE_AND:
1233                 h = match_expr_eval(opt, x->u.binary.left, bol, eol, ctx, col,
1234                                     icol, 0);
1235                 if (h || opt->columnnum) {
1236                         /*
1237                          * Don't short-circuit AND when given --column, since a
1238                          * NOT earlier in the tree may turn this into an OR. In
1239                          * this case, see the below comment.
1240                          */
1241                         h &= match_expr_eval(opt, x->u.binary.right, bol, eol,
1242                                              ctx, col, icol, 0);
1243                 }
1244                 break;
1245         case GREP_NODE_OR:
1246                 if (!(collect_hits || opt->columnnum)) {
1247                         /*
1248                          * Don't short-circuit OR when given --column (or
1249                          * collecting hits) to ensure we don't skip a later
1250                          * child that would produce an earlier match.
1251                          */
1252                         return (match_expr_eval(opt, x->u.binary.left, bol, eol,
1253                                                 ctx, col, icol, 0) ||
1254                                 match_expr_eval(opt, x->u.binary.right, bol,
1255                                                 eol, ctx, col, icol, 0));
1256                 }
1257                 h = match_expr_eval(opt, x->u.binary.left, bol, eol, ctx, col,
1258                                     icol, 0);
1259                 if (collect_hits)
1260                         x->u.binary.left->hit |= h;
1261                 h |= match_expr_eval(opt, x->u.binary.right, bol, eol, ctx, col,
1262                                      icol, collect_hits);
1263                 break;
1264         default:
1265                 die("Unexpected node type (internal error) %d", x->node);
1266         }
1267         if (collect_hits)
1268                 x->hit |= h;
1269         return h;
1270 }
1271
1272 static int match_expr(struct grep_opt *opt, char *bol, char *eol,
1273                       enum grep_context ctx, ssize_t *col,
1274                       ssize_t *icol, int collect_hits)
1275 {
1276         struct grep_expr *x = opt->pattern_expression;
1277         return match_expr_eval(opt, x, bol, eol, ctx, col, icol, collect_hits);
1278 }
1279
1280 static int match_line(struct grep_opt *opt, char *bol, char *eol,
1281                       ssize_t *col, ssize_t *icol,
1282                       enum grep_context ctx, int collect_hits)
1283 {
1284         struct grep_pat *p;
1285         int hit = 0;
1286
1287         if (opt->extended)
1288                 return match_expr(opt, bol, eol, ctx, col, icol,
1289                                   collect_hits);
1290
1291         /* we do not call with collect_hits without being extended */
1292         for (p = opt->pattern_list; p; p = p->next) {
1293                 regmatch_t tmp;
1294                 if (match_one_pattern(p, bol, eol, ctx, &tmp, 0)) {
1295                         hit |= 1;
1296                         if (!opt->columnnum) {
1297                                 /*
1298                                  * Without --column, any single match on a line
1299                                  * is enough to know that it needs to be
1300                                  * printed. With --column, scan _all_ patterns
1301                                  * to find the earliest.
1302                                  */
1303                                 break;
1304                         }
1305                         if (*col < 0 || tmp.rm_so < *col)
1306                                 *col = tmp.rm_so;
1307                 }
1308         }
1309         return hit;
1310 }
1311
1312 static int match_next_pattern(struct grep_pat *p, char *bol, char *eol,
1313                               enum grep_context ctx,
1314                               regmatch_t *pmatch, int eflags)
1315 {
1316         regmatch_t match;
1317
1318         if (!match_one_pattern(p, bol, eol, ctx, &match, eflags))
1319                 return 0;
1320         if (match.rm_so < 0 || match.rm_eo < 0)
1321                 return 0;
1322         if (pmatch->rm_so >= 0 && pmatch->rm_eo >= 0) {
1323                 if (match.rm_so > pmatch->rm_so)
1324                         return 1;
1325                 if (match.rm_so == pmatch->rm_so && match.rm_eo < pmatch->rm_eo)
1326                         return 1;
1327         }
1328         pmatch->rm_so = match.rm_so;
1329         pmatch->rm_eo = match.rm_eo;
1330         return 1;
1331 }
1332
1333 static int next_match(struct grep_opt *opt, char *bol, char *eol,
1334                       enum grep_context ctx, regmatch_t *pmatch, int eflags)
1335 {
1336         struct grep_pat *p;
1337         int hit = 0;
1338
1339         pmatch->rm_so = pmatch->rm_eo = -1;
1340         if (bol < eol) {
1341                 for (p = opt->pattern_list; p; p = p->next) {
1342                         switch (p->token) {
1343                         case GREP_PATTERN: /* atom */
1344                         case GREP_PATTERN_HEAD:
1345                         case GREP_PATTERN_BODY:
1346                                 hit |= match_next_pattern(p, bol, eol, ctx,
1347                                                           pmatch, eflags);
1348                                 break;
1349                         default:
1350                                 break;
1351                         }
1352                 }
1353         }
1354         return hit;
1355 }
1356
1357 static void show_line_header(struct grep_opt *opt, const char *name,
1358                              unsigned lno, ssize_t cno, char sign)
1359 {
1360         if (opt->heading && opt->last_shown == 0) {
1361                 output_color(opt, name, strlen(name), opt->colors[GREP_COLOR_FILENAME]);
1362                 opt->output(opt, "\n", 1);
1363         }
1364         opt->last_shown = lno;
1365
1366         if (!opt->heading && opt->pathname) {
1367                 output_color(opt, name, strlen(name), opt->colors[GREP_COLOR_FILENAME]);
1368                 output_sep(opt, sign);
1369         }
1370         if (opt->linenum) {
1371                 char buf[32];
1372                 xsnprintf(buf, sizeof(buf), "%d", lno);
1373                 output_color(opt, buf, strlen(buf), opt->colors[GREP_COLOR_LINENO]);
1374                 output_sep(opt, sign);
1375         }
1376         /*
1377          * Treat 'cno' as the 1-indexed offset from the start of a non-context
1378          * line to its first match. Otherwise, 'cno' is 0 indicating that we are
1379          * being called with a context line.
1380          */
1381         if (opt->columnnum && cno) {
1382                 char buf[32];
1383                 xsnprintf(buf, sizeof(buf), "%"PRIuMAX, (uintmax_t)cno);
1384                 output_color(opt, buf, strlen(buf), opt->colors[GREP_COLOR_COLUMNNO]);
1385                 output_sep(opt, sign);
1386         }
1387 }
1388
1389 static void show_line(struct grep_opt *opt, char *bol, char *eol,
1390                       const char *name, unsigned lno, ssize_t cno, char sign)
1391 {
1392         int rest = eol - bol;
1393         const char *match_color = NULL;
1394         const char *line_color = NULL;
1395
1396         if (opt->file_break && opt->last_shown == 0) {
1397                 if (opt->show_hunk_mark)
1398                         opt->output(opt, "\n", 1);
1399         } else if (opt->pre_context || opt->post_context || opt->funcbody) {
1400                 if (opt->last_shown == 0) {
1401                         if (opt->show_hunk_mark) {
1402                                 output_color(opt, "--", 2, opt->colors[GREP_COLOR_SEP]);
1403                                 opt->output(opt, "\n", 1);
1404                         }
1405                 } else if (lno > opt->last_shown + 1) {
1406                         output_color(opt, "--", 2, opt->colors[GREP_COLOR_SEP]);
1407                         opt->output(opt, "\n", 1);
1408                 }
1409         }
1410         if (!opt->only_matching) {
1411                 /*
1412                  * In case the line we're being called with contains more than
1413                  * one match, leave printing each header to the loop below.
1414                  */
1415                 show_line_header(opt, name, lno, cno, sign);
1416         }
1417         if (opt->color || opt->only_matching) {
1418                 regmatch_t match;
1419                 enum grep_context ctx = GREP_CONTEXT_BODY;
1420                 int ch = *eol;
1421                 int eflags = 0;
1422
1423                 if (opt->color) {
1424                         if (sign == ':')
1425                                 match_color = opt->colors[GREP_COLOR_MATCH_SELECTED];
1426                         else
1427                                 match_color = opt->colors[GREP_COLOR_MATCH_CONTEXT];
1428                         if (sign == ':')
1429                                 line_color = opt->colors[GREP_COLOR_SELECTED];
1430                         else if (sign == '-')
1431                                 line_color = opt->colors[GREP_COLOR_CONTEXT];
1432                         else if (sign == '=')
1433                                 line_color = opt->colors[GREP_COLOR_FUNCTION];
1434                 }
1435                 *eol = '\0';
1436                 while (next_match(opt, bol, eol, ctx, &match, eflags)) {
1437                         if (match.rm_so == match.rm_eo)
1438                                 break;
1439
1440                         if (opt->only_matching)
1441                                 show_line_header(opt, name, lno, cno, sign);
1442                         else
1443                                 output_color(opt, bol, match.rm_so, line_color);
1444                         output_color(opt, bol + match.rm_so,
1445                                      match.rm_eo - match.rm_so, match_color);
1446                         if (opt->only_matching)
1447                                 opt->output(opt, "\n", 1);
1448                         bol += match.rm_eo;
1449                         cno += match.rm_eo;
1450                         rest -= match.rm_eo;
1451                         eflags = REG_NOTBOL;
1452                 }
1453                 *eol = ch;
1454         }
1455         if (!opt->only_matching) {
1456                 output_color(opt, bol, rest, line_color);
1457                 opt->output(opt, "\n", 1);
1458         }
1459 }
1460
1461 int grep_use_locks;
1462
1463 /*
1464  * This lock protects access to the gitattributes machinery, which is
1465  * not thread-safe.
1466  */
1467 pthread_mutex_t grep_attr_mutex;
1468
1469 static inline void grep_attr_lock(void)
1470 {
1471         if (grep_use_locks)
1472                 pthread_mutex_lock(&grep_attr_mutex);
1473 }
1474
1475 static inline void grep_attr_unlock(void)
1476 {
1477         if (grep_use_locks)
1478                 pthread_mutex_unlock(&grep_attr_mutex);
1479 }
1480
1481 /*
1482  * Same as git_attr_mutex, but protecting the thread-unsafe object db access.
1483  */
1484 pthread_mutex_t grep_read_mutex;
1485
1486 static int match_funcname(struct grep_opt *opt, struct grep_source *gs, char *bol, char *eol)
1487 {
1488         xdemitconf_t *xecfg = opt->priv;
1489         if (xecfg && !xecfg->find_func) {
1490                 grep_source_load_driver(gs, opt->repo->index);
1491                 if (gs->driver->funcname.pattern) {
1492                         const struct userdiff_funcname *pe = &gs->driver->funcname;
1493                         xdiff_set_find_func(xecfg, pe->pattern, pe->cflags);
1494                 } else {
1495                         xecfg = opt->priv = NULL;
1496                 }
1497         }
1498
1499         if (xecfg) {
1500                 char buf[1];
1501                 return xecfg->find_func(bol, eol - bol, buf, 1,
1502                                         xecfg->find_func_priv) >= 0;
1503         }
1504
1505         if (bol == eol)
1506                 return 0;
1507         if (isalpha(*bol) || *bol == '_' || *bol == '$')
1508                 return 1;
1509         return 0;
1510 }
1511
1512 static void show_funcname_line(struct grep_opt *opt, struct grep_source *gs,
1513                                char *bol, unsigned lno)
1514 {
1515         while (bol > gs->buf) {
1516                 char *eol = --bol;
1517
1518                 while (bol > gs->buf && bol[-1] != '\n')
1519                         bol--;
1520                 lno--;
1521
1522                 if (lno <= opt->last_shown)
1523                         break;
1524
1525                 if (match_funcname(opt, gs, bol, eol)) {
1526                         show_line(opt, bol, eol, gs->name, lno, 0, '=');
1527                         break;
1528                 }
1529         }
1530 }
1531
1532 static int is_empty_line(const char *bol, const char *eol);
1533
1534 static void show_pre_context(struct grep_opt *opt, struct grep_source *gs,
1535                              char *bol, char *end, unsigned lno)
1536 {
1537         unsigned cur = lno, from = 1, funcname_lno = 0, orig_from;
1538         int funcname_needed = !!opt->funcname, comment_needed = 0;
1539
1540         if (opt->pre_context < lno)
1541                 from = lno - opt->pre_context;
1542         if (from <= opt->last_shown)
1543                 from = opt->last_shown + 1;
1544         orig_from = from;
1545         if (opt->funcbody) {
1546                 if (match_funcname(opt, gs, bol, end))
1547                         comment_needed = 1;
1548                 else
1549                         funcname_needed = 1;
1550                 from = opt->last_shown + 1;
1551         }
1552
1553         /* Rewind. */
1554         while (bol > gs->buf && cur > from) {
1555                 char *next_bol = bol;
1556                 char *eol = --bol;
1557
1558                 while (bol > gs->buf && bol[-1] != '\n')
1559                         bol--;
1560                 cur--;
1561                 if (comment_needed && (is_empty_line(bol, eol) ||
1562                                        match_funcname(opt, gs, bol, eol))) {
1563                         comment_needed = 0;
1564                         from = orig_from;
1565                         if (cur < from) {
1566                                 cur++;
1567                                 bol = next_bol;
1568                                 break;
1569                         }
1570                 }
1571                 if (funcname_needed && match_funcname(opt, gs, bol, eol)) {
1572                         funcname_lno = cur;
1573                         funcname_needed = 0;
1574                         if (opt->funcbody)
1575                                 comment_needed = 1;
1576                         else
1577                                 from = orig_from;
1578                 }
1579         }
1580
1581         /* We need to look even further back to find a function signature. */
1582         if (opt->funcname && funcname_needed)
1583                 show_funcname_line(opt, gs, bol, cur);
1584
1585         /* Back forward. */
1586         while (cur < lno) {
1587                 char *eol = bol, sign = (cur == funcname_lno) ? '=' : '-';
1588
1589                 while (*eol != '\n')
1590                         eol++;
1591                 show_line(opt, bol, eol, gs->name, cur, 0, sign);
1592                 bol = eol + 1;
1593                 cur++;
1594         }
1595 }
1596
1597 static int should_lookahead(struct grep_opt *opt)
1598 {
1599         struct grep_pat *p;
1600
1601         if (opt->extended)
1602                 return 0; /* punt for too complex stuff */
1603         if (opt->invert)
1604                 return 0;
1605         for (p = opt->pattern_list; p; p = p->next) {
1606                 if (p->token != GREP_PATTERN)
1607                         return 0; /* punt for "header only" and stuff */
1608         }
1609         return 1;
1610 }
1611
1612 static int look_ahead(struct grep_opt *opt,
1613                       unsigned long *left_p,
1614                       unsigned *lno_p,
1615                       char **bol_p)
1616 {
1617         unsigned lno = *lno_p;
1618         char *bol = *bol_p;
1619         struct grep_pat *p;
1620         char *sp, *last_bol;
1621         regoff_t earliest = -1;
1622
1623         for (p = opt->pattern_list; p; p = p->next) {
1624                 int hit;
1625                 regmatch_t m;
1626
1627                 hit = patmatch(p, bol, bol + *left_p, &m, 0);
1628                 if (!hit || m.rm_so < 0 || m.rm_eo < 0)
1629                         continue;
1630                 if (earliest < 0 || m.rm_so < earliest)
1631                         earliest = m.rm_so;
1632         }
1633
1634         if (earliest < 0) {
1635                 *bol_p = bol + *left_p;
1636                 *left_p = 0;
1637                 return 1;
1638         }
1639         for (sp = bol + earliest; bol < sp && sp[-1] != '\n'; sp--)
1640                 ; /* find the beginning of the line */
1641         last_bol = sp;
1642
1643         for (sp = bol; sp < last_bol; sp++) {
1644                 if (*sp == '\n')
1645                         lno++;
1646         }
1647         *left_p -= last_bol - bol;
1648         *bol_p = last_bol;
1649         *lno_p = lno;
1650         return 0;
1651 }
1652
1653 static int fill_textconv_grep(struct repository *r,
1654                               struct userdiff_driver *driver,
1655                               struct grep_source *gs)
1656 {
1657         struct diff_filespec *df;
1658         char *buf;
1659         size_t size;
1660
1661         if (!driver || !driver->textconv)
1662                 return grep_source_load(gs);
1663
1664         /*
1665          * The textconv interface is intimately tied to diff_filespecs, so we
1666          * have to pretend to be one. If we could unify the grep_source
1667          * and diff_filespec structs, this mess could just go away.
1668          */
1669         df = alloc_filespec(gs->path);
1670         switch (gs->type) {
1671         case GREP_SOURCE_OID:
1672                 fill_filespec(df, gs->identifier, 1, 0100644);
1673                 break;
1674         case GREP_SOURCE_FILE:
1675                 fill_filespec(df, &null_oid, 0, 0100644);
1676                 break;
1677         default:
1678                 BUG("attempt to textconv something without a path?");
1679         }
1680
1681         /*
1682          * fill_textconv is not remotely thread-safe; it may load objects
1683          * behind the scenes, and it modifies the global diff tempfile
1684          * structure.
1685          */
1686         grep_read_lock();
1687         size = fill_textconv(r, driver, df, &buf);
1688         grep_read_unlock();
1689         free_filespec(df);
1690
1691         /*
1692          * The normal fill_textconv usage by the diff machinery would just keep
1693          * the textconv'd buf separate from the diff_filespec. But much of the
1694          * grep code passes around a grep_source and assumes that its "buf"
1695          * pointer is the beginning of the thing we are searching. So let's
1696          * install our textconv'd version into the grep_source, taking care not
1697          * to leak any existing buffer.
1698          */
1699         grep_source_clear_data(gs);
1700         gs->buf = buf;
1701         gs->size = size;
1702
1703         return 0;
1704 }
1705
1706 static int is_empty_line(const char *bol, const char *eol)
1707 {
1708         while (bol < eol && isspace(*bol))
1709                 bol++;
1710         return bol == eol;
1711 }
1712
1713 static int grep_source_1(struct grep_opt *opt, struct grep_source *gs, int collect_hits)
1714 {
1715         char *bol;
1716         char *peek_bol = NULL;
1717         unsigned long left;
1718         unsigned lno = 1;
1719         unsigned last_hit = 0;
1720         int binary_match_only = 0;
1721         unsigned count = 0;
1722         int try_lookahead = 0;
1723         int show_function = 0;
1724         struct userdiff_driver *textconv = NULL;
1725         enum grep_context ctx = GREP_CONTEXT_HEAD;
1726         xdemitconf_t xecfg;
1727
1728         if (!opt->status_only && gs->name == NULL)
1729                 BUG("grep call which could print a name requires "
1730                     "grep_source.name be non-NULL");
1731
1732         if (!opt->output)
1733                 opt->output = std_output;
1734
1735         if (opt->pre_context || opt->post_context || opt->file_break ||
1736             opt->funcbody) {
1737                 /* Show hunk marks, except for the first file. */
1738                 if (opt->last_shown)
1739                         opt->show_hunk_mark = 1;
1740                 /*
1741                  * If we're using threads then we can't easily identify
1742                  * the first file.  Always put hunk marks in that case
1743                  * and skip the very first one later in work_done().
1744                  */
1745                 if (opt->output != std_output)
1746                         opt->show_hunk_mark = 1;
1747         }
1748         opt->last_shown = 0;
1749
1750         if (opt->allow_textconv) {
1751                 grep_source_load_driver(gs, opt->repo->index);
1752                 /*
1753                  * We might set up the shared textconv cache data here, which
1754                  * is not thread-safe.
1755                  */
1756                 grep_attr_lock();
1757                 textconv = userdiff_get_textconv(opt->repo, gs->driver);
1758                 grep_attr_unlock();
1759         }
1760
1761         /*
1762          * We know the result of a textconv is text, so we only have to care
1763          * about binary handling if we are not using it.
1764          */
1765         if (!textconv) {
1766                 switch (opt->binary) {
1767                 case GREP_BINARY_DEFAULT:
1768                         if (grep_source_is_binary(gs, opt->repo->index))
1769                                 binary_match_only = 1;
1770                         break;
1771                 case GREP_BINARY_NOMATCH:
1772                         if (grep_source_is_binary(gs, opt->repo->index))
1773                                 return 0; /* Assume unmatch */
1774                         break;
1775                 case GREP_BINARY_TEXT:
1776                         break;
1777                 default:
1778                         BUG("unknown binary handling mode");
1779                 }
1780         }
1781
1782         memset(&xecfg, 0, sizeof(xecfg));
1783         opt->priv = &xecfg;
1784
1785         try_lookahead = should_lookahead(opt);
1786
1787         if (fill_textconv_grep(opt->repo, textconv, gs) < 0)
1788                 return 0;
1789
1790         bol = gs->buf;
1791         left = gs->size;
1792         while (left) {
1793                 char *eol, ch;
1794                 int hit;
1795                 ssize_t cno;
1796                 ssize_t col = -1, icol = -1;
1797
1798                 /*
1799                  * look_ahead() skips quickly to the line that possibly
1800                  * has the next hit; don't call it if we need to do
1801                  * something more than just skipping the current line
1802                  * in response to an unmatch for the current line.  E.g.
1803                  * inside a post-context window, we will show the current
1804                  * line as a context around the previous hit when it
1805                  * doesn't hit.
1806                  */
1807                 if (try_lookahead
1808                     && !(last_hit
1809                          && (show_function ||
1810                              lno <= last_hit + opt->post_context))
1811                     && look_ahead(opt, &left, &lno, &bol))
1812                         break;
1813                 eol = end_of_line(bol, &left);
1814                 ch = *eol;
1815                 *eol = 0;
1816
1817                 if ((ctx == GREP_CONTEXT_HEAD) && (eol == bol))
1818                         ctx = GREP_CONTEXT_BODY;
1819
1820                 hit = match_line(opt, bol, eol, &col, &icol, ctx, collect_hits);
1821                 *eol = ch;
1822
1823                 if (collect_hits)
1824                         goto next_line;
1825
1826                 /* "grep -v -e foo -e bla" should list lines
1827                  * that do not have either, so inversion should
1828                  * be done outside.
1829                  */
1830                 if (opt->invert)
1831                         hit = !hit;
1832                 if (opt->unmatch_name_only) {
1833                         if (hit)
1834                                 return 0;
1835                         goto next_line;
1836                 }
1837                 if (hit) {
1838                         count++;
1839                         if (opt->status_only)
1840                                 return 1;
1841                         if (opt->name_only) {
1842                                 show_name(opt, gs->name);
1843                                 return 1;
1844                         }
1845                         if (opt->count)
1846                                 goto next_line;
1847                         if (binary_match_only) {
1848                                 opt->output(opt, "Binary file ", 12);
1849                                 output_color(opt, gs->name, strlen(gs->name),
1850                                              opt->colors[GREP_COLOR_FILENAME]);
1851                                 opt->output(opt, " matches\n", 9);
1852                                 return 1;
1853                         }
1854                         /* Hit at this line.  If we haven't shown the
1855                          * pre-context lines, we would need to show them.
1856                          */
1857                         if (opt->pre_context || opt->funcbody)
1858                                 show_pre_context(opt, gs, bol, eol, lno);
1859                         else if (opt->funcname)
1860                                 show_funcname_line(opt, gs, bol, lno);
1861                         cno = opt->invert ? icol : col;
1862                         if (cno < 0) {
1863                                 /*
1864                                  * A negative cno indicates that there was no
1865                                  * match on the line. We are thus inverted and
1866                                  * being asked to show all lines that _don't_
1867                                  * match a given expression. Therefore, set cno
1868                                  * to 0 to suggest the whole line matches.
1869                                  */
1870                                 cno = 0;
1871                         }
1872                         show_line(opt, bol, eol, gs->name, lno, cno + 1, ':');
1873                         last_hit = lno;
1874                         if (opt->funcbody)
1875                                 show_function = 1;
1876                         goto next_line;
1877                 }
1878                 if (show_function && (!peek_bol || peek_bol < bol)) {
1879                         unsigned long peek_left = left;
1880                         char *peek_eol = eol;
1881
1882                         /*
1883                          * Trailing empty lines are not interesting.
1884                          * Peek past them to see if they belong to the
1885                          * body of the current function.
1886                          */
1887                         peek_bol = bol;
1888                         while (is_empty_line(peek_bol, peek_eol)) {
1889                                 peek_bol = peek_eol + 1;
1890                                 peek_eol = end_of_line(peek_bol, &peek_left);
1891                         }
1892
1893                         if (match_funcname(opt, gs, peek_bol, peek_eol))
1894                                 show_function = 0;
1895                 }
1896                 if (show_function ||
1897                     (last_hit && lno <= last_hit + opt->post_context)) {
1898                         /* If the last hit is within the post context,
1899                          * we need to show this line.
1900                          */
1901                         show_line(opt, bol, eol, gs->name, lno, col + 1, '-');
1902                 }
1903
1904         next_line:
1905                 bol = eol + 1;
1906                 if (!left)
1907                         break;
1908                 left--;
1909                 lno++;
1910         }
1911
1912         if (collect_hits)
1913                 return 0;
1914
1915         if (opt->status_only)
1916                 return opt->unmatch_name_only;
1917         if (opt->unmatch_name_only) {
1918                 /* We did not see any hit, so we want to show this */
1919                 show_name(opt, gs->name);
1920                 return 1;
1921         }
1922
1923         xdiff_clear_find_func(&xecfg);
1924         opt->priv = NULL;
1925
1926         /* NEEDSWORK:
1927          * The real "grep -c foo *.c" gives many "bar.c:0" lines,
1928          * which feels mostly useless but sometimes useful.  Maybe
1929          * make it another option?  For now suppress them.
1930          */
1931         if (opt->count && count) {
1932                 char buf[32];
1933                 if (opt->pathname) {
1934                         output_color(opt, gs->name, strlen(gs->name),
1935                                      opt->colors[GREP_COLOR_FILENAME]);
1936                         output_sep(opt, ':');
1937                 }
1938                 xsnprintf(buf, sizeof(buf), "%u\n", count);
1939                 opt->output(opt, buf, strlen(buf));
1940                 return 1;
1941         }
1942         return !!last_hit;
1943 }
1944
1945 static void clr_hit_marker(struct grep_expr *x)
1946 {
1947         /* All-hit markers are meaningful only at the very top level
1948          * OR node.
1949          */
1950         while (1) {
1951                 x->hit = 0;
1952                 if (x->node != GREP_NODE_OR)
1953                         return;
1954                 x->u.binary.left->hit = 0;
1955                 x = x->u.binary.right;
1956         }
1957 }
1958
1959 static int chk_hit_marker(struct grep_expr *x)
1960 {
1961         /* Top level nodes have hit markers.  See if they all are hits */
1962         while (1) {
1963                 if (x->node != GREP_NODE_OR)
1964                         return x->hit;
1965                 if (!x->u.binary.left->hit)
1966                         return 0;
1967                 x = x->u.binary.right;
1968         }
1969 }
1970
1971 int grep_source(struct grep_opt *opt, struct grep_source *gs)
1972 {
1973         /*
1974          * we do not have to do the two-pass grep when we do not check
1975          * buffer-wide "all-match".
1976          */
1977         if (!opt->all_match)
1978                 return grep_source_1(opt, gs, 0);
1979
1980         /* Otherwise the toplevel "or" terms hit a bit differently.
1981          * We first clear hit markers from them.
1982          */
1983         clr_hit_marker(opt->pattern_expression);
1984         grep_source_1(opt, gs, 1);
1985
1986         if (!chk_hit_marker(opt->pattern_expression))
1987                 return 0;
1988
1989         return grep_source_1(opt, gs, 0);
1990 }
1991
1992 int grep_buffer(struct grep_opt *opt, char *buf, unsigned long size)
1993 {
1994         struct grep_source gs;
1995         int r;
1996
1997         grep_source_init(&gs, GREP_SOURCE_BUF, NULL, NULL, NULL);
1998         gs.buf = buf;
1999         gs.size = size;
2000
2001         r = grep_source(opt, &gs);
2002
2003         grep_source_clear(&gs);
2004         return r;
2005 }
2006
2007 void grep_source_init(struct grep_source *gs, enum grep_source_type type,
2008                       const char *name, const char *path,
2009                       const void *identifier)
2010 {
2011         gs->type = type;
2012         gs->name = xstrdup_or_null(name);
2013         gs->path = xstrdup_or_null(path);
2014         gs->buf = NULL;
2015         gs->size = 0;
2016         gs->driver = NULL;
2017
2018         switch (type) {
2019         case GREP_SOURCE_FILE:
2020                 gs->identifier = xstrdup(identifier);
2021                 break;
2022         case GREP_SOURCE_OID:
2023                 gs->identifier = oiddup(identifier);
2024                 break;
2025         case GREP_SOURCE_BUF:
2026                 gs->identifier = NULL;
2027                 break;
2028         }
2029 }
2030
2031 void grep_source_clear(struct grep_source *gs)
2032 {
2033         FREE_AND_NULL(gs->name);
2034         FREE_AND_NULL(gs->path);
2035         FREE_AND_NULL(gs->identifier);
2036         grep_source_clear_data(gs);
2037 }
2038
2039 void grep_source_clear_data(struct grep_source *gs)
2040 {
2041         switch (gs->type) {
2042         case GREP_SOURCE_FILE:
2043         case GREP_SOURCE_OID:
2044                 FREE_AND_NULL(gs->buf);
2045                 gs->size = 0;
2046                 break;
2047         case GREP_SOURCE_BUF:
2048                 /* leave user-provided buf intact */
2049                 break;
2050         }
2051 }
2052
2053 static int grep_source_load_oid(struct grep_source *gs)
2054 {
2055         enum object_type type;
2056
2057         grep_read_lock();
2058         gs->buf = read_object_file(gs->identifier, &type, &gs->size);
2059         grep_read_unlock();
2060
2061         if (!gs->buf)
2062                 return error(_("'%s': unable to read %s"),
2063                              gs->name,
2064                              oid_to_hex(gs->identifier));
2065         return 0;
2066 }
2067
2068 static int grep_source_load_file(struct grep_source *gs)
2069 {
2070         const char *filename = gs->identifier;
2071         struct stat st;
2072         char *data;
2073         size_t size;
2074         int i;
2075
2076         if (lstat(filename, &st) < 0) {
2077         err_ret:
2078                 if (errno != ENOENT)
2079                         error_errno(_("failed to stat '%s'"), filename);
2080                 return -1;
2081         }
2082         if (!S_ISREG(st.st_mode))
2083                 return -1;
2084         size = xsize_t(st.st_size);
2085         i = open(filename, O_RDONLY);
2086         if (i < 0)
2087                 goto err_ret;
2088         data = xmallocz(size);
2089         if (st.st_size != read_in_full(i, data, size)) {
2090                 error_errno(_("'%s': short read"), filename);
2091                 close(i);
2092                 free(data);
2093                 return -1;
2094         }
2095         close(i);
2096
2097         gs->buf = data;
2098         gs->size = size;
2099         return 0;
2100 }
2101
2102 static int grep_source_load(struct grep_source *gs)
2103 {
2104         if (gs->buf)
2105                 return 0;
2106
2107         switch (gs->type) {
2108         case GREP_SOURCE_FILE:
2109                 return grep_source_load_file(gs);
2110         case GREP_SOURCE_OID:
2111                 return grep_source_load_oid(gs);
2112         case GREP_SOURCE_BUF:
2113                 return gs->buf ? 0 : -1;
2114         }
2115         BUG("invalid grep_source type to load");
2116 }
2117
2118 void grep_source_load_driver(struct grep_source *gs,
2119                              struct index_state *istate)
2120 {
2121         if (gs->driver)
2122                 return;
2123
2124         grep_attr_lock();
2125         if (gs->path)
2126                 gs->driver = userdiff_find_by_path(istate, gs->path);
2127         if (!gs->driver)
2128                 gs->driver = userdiff_find_by_name("default");
2129         grep_attr_unlock();
2130 }
2131
2132 static int grep_source_is_binary(struct grep_source *gs,
2133                                  struct index_state *istate)
2134 {
2135         grep_source_load_driver(gs, istate);
2136         if (gs->driver->binary != -1)
2137                 return gs->driver->binary;
2138
2139         if (!grep_source_load(gs))
2140                 return buffer_is_binary(gs->buf, gs->size);
2141
2142         return 0;
2143 }