diffcore-pickaxe.c

   1 /*
   2  * Copyright (C) 2005 Junio C Hamano
   3  * Copyright (C) 2010 Google Inc.
   4  */
   5 #include "git-compat-util.h"
   6 #include "diff.h"
   7 #include "diffcore.h"
   8 #include "xdiff-interface.h"
   9 #include "kwset.h"
  10 #include "oidset.h"
  11 #include "pretty.h"
  12 #include "quote.h"
  13
  14 typedef int (*pickaxe_fn)(mmfile_t *one, mmfile_t *two,
  15                           struct diff_options *o,
  16                           regex_t *regexp, kwset_t kws);
  17
  18 struct diffgrep_cb {
  19         regex_t *regexp;
  20         int hit;
  21 };
  22
  23 static int diffgrep_consume(void *priv, char *line, unsigned long len)
  24 {
  25         struct diffgrep_cb *data = priv;
  26         regmatch_t regmatch;
  27
  28         if (line[0] != '+' && line[0] != '-')
  29                 return 0;
  30         if (data->hit)
  31                 BUG("Already matched in diffgrep_consume! Broken xdiff_emit_line_fn?");
  32         if (!regexec_buf(data->regexp, line + 1, len - 1, 1,
  33                          &regmatch, 0)) {
  34                 data->hit = 1;
  35                 return 1;
  36         }
  37         return 0;
  38 }
  39
  40 static int diff_grep(mmfile_t *one, mmfile_t *two,
  41                      struct diff_options *o,
  42                      regex_t *regexp, kwset_t kws UNUSED)
  43 {
  44         struct diffgrep_cb ecbdata;
  45         xpparam_t xpp;
  46         xdemitconf_t xecfg;
  47         int ret;
  48
  49         /*
  50          * We have both sides; need to run textual diff and see if
  51          * the pattern appears on added/deleted lines.
  52          */
  53         memset(&xpp, 0, sizeof(xpp));
  54         memset(&xecfg, 0, sizeof(xecfg));
  55         ecbdata.regexp = regexp;
  56         ecbdata.hit = 0;
  57         xecfg.flags = XDL_EMIT_NO_HUNK_HDR;
  58         xecfg.ctxlen = o->context;
  59         xecfg.interhunkctxlen = o->interhunkcontext;
  60
  61         /*
  62          * An xdiff error might be our "data->hit" from above. See the
  63          * comment for xdiff_emit_line_fn in xdiff-interface.h
  64          */
  65         ret = xdi_diff_outf(one, two, NULL, diffgrep_consume,
  66                             &ecbdata, &xpp, &xecfg);
  67         if (ecbdata.hit)
  68                 return 1;
  69         if (ret)
  70                 return ret;
  71         return 0;
  72 }
  73
  74 static unsigned int contains(mmfile_t *mf, regex_t *regexp, kwset_t kws,
  75                              unsigned int limit)
  76 {
  77         unsigned int cnt = 0;
  78         unsigned long sz = mf->size;
  79         const char *data = mf->ptr;
  80
  81         if (regexp) {
  82                 regmatch_t regmatch;
  83                 int flags = 0;
  84
  85                 while (sz &&
  86                        !regexec_buf(regexp, data, sz, 1, &regmatch, flags)) {
  87                         flags |= REG_NOTBOL;
  88                         data += regmatch.rm_eo;
  89                         sz -= regmatch.rm_eo;
  90                         if (sz && regmatch.rm_so == regmatch.rm_eo) {
  91                                 data++;
  92                                 sz--;
  93                         }
  94                         cnt++;
  95
  96                         if (limit && cnt == limit)
  97                                 return cnt;
  98                 }
  99
 100         } else { /* Classic exact string match */
 101                 while (sz) {
 102                         struct kwsmatch kwsm;
 103                         size_t offset = kwsexec(kws, data, sz, &kwsm);
 104                         if (offset == -1)
 105                                 break;
 106                         sz -= offset + kwsm.size[0];
 107                         data += offset + kwsm.size[0];
 108                         cnt++;
 109
 110                         if (limit && cnt == limit)
 111                                 return cnt;
 112                 }
 113         }
 114         return cnt;
 115 }
 116
 117 static int has_changes(mmfile_t *one, mmfile_t *two,
 118                        struct diff_options *o UNUSED,
 119                        regex_t *regexp, kwset_t kws)
 120 {
 121         unsigned int c1 = one ? contains(one, regexp, kws, 0) : 0;
 122         unsigned int c2 = two ? contains(two, regexp, kws, c1 + 1) : 0;
 123         return c1 != c2;
 124 }
 125
 126 static int pickaxe_match(struct diff_filepair *p, struct diff_options *o,
 127                          regex_t *regexp, kwset_t kws, pickaxe_fn fn)
 128 {
 129         struct userdiff_driver *textconv_one = NULL;
 130         struct userdiff_driver *textconv_two = NULL;
 131         mmfile_t mf1, mf2;
 132         int ret;
 133
 134         /* ignore unmerged */
 135         if (!DIFF_FILE_VALID(p->one) && !DIFF_FILE_VALID(p->two))
 136                 return 0;
 137
 138         if (o->objfind) {
 139                 return  (DIFF_FILE_VALID(p->one) &&
 140                          oidset_contains(o->objfind, &p->one->oid)) ||
 141                         (DIFF_FILE_VALID(p->two) &&
 142                          oidset_contains(o->objfind, &p->two->oid));
 143         }
 144
 145         if (o->flags.allow_textconv) {
 146                 textconv_one = get_textconv(o->repo, p->one);
 147                 textconv_two = get_textconv(o->repo, p->two);
 148         }
 149
 150         /*
 151          * If we have an unmodified pair, we know that the count will be the
 152          * same and don't even have to load the blobs. Unless textconv is in
 153          * play, _and_ we are using two different textconv filters (e.g.,
 154          * because a pair is an exact rename with different textconv attributes
 155          * for each side, which might generate different content).
 156          */
 157         if (textconv_one == textconv_two && diff_unmodified_pair(p))
 158                 return 0;
 159
 160         if ((o->pickaxe_opts & DIFF_PICKAXE_KIND_G) &&
 161             !o->flags.text &&
 162             ((!textconv_one && diff_filespec_is_binary(o->repo, p->one)) ||
 163              (!textconv_two && diff_filespec_is_binary(o->repo, p->two))))
 164                 return 0;
 165
 166         mf1.size = fill_textconv(o->repo, textconv_one, p->one, &mf1.ptr);
 167         mf2.size = fill_textconv(o->repo, textconv_two, p->two, &mf2.ptr);
 168
 169         ret = fn(&mf1, &mf2, o, regexp, kws);
 170
 171         if (textconv_one)
 172                 free(mf1.ptr);
 173         if (textconv_two)
 174                 free(mf2.ptr);
 175         diff_free_filespec_data(p->one);
 176         diff_free_filespec_data(p->two);
 177
 178         return ret;
 179 }
 180
 181 static void pickaxe(struct diff_queue_struct *q, struct diff_options *o,
 182                     regex_t *regexp, kwset_t kws, pickaxe_fn fn)
 183 {
 184         int i;
 185         struct diff_queue_struct outq;
 186
 187         DIFF_QUEUE_CLEAR(&outq);
 188
 189         if (o->pickaxe_opts & DIFF_PICKAXE_ALL) {
 190                 /* Showing the whole changeset if needle exists */
 191                 for (i = 0; i < q->nr; i++) {
 192                         struct diff_filepair *p = q->queue[i];
 193                         if (pickaxe_match(p, o, regexp, kws, fn))
 194                                 return; /* do not munge the queue */
 195                 }
 196
 197                 /*
 198                  * Otherwise we will clear the whole queue by copying
 199                  * the empty outq at the end of this function, but
 200                  * first clear the current entries in the queue.
 201                  */
 202                 for (i = 0; i < q->nr; i++)
 203                         diff_free_filepair(q->queue[i]);
 204         } else {
 205                 /* Showing only the filepairs that has the needle */
 206                 for (i = 0; i < q->nr; i++) {
 207                         struct diff_filepair *p = q->queue[i];
 208                         if (pickaxe_match(p, o, regexp, kws, fn))
 209                                 diff_q(&outq, p);
 210                         else
 211                                 diff_free_filepair(p);
 212                 }
 213         }
 214
 215         free(q->queue);
 216         *q = outq;
 217 }
 218
 219 static void regcomp_or_die(regex_t *regex, const char *needle, int cflags)
 220 {
 221         int err = regcomp(regex, needle, cflags);
 222         if (err) {
 223                 /* The POSIX.2 people are surely sick */
 224                 char errbuf[1024];
 225                 regerror(err, regex, errbuf, 1024);
 226                 die("invalid regex: %s", errbuf);
 227         }
 228 }
 229
 230 void diffcore_pickaxe(struct diff_options *o)
 231 {
 232         const char *needle = o->pickaxe;
 233         int opts = o->pickaxe_opts;
 234         regex_t regex, *regexp = NULL;
 235         kwset_t kws = NULL;
 236         pickaxe_fn fn;
 237
 238         if (opts & ~DIFF_PICKAXE_KIND_OBJFIND &&
 239             (!needle || !*needle))
 240                 BUG("should have needle under -G or -S");
 241         if (opts & (DIFF_PICKAXE_REGEX | DIFF_PICKAXE_KIND_G)) {
 242                 int cflags = REG_EXTENDED | REG_NEWLINE;
 243                 if (o->pickaxe_opts & DIFF_PICKAXE_IGNORE_CASE)
 244                         cflags |= REG_ICASE;
 245                 regcomp_or_die(&regex, needle, cflags);
 246                 regexp = &regex;
 247
 248                 if (opts & DIFF_PICKAXE_KIND_G)
 249                         fn = diff_grep;
 250                 else if (opts & DIFF_PICKAXE_REGEX)
 251                         fn = has_changes;
 252                 else
 253                         /*
 254                          * We don't need to check the combination of
 255                          * -G and --pickaxe-regex, by the time we get
 256                          * here diff.c has already died if they're
 257                          * combined. See the usage tests in
 258                          * t4209-log-pickaxe.sh.
 259                          */
 260                         BUG("unreachable");
 261         } else if (opts & DIFF_PICKAXE_KIND_S) {
 262                 if (o->pickaxe_opts & DIFF_PICKAXE_IGNORE_CASE &&
 263                     has_non_ascii(needle)) {
 264                         struct strbuf sb = STRBUF_INIT;
 265                         int cflags = REG_NEWLINE | REG_ICASE;
 266
 267                         basic_regex_quote_buf(&sb, needle);
 268                         regcomp_or_die(&regex, sb.buf, cflags);
 269                         strbuf_release(&sb);
 270                         regexp = &regex;
 271                 } else {
 272                         kws = kwsalloc(o->pickaxe_opts & DIFF_PICKAXE_IGNORE_CASE
 273                                        ? tolower_trans_tbl : NULL);
 274                         kwsincr(kws, needle, strlen(needle));
 275                         kwsprep(kws);
 276                 }
 277                 fn = has_changes;
 278         } else if (opts & DIFF_PICKAXE_KIND_OBJFIND) {
 279                 fn = NULL;
 280         } else {
 281                 BUG("unknown pickaxe_opts flag");
 282         }
 283
 284         pickaxe(&diff_queued_diff, o, regexp, kws, fn);
 285
 286         if (regexp)
 287                 regfree(regexp);
 288         if (kws)
 289                 kwsfree(kws);
 290         return;
 291 }