Andrew's git - gitweb.git/blob - diffcore-pickaxe.c

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