| 1 | #define DISABLE_SIGN_COMPARE_WARNINGS |
| 2 | |
| 3 | #include "git-compat-util.h" |
| 4 | #include "environment.h" |
| 5 | #include "gettext.h" |
| 6 | #include "config.h" |
| 7 | #include "hex.h" |
| 8 | #include "odb.h" |
| 9 | #include "strbuf.h" |
| 10 | #include "xdiff-interface.h" |
| 11 | #include "xdiff/xtypes.h" |
| 12 | #include "xdiff/xdiffi.h" |
| 13 | #include "xdiff/xutils.h" |
| 14 | |
| 15 | struct xdiff_emit_state { |
| 16 | xdiff_emit_hunk_fn hunk_fn; |
| 17 | xdiff_emit_line_fn line_fn; |
| 18 | void *consume_callback_data; |
| 19 | struct strbuf remainder; |
| 20 | }; |
| 21 | |
| 22 | static int xdiff_out_hunk(void *priv_, |
| 23 | long old_begin, long old_nr, |
| 24 | long new_begin, long new_nr, |
| 25 | const char *func, long funclen) |
| 26 | { |
| 27 | struct xdiff_emit_state *priv = priv_; |
| 28 | |
| 29 | if (priv->remainder.len) |
| 30 | BUG("xdiff emitted hunk in the middle of a line"); |
| 31 | |
| 32 | priv->hunk_fn(priv->consume_callback_data, |
| 33 | old_begin, old_nr, new_begin, new_nr, |
| 34 | func, funclen); |
| 35 | return 0; |
| 36 | } |
| 37 | |
| 38 | static int consume_one(void *priv_, char *s, unsigned long size) |
| 39 | { |
| 40 | struct xdiff_emit_state *priv = priv_; |
| 41 | char *ep; |
| 42 | while (size) { |
| 43 | unsigned long this_size; |
| 44 | int ret; |
| 45 | ep = memchr(s, '\n', size); |
| 46 | this_size = (ep == NULL) ? size : (ep - s + 1); |
| 47 | ret = priv->line_fn(priv->consume_callback_data, s, this_size); |
| 48 | if (ret) |
| 49 | return ret; |
| 50 | size -= this_size; |
| 51 | s += this_size; |
| 52 | } |
| 53 | return 0; |
| 54 | } |
| 55 | |
| 56 | static int xdiff_outf(void *priv_, mmbuffer_t *mb, int nbuf) |
| 57 | { |
| 58 | struct xdiff_emit_state *priv = priv_; |
| 59 | int i; |
| 60 | int stop = 0; |
| 61 | |
| 62 | if (!priv->line_fn) |
| 63 | return 0; |
| 64 | |
| 65 | for (i = 0; i < nbuf; i++) { |
| 66 | if (stop) |
| 67 | return 1; |
| 68 | if (mb[i].ptr[mb[i].size-1] != '\n') { |
| 69 | /* Incomplete line */ |
| 70 | strbuf_add(&priv->remainder, mb[i].ptr, mb[i].size); |
| 71 | continue; |
| 72 | } |
| 73 | |
| 74 | /* we have a complete line */ |
| 75 | if (!priv->remainder.len) { |
| 76 | stop = consume_one(priv, mb[i].ptr, mb[i].size); |
| 77 | continue; |
| 78 | } |
| 79 | strbuf_add(&priv->remainder, mb[i].ptr, mb[i].size); |
| 80 | stop = consume_one(priv, priv->remainder.buf, priv->remainder.len); |
| 81 | strbuf_reset(&priv->remainder); |
| 82 | } |
| 83 | if (stop) |
| 84 | return -1; |
| 85 | if (priv->remainder.len) { |
| 86 | stop = consume_one(priv, priv->remainder.buf, priv->remainder.len); |
| 87 | strbuf_reset(&priv->remainder); |
| 88 | } |
| 89 | if (stop) |
| 90 | return -1; |
| 91 | return 0; |
| 92 | } |
| 93 | |
| 94 | /* |
| 95 | * Trim down common substring at the end of the buffers, |
| 96 | * but end on a complete line. |
| 97 | */ |
| 98 | static void trim_common_tail(mmfile_t *a, mmfile_t *b) |
| 99 | { |
| 100 | const int blk = 1024; |
| 101 | long trimmed = 0, recovered = 0; |
| 102 | char *ap = a->size ? a->ptr + a->size : a->ptr; |
| 103 | char *bp = b->size ? b->ptr + b->size : b->ptr; |
| 104 | long smaller = (a->size < b->size) ? a->size : b->size; |
| 105 | |
| 106 | while (blk + trimmed <= smaller && !memcmp(ap - blk, bp - blk, blk)) { |
| 107 | trimmed += blk; |
| 108 | ap -= blk; |
| 109 | bp -= blk; |
| 110 | } |
| 111 | |
| 112 | while (recovered < trimmed) |
| 113 | if (ap[recovered++] == '\n') |
| 114 | break; |
| 115 | a->size -= trimmed - recovered; |
| 116 | b->size -= trimmed - recovered; |
| 117 | } |
| 118 | |
| 119 | int xdi_diff(mmfile_t *mf1, mmfile_t *mf2, xpparam_t const *xpp, xdemitconf_t const *xecfg, xdemitcb_t *xecb) |
| 120 | { |
| 121 | mmfile_t a = *mf1; |
| 122 | mmfile_t b = *mf2; |
| 123 | |
| 124 | if (mf1->size > MAX_XDIFF_SIZE || mf2->size > MAX_XDIFF_SIZE) |
| 125 | return -1; |
| 126 | |
| 127 | if (!xecfg->ctxlen && !(xecfg->flags & XDL_EMIT_FUNCCONTEXT)) |
| 128 | trim_common_tail(&a, &b); |
| 129 | |
| 130 | return xdl_diff(&a, &b, xpp, xecfg, xecb); |
| 131 | } |
| 132 | |
| 133 | int xdi_diff_outf(mmfile_t *mf1, mmfile_t *mf2, |
| 134 | xdiff_emit_hunk_fn hunk_fn, |
| 135 | xdiff_emit_line_fn line_fn, |
| 136 | void *consume_callback_data, |
| 137 | xpparam_t const *xpp, xdemitconf_t const *xecfg) |
| 138 | { |
| 139 | int ret; |
| 140 | struct xdiff_emit_state state; |
| 141 | xdemitcb_t ecb; |
| 142 | |
| 143 | memset(&state, 0, sizeof(state)); |
| 144 | state.hunk_fn = hunk_fn; |
| 145 | state.line_fn = line_fn; |
| 146 | state.consume_callback_data = consume_callback_data; |
| 147 | memset(&ecb, 0, sizeof(ecb)); |
| 148 | if (hunk_fn) |
| 149 | ecb.out_hunk = xdiff_out_hunk; |
| 150 | ecb.out_line = xdiff_outf; |
| 151 | ecb.priv = &state; |
| 152 | strbuf_init(&state.remainder, 0); |
| 153 | ret = xdi_diff(mf1, mf2, xpp, xecfg, &ecb); |
| 154 | strbuf_release(&state.remainder); |
| 155 | return ret; |
| 156 | } |
| 157 | |
| 158 | int read_mmfile(mmfile_t *ptr, const char *filename) |
| 159 | { |
| 160 | struct stat st; |
| 161 | FILE *f; |
| 162 | size_t sz; |
| 163 | |
| 164 | if (stat(filename, &st)) |
| 165 | return error_errno("Could not stat %s", filename); |
| 166 | if (!(f = fopen(filename, "rb"))) |
| 167 | return error_errno("Could not open %s", filename); |
| 168 | sz = xsize_t(st.st_size); |
| 169 | ptr->ptr = xmalloc(sz ? sz : 1); |
| 170 | if (sz && fread(ptr->ptr, sz, 1, f) != 1) { |
| 171 | fclose(f); |
| 172 | return error("Could not read %s", filename); |
| 173 | } |
| 174 | fclose(f); |
| 175 | ptr->size = sz; |
| 176 | return 0; |
| 177 | } |
| 178 | |
| 179 | void read_mmblob(mmfile_t *ptr, struct object_database *odb, |
| 180 | const struct object_id *oid) |
| 181 | { |
| 182 | unsigned long size; |
| 183 | enum object_type type; |
| 184 | |
| 185 | if (is_null_oid(oid)) { |
| 186 | ptr->ptr = xstrdup(""); |
| 187 | ptr->size = 0; |
| 188 | return; |
| 189 | } |
| 190 | |
| 191 | ptr->ptr = odb_read_object(odb, oid, &type, &size); |
| 192 | if (!ptr->ptr || type != OBJ_BLOB) |
| 193 | die("unable to read blob object %s", oid_to_hex(oid)); |
| 194 | ptr->size = size; |
| 195 | } |
| 196 | |
| 197 | #define FIRST_FEW_BYTES 8000 |
| 198 | int buffer_is_binary(const char *ptr, unsigned long size) |
| 199 | { |
| 200 | if (FIRST_FEW_BYTES < size) |
| 201 | size = FIRST_FEW_BYTES; |
| 202 | return !!memchr(ptr, 0, size); |
| 203 | } |
| 204 | |
| 205 | struct ff_regs { |
| 206 | int nr; |
| 207 | struct ff_reg { |
| 208 | regex_t re; |
| 209 | int negate; |
| 210 | } *array; |
| 211 | }; |
| 212 | |
| 213 | static long ff_regexp(const char *line, long len, |
| 214 | char *buffer, long buffer_size, void *priv) |
| 215 | { |
| 216 | struct ff_regs *regs = priv; |
| 217 | regmatch_t pmatch[2]; |
| 218 | int i; |
| 219 | int result; |
| 220 | |
| 221 | /* Exclude terminating newline (and cr) from matching */ |
| 222 | if (len > 0 && line[len-1] == '\n') { |
| 223 | if (len > 1 && line[len-2] == '\r') |
| 224 | len -= 2; |
| 225 | else |
| 226 | len--; |
| 227 | } |
| 228 | |
| 229 | for (i = 0; i < regs->nr; i++) { |
| 230 | struct ff_reg *reg = regs->array + i; |
| 231 | if (!regexec_buf(®->re, line, len, 2, pmatch, 0)) { |
| 232 | if (reg->negate) |
| 233 | return -1; |
| 234 | break; |
| 235 | } |
| 236 | } |
| 237 | if (regs->nr <= i) |
| 238 | return -1; |
| 239 | i = pmatch[1].rm_so >= 0 ? 1 : 0; |
| 240 | line += pmatch[i].rm_so; |
| 241 | result = pmatch[i].rm_eo - pmatch[i].rm_so; |
| 242 | if (result > buffer_size) |
| 243 | result = buffer_size; |
| 244 | while (result > 0 && (isspace(line[result - 1]))) |
| 245 | result--; |
| 246 | memcpy(buffer, line, result); |
| 247 | return result; |
| 248 | } |
| 249 | |
| 250 | void xdiff_set_find_func(xdemitconf_t *xecfg, const char *value, int cflags) |
| 251 | { |
| 252 | int i; |
| 253 | struct ff_regs *regs; |
| 254 | |
| 255 | xecfg->find_func = ff_regexp; |
| 256 | regs = xecfg->find_func_priv = xmalloc(sizeof(struct ff_regs)); |
| 257 | for (i = 0, regs->nr = 1; value[i]; i++) |
| 258 | if (value[i] == '\n') |
| 259 | regs->nr++; |
| 260 | ALLOC_ARRAY(regs->array, regs->nr); |
| 261 | for (i = 0; i < regs->nr; i++) { |
| 262 | struct ff_reg *reg = regs->array + i; |
| 263 | const char *ep, *expression; |
| 264 | char *buffer = NULL; |
| 265 | |
| 266 | if (!value) |
| 267 | BUG("mismatch between line count and parsing"); |
| 268 | ep = strchr(value, '\n'); |
| 269 | |
| 270 | reg->negate = (*value == '!'); |
| 271 | if (reg->negate && i == regs->nr - 1) |
| 272 | die("Last expression must not be negated: %s", value); |
| 273 | if (*value == '!') |
| 274 | value++; |
| 275 | if (ep) |
| 276 | expression = buffer = xstrndup(value, ep - value); |
| 277 | else |
| 278 | expression = value; |
| 279 | if (regcomp(®->re, expression, cflags)) |
| 280 | die("Invalid regexp to look for hunk header: %s", expression); |
| 281 | free(buffer); |
| 282 | value = ep ? ep + 1 : NULL; |
| 283 | } |
| 284 | } |
| 285 | |
| 286 | void xdiff_clear_find_func(xdemitconf_t *xecfg) |
| 287 | { |
| 288 | if (xecfg->find_func) { |
| 289 | int i; |
| 290 | struct ff_regs *regs = xecfg->find_func_priv; |
| 291 | |
| 292 | for (i = 0; i < regs->nr; i++) |
| 293 | regfree(®s->array[i].re); |
| 294 | free(regs->array); |
| 295 | free(regs); |
| 296 | xecfg->find_func = NULL; |
| 297 | xecfg->find_func_priv = NULL; |
| 298 | } |
| 299 | } |
| 300 | |
| 301 | unsigned long xdiff_hash_string(const char *s, size_t len, long flags) |
| 302 | { |
| 303 | return xdl_hash_record((uint8_t const**)&s, (uint8_t const*)s + len, flags); |
| 304 | } |
| 305 | |
| 306 | int xdiff_compare_lines(const char *l1, long s1, |
| 307 | const char *l2, long s2, long flags) |
| 308 | { |
| 309 | return xdl_recmatch(l1, s1, l2, s2, flags); |
| 310 | } |
| 311 | |
| 312 | int parse_conflict_style_name(const char *value) |
| 313 | { |
| 314 | if (!strcmp(value, "diff3")) |
| 315 | return XDL_MERGE_DIFF3; |
| 316 | else if (!strcmp(value, "zdiff3")) |
| 317 | return XDL_MERGE_ZEALOUS_DIFF3; |
| 318 | else if (!strcmp(value, "merge")) |
| 319 | return 0; |
| 320 | /* |
| 321 | * Please update _git_checkout() in git-completion.bash when |
| 322 | * you add new merge config |
| 323 | */ |
| 324 | else |
| 325 | return -1; |
| 326 | } |
| 327 | |
| 328 | const char *conflict_style_name(int style) |
| 329 | { |
| 330 | switch (style) { |
| 331 | case XDL_MERGE_DIFF3: |
| 332 | return "diff3"; |
| 333 | case XDL_MERGE_ZEALOUS_DIFF3: |
| 334 | return "zdiff3"; |
| 335 | default: |
| 336 | return "merge"; |
| 337 | } |
| 338 | } |
| 339 | |
| 340 | int git_xmerge_style = -1; |
| 341 | |
| 342 | int git_xmerge_config(const char *var, const char *value, |
| 343 | const struct config_context *ctx, void *cb) |
| 344 | { |
| 345 | if (!strcmp(var, "merge.conflictstyle")) { |
| 346 | if (!value) |
| 347 | return config_error_nonbool(var); |
| 348 | git_xmerge_style = parse_conflict_style_name(value); |
| 349 | if (git_xmerge_style == -1) |
| 350 | return error(_("unknown style '%s' given for '%s'"), |
| 351 | value, var); |
| 352 | return 0; |
| 353 | } |
| 354 | return git_default_config(var, value, ctx, cb); |
| 355 | } |