1 /* Shared definitions for GNU DIFF 2 3 Copyright (C) 1988-1989, 1991-1995, 1998, 2001-2002, 2004, 2009-2011 Free 4 Software Foundation, Inc. 5 6 This file is part of GNU DIFF. 7 8 This program is free software: you can redistribute it and/or modify 9 it under the terms of the GNU General Public License as published by 10 the Free Software Foundation, either version 3 of the License, or 11 (at your option) any later version. 12 13 This program is distributed in the hope that it will be useful, 14 but WITHOUT ANY WARRANTY; without even the implied warranty of 15 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the 16 GNU General Public License for more details. 17 18 You should have received a copy of the GNU General Public License 19 along with this program. If not, see <http://www.gnu.org/licenses/>. */ 20 21 #include "system.h" 22 #include <regex.h> 23 #include <stdio.h> 24 #include <unlocked-io.h> 25 26 /* What kind of changes a hunk contains. */ 27 enum changes 28 { 29 /* No changes: lines common to both files. */ 30 UNCHANGED, 31 32 /* Deletes only: lines taken from just the first file. */ 33 OLD, 34 35 /* Inserts only: lines taken from just the second file. */ 36 NEW, 37 38 /* Both deletes and inserts: a hunk containing both old and new lines. */ 39 CHANGED 40 }; 41 42 /* Variables for command line options */ 43 44 #ifndef GDIFF_MAIN 45 # define XTERN extern 46 #else 47 # define XTERN 48 #endif 49 50 enum output_style 51 { 52 /* No output style specified. */ 53 OUTPUT_UNSPECIFIED, 54 55 /* Default output style. */ 56 OUTPUT_NORMAL, 57 58 /* Output the differences with lines of context before and after (-c). */ 59 OUTPUT_CONTEXT, 60 61 /* Output the differences in a unified context diff format (-u). */ 62 OUTPUT_UNIFIED, 63 64 /* Output the differences as commands suitable for `ed' (-e). */ 65 OUTPUT_ED, 66 67 /* Output the diff as a forward ed script (-f). */ 68 OUTPUT_FORWARD_ED, 69 70 /* Like -f, but output a count of changed lines in each "command" (-n). */ 71 OUTPUT_RCS, 72 73 /* Output merged #ifdef'd file (-D). */ 74 OUTPUT_IFDEF, 75 76 /* Output sdiff style (-y). */ 77 OUTPUT_SDIFF 78 }; 79 80 /* True for output styles that are robust, 81 i.e. can handle a file that ends in a non-newline. */ 82 #define ROBUST_OUTPUT_STYLE(S) ((S) != OUTPUT_ED && (S) != OUTPUT_FORWARD_ED) 83 84 XTERN enum output_style output_style; 85 86 /* Nonzero if output cannot be generated for identical files. */ 87 XTERN bool no_diff_means_no_output; 88 89 /* Number of lines of context to show in each set of diffs. 90 This is zero when context is not to be shown. */ 91 XTERN lin context; 92 93 /* Consider all files as text files (-a). 94 Don't interpret codes over 0177 as implying a "binary file". */ 95 XTERN bool text; 96 97 /* Number of lines to keep in identical prefix and suffix. */ 98 XTERN lin horizon_lines; 99 100 /* The significance of white space during comparisons. */ 101 enum DIFF_white_space 102 { 103 /* All white space is significant (the default). */ 104 IGNORE_NO_WHITE_SPACE, 105 106 /* Ignore changes due to tab expansion (-E). */ 107 IGNORE_TAB_EXPANSION, 108 109 /* Ignore changes in trailing horizontal white space (-Z). */ 110 IGNORE_TRAILING_SPACE, 111 112 /* IGNORE_TAB_EXPANSION and IGNORE_TRAILING_SPACE are a special case 113 because they are independent and can be ORed together, yielding 114 IGNORE_TAB_EXPANSION_AND_TRAILING_SPACE. */ 115 IGNORE_TAB_EXPANSION_AND_TRAILING_SPACE, 116 117 /* Ignore changes in horizontal white space (-b). */ 118 IGNORE_SPACE_CHANGE, 119 120 /* Ignore all horizontal white space (-w). */ 121 IGNORE_ALL_SPACE 122 }; 123 XTERN enum DIFF_white_space ignore_white_space; 124 125 /* Ignore changes that affect only blank lines (-B). */ 126 XTERN bool ignore_blank_lines; 127 128 /* Files can be compared byte-by-byte, as if they were binary. 129 This depends on various options. */ 130 XTERN bool files_can_be_treated_as_binary; 131 132 /* Ignore differences in case of letters (-i). */ 133 XTERN bool ignore_case; 134 135 /* Ignore differences in case of letters in file names. */ 136 XTERN bool ignore_file_name_case; 137 138 /* File labels for `-c' output headers (--label). */ 139 XTERN char *file_label[2]; 140 141 /* Regexp to identify function-header lines (-F). */ 142 XTERN struct re_pattern_buffer function_regexp; 143 144 /* Ignore changes that affect only lines matching this regexp (-I). */ 145 XTERN struct re_pattern_buffer ignore_regexp; 146 147 /* Say only whether files differ, not how (-q). */ 148 XTERN bool brief; 149 150 /* Expand tabs in the output so the text lines up properly 151 despite the characters added to the front of each line (-t). */ 152 XTERN bool expand_tabs; 153 154 /* Number of columns between tab stops. */ 155 XTERN size_t tabsize; 156 157 /* Use a tab in the output, rather than a space, before the text of an 158 input line, so as to keep the proper alignment in the input line 159 without changing the characters in it (-T). */ 160 XTERN bool initial_tab; 161 162 /* Do not output an initial space or tab before the text of an empty line. */ 163 XTERN bool suppress_blank_empty; 164 165 /* Remove trailing carriage returns from input. */ 166 XTERN bool strip_trailing_cr; 167 168 /* In directory comparison, specify file to start with (-S). 169 This is used for resuming an aborted comparison. 170 All file names less than this name are ignored. */ 171 XTERN char const *starting_file; 172 173 /* Pipe each file's output through pr (-l). */ 174 XTERN bool paginate; 175 176 /* Line group formats for unchanged, old, new, and changed groups. */ 177 XTERN char const *group_format[CHANGED + 1]; 178 179 /* Line formats for unchanged, old, and new lines. */ 180 XTERN char const *line_format[NEW + 1]; 181 182 /* If using OUTPUT_SDIFF print extra information to help the sdiff filter. */ 183 XTERN bool sdiff_merge_assist; 184 185 /* Tell OUTPUT_SDIFF to show only the left version of common lines. */ 186 XTERN bool left_column; 187 188 /* Tell OUTPUT_SDIFF to not show common lines. */ 189 XTERN bool suppress_common_lines; 190 191 /* The half line width and column 2 offset for OUTPUT_SDIFF. */ 192 XTERN size_t sdiff_half_width; 193 XTERN size_t sdiff_column2_offset; 194 195 /* String containing all the command options diff received, 196 with spaces between and at the beginning but none at the end. 197 If there were no options given, this string is empty. */ 198 XTERN char *switch_string; 199 200 /* Use heuristics for better speed with large files with a small 201 density of changes. */ 202 XTERN bool speed_large_files; 203 204 /* Patterns that match file names to be excluded. */ 205 XTERN struct exclude *excluded; 206 207 /* Don't discard lines. This makes things slower (sometimes much 208 slower) but will find a guaranteed minimal set of changes. */ 209 XTERN bool minimal; 210 211 /* The strftime format to use for time strings. */ 212 XTERN char const *time_format; 213 214 /* The result of comparison is an "edit script": a chain of `struct change'. 215 Each `struct change' represents one place where some lines are deleted 216 and some are inserted. 217 218 LINE0 and LINE1 are the first affected lines in the two files (origin 0). 219 DELETED is the number of lines deleted here from file 0. 220 INSERTED is the number of lines inserted here in file 1. 221 222 If DELETED is 0 then LINE0 is the number of the line before 223 which the insertion was done; vice versa for INSERTED and LINE1. */ 224 225 struct change 226 { 227 struct change *link; /* Previous or next edit command */ 228 lin inserted; /* # lines of file 1 changed here. */ 229 lin deleted; /* # lines of file 0 changed here. */ 230 lin line0; /* Line number of 1st deleted line. */ 231 lin line1; /* Line number of 1st inserted line. */ 232 bool ignore; /* Flag used in context.c. */ 233 }; 234 235 /* Structures that describe the input files. */ 236 237 /* Data on one input file being compared. */ 238 239 struct file_data { 240 int desc; /* File descriptor */ 241 char const *name; /* File name */ 242 struct stat stat; /* File status */ 243 244 /* Buffer in which text of file is read. */ 245 word *buffer; 246 247 /* Allocated size of buffer, in bytes. Always a multiple of 248 sizeof *buffer. */ 249 size_t bufsize; 250 251 /* Number of valid bytes now in the buffer. */ 252 size_t buffered; 253 254 /* Array of pointers to lines in the file. */ 255 char const **linbuf; 256 257 /* linbuf_base <= buffered_lines <= valid_lines <= alloc_lines. 258 linebuf[linbuf_base ... buffered_lines - 1] are possibly differing. 259 linebuf[linbuf_base ... valid_lines - 1] contain valid data. 260 linebuf[linbuf_base ... alloc_lines - 1] are allocated. */ 261 lin linbuf_base, buffered_lines, valid_lines, alloc_lines; 262 263 /* Pointer to end of prefix of this file to ignore when hashing. */ 264 char const *prefix_end; 265 266 /* Count of lines in the prefix. 267 There are this many lines in the file before linbuf[0]. */ 268 lin prefix_lines; 269 270 /* Pointer to start of suffix of this file to ignore when hashing. */ 271 char const *suffix_begin; 272 273 /* Vector, indexed by line number, containing an equivalence code for 274 each line. It is this vector that is actually compared with that 275 of another file to generate differences. */ 276 lin *equivs; 277 278 /* Vector, like the previous one except that 279 the elements for discarded lines have been squeezed out. */ 280 lin *undiscarded; 281 282 /* Vector mapping virtual line numbers (not counting discarded lines) 283 to real ones (counting those lines). Both are origin-0. */ 284 lin *realindexes; 285 286 /* Total number of nondiscarded lines. */ 287 lin nondiscarded_lines; 288 289 /* Vector, indexed by real origin-0 line number, 290 containing 1 for a line that is an insertion or a deletion. 291 The results of comparison are stored here. */ 292 char *changed; 293 294 /* 1 if file ends in a line with no final newline. */ 295 bool missing_newline; 296 297 /* 1 if at end of file. */ 298 bool eof; 299 300 /* 1 more than the maximum equivalence value used for this or its 301 sibling file. */ 302 lin equiv_max; 303 }; 304 305 /* The file buffer, considered as an array of bytes rather than 306 as an array of words. */ 307 #define FILE_BUFFER(f) ((char *) (f)->buffer) 308 309 /* Data on two input files being compared. */ 310 311 struct comparison 312 { 313 struct file_data file[2]; 314 struct comparison const *parent; /* parent, if a recursive comparison */ 315 }; 316 317 /* Describe the two files currently being compared. */ 318 319 XTERN struct file_data files[2]; 320 321 /* Stdio stream to output diffs to. */ 322 323 XTERN FILE *outfile; 324 325 /* Declare various functions. */ 326 327 /* analyze.c */ 328 extern int diff_2_files (struct comparison *); 329 330 /* context.c */ 331 extern void print_context_header (struct file_data[], bool); 332 extern void print_context_script (struct change *, bool); 333 334 /* dir.c */ 335 extern int diff_dirs (struct comparison const *, 336 int (*) (struct comparison const *, 337 char const *, char const *)); 338 extern char *find_dir_file_pathname (char const *, char const *); 339 340 /* ed.c */ 341 extern void print_ed_script (struct change *); 342 extern void pr_forward_ed_script (struct change *); 343 344 /* ifdef.c */ 345 extern void print_ifdef_script (struct change *); 346 347 /* io.c */ 348 extern void file_block_read (struct file_data *, size_t); 349 extern bool read_files (struct file_data[], bool); 350 351 /* normal.c */ 352 extern void print_normal_script (struct change *); 353 354 /* rcs.c */ 355 extern void print_rcs_script (struct change *); 356 357 /* side.c */ 358 extern void print_sdiff_script (struct change *); 359 360 /* util.c */ 361 extern char const change_letter[4]; 362 extern char const pr_program[]; 363 extern char *concat (char const *, char const *, char const *); 364 extern bool lines_differ (char const *, char const *); 365 extern lin translate_line_number (struct file_data const *, lin); 366 extern struct change *find_change (struct change *); 367 extern struct change *find_reverse_change (struct change *); 368 extern void *zalloc (size_t); 369 extern enum changes analyze_hunk (struct change *, lin *, lin *, lin *, lin *); 370 extern void begin_output (void); 371 extern void debug_script (struct change *); 372 extern void fatal (char const *) __attribute__((noreturn)); 373 extern void finish_output (void); 374 extern void message (char const *, char const *, char const *); 375 extern void message5 (char const *, char const *, char const *, 376 char const *, char const *); 377 extern void output_1_line (char const *, char const *, char const *, 378 char const *); 379 extern void perror_with_name (char const *); 380 extern void pfatal_with_name (char const *) __attribute__((noreturn)); 381 extern void print_1_line (char const *, char const * const *); 382 extern void print_message_queue (void); 383 extern void print_number_range (char, struct file_data *, lin, lin); 384 extern void print_script (struct change *, struct change * (*) (struct change *), 385 void (*) (struct change *)); 386 extern void setup_output (char const *, char const *, bool); 387 extern void translate_range (struct file_data const *, lin, lin, 388 long int *, long int *); 389