1 /* Shared definitions for GNU DIFF 2 3 Copyright (C) 1988-1989, 1991-1995, 1998, 2001-2002, 2004, 2009-2010 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 XTERN enum 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 horizontal white space (-b). */ 110 IGNORE_SPACE_CHANGE, 111 112 /* Ignore all horizontal white space (-w). */ 113 IGNORE_ALL_SPACE 114 } ignore_white_space; 115 116 /* Ignore changes that affect only blank lines (-B). */ 117 XTERN bool ignore_blank_lines; 118 119 /* Files can be compared byte-by-byte, as if they were binary. 120 This depends on various options. */ 121 XTERN bool files_can_be_treated_as_binary; 122 123 /* Ignore differences in case of letters (-i). */ 124 XTERN bool ignore_case; 125 126 /* Ignore differences in case of letters in file names. */ 127 XTERN bool ignore_file_name_case; 128 129 /* File labels for `-c' output headers (--label). */ 130 XTERN char *file_label[2]; 131 132 /* Regexp to identify function-header lines (-F). */ 133 XTERN struct re_pattern_buffer function_regexp; 134 135 /* Ignore changes that affect only lines matching this regexp (-I). */ 136 XTERN struct re_pattern_buffer ignore_regexp; 137 138 /* Say only whether files differ, not how (-q). */ 139 XTERN bool brief; 140 141 /* Expand tabs in the output so the text lines up properly 142 despite the characters added to the front of each line (-t). */ 143 XTERN bool expand_tabs; 144 145 /* Number of columns between tab stops. */ 146 XTERN size_t tabsize; 147 148 /* Use a tab in the output, rather than a space, before the text of an 149 input line, so as to keep the proper alignment in the input line 150 without changing the characters in it (-T). */ 151 XTERN bool initial_tab; 152 153 /* Do not output an initial space or tab before the text of an empty line. */ 154 XTERN bool suppress_blank_empty; 155 156 /* Remove trailing carriage returns from input. */ 157 XTERN bool strip_trailing_cr; 158 159 /* In directory comparison, specify file to start with (-S). 160 This is used for resuming an aborted comparison. 161 All file names less than this name are ignored. */ 162 XTERN char const *starting_file; 163 164 /* Pipe each file's output through pr (-l). */ 165 XTERN bool paginate; 166 167 /* Line group formats for unchanged, old, new, and changed groups. */ 168 XTERN char const *group_format[CHANGED + 1]; 169 170 /* Line formats for unchanged, old, and new lines. */ 171 XTERN char const *line_format[NEW + 1]; 172 173 /* If using OUTPUT_SDIFF print extra information to help the sdiff filter. */ 174 XTERN bool sdiff_merge_assist; 175 176 /* Tell OUTPUT_SDIFF to show only the left version of common lines. */ 177 XTERN bool left_column; 178 179 /* Tell OUTPUT_SDIFF to not show common lines. */ 180 XTERN bool suppress_common_lines; 181 182 /* The half line width and column 2 offset for OUTPUT_SDIFF. */ 183 XTERN size_t sdiff_half_width; 184 XTERN size_t sdiff_column2_offset; 185 186 /* String containing all the command options diff received, 187 with spaces between and at the beginning but none at the end. 188 If there were no options given, this string is empty. */ 189 XTERN char *switch_string; 190 191 /* Use heuristics for better speed with large files with a small 192 density of changes. */ 193 XTERN bool speed_large_files; 194 195 /* Patterns that match file names to be excluded. */ 196 XTERN struct exclude *excluded; 197 198 /* Don't discard lines. This makes things slower (sometimes much 199 slower) but will find a guaranteed minimal set of changes. */ 200 XTERN bool minimal; 201 202 /* The strftime format to use for time strings. */ 203 XTERN char const *time_format; 204 205 /* The result of comparison is an "edit script": a chain of `struct change'. 206 Each `struct change' represents one place where some lines are deleted 207 and some are inserted. 208 209 LINE0 and LINE1 are the first affected lines in the two files (origin 0). 210 DELETED is the number of lines deleted here from file 0. 211 INSERTED is the number of lines inserted here in file 1. 212 213 If DELETED is 0 then LINE0 is the number of the line before 214 which the insertion was done; vice versa for INSERTED and LINE1. */ 215 216 struct change 217 { 218 struct change *link; /* Previous or next edit command */ 219 lin inserted; /* # lines of file 1 changed here. */ 220 lin deleted; /* # lines of file 0 changed here. */ 221 lin line0; /* Line number of 1st deleted line. */ 222 lin line1; /* Line number of 1st inserted line. */ 223 bool ignore; /* Flag used in context.c. */ 224 }; 225 226 /* Structures that describe the input files. */ 227 228 /* Data on one input file being compared. */ 229 230 struct file_data { 231 int desc; /* File descriptor */ 232 char const *name; /* File name */ 233 struct stat stat; /* File status */ 234 235 /* Buffer in which text of file is read. */ 236 word *buffer; 237 238 /* Allocated size of buffer, in bytes. Always a multiple of 239 sizeof *buffer. */ 240 size_t bufsize; 241 242 /* Number of valid bytes now in the buffer. */ 243 size_t buffered; 244 245 /* Array of pointers to lines in the file. */ 246 char const **linbuf; 247 248 /* linbuf_base <= buffered_lines <= valid_lines <= alloc_lines. 249 linebuf[linbuf_base ... buffered_lines - 1] are possibly differing. 250 linebuf[linbuf_base ... valid_lines - 1] contain valid data. 251 linebuf[linbuf_base ... alloc_lines - 1] are allocated. */ 252 lin linbuf_base, buffered_lines, valid_lines, alloc_lines; 253 254 /* Pointer to end of prefix of this file to ignore when hashing. */ 255 char const *prefix_end; 256 257 /* Count of lines in the prefix. 258 There are this many lines in the file before linbuf[0]. */ 259 lin prefix_lines; 260 261 /* Pointer to start of suffix of this file to ignore when hashing. */ 262 char const *suffix_begin; 263 264 /* Vector, indexed by line number, containing an equivalence code for 265 each line. It is this vector that is actually compared with that 266 of another file to generate differences. */ 267 lin *equivs; 268 269 /* Vector, like the previous one except that 270 the elements for discarded lines have been squeezed out. */ 271 lin *undiscarded; 272 273 /* Vector mapping virtual line numbers (not counting discarded lines) 274 to real ones (counting those lines). Both are origin-0. */ 275 lin *realindexes; 276 277 /* Total number of nondiscarded lines. */ 278 lin nondiscarded_lines; 279 280 /* Vector, indexed by real origin-0 line number, 281 containing 1 for a line that is an insertion or a deletion. 282 The results of comparison are stored here. */ 283 char *changed; 284 285 /* 1 if file ends in a line with no final newline. */ 286 bool missing_newline; 287 288 /* 1 if at end of file. */ 289 bool eof; 290 291 /* 1 more than the maximum equivalence value used for this or its 292 sibling file. */ 293 lin equiv_max; 294 }; 295 296 /* The file buffer, considered as an array of bytes rather than 297 as an array of words. */ 298 #define FILE_BUFFER(f) ((char *) (f)->buffer) 299 300 /* Data on two input files being compared. */ 301 302 struct comparison 303 { 304 struct file_data file[2]; 305 struct comparison const *parent; /* parent, if a recursive comparison */ 306 }; 307 308 /* Describe the two files currently being compared. */ 309 310 XTERN struct file_data files[2]; 311 312 /* Stdio stream to output diffs to. */ 313 314 XTERN FILE *outfile; 315 316 /* Declare various functions. */ 317 318 /* analyze.c */ 319 int diff_2_files (struct comparison *); 320 321 /* context.c */ 322 void print_context_header (struct file_data[], bool); 323 void print_context_script (struct change *, bool); 324 325 /* dir.c */ 326 int diff_dirs (struct comparison const *, int (*) (struct comparison const *, char const *, char const *)); 327 328 /* ed.c */ 329 void print_ed_script (struct change *); 330 void pr_forward_ed_script (struct change *); 331 332 /* ifdef.c */ 333 void print_ifdef_script (struct change *); 334 335 /* io.c */ 336 void file_block_read (struct file_data *, size_t); 337 bool read_files (struct file_data[], bool); 338 339 /* normal.c */ 340 void print_normal_script (struct change *); 341 342 /* rcs.c */ 343 void print_rcs_script (struct change *); 344 345 /* side.c */ 346 void print_sdiff_script (struct change *); 347 348 /* util.c */ 349 extern char const change_letter[4]; 350 extern char const pr_program[]; 351 char *concat (char const *, char const *, char const *); 352 char *dir_file_pathname (char const *, char const *); 353 bool lines_differ (char const *, char const *); 354 lin translate_line_number (struct file_data const *, lin); 355 struct change *find_change (struct change *); 356 struct change *find_reverse_change (struct change *); 357 void *zalloc (size_t); 358 enum changes analyze_hunk (struct change *, lin *, lin *, lin *, lin *); 359 void begin_output (void); 360 void debug_script (struct change *); 361 void fatal (char const *) __attribute__((noreturn)); 362 void finish_output (void); 363 void message (char const *, char const *, char const *); 364 void message5 (char const *, char const *, char const *, char const *, char const *); 365 void output_1_line (char const *, char const *, char const *, char const *); 366 void perror_with_name (char const *); 367 void pfatal_with_name (char const *) __attribute__((noreturn)); 368 void print_1_line (char const *, char const * const *); 369 void print_message_queue (void); 370 void print_number_range (char, struct file_data *, lin, lin); 371 void print_script (struct change *, struct change * (*) (struct change *), void (*) (struct change *)); 372 void setup_output (char const *, char const *, bool); 373 void translate_range (struct file_data const *, lin, lin, long int *, long int *); 374