gnu/dist/diffutils/src/diff.h

   1 /*      $NetBSD$        */
   2
   3 /* Shared definitions for GNU DIFF
   4
   5    Copyright (C) 1988, 1989, 1991, 1992, 1993, 1994, 1995, 1998, 2001,
   6    2002 Free Software Foundation, Inc.
   7
   8    This file is part of GNU DIFF.
   9
  10    GNU DIFF is free software; you can redistribute it and/or modify
  11    it under the terms of the GNU General Public License as published by
  12    the Free Software Foundation; either version 2, or (at your option)
  13    any later version.
  14
  15    GNU DIFF is distributed in the hope that it will be useful,
  16    but WITHOUT ANY WARRANTY; without even the implied warranty of
  17    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  18    GNU General Public License for more details.
  19
  20    You should have received a copy of the GNU General Public License
  21    along with this program; see the file COPYING.
  22    If not, write to the Free Software Foundation,
  23    59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.  */
  24
  25 #include "system.h"
  26 #include <stdio.h>
  27
  28 #define TAB_WIDTH 8
  29
  30 /* What kind of changes a hunk contains.  */
  31 enum changes
  32 {
  33   /* No changes: lines common to both files.  */
  34   UNCHANGED,
  35
  36   /* Deletes only: lines taken from just the first file.  */
  37   OLD,
  38
  39   /* Inserts only: lines taken from just the second file.  */
  40   NEW,
  41
  42   /* Both deletes and inserts: a hunk containing both old and new lines.  */
  43   CHANGED
  44 };
  45 \f
  46 /* Variables for command line options */
  47
  48 #ifndef GDIFF_MAIN
  49 # define XTERN extern
  50 #else
  51 # define XTERN
  52 #endif
  53
  54 enum output_style
  55 {
  56   /* No output style specified.  */
  57   OUTPUT_UNSPECIFIED,
  58
  59   /* Default output style.  */
  60   OUTPUT_NORMAL,
  61
  62   /* Output the differences with lines of context before and after (-c).  */
  63   OUTPUT_CONTEXT,
  64
  65   /* Output the differences in a unified context diff format (-u).  */
  66   OUTPUT_UNIFIED,
  67
  68   /* Output the differences as commands suitable for `ed' (-e).  */
  69   OUTPUT_ED,
  70
  71   /* Output the diff as a forward ed script (-f).  */
  72   OUTPUT_FORWARD_ED,
  73
  74   /* Like -f, but output a count of changed lines in each "command" (-n).  */
  75   OUTPUT_RCS,
  76
  77   /* Output merged #ifdef'd file (-D).  */
  78   OUTPUT_IFDEF,
  79
  80   /* Output sdiff style (-y).  */
  81   OUTPUT_SDIFF
  82 };
  83
  84 /* True for output styles that are robust,
  85    i.e. can handle a file that ends in a non-newline.  */
  86 #define ROBUST_OUTPUT_STYLE(S) ((S) != OUTPUT_ED && (S) != OUTPUT_FORWARD_ED)
  87
  88 XTERN enum output_style output_style;
  89
  90 /* Nonzero if output cannot be generated for identical files.  */
  91 XTERN bool no_diff_means_no_output;
  92
  93 /* Number of lines of context to show in each set of diffs.
  94    This is zero when context is not to be shown.  */
  95 XTERN lin context;
  96
  97 /* Consider all files as text files (-a).
  98    Don't interpret codes over 0177 as implying a "binary file".  */
  99 XTERN bool text;
 100
 101 /* Number of lines to keep in identical prefix and suffix.  */
 102 XTERN lin horizon_lines;
 103
 104 /* The significance of white space during comparisons.  */
 105 XTERN enum
 106 {
 107   /* All white space is significant (the default).  */
 108   IGNORE_NO_WHITE_SPACE,
 109
 110   /* Ignore changes due to tab expansion (-E).  */
 111   IGNORE_TAB_EXPANSION,
 112
 113   /* Ignore changes in horizontal white space (-b).  */
 114   IGNORE_SPACE_CHANGE,
 115
 116   /* Ignore all horizontal white space (-w).  */
 117   IGNORE_ALL_SPACE
 118 } ignore_white_space;
 119
 120 /* Ignore changes that affect only blank lines (-B).  */
 121 XTERN bool ignore_blank_lines;
 122
 123 /* Files can be compared byte-by-byte, as if they were binary.
 124    This depends on various options.  */
 125 XTERN bool files_can_be_treated_as_binary;
 126
 127 /* Ignore differences in case of letters (-i).  */
 128 XTERN bool ignore_case;
 129
 130 /* Ignore differences in case of letters in file names.  */
 131 XTERN bool ignore_file_name_case;
 132
 133 /* File labels for `-c' output headers (--label).  */
 134 XTERN char *file_label[2];
 135
 136 /* Regexp to identify function-header lines (-F).  */
 137 XTERN struct re_pattern_buffer function_regexp;
 138
 139 /* Ignore changes that affect only lines matching this regexp (-I).  */
 140 XTERN struct re_pattern_buffer ignore_regexp;
 141
 142 /* Say only whether files differ, not how (-q).  */
 143 XTERN bool brief;
 144
 145 /* Expand tabs in the output so the text lines up properly
 146    despite the characters added to the front of each line (-t).  */
 147 XTERN bool expand_tabs;
 148
 149 /* Use a tab in the output, rather than a space, before the text of an
 150    input line, so as to keep the proper alignment in the input line
 151    without changing the characters in it (-T).  */
 152 XTERN bool initial_tab;
 153
 154 /* Remove trailing carriage returns from input.  */
 155 XTERN bool strip_trailing_cr;
 156
 157 /* In directory comparison, specify file to start with (-S).
 158    This is used for resuming an aborted comparison.
 159    All file names less than this name are ignored.  */
 160 XTERN char const *starting_file;
 161
 162 /* Pipe each file's output through pr (-l).  */
 163 XTERN bool paginate;
 164
 165 /* Line group formats for unchanged, old, new, and changed groups.  */
 166 XTERN char const *group_format[CHANGED + 1];
 167
 168 /* Line formats for unchanged, old, and new lines.  */
 169 XTERN char const *line_format[NEW + 1];
 170
 171 /* If using OUTPUT_SDIFF print extra information to help the sdiff filter.  */
 172 XTERN bool sdiff_merge_assist;
 173
 174 /* Tell OUTPUT_SDIFF to show only the left version of common lines.  */
 175 XTERN bool left_column;
 176
 177 /* Tell OUTPUT_SDIFF to not show common lines.  */
 178 XTERN bool suppress_common_lines;
 179
 180 /* The half line width and column 2 offset for OUTPUT_SDIFF.  */
 181 XTERN unsigned int sdiff_half_width;
 182 XTERN unsigned int sdiff_column2_offset;
 183
 184 /* String containing all the command options diff received,
 185    with spaces between and at the beginning but none at the end.
 186    If there were no options given, this string is empty.  */
 187 XTERN char *switch_string;
 188
 189 /* Use heuristics for better speed with large files with a small
 190    density of changes.  */
 191 XTERN bool speed_large_files;
 192
 193 /* Patterns that match file names to be excluded.  */
 194 XTERN struct exclude *excluded;
 195
 196 /* Don't discard lines.  This makes things slower (sometimes much
 197    slower) but will find a guaranteed minimal set of changes.  */
 198 XTERN bool minimal;
 199
 200 /* Name of program the user invoked (for error messages).  */
 201 XTERN char *program_name;
 202
 203 /* The strftime format to use for time strings.  */
 204 XTERN char const *time_format;
 205 \f
 206 /* The result of comparison is an "edit script": a chain of `struct change'.
 207    Each `struct change' represents one place where some lines are deleted
 208    and some are inserted.
 209
 210    LINE0 and LINE1 are the first affected lines in the two files (origin 0).
 211    DELETED is the number of lines deleted here from file 0.
 212    INSERTED is the number of lines inserted here in file 1.
 213
 214    If DELETED is 0 then LINE0 is the number of the line before
 215    which the insertion was done; vice versa for INSERTED and LINE1.  */
 216
 217 struct change
 218 {
 219   struct change *link;          /* Previous or next edit command  */
 220   lin inserted;                 /* # lines of file 1 changed here.  */
 221   lin deleted;                  /* # lines of file 0 changed here.  */
 222   lin line0;                    /* Line number of 1st deleted line.  */
 223   lin line1;                    /* Line number of 1st inserted line.  */
 224   bool ignore;                  /* Flag used in context.c.  */
 225 };
 226 \f
 227 /* Structures that describe the input files.  */
 228
 229 /* Data on one input file being compared.  */
 230
 231 struct file_data {
 232     int             desc;       /* File descriptor  */
 233     char const      *name;      /* File name  */
 234     struct stat     stat;       /* File status */
 235
 236     /* Buffer in which text of file is read.  */
 237     word *buffer;
 238
 239     /* Allocated size of buffer, in bytes.  Always a multiple of
 240        sizeof *buffer.  */
 241     size_t bufsize;
 242
 243     /* Number of valid bytes now in the buffer.  */
 244     size_t buffered;
 245
 246     /* Array of pointers to lines in the file.  */
 247     char const **linbuf;
 248
 249     /* linbuf_base <= buffered_lines <= valid_lines <= alloc_lines.
 250        linebuf[linbuf_base ... buffered_lines - 1] are possibly differing.
 251        linebuf[linbuf_base ... valid_lines - 1] contain valid data.
 252        linebuf[linbuf_base ... alloc_lines - 1] are allocated.  */
 253     lin linbuf_base, buffered_lines, valid_lines, alloc_lines;
 254
 255     /* Pointer to end of prefix of this file to ignore when hashing.  */
 256     char const *prefix_end;
 257
 258     /* Count of lines in the prefix.
 259        There are this many lines in the file before linbuf[0].  */
 260     lin prefix_lines;
 261
 262     /* Pointer to start of suffix of this file to ignore when hashing.  */
 263     char const *suffix_begin;
 264
 265     /* Vector, indexed by line number, containing an equivalence code for
 266        each line.  It is this vector that is actually compared with that
 267        of another file to generate differences.  */
 268     lin *equivs;
 269
 270     /* Vector, like the previous one except that
 271        the elements for discarded lines have been squeezed out.  */
 272     lin *undiscarded;
 273
 274     /* Vector mapping virtual line numbers (not counting discarded lines)
 275        to real ones (counting those lines).  Both are origin-0.  */
 276     lin *realindexes;
 277
 278     /* Total number of nondiscarded lines.  */
 279     lin nondiscarded_lines;
 280
 281     /* Vector, indexed by real origin-0 line number,
 282        containing TRUE for a line that is an insertion or a deletion.
 283        The results of comparison are stored here.  */
 284     bool *changed;
 285
 286     /* 1 if file ends in a line with no final newline.  */
 287     bool missing_newline;
 288
 289     /* 1 if at end of file.  */
 290     bool eof;
 291
 292     /* 1 more than the maximum equivalence value used for this or its
 293        sibling file.  */
 294     lin equiv_max;
 295 };
 296
 297 /* The file buffer, considered as an array of bytes rather than
 298    as an array of words.  */
 299 #define FILE_BUFFER(f) ((char *) (f)->buffer)
 300
 301 /* Data on two input files being compared.  */
 302
 303 struct comparison
 304   {
 305     struct file_data file[2];
 306     struct comparison const *parent;  /* parent, if a recursive comparison */
 307   };
 308
 309 /* Describe the two files currently being compared.  */
 310
 311 XTERN struct file_data files[2];
 312 \f
 313 /* Stdio stream to output diffs to.  */
 314
 315 XTERN FILE *outfile;
 316 \f
 317 /* Declare various functions.  */
 318
 319 /* analyze.c */
 320 int diff_2_files (struct comparison *);
 321
 322 /* context.c */
 323 void print_context_header (struct file_data[], bool);
 324 void print_context_script (struct change *, bool);
 325
 326 /* dir.c */
 327 int diff_dirs (struct comparison const *, int (*) (struct comparison const *, char const *, char const *));
 328
 329 /* ed.c */
 330 void print_ed_script (struct change *);
 331 void pr_forward_ed_script (struct change *);
 332
 333 /* ifdef.c */
 334 void print_ifdef_script (struct change *);
 335
 336 /* io.c */
 337 void file_block_read (struct file_data *, size_t);
 338 bool read_files (struct file_data[], bool);
 339
 340 /* normal.c */
 341 void print_normal_script (struct change *);
 342
 343 /* rcs.c */
 344 void print_rcs_script (struct change *);
 345
 346 /* side.c */
 347 void print_sdiff_script (struct change *);
 348
 349 /* util.c */
 350 extern char const change_letter[4];
 351 extern char const pr_program[];
 352 char *concat (char const *, char const *, char const *);
 353 char *dir_file_pathname (char const *, char const *);
 354 bool lines_differ (char const *, char const *);
 355 lin translate_line_number (struct file_data const *, lin);
 356 struct change *find_change (struct change *);
 357 struct change *find_reverse_change (struct change *);
 358 void *zalloc (size_t);
 359 enum changes analyze_hunk (struct change *, lin *, lin *, lin *, lin *);
 360 void begin_output (void);
 361 void debug_script (struct change *);
 362 void fatal (char const *) __attribute__((noreturn));
 363 void finish_output (void);
 364 void message (char const *, char const *, char const *);
 365 void message5 (char const *, char const *, char const *, char const *, char const *);
 366 void output_1_line (char const *, char const *, char const *, char const *);
 367 void perror_with_name (char const *);
 368 void pfatal_with_name (char const *) __attribute__((noreturn));
 369 void print_1_line (char const *, char const * const *);
 370 void print_message_queue (void);
 371 void print_number_range (char, struct file_data *, lin, lin);
 372 void print_script (struct change *, struct change * (*) (struct change *), void (*) (struct change *));
 373 void setup_output (char const *, char const *, bool);
 374 void translate_range (struct file_data const *, lin, lin, long *, long *);
 375
 376 /* version.c */
 377 extern char const version_string[];