contrib/diffutils-2.8/src/diff.h

   1 /* Shared definitions for GNU DIFF
   2
   3    Copyright (C) 1988, 1989, 1991, 1992, 1993, 1994, 1995, 1998, 2001,
   4    2002, 2004 Free Software Foundation, Inc.
   5
   6    This file is part of GNU DIFF.
   7
   8    GNU DIFF 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 2, or (at your option)
  11    any later version.
  12
  13    GNU DIFF 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; see the file COPYING.
  20    If not, write to the Free Software Foundation,
  21    59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.  */
  22
  23 #include "system.h"
  24 #include <regex.h>
  25 #include <stdio.h>
  26 #include <unlocked-io.h>
  27
  28 /* What kind of changes a hunk contains.  */
  29 enum changes
  30 {
  31   /* No changes: lines common to both files.  */
  32   UNCHANGED,
  33
  34   /* Deletes only: lines taken from just the first file.  */
  35   OLD,
  36
  37   /* Inserts only: lines taken from just the second file.  */
  38   NEW,
  39
  40   /* Both deletes and inserts: a hunk containing both old and new lines.  */
  41   CHANGED
  42 };
  43 \f
  44 /* Variables for command line options */
  45
  46 #ifndef GDIFF_MAIN
  47 # define XTERN extern
  48 #else
  49 # define XTERN
  50 #endif
  51
  52 enum output_style
  53 {
  54   /* No output style specified.  */
  55   OUTPUT_UNSPECIFIED,
  56
  57   /* Default output style.  */
  58   OUTPUT_NORMAL,
  59
  60   /* Output the differences with lines of context before and after (-c).  */
  61   OUTPUT_CONTEXT,
  62
  63   /* Output the differences in a unified context diff format (-u).  */
  64   OUTPUT_UNIFIED,
  65
  66   /* Output the differences as commands suitable for `ed' (-e).  */
  67   OUTPUT_ED,
  68
  69   /* Output the diff as a forward ed script (-f).  */
  70   OUTPUT_FORWARD_ED,
  71
  72   /* Like -f, but output a count of changed lines in each "command" (-n).  */
  73   OUTPUT_RCS,
  74
  75   /* Output merged #ifdef'd file (-D).  */
  76   OUTPUT_IFDEF,
  77
  78   /* Output sdiff style (-y).  */
  79   OUTPUT_SDIFF
  80 };
  81
  82 /* True for output styles that are robust,
  83    i.e. can handle a file that ends in a non-newline.  */
  84 #define ROBUST_OUTPUT_STYLE(S) ((S) != OUTPUT_ED && (S) != OUTPUT_FORWARD_ED)
  85
  86 XTERN enum output_style output_style;
  87
  88 /* Nonzero if output cannot be generated for identical files.  */
  89 XTERN bool no_diff_means_no_output;
  90
  91 /* Number of lines of context to show in each set of diffs.
  92    This is zero when context is not to be shown.  */
  93 XTERN lin context;
  94
  95 /* Consider all files as text files (-a).
  96    Don't interpret codes over 0177 as implying a "binary file".  */
  97 XTERN bool text;
  98
  99 /* Number of lines to keep in identical prefix and suffix.  */
 100 XTERN lin horizon_lines;
 101
 102 /* The significance of white space during comparisons.  */
 103 XTERN enum
 104 {
 105   /* All white space is significant (the default).  */
 106   IGNORE_NO_WHITE_SPACE,
 107
 108   /* Ignore changes due to tab expansion (-E).  */
 109   IGNORE_TAB_EXPANSION,
 110
 111   /* Ignore changes in horizontal white space (-b).  */
 112   IGNORE_SPACE_CHANGE,
 113
 114   /* Ignore all horizontal white space (-w).  */
 115   IGNORE_ALL_SPACE
 116 } ignore_white_space;
 117
 118 /* Ignore changes that affect only blank lines (-B).  */
 119 XTERN bool ignore_blank_lines;
 120
 121 /* Files can be compared byte-by-byte, as if they were binary.
 122    This depends on various options.  */
 123 XTERN bool files_can_be_treated_as_binary;
 124
 125 /* Ignore differences in case of letters (-i).  */
 126 XTERN bool ignore_case;
 127
 128 /* Ignore differences in case of letters in file names.  */
 129 XTERN bool ignore_file_name_case;
 130
 131 /* File labels for `-c' output headers (--label).  */
 132 XTERN char *file_label[2];
 133
 134 /* Regexp to identify function-header lines (-F).  */
 135 XTERN struct re_pattern_buffer function_regexp;
 136
 137 /* Ignore changes that affect only lines matching this regexp (-I).  */
 138 XTERN struct re_pattern_buffer ignore_regexp;
 139
 140 /* Say only whether files differ, not how (-q).  */
 141 XTERN bool brief;
 142
 143 /* Expand tabs in the output so the text lines up properly
 144    despite the characters added to the front of each line (-t).  */
 145 XTERN bool expand_tabs;
 146
 147 /* Number of columns between tab stops.  */
 148 XTERN size_t tabsize;
 149
 150 /* Use a tab in the output, rather than a space, before the text of an
 151    input line, so as to keep the proper alignment in the input line
 152    without changing the characters in it (-T).  */
 153 XTERN bool initial_tab;
 154
 155 /* Remove trailing carriage returns from input.  */
 156 XTERN bool strip_trailing_cr;
 157
 158 /* In directory comparison, specify file to start with (-S).
 159    This is used for resuming an aborted comparison.
 160    All file names less than this name are ignored.  */
 161 XTERN char const *starting_file;
 162
 163 /* Pipe each file's output through pr (-l).  */
 164 XTERN bool paginate;
 165
 166 /* Line group formats for unchanged, old, new, and changed groups.  */
 167 XTERN char const *group_format[CHANGED + 1];
 168
 169 /* Line formats for unchanged, old, and new lines.  */
 170 XTERN char const *line_format[NEW + 1];
 171
 172 /* If using OUTPUT_SDIFF print extra information to help the sdiff filter.  */
 173 XTERN bool sdiff_merge_assist;
 174
 175 /* Tell OUTPUT_SDIFF to show only the left version of common lines.  */
 176 XTERN bool left_column;
 177
 178 /* Tell OUTPUT_SDIFF to not show common lines.  */
 179 XTERN bool suppress_common_lines;
 180
 181 /* The half line width and column 2 offset for OUTPUT_SDIFF.  */
 182 XTERN size_t sdiff_half_width;
 183 XTERN size_t sdiff_column2_offset;
 184
 185 /* String containing all the command options diff received,
 186    with spaces between and at the beginning but none at the end.
 187    If there were no options given, this string is empty.  */
 188 XTERN char *switch_string;
 189
 190 /* Use heuristics for better speed with large files with a small
 191    density of changes.  */
 192 XTERN bool speed_large_files;
 193
 194 /* Patterns that match file names to be excluded.  */
 195 XTERN struct exclude *excluded;
 196
 197 /* Don't discard lines.  This makes things slower (sometimes much
 198    slower) but will find a guaranteed minimal set of changes.  */
 199 XTERN bool minimal;
 200
 201 /* Name of program the user invoked (for error messages).  */
 202 XTERN char *program_name;
 203
 204 /* The strftime format to use for time strings.  */
 205 XTERN char const *time_format;
 206 \f
 207 /* The result of comparison is an "edit script": a chain of `struct change'.
 208    Each `struct change' represents one place where some lines are deleted
 209    and some are inserted.
 210
 211    LINE0 and LINE1 are the first affected lines in the two files (origin 0).
 212    DELETED is the number of lines deleted here from file 0.
 213    INSERTED is the number of lines inserted here in file 1.
 214
 215    If DELETED is 0 then LINE0 is the number of the line before
 216    which the insertion was done; vice versa for INSERTED and LINE1.  */
 217
 218 struct change
 219 {
 220   struct change *link;          /* Previous or next edit command  */
 221   lin inserted;                 /* # lines of file 1 changed here.  */
 222   lin deleted;                  /* # lines of file 0 changed here.  */
 223   lin line0;                    /* Line number of 1st deleted line.  */
 224   lin line1;                    /* Line number of 1st inserted line.  */
 225   bool ignore;                  /* Flag used in context.c.  */
 226 };
 227 \f
 228 /* Structures that describe the input files.  */
 229
 230 /* Data on one input file being compared.  */
 231
 232 struct file_data {
 233     int             desc;       /* File descriptor  */
 234     char const      *name;      /* File name  */
 235     struct stat     stat;       /* File status */
 236
 237     /* Buffer in which text of file is read.  */
 238     word *buffer;
 239
 240     /* Allocated size of buffer, in bytes.  Always a multiple of
 241        sizeof *buffer.  */
 242     size_t bufsize;
 243
 244     /* Number of valid bytes now in the buffer.  */
 245     size_t buffered;
 246
 247     /* Array of pointers to lines in the file.  */
 248     char const **linbuf;
 249
 250     /* linbuf_base <= buffered_lines <= valid_lines <= alloc_lines.
 251        linebuf[linbuf_base ... buffered_lines - 1] are possibly differing.
 252        linebuf[linbuf_base ... valid_lines - 1] contain valid data.
 253        linebuf[linbuf_base ... alloc_lines - 1] are allocated.  */
 254     lin linbuf_base, buffered_lines, valid_lines, alloc_lines;
 255
 256     /* Pointer to end of prefix of this file to ignore when hashing.  */
 257     char const *prefix_end;
 258
 259     /* Count of lines in the prefix.
 260        There are this many lines in the file before linbuf[0].  */
 261     lin prefix_lines;
 262
 263     /* Pointer to start of suffix of this file to ignore when hashing.  */
 264     char const *suffix_begin;
 265
 266     /* Vector, indexed by line number, containing an equivalence code for
 267        each line.  It is this vector that is actually compared with that
 268        of another file to generate differences.  */
 269     lin *equivs;
 270
 271     /* Vector, like the previous one except that
 272        the elements for discarded lines have been squeezed out.  */
 273     lin *undiscarded;
 274
 275     /* Vector mapping virtual line numbers (not counting discarded lines)
 276        to real ones (counting those lines).  Both are origin-0.  */
 277     lin *realindexes;
 278
 279     /* Total number of nondiscarded lines.  */
 280     lin nondiscarded_lines;
 281
 282     /* Vector, indexed by real origin-0 line number,
 283        containing 1 for a line that is an insertion or a deletion.
 284        The results of comparison are stored here.  */
 285     char *changed;
 286
 287     /* 1 if file ends in a line with no final newline.  */
 288     bool missing_newline;
 289
 290     /* 1 if at end of file.  */
 291     bool eof;
 292
 293     /* 1 more than the maximum equivalence value used for this or its
 294        sibling file.  */
 295     lin equiv_max;
 296 };
 297
 298 /* The file buffer, considered as an array of bytes rather than
 299    as an array of words.  */
 300 #define FILE_BUFFER(f) ((char *) (f)->buffer)
 301
 302 /* Data on two input files being compared.  */
 303
 304 struct comparison
 305   {
 306     struct file_data file[2];
 307     struct comparison const *parent;  /* parent, if a recursive comparison */
 308   };
 309
 310 /* Describe the two files currently being compared.  */
 311
 312 XTERN struct file_data files[2];
 313 \f
 314 /* Stdio stream to output diffs to.  */
 315
 316 XTERN FILE *outfile;
 317 \f
 318 /* Declare various functions.  */
 319
 320 /* analyze.c */
 321 int diff_2_files (struct comparison *);
 322
 323 /* context.c */
 324 void print_context_header (struct file_data[], bool);
 325 void print_context_script (struct change *, bool);
 326
 327 /* dir.c */
 328 int diff_dirs (struct comparison const *, int (*) (struct comparison const *, char const *, char const *));
 329
 330 /* ed.c */
 331 void print_ed_script (struct change *);
 332 void pr_forward_ed_script (struct change *);
 333
 334 /* ifdef.c */
 335 void print_ifdef_script (struct change *);
 336
 337 /* io.c */
 338 void file_block_read (struct file_data *, size_t);
 339 bool read_files (struct file_data[], bool);
 340
 341 /* normal.c */
 342 void print_normal_script (struct change *);
 343
 344 /* rcs.c */
 345 void print_rcs_script (struct change *);
 346
 347 /* side.c */
 348 void print_sdiff_script (struct change *);
 349
 350 /* util.c */
 351 extern char const change_letter[4];
 352 extern char const pr_program[];
 353 char *concat (char const *, char const *, char const *);
 354 char *dir_file_pathname (char const *, char const *);
 355 bool lines_differ (char const *, char const *);
 356 lin translate_line_number (struct file_data const *, lin);
 357 struct change *find_change (struct change *);
 358 struct change *find_reverse_change (struct change *);
 359 void *zalloc (size_t);
 360 enum changes analyze_hunk (struct change *, lin *, lin *, lin *, lin *);
 361 void begin_output (void);
 362 void debug_script (struct change *);
 363 void fatal (char const *) __attribute__((noreturn));
 364 void finish_output (void);
 365 void message (char const *, char const *, char const *);
 366 void message5 (char const *, char const *, char const *, char const *, char const *);
 367 void output_1_line (char const *, char const *, char const *, char const *);
 368 void perror_with_name (char const *);
 369 void pfatal_with_name (char const *) __attribute__((noreturn));
 370 void print_1_line (char const *, char const * const *);
 371 void print_message_queue (void);
 372 void print_number_range (char, struct file_data *, lin, lin);
 373 void print_script (struct change *, struct change * (*) (struct change *), void (*) (struct change *));
 374 void setup_output (char const *, char const *, bool);
 375 void translate_range (struct file_data const *, lin, lin, long int *, long int *);