contrib/diffutils/src/diff.h

   1 /* Shared definitions for GNU DIFF
   2
   3    Copyright (C) 1988-1989, 1991-1995, 1998, 2001-2002, 2004, 2009-2013,
   4    2015-2018 Free 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 /* When colors should be used in the output.  */
  43 enum colors_style
  44 {
  45   /* Never output colors.  */
  46   NEVER,
  47
  48   /* Output colors if the output is a terminal.  */
  49   AUTO,
  50
  51   /* Always output colors.  */
  52   ALWAYS,
  53 };
  54 \f
  55 /* Variables for command line options */
  56
  57 #ifndef GDIFF_MAIN
  58 # define XTERN extern
  59 #else
  60 # define XTERN
  61 #endif
  62
  63 enum output_style
  64 {
  65   /* No output style specified.  */
  66   OUTPUT_UNSPECIFIED,
  67
  68   /* Default output style.  */
  69   OUTPUT_NORMAL,
  70
  71   /* Output the differences with lines of context before and after (-c).  */
  72   OUTPUT_CONTEXT,
  73
  74   /* Output the differences in a unified context diff format (-u).  */
  75   OUTPUT_UNIFIED,
  76
  77   /* Output the differences as commands suitable for 'ed' (-e).  */
  78   OUTPUT_ED,
  79
  80   /* Output the diff as a forward ed script (-f).  */
  81   OUTPUT_FORWARD_ED,
  82
  83   /* Like -f, but output a count of changed lines in each "command" (-n).  */
  84   OUTPUT_RCS,
  85
  86   /* Output merged #ifdef'd file (-D).  */
  87   OUTPUT_IFDEF,
  88
  89   /* Output sdiff style (-y).  */
  90   OUTPUT_SDIFF
  91 };
  92
  93 /* True for output styles that are robust,
  94    i.e. can handle a file that ends in a non-newline.  */
  95 #define ROBUST_OUTPUT_STYLE(S) ((S) != OUTPUT_ED && (S) != OUTPUT_FORWARD_ED)
  96
  97 XTERN enum output_style output_style;
  98
  99 /* Define the current color context used to print a line.  */
 100 XTERN enum colors_style colors_style;
 101
 102 /* Nonzero if output cannot be generated for identical files.  */
 103 XTERN bool no_diff_means_no_output;
 104
 105 /* Number of lines of context to show in each set of diffs.
 106    This is zero when context is not to be shown.  */
 107 XTERN lin context;
 108
 109 /* Consider all files as text files (-a).
 110    Don't interpret codes over 0177 as implying a "binary file".  */
 111 XTERN bool text;
 112
 113 /* Number of lines to keep in identical prefix and suffix.  */
 114 XTERN lin horizon_lines;
 115
 116 /* The significance of white space during comparisons.  */
 117 enum DIFF_white_space
 118 {
 119   /* All white space is significant (the default).  */
 120   IGNORE_NO_WHITE_SPACE,
 121
 122   /* Ignore changes due to tab expansion (-E).  */
 123   IGNORE_TAB_EXPANSION,
 124
 125   /* Ignore changes in trailing horizontal white space (-Z).  */
 126   IGNORE_TRAILING_SPACE,
 127
 128   /* IGNORE_TAB_EXPANSION and IGNORE_TRAILING_SPACE are a special case
 129      because they are independent and can be ORed together, yielding
 130      IGNORE_TAB_EXPANSION_AND_TRAILING_SPACE.  */
 131   IGNORE_TAB_EXPANSION_AND_TRAILING_SPACE,
 132
 133   /* Ignore changes in horizontal white space (-b).  */
 134   IGNORE_SPACE_CHANGE,
 135
 136   /* Ignore all horizontal white space (-w).  */
 137   IGNORE_ALL_SPACE
 138 };
 139 XTERN enum DIFF_white_space ignore_white_space;
 140
 141 /* Ignore changes that affect only blank lines (-B).  */
 142 XTERN bool ignore_blank_lines;
 143
 144 /* Files can be compared byte-by-byte, as if they were binary.
 145    This depends on various options.  */
 146 XTERN bool files_can_be_treated_as_binary;
 147
 148 /* Ignore differences in case of letters (-i).  */
 149 XTERN bool ignore_case;
 150
 151 /* Ignore differences in case of letters in file names.  */
 152 XTERN bool ignore_file_name_case;
 153
 154 /* Act on symbolic links themselves rather than on their target
 155    (--no-dereference).  */
 156 XTERN bool no_dereference_symlinks;
 157
 158 /* File labels for '-c' output headers (--label).  */
 159 XTERN char *file_label[2];
 160
 161 /* Regexp to identify function-header lines (-F).  */
 162 XTERN struct re_pattern_buffer function_regexp;
 163
 164 /* Ignore changes that affect only lines matching this regexp (-I).  */
 165 XTERN struct re_pattern_buffer ignore_regexp;
 166
 167 /* Say only whether files differ, not how (-q).  */
 168 XTERN bool brief;
 169
 170 /* Expand tabs in the output so the text lines up properly
 171    despite the characters added to the front of each line (-t).  */
 172 XTERN bool expand_tabs;
 173
 174 /* Number of columns between tab stops.  */
 175 XTERN size_t tabsize;
 176
 177 /* Use a tab in the output, rather than a space, before the text of an
 178    input line, so as to keep the proper alignment in the input line
 179    without changing the characters in it (-T).  */
 180 XTERN bool initial_tab;
 181
 182 /* Do not output an initial space or tab before the text of an empty line.  */
 183 XTERN bool suppress_blank_empty;
 184
 185 /* Remove trailing carriage returns from input.  */
 186 XTERN bool strip_trailing_cr;
 187
 188 /* In directory comparison, specify file to start with (-S).
 189    This is used for resuming an aborted comparison.
 190    All file names less than this name are ignored.  */
 191 XTERN char const *starting_file;
 192
 193 /* Pipe each file's output through pr (-l).  */
 194 XTERN bool paginate;
 195
 196 /* Line group formats for unchanged, old, new, and changed groups.  */
 197 XTERN char const *group_format[CHANGED + 1];
 198
 199 /* Line formats for unchanged, old, and new lines.  */
 200 XTERN char const *line_format[NEW + 1];
 201
 202 /* If using OUTPUT_SDIFF print extra information to help the sdiff filter.  */
 203 XTERN bool sdiff_merge_assist;
 204
 205 /* Tell OUTPUT_SDIFF to show only the left version of common lines.  */
 206 XTERN bool left_column;
 207
 208 /* Tell OUTPUT_SDIFF to not show common lines.  */
 209 XTERN bool suppress_common_lines;
 210
 211 /* The half line width and column 2 offset for OUTPUT_SDIFF.  */
 212 XTERN size_t sdiff_half_width;
 213 XTERN size_t sdiff_column2_offset;
 214
 215 /* String containing all the command options diff received,
 216    with spaces between and at the beginning but none at the end.
 217    If there were no options given, this string is empty.  */
 218 XTERN char *switch_string;
 219
 220 /* Use heuristics for better speed with large files with a small
 221    density of changes.  */
 222 XTERN bool speed_large_files;
 223
 224 /* Patterns that match file names to be excluded.  */
 225 XTERN struct exclude *excluded;
 226
 227 /* Don't discard lines.  This makes things slower (sometimes much
 228    slower) but will find a guaranteed minimal set of changes.  */
 229 XTERN bool minimal;
 230
 231 /* The strftime format to use for time strings.  */
 232 XTERN char const *time_format;
 233 \f
 234 /* The result of comparison is an "edit script": a chain of 'struct change'.
 235    Each 'struct change' represents one place where some lines are deleted
 236    and some are inserted.
 237
 238    LINE0 and LINE1 are the first affected lines in the two files (origin 0).
 239    DELETED is the number of lines deleted here from file 0.
 240    INSERTED is the number of lines inserted here in file 1.
 241
 242    If DELETED is 0 then LINE0 is the number of the line before
 243    which the insertion was done; vice versa for INSERTED and LINE1.  */
 244
 245 struct change
 246 {
 247   struct change *link;          /* Previous or next edit command  */
 248   lin inserted;                 /* # lines of file 1 changed here.  */
 249   lin deleted;                  /* # lines of file 0 changed here.  */
 250   lin line0;                    /* Line number of 1st deleted line.  */
 251   lin line1;                    /* Line number of 1st inserted line.  */
 252   bool ignore;                  /* Flag used in context.c.  */
 253 };
 254 \f
 255 /* Structures that describe the input files.  */
 256
 257 /* Data on one input file being compared.  */
 258
 259 struct file_data {
 260     int             desc;       /* File descriptor  */
 261     char const      *name;      /* File name  */
 262     struct stat     stat;       /* File status */
 263
 264     /* Buffer in which text of file is read.  */
 265     word *buffer;
 266
 267     /* Allocated size of buffer, in bytes.  Always a multiple of
 268        sizeof *buffer.  */
 269     size_t bufsize;
 270
 271     /* Number of valid bytes now in the buffer.  */
 272     size_t buffered;
 273
 274     /* Array of pointers to lines in the file.  */
 275     char const **linbuf;
 276
 277     /* linbuf_base <= buffered_lines <= valid_lines <= alloc_lines.
 278        linebuf[linbuf_base ... buffered_lines - 1] are possibly differing.
 279        linebuf[linbuf_base ... valid_lines - 1] contain valid data.
 280        linebuf[linbuf_base ... alloc_lines - 1] are allocated.  */
 281     lin linbuf_base, buffered_lines, valid_lines, alloc_lines;
 282
 283     /* Pointer to end of prefix of this file to ignore when hashing.  */
 284     char const *prefix_end;
 285
 286     /* Count of lines in the prefix.
 287        There are this many lines in the file before linbuf[0].  */
 288     lin prefix_lines;
 289
 290     /* Pointer to start of suffix of this file to ignore when hashing.  */
 291     char const *suffix_begin;
 292
 293     /* Vector, indexed by line number, containing an equivalence code for
 294        each line.  It is this vector that is actually compared with that
 295        of another file to generate differences.  */
 296     lin *equivs;
 297
 298     /* Vector, like the previous one except that
 299        the elements for discarded lines have been squeezed out.  */
 300     lin *undiscarded;
 301
 302     /* Vector mapping virtual line numbers (not counting discarded lines)
 303        to real ones (counting those lines).  Both are origin-0.  */
 304     lin *realindexes;
 305
 306     /* Total number of nondiscarded lines.  */
 307     lin nondiscarded_lines;
 308
 309     /* Vector, indexed by real origin-0 line number,
 310        containing 1 for a line that is an insertion or a deletion.
 311        The results of comparison are stored here.  */
 312     char *changed;
 313
 314     /* 1 if file ends in a line with no final newline.  */
 315     bool missing_newline;
 316
 317     /* 1 if at end of file.  */
 318     bool eof;
 319
 320     /* 1 more than the maximum equivalence value used for this or its
 321        sibling file.  */
 322     lin equiv_max;
 323 };
 324
 325 /* The file buffer, considered as an array of bytes rather than
 326    as an array of words.  */
 327 #define FILE_BUFFER(f) ((char *) (f)->buffer)
 328
 329 /* Data on two input files being compared.  */
 330
 331 struct comparison
 332   {
 333     struct file_data file[2];
 334     struct comparison const *parent;  /* parent, if a recursive comparison */
 335   };
 336
 337 /* Describe the two files currently being compared.  */
 338
 339 XTERN struct file_data files[2];
 340 \f
 341 /* Stdio stream to output diffs to.  */
 342
 343 XTERN FILE *outfile;
 344 \f
 345 /* Declare various functions.  */
 346
 347 /* analyze.c */
 348 extern int diff_2_files (struct comparison *);
 349
 350 /* context.c */
 351 extern void print_context_header (struct file_data[], char const * const *, bool);
 352 extern void print_context_script (struct change *, bool);
 353
 354 /* dir.c */
 355 extern int diff_dirs (struct comparison const *,
 356                       int (*) (struct comparison const *,
 357                                char const *, char const *));
 358 extern char *find_dir_file_pathname (char const *, char const *);
 359
 360 /* ed.c */
 361 extern void print_ed_script (struct change *);
 362 extern void pr_forward_ed_script (struct change *);
 363
 364 /* ifdef.c */
 365 extern void print_ifdef_script (struct change *);
 366
 367 /* io.c */
 368 extern void file_block_read (struct file_data *, size_t);
 369 extern bool read_files (struct file_data[], bool);
 370
 371 /* normal.c */
 372 extern void print_normal_script (struct change *);
 373
 374 /* rcs.c */
 375 extern void print_rcs_script (struct change *);
 376
 377 /* side.c */
 378 extern void print_sdiff_script (struct change *);
 379
 380 /* util.c */
 381 extern char const change_letter[4];
 382 extern char const pr_program[];
 383 extern char *concat (char const *, char const *, char const *);
 384 extern bool lines_differ (char const *, char const *) _GL_ATTRIBUTE_PURE;
 385 extern lin translate_line_number (struct file_data const *, lin);
 386 extern struct change *find_change (struct change *);
 387 extern struct change *find_reverse_change (struct change *);
 388 extern void *zalloc (size_t);
 389 extern enum changes analyze_hunk (struct change *, lin *, lin *, lin *, lin *);
 390 extern void begin_output (void);
 391 extern void debug_script (struct change *);
 392 extern void fatal (char const *) __attribute__((noreturn));
 393 extern void finish_output (void);
 394 extern void message (char const *, char const *, char const *);
 395 extern void message5 (char const *, char const *, char const *,
 396                       char const *, char const *);
 397 extern void output_1_line (char const *, char const *, char const *,
 398                            char const *);
 399 extern void perror_with_name (char const *);
 400 extern void pfatal_with_name (char const *) __attribute__((noreturn));
 401 extern void print_1_line (char const *, char const * const *);
 402 extern void print_1_line_nl (char const *, char const * const *, bool);
 403 extern void print_message_queue (void);
 404 extern void print_number_range (char, struct file_data *, lin, lin);
 405 extern void print_script (struct change *, struct change * (*) (struct change *),
 406                           void (*) (struct change *));
 407 extern void setup_output (char const *, char const *, bool);
 408 extern void translate_range (struct file_data const *, lin, lin,
 409                              printint *, printint *);
 410
 411 enum color_context
 412 {
 413   HEADER_CONTEXT,
 414   ADD_CONTEXT,
 415   DELETE_CONTEXT,
 416   RESET_CONTEXT,
 417   LINE_NUMBER_CONTEXT,
 418 };
 419
 420 XTERN bool presume_output_tty;
 421
 422 extern void set_color_context (enum color_context color_context);
 423 extern void set_color_palette (char const *palette);