1 | /* Shared definitions for GNU DIFF
|
---|
2 |
|
---|
3 | Copyright (C) 1988, 1989, 1991, 1992, 1993, 1994, 1995, 1998, 2001,
|
---|
4 | 2002 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 <stdio.h>
|
---|
25 |
|
---|
26 | #define TAB_WIDTH 8
|
---|
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 | |
---|
44 |
|
---|
45 | /* Variables for command line options */
|
---|
46 |
|
---|
47 | #ifndef GDIFF_MAIN
|
---|
48 | # define XTERN extern
|
---|
49 | #else
|
---|
50 | # define XTERN
|
---|
51 | #endif
|
---|
52 |
|
---|
53 | enum output_style
|
---|
54 | {
|
---|
55 | /* No output style specified. */
|
---|
56 | OUTPUT_UNSPECIFIED,
|
---|
57 |
|
---|
58 | /* Default output style. */
|
---|
59 | OUTPUT_NORMAL,
|
---|
60 |
|
---|
61 | /* Output the differences with lines of context before and after (-c). */
|
---|
62 | OUTPUT_CONTEXT,
|
---|
63 |
|
---|
64 | /* Output the differences in a unified context diff format (-u). */
|
---|
65 | OUTPUT_UNIFIED,
|
---|
66 |
|
---|
67 | /* Output the differences as commands suitable for `ed' (-e). */
|
---|
68 | OUTPUT_ED,
|
---|
69 |
|
---|
70 | /* Output the diff as a forward ed script (-f). */
|
---|
71 | OUTPUT_FORWARD_ED,
|
---|
72 |
|
---|
73 | /* Like -f, but output a count of changed lines in each "command" (-n). */
|
---|
74 | OUTPUT_RCS,
|
---|
75 |
|
---|
76 | /* Output merged #ifdef'd file (-D). */
|
---|
77 | OUTPUT_IFDEF,
|
---|
78 |
|
---|
79 | /* Output sdiff style (-y). */
|
---|
80 | OUTPUT_SDIFF
|
---|
81 | };
|
---|
82 |
|
---|
83 | /* True for output styles that are robust,
|
---|
84 | i.e. can handle a file that ends in a non-newline. */
|
---|
85 | #define ROBUST_OUTPUT_STYLE(S) ((S) != OUTPUT_ED && (S) != OUTPUT_FORWARD_ED)
|
---|
86 |
|
---|
87 | XTERN enum output_style output_style;
|
---|
88 |
|
---|
89 | /* Nonzero if output cannot be generated for identical files. */
|
---|
90 | XTERN bool no_diff_means_no_output;
|
---|
91 |
|
---|
92 | /* Number of lines of context to show in each set of diffs.
|
---|
93 | This is zero when context is not to be shown. */
|
---|
94 | XTERN lin context;
|
---|
95 |
|
---|
96 | /* Consider all files as text files (-a).
|
---|
97 | Don't interpret codes over 0177 as implying a "binary file". */
|
---|
98 | XTERN bool text;
|
---|
99 |
|
---|
100 | /* Number of lines to keep in identical prefix and suffix. */
|
---|
101 | XTERN lin horizon_lines;
|
---|
102 |
|
---|
103 | /* The significance of white space during comparisons. */
|
---|
104 | XTERN enum
|
---|
105 | {
|
---|
106 | /* All white space is significant (the default). */
|
---|
107 | IGNORE_NO_WHITE_SPACE,
|
---|
108 |
|
---|
109 | /* Ignore changes due to tab expansion (-E). */
|
---|
110 | IGNORE_TAB_EXPANSION,
|
---|
111 |
|
---|
112 | /* Ignore changes in horizontal white space (-b). */
|
---|
113 | IGNORE_SPACE_CHANGE,
|
---|
114 |
|
---|
115 | /* Ignore all horizontal white space (-w). */
|
---|
116 | IGNORE_ALL_SPACE
|
---|
117 | } ignore_white_space;
|
---|
118 |
|
---|
119 | /* Ignore changes that affect only blank lines (-B). */
|
---|
120 | XTERN bool ignore_blank_lines;
|
---|
121 |
|
---|
122 | /* Files can be compared byte-by-byte, as if they were binary.
|
---|
123 | This depends on various options. */
|
---|
124 | XTERN bool files_can_be_treated_as_binary;
|
---|
125 |
|
---|
126 | /* Ignore differences in case of letters (-i). */
|
---|
127 | XTERN bool ignore_case;
|
---|
128 |
|
---|
129 | /* Ignore differences in case of letters in file names. */
|
---|
130 | XTERN bool ignore_file_name_case;
|
---|
131 |
|
---|
132 | /* File labels for `-c' output headers (--label). */
|
---|
133 | XTERN char *file_label[2];
|
---|
134 |
|
---|
135 | /* Regexp to identify function-header lines (-F). */
|
---|
136 | XTERN struct re_pattern_buffer function_regexp;
|
---|
137 |
|
---|
138 | /* Ignore changes that affect only lines matching this regexp (-I). */
|
---|
139 | XTERN struct re_pattern_buffer ignore_regexp;
|
---|
140 |
|
---|
141 | /* Say only whether files differ, not how (-q). */
|
---|
142 | XTERN bool brief;
|
---|
143 |
|
---|
144 | /* Expand tabs in the output so the text lines up properly
|
---|
145 | despite the characters added to the front of each line (-t). */
|
---|
146 | XTERN bool expand_tabs;
|
---|
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 | /* Remove trailing carriage returns from input. */
|
---|
154 | XTERN bool strip_trailing_cr;
|
---|
155 |
|
---|
156 | /* In directory comparison, specify file to start with (-S).
|
---|
157 | This is used for resuming an aborted comparison.
|
---|
158 | All file names less than this name are ignored. */
|
---|
159 | XTERN char const *starting_file;
|
---|
160 |
|
---|
161 | /* Pipe each file's output through pr (-l). */
|
---|
162 | XTERN bool paginate;
|
---|
163 |
|
---|
164 | /* Line group formats for unchanged, old, new, and changed groups. */
|
---|
165 | XTERN char const *group_format[CHANGED + 1];
|
---|
166 |
|
---|
167 | /* Line formats for unchanged, old, and new lines. */
|
---|
168 | XTERN char const *line_format[NEW + 1];
|
---|
169 |
|
---|
170 | /* If using OUTPUT_SDIFF print extra information to help the sdiff filter. */
|
---|
171 | XTERN bool sdiff_merge_assist;
|
---|
172 |
|
---|
173 | /* Tell OUTPUT_SDIFF to show only the left version of common lines. */
|
---|
174 | XTERN bool left_column;
|
---|
175 |
|
---|
176 | /* Tell OUTPUT_SDIFF to not show common lines. */
|
---|
177 | XTERN bool suppress_common_lines;
|
---|
178 |
|
---|
179 | /* The half line width and column 2 offset for OUTPUT_SDIFF. */
|
---|
180 | XTERN unsigned int sdiff_half_width;
|
---|
181 | XTERN unsigned int sdiff_column2_offset;
|
---|
182 |
|
---|
183 | /* String containing all the command options diff received,
|
---|
184 | with spaces between and at the beginning but none at the end.
|
---|
185 | If there were no options given, this string is empty. */
|
---|
186 | XTERN char *switch_string;
|
---|
187 |
|
---|
188 | /* Use heuristics for better speed with large files with a small
|
---|
189 | density of changes. */
|
---|
190 | XTERN bool speed_large_files;
|
---|
191 |
|
---|
192 | /* Patterns that match file names to be excluded. */
|
---|
193 | XTERN struct exclude *excluded;
|
---|
194 |
|
---|
195 | /* Don't discard lines. This makes things slower (sometimes much
|
---|
196 | slower) but will find a guaranteed minimal set of changes. */
|
---|
197 | XTERN bool minimal;
|
---|
198 |
|
---|
199 | /* Name of program the user invoked (for error messages). */
|
---|
200 | XTERN char *program_name;
|
---|
201 |
|
---|
202 | /* The strftime format to use for time strings. */
|
---|
203 | XTERN char const *time_format;
|
---|
204 | |
---|
205 |
|
---|
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 | |
---|
227 |
|
---|
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 TRUE for a line that is an insertion or a deletion.
|
---|
284 | The results of comparison are stored here. */
|
---|
285 | bool *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 | |
---|
314 |
|
---|
315 | /* Stdio stream to output diffs to. */
|
---|
316 |
|
---|
317 | XTERN FILE *outfile;
|
---|
318 | |
---|
319 |
|
---|
320 | /* Declare various functions. */
|
---|
321 |
|
---|
322 | /* analyze.c */
|
---|
323 | int diff_2_files (struct comparison *);
|
---|
324 |
|
---|
325 | /* context.c */
|
---|
326 | void print_context_header (struct file_data[], bool);
|
---|
327 | void print_context_script (struct change *, bool);
|
---|
328 |
|
---|
329 | /* dir.c */
|
---|
330 | int diff_dirs (struct comparison const *, int (*) (struct comparison const *, char const *, char const *));
|
---|
331 |
|
---|
332 | /* ed.c */
|
---|
333 | void print_ed_script (struct change *);
|
---|
334 | void pr_forward_ed_script (struct change *);
|
---|
335 |
|
---|
336 | /* ifdef.c */
|
---|
337 | void print_ifdef_script (struct change *);
|
---|
338 |
|
---|
339 | /* io.c */
|
---|
340 | void file_block_read (struct file_data *, size_t);
|
---|
341 | bool read_files (struct file_data[], bool);
|
---|
342 |
|
---|
343 | /* normal.c */
|
---|
344 | void print_normal_script (struct change *);
|
---|
345 |
|
---|
346 | /* rcs.c */
|
---|
347 | void print_rcs_script (struct change *);
|
---|
348 |
|
---|
349 | /* side.c */
|
---|
350 | void print_sdiff_script (struct change *);
|
---|
351 |
|
---|
352 | /* util.c */
|
---|
353 | extern char const change_letter[4];
|
---|
354 | extern char const pr_program[];
|
---|
355 | char *concat (char const *, char const *, char const *);
|
---|
356 | char *dir_file_pathname (char const *, char const *);
|
---|
357 | bool lines_differ (char const *, char const *);
|
---|
358 | lin translate_line_number (struct file_data const *, lin);
|
---|
359 | struct change *find_change (struct change *);
|
---|
360 | struct change *find_reverse_change (struct change *);
|
---|
361 | void *zalloc (size_t);
|
---|
362 | enum changes analyze_hunk (struct change *, lin *, lin *, lin *, lin *);
|
---|
363 | void begin_output (void);
|
---|
364 | void debug_script (struct change *);
|
---|
365 | void fatal (char const *) __attribute__((noreturn));
|
---|
366 | void finish_output (void);
|
---|
367 | void message (char const *, char const *, char const *);
|
---|
368 | void message5 (char const *, char const *, char const *, char const *, char const *);
|
---|
369 | void output_1_line (char const *, char const *, char const *, char const *);
|
---|
370 | void perror_with_name (char const *);
|
---|
371 | void pfatal_with_name (char const *) __attribute__((noreturn));
|
---|
372 | void print_1_line (char const *, char const * const *);
|
---|
373 | void print_message_queue (void);
|
---|
374 | void print_number_range (char, struct file_data *, lin, lin);
|
---|
375 | void print_script (struct change *, struct change * (*) (struct change *), void (*) (struct change *));
|
---|
376 | void setup_output (char const *, char const *, bool);
|
---|
377 | void translate_range (struct file_data const *, lin, lin, long *, long *);
|
---|
378 |
|
---|
379 | /* version.c */
|
---|
380 | extern char const version_string[];
|
---|