1PATCH(1) General Commands Manual PATCH(1)
2
6 patch - apply a diff file to an original
7
9 patch [options] [originalfile [patchfile]]
10 but usually just
12 patch -pnum <patchfile
14
16 patch takes a patch file patchfile containing a difference listing pro‐
17 duced by the diff program and applies those differences to one or more
18 original files, producing patched versions. Normally the patched ver‐
19 sions are put in place of the originals. Backups can be made; see the
20 -b or --backup option. The names of the files to be patched are usu‐
21 ally taken from the patch file, but if there's just one file to be
22 patched it can be specified on the command line as originalfile.
23 Upon startup, patch attempts to determine the type of the diff listing,
25 unless overruled by a -c (--context), -e (--ed), -n (--normal), or -u
26 (--unified) option. Context diffs (old-style, new-style, and unified)
27 and normal diffs are applied by the patch program itself, while ed
28 diffs are simply fed to the ed(1) editor via a pipe.
29 patch tries to skip any leading garbage, apply the diff, and then skip
31 any trailing garbage. Thus you could feed an article or message con‐
32 taining a diff listing to patch, and it should work. If the entire
33 diff is indented by a consistent amount, if lines end in CRLF, or if a
34 diff is encapsulated one or more times by prepending "- " to lines
35 starting with "-" as specified by Internet RFC 934, this is taken into
36 account. After removing indenting or encapsulation, lines beginning
37 with # are ignored, as they are considered to be comments.
38 With context diffs, and to a lesser extent with normal diffs, patch can
40 detect when the line numbers mentioned in the patch are incorrect, and
41 attempts to find the correct place to apply each hunk of the patch. As
42 a first guess, it takes the line number mentioned for the hunk, plus or
43 minus any offset used in applying the previous hunk. If that is not
44 the correct place, patch scans both forwards and backwards for a set of
45 lines matching the context given in the hunk. First patch looks for a
46 place where all lines of the context match. If no such place is found,
47 and it's a context diff, and the maximum fuzz factor is set to 1 or
48 more, then another scan takes place ignoring the first and last line of
49 context. If that fails, and the maximum fuzz factor is set to 2 or
50 more, the first two and last two lines of context are ignored, and
51 another scan is made. (The default maximum fuzz factor is 2.)
52 Hunks with less prefix context than suffix context (after applying
54 fuzz) must apply at the start of the file if their first line number
55 is 1. Hunks with more prefix context than suffix context (after apply‐
56 ing fuzz) must apply at the end of the file.
57 If patch cannot find a place to install that hunk of the patch, it puts
59 the hunk out to a reject file, which normally is the name of the output
60 file plus a .rej suffix, or # if .rej would generate a file name that
61 is too long (if even appending the single character # makes the file
62 name too long, then # replaces the file name's last character).
63 The rejected hunk comes out in unified or context diff format. If the
65 input was a normal diff, many of the contexts are simply null. The
66 line numbers on the hunks in the reject file may be different than in
67 the patch file: they reflect the approximate location patch thinks the
68 failed hunks belong in the new file rather than the old one.
69 As each hunk is completed, you are told if the hunk failed, and if so
71 which line (in the new file) patch thought the hunk should go on. If
72 the hunk is installed at a different line from the line number speci‐
73 fied in the diff, you are told the offset. A single large offset may
74 indicate that a hunk was installed in the wrong place. You are also
75 told if a fuzz factor was used to make the match, in which case you
76 should also be slightly suspicious. If the --verbose option is given,
77 you are also told about hunks that match exactly.
78 If no original file origfile is specified on the command line, patch
80 tries to figure out from the leading garbage what the name of the file
81 to edit is, using the following rules.
82 First, patch takes an ordered list of candidate file names as follows:
84 · If the header is that of a context diff, patch takes the old and new
86 file names in the header. A name is ignored if it does not have
87 enough slashes to satisfy the -pnum or --strip=num option. The name
88 /dev/null is also ignored.
89 · If there is an Index: line in the leading garbage and if either the
91 old and new names are both absent or if patch is conforming to
92 POSIX, patch takes the name in the Index: line.
93 · For the purpose of the following rules, the candidate file names are
95 considered to be in the order (old, new, index), regardless of the
96 order that they appear in the header.
97 Then patch selects a file name from the candidate list as follows:
99 · If some of the named files exist, patch selects the first name if
101 conforming to POSIX, and the best name otherwise.
102 · If patch is not ignoring RCS, ClearCase, Perforce, and SCCS (see the
104 -g num or --get=num option), and no named files exist but an RCS,
105 ClearCase, Perforce, or SCCS master is found, patch selects the
106 first named file with an RCS, ClearCase, Perforce, or SCCS master.
107 · If no named files exist, no RCS, ClearCase, Perforce, or SCCS master
109 was found, some names are given, patch is not conforming to POSIX,
110 and the patch appears to create a file, patch selects the best name
111 requiring the creation of the fewest directories.
112 · If no file name results from the above heuristics, you are asked for
114 the name of the file to patch, and patch selects that name.
115 To determine the best of a nonempty list of file names, patch first
117 takes all the names with the fewest path name components; of those, it
118 then takes all the names with the shortest basename; of those, it then
119 takes all the shortest names; finally, it takes the first remaining
120 name.
121 Additionally, if the leading garbage contains a Prereq: line, patch
123 takes the first word from the prerequisites line (normally a version
124 number) and checks the original file to see if that word can be found.
125 If not, patch asks for confirmation before proceeding.
126 The upshot of all this is that you should be able to say, while in a
128 news interface, something like the following:
129 | patch -d /usr/src/local/blurfl
131 and patch a file in the blurfl directory directly from the article con‐
133 taining the patch.
134 If the patch file contains more than one patch, patch tries to apply
136 each of them as if they came from separate patch files. This means,
137 among other things, that it is assumed that the name of the file to
138 patch must be determined for each diff listing, and that the garbage
139 before each diff listing contains interesting things such as file names
140 and revision level, as mentioned previously.
141
143 -b or --backup
144 Make backup files. That is, when patching a file, rename or copy
145 the original instead of removing it. When backing up a file that
146 does not exist, an empty, unreadable backup file is created as a
147 placeholder to represent the nonexistent file. See the -V or --ver‐
148 sion-control option for details about how backup file names are
149 determined.
150 --backup-if-mismatch
152 Back up a file if the patch does not match the file exactly and if
153 backups are not otherwise requested. This is the default unless
154 patch is conforming to POSIX.
155 --no-backup-if-mismatch
157 Do not back up a file if the patch does not match the file exactly
158 and if backups are not otherwise requested. This is the default if
159 patch is conforming to POSIX.
160 -B pref or --prefix=pref
162 Use the simple method to determine backup file names (see the -V
163 method or --version-control method option), and append pref to a
164 file name when generating its backup file name. For example, with
165 -B /junk/ the simple backup file name for src/patch/util.c is
166 /junk/src/patch/util.c.
167 --binary
169 Write all files in binary mode, except for standard output and
170 /dev/tty. When reading, disable the heuristic for transforming CRLF
171 line endings into LF line endings. This option is needed on POSIX
172 systems when applying patches generated on non-POSIX systems to non-
173 POSIX files. (On POSIX systems, file reads and writes never trans‐
174 form line endings. On Windows, reads and writes do transform line
175 endings by default, and patches should be generated by diff --binary
176 when line endings are significant.)
177 -c or --context
179 Interpret the patch file as a ordinary context diff.
180 -d dir or --directory=dir
182 Change to the directory dir immediately, before doing anything else.
183 -D define or --ifdef=define
185 Use the #ifdef ... #endif construct to mark changes, with define as
186 the differentiating symbol.
187 --dry-run
189 Print the results of applying the patches without actually changing
190 any files.
191 -e or --ed
193 Interpret the patch file as an ed script.
194 -E or --remove-empty-files
196 Remove output files that are empty after the patches have been
197 applied. Normally this option is unnecessary, since patch can exam‐
198 ine the time stamps on the header to determine whether a file should
199 exist after patching. However, if the input is not a context diff
200 or if patch is conforming to POSIX, patch does not remove empty
201 patched files unless this option is given. When patch removes a
202 file, it also attempts to remove any empty ancestor directories.
203 -f or --force
205 Assume that the user knows exactly what he or she is doing, and do
206 not ask any questions. Skip patches whose headers do not say which
207 file is to be patched; patch files even though they have the wrong
208 version for the Prereq: line in the patch; and assume that patches
209 are not reversed even if they look like they are. This option does
210 not suppress commentary; use -s for that.
211 -F num or --fuzz=num
213 Set the maximum fuzz factor. This option only applies to diffs that
214 have context, and causes patch to ignore up to that many lines of
215 context in looking for places to install a hunk. Note that a larger
216 fuzz factor increases the odds of a faulty patch. The default fuzz
217 factor is 2. A fuzz factor greater than or equal to the number of
218 lines of context in the context diff, ordinarily 3, ignores all con‐
219 text.
220 -g num or --get=num
222 This option controls patch's actions when a file is under RCS or
223 SCCS control, and does not exist or is read-only and matches the
224 default version, or when a file is under ClearCase or Perforce con‐
225 trol and does not exist. If num is positive, patch gets (or checks
226 out) the file from the revision control system; if zero, patch
227 ignores RCS, ClearCase, Perforce, and SCCS and does not get the
228 file; and if negative, patch asks the user whether to get the file.
229 The default value of this option is given by the value of the
230 PATCH_GET environment variable if it is set; if not, the default
231 value is zero.
232 --help
234 Print a summary of options and exit.
235 -i patchfile or --input=patchfile
237 Read the patch from patchfile. If patchfile is -, read from stan‐
238 dard input, the default.
239 -l or --ignore-whitespace
241 Match patterns loosely, in case tabs or spaces have been munged in
242 your files. Any sequence of one or more blanks in the patch file
243 matches any sequence in the original file, and sequences of blanks
244 at the ends of lines are ignored. Normal characters must still
245 match exactly. Each line of the context must still match a line in
246 the original file.
247 --merge or --merge=merge or --merge=diff3
249 Merge a patch file into the original files similar to diff3(1) or
250 merge(1). If a conflict is found, patch outputs a warning and
251 brackets the conflict with <<<<<<< and >>>>>>> lines. A typical
252 conflict will look like this:
253 <<<<<<<
255 lines from the original file
256 |||||||
257 original lines from the patch
258 =======
259 new lines from the patch
260 >>>>>>>
261 The optional argument of --merge determines the output format for
263 conflicts: the diff3 format shows the ||||||| section with the orig‐
264 inal lines from the patch; in the merge format, this section is
265 missing. The merge format is the default.
266 This option implies --forward and does not take the --fuzz=num
268 option into account.
269 -n or --normal
271 Interpret the patch file as a normal diff.
272 -N or --forward
274 Ignore patches that seem to be reversed or already applied. See
275 also -R.
276 -o outfile or --output=outfile
278 Send output to outfile instead of patching files in place. Do not
279 use this option if outfile is one of the files to be patched. When
280 outfile is -, send output to standard output, and send any messages
281 that would usually go to standard output to standard error.
282 -pnum or --strip=num
284 Strip the smallest prefix containing num leading slashes from each
285 file name found in the patch file. A sequence of one or more adja‐
286 cent slashes is counted as a single slash. This controls how file
287 names found in the patch file are treated, in case you keep your
288 files in a different directory than the person who sent out the
289 patch. For example, supposing the file name in the patch file was
290 /u/howard/src/blurfl/blurfl.c
292 setting -p0 gives the entire file name unmodified, -p1 gives
294 u/howard/src/blurfl/blurfl.c
296 without the leading slash, -p4 gives
298 blurfl/blurfl.c
300 and not specifying -p at all just gives you blurfl.c. Whatever you
302 end up with is looked for either in the current directory, or the
303 directory specified by the -d option.
304 --posix
306 Conform more strictly to the POSIX standard, as follows.
307 · Take the first existing file from the list (old, new, index) when
309 intuiting file names from diff headers.
310 · Do not remove files that are empty after patching.
312 · Do not ask whether to get files from RCS, ClearCase, Perforce, or
314 SCCS.
315 · Require that all options precede the files in the command line.
317 · Do not backup files when there is a mismatch.
319 --quoting-style=word
321 Use style word to quote output names. The word should be one of the
322 following:
323 literal
325 Output names as-is.
326 shell Quote names for the shell if they contain shell metacharac‐
328 ters or would cause ambiguous output.
329 shell-always
331 Quote names for the shell, even if they would normally not
332 require quoting.
333 c Quote names as for a C language string.
335 escape Quote as with c except omit the surrounding double-quote
337 characters.
338 You can specify the default value of the --quoting-style option with
340 the environment variable QUOTING_STYLE. If that environment vari‐
341 able is not set, the default value is shell.
342 -r rejectfile or --reject-file=rejectfile
344 Put rejects into rejectfile instead of the default .rej file. When
345 rejectfile is -, discard rejects.
346 -R or --reverse
348 Assume that this patch was created with the old and new files
349 swapped. (Yes, I'm afraid that does happen occasionally, human
350 nature being what it is.) patch attempts to swap each hunk around
351 before applying it. Rejects come out in the swapped format. The -R
352 option does not work with ed diff scripts because there is too lit‐
353 tle information to reconstruct the reverse operation.
354 If the first hunk of a patch fails, patch reverses the hunk to see
356 if it can be applied that way. If it can, you are asked if you want
357 to have the -R option set. If it can't, the patch continues to be
358 applied normally. (Note: this method cannot detect a reversed patch
359 if it is a normal diff and if the first command is an append (i.e.
360 it should have been a delete) since appends always succeed, due to
361 the fact that a null context matches anywhere. Luckily, most
362 patches add or change lines rather than delete them, so most
363 reversed normal diffs begin with a delete, which fails, triggering
364 the heuristic.)
365 --reject-format=format
367 Produce reject files in the specified format (either context or uni‐
368 fied). Without this option, rejected hunks come out in unified diff
369 format if the input patch was of that format, otherwise in ordinary
370 context diff form.
371 -s or --silent or --quiet
373 Work silently, unless an error occurs.
374 -t or --batch
376 Suppress questions like -f, but make some different assumptions:
377 skip patches whose headers do not contain file names (the same as
378 -f); skip patches for which the file has the wrong version for the
379 Prereq: line in the patch; and assume that patches are reversed if
380 they look like they are.
381 -T or --set-time
383 Set the modification and access times of patched files from time
384 stamps given in context diff headers, assuming that the context diff
385 headers use local time. This option is not recommended, because
386 patches using local time cannot easily be used by people in other
387 time zones, and because local time stamps are ambiguous when local
388 clocks move backwards during daylight-saving time adjustments.
389 Instead of using this option, generate patches with UTC and use the
390 -Z or --set-utc option instead.
391 -u or --unified
393 Interpret the patch file as a unified context diff.
394 -v or --version
396 Print out patch's revision header and patch level, and exit.
397 -V method or --version-control=method
399 Use method to determine backup file names. The method can also be
400 given by the PATCH_VERSION_CONTROL (or, if that's not set, the VER‐
401 SION_CONTROL) environment variable, which is overridden by this
402 option. The method does not affect whether backup files are made;
403 it affects only the names of any backup files that are made.
404 The value of method is like the GNU Emacs `version-control' vari‐
406 able; patch also recognizes synonyms that are more descriptive. The
407 valid values for method are (unique abbreviations are accepted):
408 existing or nil
410 Make numbered backups of files that already have them, otherwise
411 simple backups. This is the default.
412 numbered or t
414 Make numbered backups. The numbered backup file name for F is
415 F.~N~ where N is the version number.
416 simple or never
418 Make simple backups. The -B or --prefix, -Y or --basename-pre‐
419 fix, and -z or --suffix options specify the simple backup file
420 name. If none of these options are given, then a simple backup
421 suffix is used; it is the value of the SIMPLE_BACKUP_SUFFIX envi‐
422 ronment variable if set, and is .orig otherwise.
423 With numbered or simple backups, if the backup file name is too
425 long, the backup suffix ~ is used instead; if even appending ~ would
426 make the name too long, then ~ replaces the last character of the
427 file name.
428 --verbose
430 Output extra information about the work being done.
431 -x num or --debug=num
433 Set internal debugging flags of interest only to patch patchers.
434 -Y pref or --basename-prefix=pref
436 Use the simple method to determine backup file names (see the -V
437 method or --version-control method option), and prefix pref to the
438 basename of a file name when generating its backup file name. For
439 example, with -Y .del/ the simple backup file name for
440 src/patch/util.c is src/patch/.del/util.c.
441 -z suffix or --suffix=suffix
443 Use the simple method to determine backup file names (see the -V
444 method or --version-control method option), and use suffix as the
445 suffix. For example, with -z - the backup file name for
446 src/patch/util.c is src/patch/util.c-.
447 -Z or --set-utc
449 Set the modification and access times of patched files from time
450 stamps given in context diff headers, assuming that the context diff
451 headers use Coordinated Universal Time (UTC, often known as GMT).
452 Also see the -T or --set-time option.
453 The -Z or --set-utc and -T or --set-time options normally refrain
455 from setting a file's time if the file's original time does not
456 match the time given in the patch header, or if its contents do not
457 match the patch exactly. However, if the -f or --force option is
458 given, the file time is set regardless.
459 Due to the limitations of diff output format, these options cannot
461 update the times of files whose contents have not changed. Also, if
462 you use these options, you should remove (e.g. with make clean) all
463 files that depend on the patched files, so that later invocations of
464 make do not get confused by the patched files' times.
465
467 PATCH_GET
468 This specifies whether patch gets missing or read-only files from
469 RCS, ClearCase, Perforce, or SCCS by default; see the -g or --get
470 option.
471 POSIXLY_CORRECT
473 If set, patch conforms more strictly to the POSIX standard by
474 default: see the --posix option.
475 QUOTING_STYLE
477 Default value of the --quoting-style option.
478 SIMPLE_BACKUP_SUFFIX
480 Extension to use for simple backup file names instead of .orig.
481 TMPDIR, TMP, TEMP
483 Directory to put temporary files in; patch uses the first environ‐
484 ment variable in this list that is set. If none are set, the
485 default is system-dependent; it is normally /tmp on Unix hosts.
486 VERSION_CONTROL or PATCH_VERSION_CONTROL
488 Selects version control style; see the -v or --version-control
489 option.
490
492 $TMPDIR/p*
493 temporary files
494 /dev/tty
496 controlling terminal; used to get answers to questions asked of the
497 user
498
500 diff(1), ed(1), merge(1).
501 Marshall T. Rose and Einar A. Stefferud, Proposed Standard for Message
503 Encapsulation, Internet RFC 934 <URL:ftp://ftp.isi.edu/in-
504 notes/rfc934.txt> (1985-01).
505
507 There are several things you should bear in mind if you are going to be
508 sending out patches.
509 Create your patch systematically. A good method is the command
511 diff -Naur old new where old and new identify the old and new directo‐
512 ries. The names old and new should not contain any slashes. The diff
513 command's headers should have dates and times in Universal Time using
514 traditional Unix format, so that patch recipients can use the -Z or
515 --set-utc option. Here is an example command, using Bourne shell syn‐
516 tax:
517 LC_ALL=C TZ=UTC0 diff -Naur gcc-2.7 gcc-2.8
519 Tell your recipients how to apply the patch by telling them which
521 directory to cd to, and which patch options to use. The option string
522 -Np1 is recommended. Test your procedure by pretending to be a recipi‐
523 ent and applying your patch to a copy of the original files.
524 You can save people a lot of grief by keeping a patchlevel.h file which
526 is patched to increment the patch level as the first diff in the patch
527 file you send out. If you put a Prereq: line in with the patch, it
528 won't let them apply patches out of order without some warning.
529 You can create a file by sending out a diff that compares /dev/null or
531 an empty file dated the Epoch (1970-01-01 00:00:00 UTC) to the file you
532 want to create. This only works if the file you want to create doesn't
533 exist already in the target directory. Conversely, you can remove a
534 file by sending out a context diff that compares the file to be deleted
535 with an empty file dated the Epoch. The file will be removed unless
536 patch is conforming to POSIX and the -E or --remove-empty-files option
537 is not given. An easy way to generate patches that create and remove
538 files is to use GNU diff's -N or --new-file option.
539 If the recipient is supposed to use the -pN option, do not send output
541 that looks like this:
542 diff -Naur v2.0.29/prog/README prog/README
544 --- v2.0.29/prog/README Mon Mar 10 15:13:12 1997
545 +++ prog/README Mon Mar 17 14:58:22 1997
546 because the two file names have different numbers of slashes, and dif‐
548 ferent versions of patch interpret the file names differently. To
549 avoid confusion, send output that looks like this instead:
550 diff -Naur v2.0.29/prog/README v2.0.30/prog/README
552 --- v2.0.29/prog/README Mon Mar 10 15:13:12 1997
553 +++ v2.0.30/prog/README Mon Mar 17 14:58:22 1997
554 Avoid sending patches that compare backup file names like README.orig,
557 since this might confuse patch into patching a backup file instead of
558 the real file. Instead, send patches that compare the same base file
559 names in different directories, e.g. old/README and new/README.
560 Take care not to send out reversed patches, since it makes people won‐
562 der whether they already applied the patch.
563 Try not to have your patch modify derived files (e.g. the file config‐
565 ure where there is a line configure: configure.in in your makefile),
566 since the recipient should be able to regenerate the derived files any‐
567 way. If you must send diffs of derived files, generate the diffs using
568 UTC, have the recipients apply the patch with the -Z or --set-utc
569 option, and have them remove any unpatched files that depend on patched
570 files (e.g. with make clean).
571 While you may be able to get away with putting 582 diff listings into
573 one file, it may be wiser to group related patches into separate files
574 in case something goes haywire.
575
577 Diagnostics generally indicate that patch couldn't parse your patch
578 file.
579 If the --verbose option is given, the message Hmm... indicates that
581 there is unprocessed text in the patch file and that patch is attempt‐
582 ing to intuit whether there is a patch in that text and, if so, what
583 kind of patch it is.
584 patch's exit status is 0 if all hunks are applied successfully, 1 if
586 some hunks cannot be applied or there were merge conflicts, and 2 if
587 there is more serious trouble. When applying a set of patches in a
588 loop it behooves you to check this exit status so you don't apply a
589 later patch to a partially patched file.
590
592 Context diffs cannot reliably represent the creation or deletion of
593 empty files, empty directories, or special files such as symbolic
594 links. Nor can they represent changes to file metadata like ownership,
595 permissions, or whether one file is a hard link to another. If changes
596 like these are also required, separate instructions (e.g. a shell
597 script) to accomplish them should accompany the patch.
598 patch cannot tell if the line numbers are off in an ed script, and can
600 detect bad line numbers in a normal diff only when it finds a change or
601 deletion. A context diff using fuzz factor 3 may have the same prob‐
602 lem. You should probably do a context diff in these cases to see if
603 the changes made sense. Of course, compiling without errors is a
604 pretty good indication that the patch worked, but not always.
605 patch usually produces the correct results, even when it has to do a
607 lot of guessing. However, the results are guaranteed to be correct
608 only when the patch is applied to exactly the same version of the file
609 that the patch was generated from.
610
612 The POSIX standard specifies behavior that differs from patch's tradi‐
613 tional behavior. You should be aware of these differences if you must
614 interoperate with patch versions 2.1 and earlier, which do not conform
615 to POSIX.
616 · In traditional patch, the -p option's operand was optional, and a
618 bare -p was equivalent to -p0. The -p option now requires an oper‐
619 and, and -p 0 is now equivalent to -p0. For maximum compatibility,
620 use options like -p0 and -p1.
621 Also, traditional patch simply counted slashes when stripping path
623 prefixes; patch now counts pathname components. That is, a sequence
624 of one or more adjacent slashes now counts as a single slash. For
625 maximum portability, avoid sending patches containing // in file
626 names.
627 · In traditional patch, backups were enabled by default. This behav‐
629 ior is now enabled with the -b or --backup option.
630 Conversely, in POSIX patch, backups are never made, even when there
632 is a mismatch. In GNU patch, this behavior is enabled with the
633 --no-backup-if-mismatch option, or by conforming to POSIX with the
634 --posix option or by setting the POSIXLY_CORRECT environment vari‐
635 able.
636 The -b suffix option of traditional patch is equivalent to the
638 -b -z suffix options of GNU patch.
639 · Traditional patch used a complicated (and incompletely documented)
641 method to intuit the name of the file to be patched from the patch
642 header. This method did not conform to POSIX, and had a few
643 gotchas. Now patch uses a different, equally complicated (but bet‐
644 ter documented) method that is optionally POSIX-conforming; we hope
645 it has fewer gotchas. The two methods are compatible if the file
646 names in the context diff header and the Index: line are all identi‐
647 cal after prefix-stripping. Your patch is normally compatible if
648 each header's file names all contain the same number of slashes.
649 · When traditional patch asked the user a question, it sent the ques‐
651 tion to standard error and looked for an answer from the first file
652 in the following list that was a terminal: standard error, standard
653 output, /dev/tty, and standard input. Now patch sends questions to
654 standard output and gets answers from /dev/tty. Defaults for some
655 answers have been changed so that patch never goes into an infinite
656 loop when using default answers.
657 · Traditional patch exited with a status value that counted the number
659 of bad hunks, or with status 1 if there was real trouble. Now patch
660 exits with status 1 if some hunks failed, or with 2 if there was
661 real trouble.
662 · Limit yourself to the following options when sending instructions
664 meant to be executed by anyone running GNU patch, traditional patch,
665 or a patch that conforms to POSIX. Spaces are significant in the
666 following list, and operands are required.
667 -c
669 -d dir
670 -D define
671 -e
672 -l
673 -n
674 -N
675 -o outfile
676 -pnum
677 -R
678 -r rejectfile
679
681 Please report bugs via email to <bug-patch@gnu.org>.
682 If code has been duplicated (for instance with #ifdef OLDCODE ... #else
684 ... #endif), patch is incapable of patching both versions, and, if it
685 works at all, will likely patch the wrong one, and tell you that it
686 succeeded to boot.
687 If you apply a patch you've already applied, patch thinks it is a
689 reversed patch, and offers to un-apply the patch. This could be con‐
690 strued as a feature.
691 Computing how to merge a hunk is significantly harder than using the
693 standard fuzzy algorithm. Bigger hunks, more context, a bigger offset
694 from the original location, and a worse match all slow the algorithm
695 down.
696
698 Copyright (C) 1984, 1985, 1986, 1988 Larry Wall.
699 Copyright (C) 1989, 1990, 1991, 1992, 1993, 1994, 1995, 1996, 1997,
700 1998, 1999, 2000, 2001, 2002, 2009 Free Software Foundation, Inc.
701 Permission is granted to make and distribute verbatim copies of this
703 manual provided the copyright notice and this permission notice are
704 preserved on all copies.
705 Permission is granted to copy and distribute modified versions of this
707 manual under the conditions for verbatim copying, provided that the
708 entire resulting derived work is distributed under the terms of a per‐
709 mission notice identical to this one.
710 Permission is granted to copy and distribute translations of this man‐
712 ual into another language, under the above conditions for modified ver‐
713 sions, except that this permission notice may be included in transla‐
714 tions approved by the copyright holders instead of in the original Eng‐
715 lish.
716
718 Larry Wall wrote the original version of patch. Paul Eggert removed
719 patch's arbitrary limits; added support for binary files, setting file
720 times, and deleting files; and made it conform better to POSIX. Other
721 contributors include Wayne Davison, who added unidiff support, and
722 David MacKenzie, who added configuration and backup support. Andreas
723 Grünbacher added support for merging.
724 GNU PATCH(1)