PATCH(1) User Commands PATCH(1)

patch - a program for applying a diff file to an original

patch [options] orig

spatch [options] orig

opatch [options] orig patchfile [+ [options] orig]

sccspatch [options] orig

patch [-blNR] [-c | -e | -n | -u] [-d dir] [-D define]
[-i patchfile] [-o outfile] [-p num] [-r rejectfile]
[file]

patch <patchfile

Patch will take a patch file containing any of the four forms of difference listing produced by the diff program and apply those differences to an original file, producing a patched version. By default, the patched version is put in place of the original and no backup is made. The -b switch tells patch to create a backup file. You may also specify where you want the output to go with a -o switch. If patchfile is omitted, or is a hyphen, the patch will be read from standard input.

patch attempts to determine the type of the diff listing, unless overruled by a -c, -e, or -n option.

If the patch file contains more than one patch, patch attempts to apply each of them as if they came from separate patch files. (In this case the name of the patch file must be determinable for each diff listing.)

By default, this version of patch is in strict POSIX mode and does not permit extensions. If patch is called as opatch, it behaves like the original patch utility from Larry Wall (except for filename determination) but with new additional features enabled. The names sccspatch and spatch may be used to select POSIX patch behavior with extensions from Larry Wall and extensions from recent development.

Patch recognizes the following switches:

Save a copy of the original contents of each modified file, before the differences are applied, in a file of the same name with the suffix .orig appended to it. If the file already exists, it is overwritten. If multiple patches are applied to the same file, the .orig file is written only for the first patch. When the -o outfile option is also specified, file .orig is not created but, if outfile already exists, outfile .orig is created.

The original behavior of the -b option as found in the original patch from Larry Wall is available with the -z option.

causes the next argument to be interpreted as a prefix to the backup file name. If this argument is specified any argument from -z will be ignored. For example, with -B /old/ the backup file name for src/patch/util.c is /old/src/patch/util.c.

This argument is an extension to Larry Wall's patch v2.0.1.4, patchlevel 8, made by M. Greim (greim@sbsvax.uucp).

This option is not available in strict POSIX mode.

forces patch to interpret the patch file as a context diff.
causes patch to interpret the next argument as a directory, and cd to it before doing anything else.
causes patch to use the #ifdef define...#endif construct to mark changes. The argument following -D will be used as the differentiating symbol.
Remove output files that are empty after patching. If the patch is in the unified context diff format and the diff header contains a correct zero time stamp for the removed file, this option is not needed. A unified context diff created by diff -Nur marks removed files by including a time that refers to the beginning of the year 1970 UTC like in the following example diff output header: 

--- old/Makefile          2011-01-25 12:08:18 +0100
+++ new/Makefile          1970-01-01 01:00:00 +0100
@@ -1,18 +0,0 @@
The zero time stamp is only recognized in case the time stamp is printed in the format required by POSIX for unified diffs as in the example above or with added sub second time information.

This option is not available in strict POSIX mode.

forces patch to interpret the patch file as an ed script.
forces patch to assume that the user knows exactly what he or she is doing, and to not ask any questions. It does not suppress informational messages, however. Use -s for that.

This option is not available in strict POSIX mode.

sets the maximum fuzz factor. This switch only applies to context diffs, and causes patch to ignore up to that many lines in looking for places to install a hunk. Note that a larger fuzz factor increases the odds of a faulty patch. The default fuzz factor is 2, and it may not be set to more than the number of lines of context in the context diff, ordinarily 3.

This option is not available in strict POSIX mode.

Reads the patch information from the file named by the path name patchfile, rather than the standard input.
causes the pattern matching to be done loosely, in case the tabs and spaces have been munged in your input file. Any sequence of whitespace in the pattern line will match any sequence in the input file. Normal characters must still match exactly. Each line of the context must still match a line in the input file.
forces patch to interpret the patch file as a normal diff.
causes patch to ignore patches that it thinks are reversed or already applied. See also -R .
causes the next argument to be interpreted as the output file name.
sets the pathname strip count, which controls how pathnames found in the patch file are treated, in case the you keep your files in a different directory than the person who sent out the patch. The strip count specifies how many slashes are to be stripped from the front of the pathname. (Any intervening directory names also go away.) For example, supposing the filename in the patch file was

/u/howard/src/blurfl/blurfl.c

setting -p0 gives the entire pathname unmodified, -p1 gives

u/howard/src/blurfl/blurfl.c

without the leading slash, -p4 gives

blurfl/blurfl.c

and not specifying -p at all just gives you "blurfl.c". Whatever you end up with is looked for either in the current directory, or the directory specified by the -d switch.

When in POSIX mode, -p always must have an argument, in non-POSIX mode -p may be used as an alias for -p0.

causes the next argument to be interpreted as the reject file name. In the default case, the reject file has the same name as the output file, with the suffix .rej appended to it.
tells patch that this patch was created with the old and new files swapped. (Yes, I'm afraid that does happen occasionally, human nature being what it is.) Patch will attempt to swap each hunk around before applying it. Rejects will come out in the swapped format. The -R switch will not work with ed diff scripts because there is too little information to reconstruct the reverse operation.

If the first hunk of a patch fails, patch will reverse the hunk to see if it can be applied that way. If it can, you will be asked if you want to have the -R switch set. If it can't, the patch will continue to be applied normally. (Note: this method cannot detect a reversed patch if it is a normal diff and if the first command is an append (i.e. it should have been a delete) since appends always succeed, due to the fact that a null context will match anywhere. Luckily, most patches add or change lines rather than delete them, so most reversed normal diffs will begin with a delete, which will fail, triggering the heuristic.)

makes patch do its work silently, unless an error occurs.

This option is not available in strict POSIX mode.

causes patch to ignore this patch from the patch file, but continue on looking for the next patch in the file. Thus

patch -S + -S + <patchfile

will ignore the first and second of three patches.

This option is not available in strict POSIX mode.

Set the modification and access times of patched files from the time stamps read from the context diff headers.

If these timestamps do not include a timezone, the time is assumed to be in local time format. Since this assumption may be wrong, it is recommended to create patches with diff -u in order to enforce a timezone in the header.

If there is a content mismatch or a mismatch with the old time stamp, the new time stamp is not set. Use -f to enforce to set the time stamp in such a case.

This option is not available in strict POSIX mode.

forces patch to interpret the patch file as a unified context diff (a unidiff).
causes patch to print out it's revision header and compatiblity status.

This option is not available in strict POSIX mode.

causes patch to be more verbose.
causes patch to print out it's revision header and compatiblity status.

Causes patch to permit enhancements from historic and recent versions in POSIX mode.
Causes patch to implement compatibility for Larry Wall's patch-2.0.

This currently only applies to the way patch selects the filename from the patchfile. By default, patch uses the filename determination algorithm from POSIX that defaults to the original filename from the diff header. The algorithm from patch-2.0 defaults to the shorter name, assuming that filename.orig is used as original filename and filename is used as new filename.

Causes patch to go into strict POSIX mode. This disables the effect of a previous -W+ and -Wall option.
Causes patch to revert the effect of a previous -W+ and -Wposix option.
sets internal debugging flags, and is of interest only to patch patchers.

This option is not available in strict POSIX mode.

causes the next argument to be interpreted as the backup extension, to be used in place of ".orig" or "~".

This option is not available in strict POSIX mode.

Set the modification and access times of patched files from the time stamps read from the context diff headers.

If these timestamps do not include a timezone, the time is assumed to be in UTC format. Since this assumption may be wrong, it is recommended to create patches with diff -u in order to enforce a timezone in the header.

If there is a content mismatch or a mismatch with the old time stamp, the new time stamp is not set. Use -f to enforce to set the time stamp in such a case.

This option is not available in strict POSIX mode.

Upon startup, patch will attempt to determine the type of the diff listing, unless over-ruled by a -c, -e, -n, or -u switch. Context diffs (old-style, new-style, and unified) and normal diffs are applied by the patch program itself, while ed diffs are simply fed to the ed editor via a pipe.

Patch will try to skip any leading garbage, apply the diff, and then skip any trailing garbage. Thus you could feed an article or message containing a diff listing to patch, and it should work. If the entire diff is indented by a consistent amount, this will be taken into account.

With context diffs, and to a lesser extent with normal diffs, patch can detect when the line numbers mentioned in the patch are incorrect, and will attempt to find the correct place to apply each hunk of the patch. As a first guess, it takes the line number mentioned for the hunk, plus or minus any offset used in applying the previous hunk. If that is not the correct place, patch will scan both forwards and backwards for a set of lines matching the context given in the hunk. First patch looks for a place where all lines of the context match. If no such place is found, and it's a context diff, and the maximum fuzz factor is set to 1 or more, then another scan takes place ignoring the first and last line of context. If that fails, and the maximum fuzz factor is set to 2 or more, the first two and last two lines of context are ignored, and another scan is made. (The default maximum fuzz factor is 2.) If patch cannot find a place to install that hunk of the patch, it will put the hunk out to a reject file, which normally is the name of the output file plus ".rej" or "#" . (Note that the rejected hunk will come out in context diff form whether the input patch was a context diff or a normal diff. If the input was a normal diff, many of the contexts will simply be null.) The line numbers on the hunks in the reject file may be different than in the patch file: they reflect the approximate location patch thinks the failed hunks belong in the new file rather than the old one.

As each hunk is completed, you will be told whether the hunk succeeded or failed, and which line (in the new file) patch thought the hunk should go on. If this is different from the line number specified in the diff you will be told the offset. A single large offset MAY be an indication that a hunk was installed in the wrong place. You will also be told if a fuzz factor was used to make the match, in which case you should also be slightly suspicious.

If no original file is specified on the command line, patch will try to figure out from the leading garbage what the name of the file to edit is. The following formats are recognized.

A context diff includes the following headers:


"*** filename timestamp"
"--- filename timestamp"

The first line refers to the filename and timestamp of the original file and the second line refers to the changed file.

A unified context diff includes the following headers:


"--- filename timestamp"
"+++ filename timestamp"

The first line refers to the filename and timestamp of the original file and the second line refers to the changed file.


"Index:filename"

May be added in order to give the filename in case if a normal diff or an ed script.

If not in strict POSIX mode,


"Prereq: version-id"

is recognized and used to match against a version specific string in the file to patch.

If no original file is specified on the command line, patch will use the following search preference:

1.
In the header of a context diff, the filename is found from lines beginning with "***" or "---" (if the patch is a unified context diff).
2.
In the header of a context diff, the filename is found from lines beginning with "---" or "+++" (if the patch is a unified context diff).

If the option -Wall was specified, the shortest name of an existing file from the two above wins.

3.
If the diff type is not context and there is an "Index:" line in the leading garbage, patch will try to use the filename from that line. The context diff header takes precedence over an Index line.
4.
If the original file cannot be found, but a suitable SCCS or RCS file is handy, patch will attempt to get or check out the file.
5.
If no filename can be intuited from the leading garbage, you will be asked for the name of the file to patch via the controlling terminal (e.g. via stderr or /dev/tty).

Additionally, if not in strict POSIX mode and if the leading garbage contains a "Prereq: " line, patch will take the first word from the prerequisites line (normally a version number) and check the input file to see if that word can be found. If not, patch will ask for confirmation before proceeding.

The upshot of all this is that you should be able to say, while in a news interface, the following:

| patch -d /usr/src/local/blurfl

and patch a file in the blurfl directory directly from the article containing the patch.

If the patch file contains more than one patch, patch will try to apply each of them as if they came from separate patch files. This means, among other things, that it is assumed that the name of the file to patch must be determined for each diff listing, and that the garbage before each diff listing will be examined for interesting things such as filenames and revision level, as mentioned previously. If patch is not in POSIX mode, you can give switches (and another original file name) for the second and subsequent patches by separating the corresponding argument lists by a '+'. (The argument list for a second or subsequent patch may not specify a new patch file, however.)

If present, patch operates in POSIX mode but also recognizes the additional non-POSIX options and permits non-POSIX behavior like removing files that have been marked removed in the diff -Nur output. See -E option for more information. This behavior is implemented to make patch more compatible to the gpatch program.

See environ(5) for descriptions of the following environment variables that affect the execution of patch: LANG, LC_ALL, LC_COLLATE, LC_CTYPE, LC_MESSAGES, LC_TIME, and NLSPATH.

/tmp/patch*

The following exit values are returned:

0
Successful completion.

1
One or more lines were written to a reject file.

>1
An error occurred.

See attributes(5) for descriptions of the following attributes:

ATTRIBUTE TYPE ATTRIBUTE VALUE
Availability SUNWcsu
Interface Stability Standard

diff(1), ed(1) attributes(5), environ(5), standards(5)

There are several things you should bear in mind if you are going to be sending out patches.

First, you can save people a lot of grief by keeping a patchlevel.h file which is patched to increment the patch level as the first diff in the patch file you send out.

If you put a Prereq: line in with the patch, it won't let them apply patches out of order without some warning.

Second, make sure you've specified the filenames right, either in a context diff header, or with an Index: line. If you are patching something in a subdirectory, be sure to tell the patch user to specify a -p switch as needed.

Third, you can create a file by sending out a diff that compares a null file to the file you want to create. This will only work if the file you want to create doesn't exist already in the target directory.

Fourth, take care not to send out reversed patches, since it makes people wonder whether they already applied the patch.

Fifth, while you may be able to get away with putting 582 diff listings into one file, it is probably wiser to group related patches into separate files in case something goes haywire.

Too many to list here, but generally indicative that patch couldn't parse your patch file.

The message "Hmm..." indicates that there is unprocessed text in the patch file and that patch is attempting to intuit whether there is a patch in that text and, if so, what kind of patch it is.

When applying a set of patches in a loop it behooves you to check the exit status so you don't apply a later patch to a partially patched file.

Patch cannot tell if the line numbers are off in an ed script, and can only detect bad line numbers in a normal diff when it finds a "change" or a "delete" command. A context diff using fuzz factor 3 may have the same problem. Until a suitable interactive interface is added, you should probably do a context diff in these cases to see if the changes made sense. Of course, compiling without errors is a pretty good indication that the patch worked, but not always.

Patch usually produces the correct results, even when it has to do a lot of guessing. However, the results are guaranteed to be correct only when the patch is applied to exactly the same version of the file that the patch was generated from.

The POSIX standard introduced some small differences to the traditional behavior of patch. The behavior of gpatch (GNU patch) differs from both, the traditional patch and POSIX patch.

If you like to write portable patch scripts, you should be aware of the differences:

•
Traditional patch by default created backups with the .orig suffix and the -b option allowed to define a different suffix. Modern patch implementations by default do not create backup files and allow to enable backups with the -b option that does not permit a parameter with POSIX.

The -b suffix behavior of the traditional patch is equivalent to -b -z suffix with modern patch implementations in case that enhancements over POSIX are permitted.

•
Patch writes informational messages to stderr and does not use stdout, while gpatch uses stdout for informational messages.
•
Gpatch changed some of the informational messages; you cannot expect to get the same messages as with patch.

For portable scripts, do not rely on informational messages.

•
In traditional patch, the -p option did not need to have an operand. With POSIX patch, -p always must have an operand. The option -p without argument in traditional patch is equivalent to -p 0 in POSIX patch.

For portability, use the -p option with no space to the following argument, e.g. -p1.

•
The option -i patchfile is only supported by POSIX patch, but not by traditional patch or gpatch.

•
Traditional patch used several different and partly undocumented methods to determine the name of the file to be patched from the patchfile that could e.g. result to create an incorrect new file with new filename if a line was inserted at the beginning of an existing file. This patch and POSIX patch use the POSIX method for Filename Determination and only permit to use the the special Index: line.

Since gpatch by default ignores the POSIX Filename Determination rules, gpatch by default will use wrong filenames from patches that have been created with POSIX rules.

The only way to make sure that a patch file is interpreted correctly by all patch implementations is to use the original filename for the first diff argument and to make the first diff argument shorter than the second argument. Alternatively use the same filenames for both arguments but have a different directory prefix that is stripped off by the -p option.

•
Gpatch uses a different method to read replies to questions to the user than patch does. Do not rely on patch to read from stderr.
•
Traditional patch used an exit code value that reflected the number of failed hunks and an exit code 1 if there was one failed hunk or any other problem. POSIX patch uses exit code 1 if there was any failed hunk and an exit code > 1 if other problems occurred.
•
Patch does not implement support for long options, while gpatch documents and implements various long options.

For maximum portability, limit yourself to the following options:

Never use long options. Use a shorter name for the original filename in the diff header of the patchfile.

Could be smarter about partial matches, excessively deviant offsets and swapped code, but that would take an extra pass.

If code has been duplicated (for instance with #ifdef OLDCODE ... #else ... #endif), patch is incapable of patching both versions, and, if it works at all, will likely patch the wrong one, and tell you that it succeeded to boot.

If you apply a patch you've already applied, patch will think it is a reversed patch, and offer to un-apply the patch. This could be construed as a feature.

Larry Wall <larry@wall.org> wrote the original version of patch.

Wayne Davison added unidiff support.

Joerg Schilling added modern portability code, code to support arbitrarily long lines, large file support and code to support POSIX compliance.

2018/03/01 OmniOS