.\" patch man page
.de Id
.ds Dt \\$4
..
.Id $Id: patch.man,v 1.16 1997/05/30 08:03:48 eggert Exp $
.ds = \-\^\-
.de Sp
.if t .sp .3
.if n .sp
..
.TH PATCH 1 \*(Dt GNU
.ta 3n
.SH NAME
patch \- apply a diff file to an original
.SH SYNOPSIS
.B patch
.RI [ options ]
.RI [ originalfile
.RI [ patchfile ]]
.Sp
but usually just
.Sp
.BI "patch \-p" "number"
.BI < patchfile
.SH DESCRIPTION
.B patch
takes a patch file
.I patchfile
containing a difference listing produced by the
.B diff
program and applies those differences to one or more original files,
producing patched versions.
Normally the patched versions are put in place of the originals.
Backups can be made; see the
.B \-V
or
.B \*=version\-control
option.
The names of the files to be patched are usually taken from the patch file,
but if there's just one file to be patched it can specified on the
command line as
.IR originalfile .
.PP
Upon startup, patch attempts to determine the type of the diff listing,
unless overruled by a
\fB\-c\fP (\fB\*=context\fP),
\fB\-e\fP (\fB\*=ed\fP),
\fB\-n\fP (\fB\*=normal\fP),
or
\fB\-u\fP (\fB\*=unified\fP)
option.
Context diffs (old-style, new-style, and unified) and
normal diffs are applied by the
.B patch
program itself, while
.B ed
diffs are simply fed to the
.BR ed (1)
editor via a pipe.
.PP
.B patch
tries 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
.BR patch ,
and it should work.
If the entire diff is indented by a consistent amount,
this is taken into account.
.PP
With context diffs, and to a lesser extent with normal diffs,
.B patch
can detect when the line numbers mentioned in the patch are incorrect,
and attempts 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,
.B patch
scans both forwards and backwards for a set of lines matching the context
given in the hunk.
First
.B 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
.B patch
cannot find a place to install that hunk of the patch, it puts the
hunk out to a reject file, which normally is the name of the output file
plus a
.B \&.rej
suffix, or
.B #
if
.B \&.rej
would generate a file name that is too long
(if even appending the single character
.B #
makes the file name too long, then
.B #
replaces the file name's last character).
(The rejected hunk comes out in ordinary context diff form regardless of
the input patch's form.
If the input was a normal diff, many of the contexts are simply 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.
.PP
As each hunk is completed, you are told if the hunk
failed, and if so which line (in the new file)
.B patch
thought the hunk should go on.
If the
.B \*=verbose
option is given, you are also told about hunks that succeeded.
If the hunk is installed at a different line
from the line number specified in the diff you
are told the offset.
A single large offset
.I may
indicate that a hunk was installed in the
wrong place.
You are also told if a fuzz factor was used to make the match, in which
case you should also be slightly suspicious.
.PP
If no original file
.I origfile
is specified on the command line,
.B patch
tries to figure out from the leading garbage what the name of the file
to edit is.
If the header is that of a context diff,
.B patch
takes the old and new file names in the header,
and if there is an
.B Index:\&
line in the leading garbage,
.B patch
obtains the name in that line; any
.B /dev/null
names are ignored.
These names are considered to be in the order (old, new, index),
regardless of the order that they appear in the header.
If some of the named files exist,
.B patch
uses the first name if the
.B POSIXLY_CORRECT
environment variable is set, and the best name otherwise.
If no named files exist, some names are given,
.B POSIXLY_CORRECT
is not set, and the patch appears to create a file,
.B patch
uses the best name requiring the creation of the fewest directories.
If no file name results from the above heuristics, you are asked
for the name of the file to patch.
To determine the best of a nonempty list of file names,
.B patch
first takes all the names with the fewest path name components;
of those, it then takes all the names with the shortest basename;
of those, it then takes all the shortest names;
finally, it takes the first remaining name.
.PP
Additionally, if the leading garbage contains a
.B Prereq:\&
line,
.B patch
takes the first word from the prerequisites line (normally a version
number) and checks the original file to see if that word can be found.
If not,
.B patch
asks for confirmation before proceeding.
.PP
If an
\s-1RCS\s0 file is handy,
and the original file cannot be found
or is read-only and matches the default version,
and if version control
(see the
.B \-V
or
.B \*=version\-control
option)
is set to
.BR existing ,
.B patch
attempts to get and lock the file.
\s-1SCCS\s0 is treated in a similar way.
.PP
The upshot of all this is that you should be able to say, while in a news
interface, something like the following:
.Sp
	\fB| patch \-d /usr/src/local/blurfl\fP
.Sp
and patch a file in the
.B blurfl
directory directly from the article containing
the patch.
.PP
If the patch file contains more than one patch,
.B patch
tries 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
contains interesting things such as file names and revision level, as
mentioned previously.
.SH OPTIONS
.TP 3
\fB\-b\fP  or  \fB\*=backup\fP
Make backup files.
This option is equivalent to
.BR \*=version\-control=simple ;
see the
.B \-V
or
.B \*=version\-control
option for details.
In older versions of
.BR patch ,
this option had an argument specifying the simple backup suffix;
this argument has been moved to the
.B \-z
option.
.TP
\fB\-B\fP \fIpref\fP  or  \fB\*=prefix=\fP\fIpref\fP
Prefix simple backup file names with
.IR pref .
For example, with
.B "\-B /junk/"
the simple backup file name for
.B src/patch/util.c
is
.BR /junk/src/patch/util.c .
.TP
\fB\*=binary\fP
Read and write all files in binary mode,
except for standard output and
.BR /dev/tty .
This option has no effect on \s-1POSIX\s0-compliant systems.
On systems like \s-1DOS\s0 where this option makes a difference,
the patch should be generated by
.BR "diff\ \-a\ \*=binary" .
.TP
\fB\-c\fP  or  \fB\*=context\fP
Interpret the patch file as a ordinary context diff.
.TP
\fB\*=verbose\fP
Output extra information about the work being done.
.TP
\fB\-d\fP \fIdir\fP  or  \fB\*=directory=\fP\fIdir\fP
Change to the directory
.I dir
immediately, before doing
anything else.
.TP
\fB\-D\fP \fIsym\fP  or  \fB\*=ifdef=\fP\fIsym\fP
Use the
.BR #ifdef " .\|.\|. " #endif
construct to mark changes, with
.I sym
as the differentiating symbol.
.TP
.B "\*=dry\-run"
Print the results of applying the patches without actually changing any files.
.TP
\fB\-e\fP  or  \fB\*=ed\fP
Interpret the patch file as an
.B ed
script.
.TP
\fB\-E\fP  or  \fB\*=remove\-empty\-files\fP
Remove output files that are empty after the patches have been applied.
Normally this option is unnecessary, since
.B patch
can examine the timestamps on the header to determine whether a file
should exist after patching.
However, if the input is not a context diff or if the
.B POSIXLY_CORRECT
environment variable is set,
.B patch
does not remove empty patched files unless this option is given.
When
.B patch
removes a file, it also attempts to remove any empty ancestor directories.
.TP
\fB\-f\fP  or  \fB\*=force\fP
Assume that the user knows exactly what he or she is doing, and do not
ask any questions.  Skip patches whose headers
do not say which file is to be patched; patch files even though they have the
wrong version for the
.B Prereq:\&
line in the patch; and assume that
patches are not reversed even if they look like they are.
This option does not suppress commentary; use
.B \-s
for that.
.TP
\fB\-F\fP \fInumber\fP  or  \fB\*=fuzz=\fP\fInumber\fP
Set the maximum fuzz factor.
This option only applies to diffs that have context, and causes
.B 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.
.TP
\fB\-g\fP  or  \fB\*=get\fP
If a file does not exist or is read-only and matches the default version,
get it from \s-1RCS\s0 if it is under \s-2RCS\s0 control;
similarly for \s-1SCCS\s0.
If the
.B PATCH_GET
environment variable is set, this is the default.
.TP
\fB\-G\fP  or  \fB\*=no\-get\fP
Do not get files from \s-1RCS\s0 or \s-2SCCS\s0.
This is the default unless the
.B PATCH_GET
environment variable is set.
.TP
.B "\*=help"
Print a summary of options and exit.
.TP
\fB\-i\fP \fIpatchfile\fP  or  \fB\*=input=\fP\fIpatchfile\fP
Read the patch from
.IR patchfile .
If
.I patchfile
is
.BR \- ,
read from standard input, the default.
.TP
\fB\-l\fP  or  \fB\*=ignore\-whitespace\fP
Match patterns loosely, in case tabs or spaces
have been munged in your files.
Any sequence of one or more blanks in the patch file matches any sequence
in the original file, and sequences of blanks at the ends of lines are ignored.
Normal characters must still match exactly.
Each line of the context must still match a line in the original file.
.TP
\fB\-n\fP  or  \fB\*=normal\fP
Interpret the patch file as a normal diff.
.TP
\fB\-N\fP  or  \fB\*=forward\fP
Ignore patches that seem to be reversed or already applied.
See also
.BR \-R .
.TP
\fB\-o\fP \fIfile\fP  or  \fB\*=output=\fP\fIfile\fP
Send output to
.I file
instead of patching files in place.
.TP
\fB\-p\fP\fInumber\fP  or  \fB\*=strip\fP\fB=\fP\fInumber\fP
Strip the smallest prefix containing
.I number
leading slashes from each file name found in the patch file.
A sequence of one or more adjacent slashes is counted as a single slash.
This controls how file names found in the patch file are treated, in case
you keep your files in a different directory than the person who sent
out the patch.
For example, supposing the file name in the patch file was
.Sp
	\fB/u/howard/src/blurfl/blurfl.c\fP
.Sp
setting
.B \-p0
gives the entire file name unmodified,
.B \-p1
gives
.Sp
	\fBu/howard/src/blurfl/blurfl.c\fP
.Sp
without the leading slash,
.B \-p4
gives
.Sp
	\fBblurfl/blurfl.c\fP
.Sp
and not specifying
.B \-p
at all just gives you \fBblurfl.c\fP.
Whatever you end up with is looked for either in the current directory,
or the directory specified by the
.B \-d
option.
With \s-1GNU\s0
.BR patch ,
the two-argument
.BI "\-p " N
form of this option is equivalent to one-argument
.BI \-p N
form, but this is not true of traditional
.BR patch ,
so the one-argument form is recommended for portability.
.TP
\fB\-r\fP \fIfile\fP  or  \fB\*=reject\-file=\fP\fIfile\fP
Put rejects into
.I file
instead of the default
.B \&.rej
file.
.TP
\fB\-R\fP  or  \fB\*=reverse\fP
Assume 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.)
.B patch
attempts to swap each hunk around before applying it.
Rejects come out in the swapped format.
The
.B \-R
option does not work with
.B ed
diff scripts because there is too little
information to reconstruct the reverse operation.
.Sp
If the first hunk of a patch fails,
.B patch
reverses the hunk to see if it can be applied that way.
If it can, you are asked if you want to have the
.B \-R
option set.
If it can't, the patch continues 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 matches
anywhere.
Luckily, most patches add or change lines rather than delete them, so most
reversed normal diffs begin with a delete, which fails, triggering
the heuristic.)
.TP
\fB\-s\fP  or  \fB\*=silent\fP  or  \fB\*=quiet\fP
Work silently, unless an error occurs.
.TP
\fB\-t\fP  or  \fB\*=batch\fP
Suppress questions like
.BR \-f ,
but make some different assumptions:
skip patches whose headers do not contain file names (the same as \fB\-f\fP);
skip patches for which the file has the wrong version for the
.B Prereq:\&
line
in the patch; and assume that patches are reversed if they look like
they are.
.TP
\fB\-u\fP  or  \fB\*=unified\fP
Interpret the patch file as a unified context diff.
.TP
\fB\-v\fP  or  \fB\*=version\fP
Print out
.BR patch 's
revision header and patch level, and exit.
.TP
\fB\-V\fP \fImethod\fP  or  \fB\*=version\-control=\fP\fImethod\fP
Use
.I method
when creating
backup file names.  The type of backups made can also be given in the
.B PATCH_VERSION_CONTROL
(or, if that's not set, the
.BR VERSION_CONTROL )
environment variable, which is overridden by this option.
.Sp
The value of
.I method
is like the \s-1GNU\s0
Emacs `version-control' variable;
.B patch
also recognizes synonyms that
are more descriptive.  The valid values for
.I method
are (unique abbreviations are
accepted):
.RS
.TP 3
\fBexisting\fP  or  \fBnil\fP
Make numbered backups of files that already have them,
otherwise simple backups.
.TP
\fBnone\fP
Do not make backups.
This is the default.
.TP
\fBnumbered\fP  or  \fBt\fP
Make numbered backups.  The numbered backup file name for
.I F
is
.IB F .~ N ~
where
.I N
is the version number.
.TP
\fBsimple\fP  or  \fBnever\fP
Make simple backups.  That is, when patching a file
.IR F ,
rename or copy the original instead of removing it.
The
.B \-B
or
.BR \*=prefix ,
.B \-y
or
.BR \*=basename\-prefix ,
and
.B \-z
or
.BR \*=suffix
options specify the simple backup file name.
If none of these options are given, then a simple backup suffix is used;
it is the value of the
.B SIMPLE_BACKUP_SUFFIX
environment variable if set, and is
.B \&.orig
otherwise.
.PP
With numbered or simple backups,
if the backup file name is too long, the backup suffix
.B ~
is used instead; if even appending
.B ~
would make the name too long, then
.B ~
replaces the last character of the file name.
.RE
.TP
\fB\-x\fP \fInumber\fP  or  \fB\*=debug=\fP\fInumber\fP
Set internal debugging flags of interest only to
.B patch
patchers.
.TP
\fB\-y\fP \fIpref\fP  or  \fB\*=basename\-prefix=\fP\fIpref\fP
Prefix the basename of the simple backup file name with
.IR pref .
For example, with
.B "\-y .del/"
the backup file name for
.B src/patch/util.c
is
.BR src/patch/.del/util.c .
.TP
\fB\-z\fP \fIsuff\fP  or  \fB\*=suffix=\fP\fIsuff\fP
Use
.I suff
as the simple backup suffix.
The backup extension may also be specified by the
.B SIMPLE_BACKUP_SUFFIX
environment variable, which is overridden by this option.
.SH ENVIRONMENT
.TP 3
.B POSIXLY_CORRECT
If set,
.B patch
conforms more strictly to the \s-1POSIX\s0 standard:
it takes the first existing file when intuiting file names from diff headers,
it does not remove files that are empty after patching,
and it requires that all options precede the
files in the command line.
.TP
.B SIMPLE_BACKUP_SUFFIX
Extension to use for simple backup file names instead of
.BR \&.orig .
.TP
\fBTMPDIR\fP, \fBTMP\fP, \fBTEMP\fP
Directory to put temporary files in;
.B patch
uses the first environment variable in this list that is set.
If none are set, the default is system-dependent;
it is normally
.B /tmp
on Unix hosts.
.TP
\fBPATCH_VERSION_CONTROL\fP or \fBVERSION_CONTROL\fP
Selects version control style; see the
.B \-v
or
.B \*=version\-control
option.
.TP
\fBPATCH_GET\fP
If set,
.B patch
gets missing or read-only files from \s-1RCS\s0 or \s-1SCCS\s0
by default; see the
.B \-g
or
.B \*= get
and the
.B \-G
or
.B \*= no\-get
options.
.SH FILES
.TP 3
.IB $TMPDIR "/p\(**"
temporary files
.TP
.B /dev/tty
console; used to get answers to questions asked of the user
.SH "SEE ALSO"
.BR diff (1),
.BR ed (1)
.SH "NOTES FOR PATCH SENDERS"
There are several things you should bear in mind if you are going to
be sending out patches.
.PP
Create your patch systematically.
A good method is the command
.BI "diff\ \-Naur\ " "old\ new"
where
.I old
and
.I new
identify the old and new directories.
The names
.I old
and
.I new
should not contain any slashes.
Here is an example:
.Sp
	\fBdiff \-Naur version\-2.2 version\-2.3\fP
.PP
Tell your recipients how to apply the patch
by telling them which directory to
.B cd
to, and which
.B patch
options to use.  The option string
.B "\-Np1"
is recommended.
Test your procedure by pretending to be a recipient and applying
your patch to a copy of the original files.
.PP
You can save people a lot of grief by keeping a
.B 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
.B Prereq:\&
line in with the patch, it won't let them apply
patches out of order without some warning.
.PP
Make sure you've specified the file names right, either in a
context diff header, or with an
.B Index:\&
line.
.PP
You can create a file by sending out a diff that compares an
empty file (such as
.BR /dev/null )
to the file you want to create.
This only works if the file you want to create doesn't exist already in
the target directory.
Conversely, you can remove a file by sending out a diff that compares the
file to be deleted with an empty file.
The file will be removed unless the
.B POSIXLY_CORRECT
environment variable is set and the
.B \-E
or
.B \*=remove\-empty\-files
option is not given.
An easy way to generate patches that create and remove files
is to use \s-1GNU\s0
.BR diff 's
.B \*=new\-file
option.
.PP
If the recipient is supposed to use the
.BI \-p N
option, do not send output that looks like this:
.Sp
.ft B
.ne 3
	diff \-Naur v2.0.29/prog/README prog/README
.br
	\-\^\-\^\- v2.0.29/prog/README   Mon Mar 10 15:13:12 1997
.br
	+\^+\^+ prog/README   Mon Mar 17 14:58:22 1997
.ft
.Sp
because the two file names have different numbers of slashes,
and different versions of
.B patch
interpret the file names differently.
To avoid confusion, send output that looks like this instead:
.Sp
.ft B
.ne 3
	diff \-Naur v2.0.29/prog/README v2.0.30/prog/README
.br
	\-\^\-\^\- v2.0.29/prog/README   Mon Mar 10 15:13:12 1997
.br
	+\^+\^+ v2.0.30/prog/README   Mon Mar 17 14:58:22 1997
.ft
.Sp
.PP
Avoid sending patches that compare backup file names like
.BR README.orig ,
since this might confuse
.B patch
into patching a backup file instead of the real file.
Instead, send patches that compare the same base file names
in different directories, e.g.\&
.B old/README
and
.BR new/README .
.PP
Take care not to send out reversed patches, since it makes people wonder
whether they already applied the patch.
.PP
Try not to have your patch modify derived files
(e.g. the file
.B configure
where there is a line
.B "configure: configure.in"
in your makefile), since the recipient should be
able to regenerate the derived files anyway.
If you must send diffs of derived files,
ensure that the diffs for each derived file
follow the diffs for the files that it depends on,
so that the dependencies will be preserved as
.B patch
updates the files one by one.
Here is a sample shell script that output patches in an order that
should preserve dependencies:
.nf
.Sp
.ft B
.in +3n
#! /bin/sh
.Sp
.ne 2
old=${1?}
new=${2?}
.Sp
.ne 10
fs=
for f in `
  diff \-Nqr $old $new |
  sed \-e "s,.\(** $new/,," \-e 's, differ$,,'`
do
  if [ \-f $new/$f ]
  then fs="$fs $f"
  else diff \-au $old/$f /dev/null
  fi
done
.Sp
.ne 10
case $fs in
?\(**)
  for f in `cd $new; ls \-rt $fs`
  do
    if [ \-f $old/$f ]
    then diff \-au $old/$f $new/$f
    else diff \-au /dev/null $new/$f
    fi
  done
esac
.in
.ft
.fi
.PP
While you may be able to get away with putting 582 diff listings into
one file, it may be wiser to group related patches into separate files in
case something goes haywire.
.SH DIAGNOSTICS
Diagnostics generally indicate that
.B patch
couldn't parse your patch file.
.PP
If the
.B \*=verbose
option is given, the message
.B Hmm.\|.\|.\&
indicates that there is unprocessed text in
the patch file and that
.B patch
is attempting to intuit whether there is a patch in that text and, if so,
what kind of patch it is.
.PP
.BR patch 's
exit status is
0 if all hunks are applied successfully,
1 if some hunks cannot be applied,
and 2 if there is more serious trouble.
When applying a set of patches in a loop it behooves you to check this
exit status so you don't apply a later patch to a partially patched file.
.SH CAVEATS
Context diffs cannot reliably represent the creation or deletion of
empty files, empty directories, or special files such as symbolic links.
Nor can they represent changes to file metadata like ownership, permissions,
or whether one file is a hard link to another.
If changes like these are also required, separate instructions
(e.g. a shell script) to accomplish them should accompany the patch.
.PP
.B patch
cannot tell if the line numbers are off in an
.B ed
script, and can detect
bad line numbers in a normal diff only when it finds a change or deletion.
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.
.PP
.B 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.
.SH BUGS
.B patch
could be smarter about partial matches, excessively deviant offsets and
swapped code, but that would take an extra pass.
.PP
If code has been duplicated (for instance with
\fB#ifdef OLDCODE\fP .\|.\|. \fB#else .\|.\|. #endif\fP),
.B 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.
.PP
If you apply a patch you've already applied,
.B patch
thinks it is a reversed patch, and offers to un-apply the patch.
This could be construed as a feature.
.SH COPYING
Copyright
.if t \(co
1984, 1985, 1986, 1988 Larry Wall.
.br
Copyright
.if t \(co
1997 Free Software Foundation, Inc.
.PP
Permission is granted to make and distribute verbatim copies of
this manual provided the copyright notice and this permission notice
are preserved on all copies.
.PP
Permission is granted to copy and distribute modified versions of this
manual under the conditions for verbatim copying, provided that the
entire resulting derived work is distributed under the terms of a
permission notice identical to this one.
.PP
Permission is granted to copy and distribute translations of this
manual into another language, under the above conditions for modified
versions, except that this permission notice may be included in
translations approved by the copyright holders instead of in
the original English.
.SH AUTHORS
Larry Wall wrote the original version of
.BR patch .
Paul Eggert removed
.BR patch 's
arbitrary limits, added support for binary files,
and made it conform better to \s-1POSIX\s0.
Other contributors include Wayne Davison, who added unidiff support,
and David MacKenzie, who added configuration and backup support.