patch

Updated: April 19, 2023

Apply a diff file to an original file

Syntax:

patch [-bCcEeflNnRstuv] [-B backup-prefix] [-D symbol] [-d directory]
      [-F max-fuzz] [-i patchfile] [-o out-file] [-p strip-count]
      [-r rej-name] [-V t | nil | never] [-x number] [-z backup-ext]
      [--posix] [origfile [patchfile]]
patch <patchfile

Runs on:

QNX Neutrino

Description:


     The patch utility will take a patch file containing any of the four forms
     of difference listing produced by the diff(1) program and apply those diff-
     erences to an original file, producing a patched version.  If patchfile is
     omitted, or is a hyphen, the patch will be read from the standard input.

     The patch utility will attempt to determine the type of the diff listing,
     unless overruled by a -c, -e, -n, or -u option.  Context diffs (old-style,
     new-style, and unified) and normal diffs are applied directly by the patch
     program itself, whereas ed diffs are simply fed to the ed(1) editor via a
     pipe.

     If the patchfile 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 file
     names and revision level (see the section on Filename Determination
     below).

OPTIONS

     -B backup-prefix, --prefix backup-prefix
             Causes the next argument to be interpreted as a prefix to the
             backup file name.  If this argument is specified, any argument to
             -z will be ignored.

     -b, --backup
             Saves a backup copy of the file before it is modified.  By default
             the original file is saved with a backup extension of ".orig"
             unless the file already has a numbered backup, in which case a
             numbered backup is made.  This is equivalent to specifying "-V
             existing".  This option is currently the default, unless --posix
             is specified.

     -C, --check
             Checks that the patch would apply cleanly, but does not modify
             anything.

     -c, --context
             Forces patch to interpret the patch file as a context diff.

     -D symbol, --ifdef symbol
             Causes patch to use the "#ifdef...#endif" construct to mark
             changes.  The argument following will be used as the differenti-
             ating symbol.  Note that, unlike the C compiler, there must be a
             space between the -D and the argument.

     -d directory, --directory directory
             Causes patch to interpret the next argument as a directory, and
             change the working directory to it before doing anything else.

     -E, --remove-empty-files
             Causes patch to remove output files that are empty after the
             patches have been applied.  This option is useful when applying
             patches that create or remove files.

     -e, --ed
             Forces patch to interpret the patch file as an ed(1) script.

     -F max-fuzz, --fuzz max-fuzz
             Sets the maximum fuzz factor.  This option only applies to con-
             text 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.

     -f, --force
             Forces patch to assume that the user knows exactly what he or she
             is doing, and to not ask any questions.  It assumes the follow-
             ing: skip patches for which a file to patch cannot be found;
             patch files even though they have the wrong version for the
             "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 -s for that.

     -i patchfile, --input patchfile
             Causes the next argument to be interpreted as the input file name
             (i.e. a patchfile).  This option may be specified multiple times.

     -l, --ignore-whitespace
             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.

     -N, --forward
             Causes patch to ignore patches that it thinks are reversed or
             already applied.  See also -R.

     -n, --normal
             Forces patch to interpret the patch file as a normal diff.

     -o out-file, --output out-file
             Causes the next argument to be interpreted as the output file
             name.

     -p strip-count, --strip strip-count
             Sets the pathname strip count, which controls how pathnames 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.  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 file name 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

             Not specifying -p at all just gives you blurfl.c, unless all of
             the directories in the leading path (u/howard/src/blurfl) exist
             and that path is relative, in which case you get the entire path-
             name unmodified.  Whatever you end up with is looked for either
             in the current directory, or the directory specified by the -d
             option.

     -R, --reverse
             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.)  The patch utility will attempt to
             swap each hunk around before applying it.  Rejects will come out
             in the swapped format.  The -R option 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 option set.  If it cannot, 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.)

     -r rej-name, --reject-file rej-name
             Causes the next argument to be interpreted as the reject file
             name.

     -s, --quiet, --silent
             Makes patch do its work silently, unless an error occurs.

     -t, --batch
             Similar to -f, in that it suppresses questions, but makes some
             different assumptions: skip patches for which a file to patch
             cannot be found (the same as -f); skip patches for which the file
             has the wrong version for the "Prereq:" line in the patch; and
             assume that patches are reversed if they look like they are.

     -u, --unified
             Forces patch to interpret the patch file as a unified context
             diff (a unidiff).

     -V t | nil | never, --version-control t | nil | never
             Causes the next argument to be interpreted as a method for creat-
             ing backup file names.  The type of backups made can also be
             given in the PATCH_VERSION_CONTROL or VERSION_CONTROL
             environment variables, which are overridden by this option.  
             The -B option overrides this option, causing the prefix to always
             be used for making backup file names.  The values of the
             PATCH_VERSION_CONTROL and VERSION_CONTROL environment
             variables and the argument to the -V option are like the GNU Emacs
             “version-control” variable; they also recognize synonyms that
             are more descriptive.  The valid values are (unique abbreviations
             are accepted):

                   t, numbered
                           Always make numbered backups.

                   nil, existing
                           Make numbered backups of files that already have
                           them, simple backups of the others.

                   never, simple
                           Always make simple backups.

     -v, --version
             Causes patch to print out its revision header and patch level.

     -x number, --debug number
             Sets internal debugging flags, and is of interest only to patch
             patchers.

     -z backup-ext, --suffix backup-ext
             Causes the next argument to be interpreted as the backup exten-
             sion, to be used in place of ".orig".

     --posix
             Enables strict conformance, specifically:

             1.   Backup files are not created unless the -b option is speci-
                  fied.

             2.   If unspecified, the file name used is the first of the old,
                  new and index files that exists.

Patch Application
     The patch utility 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 con-
     text.  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 out-
     put file plus ".rej".  (Note that the rejected hunk will come out in con-
     text 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 dif-
     ferent 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.

Filename Determination
     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.
     When checking a prospective file name, pathname components are stripped
     as specified by the -p option and the file's existence and writability
     are checked relative to the current working directory (or the directory
     specified by the -d option).

     If the diff is a context or unified diff, patch is able to determine the
     old and new file names from the diff header.  For context diffs, the
     “old” file is specified in the line beginning with "***" and the
     “new” file is specified in the line beginning with "---".  For a uni-
     fied diff, the “old” file is specified in the line beginning with "---"
     and the “new” file is specified in the line beginning with "+++".  If
     there is an "Index:" line in the leading garbage (regardless of the diff
     type), patch will use the file name from that line as the “index” file.

     The patch utility will choose the file name by performing the following
     steps, with the first match used:

     1.   If patch is operating in strict mode, the first of the “old”,
          “new” and “index” file names that exist is used.  Otherwise,
          patch will examine either the “old” and “new” file names or, for
          a non-context diff, the “index” file name, and choose the file
          name with the fewest path components, the shortest basename,
          and the shortest total file name length (in that order).

     2.   If no file exists, patch checks for the existence of the files in an
          SCCS or RCS directory (using the appropriate prefix or suffix) using
          the criteria specified above.  If found, patch will attempt to get
          or check out the file.

     3.   If no suitable file was found to patch, the patch file is a context
          or unified diff, and the old file was zero length, the new file name
          is created and used.

     4.   If the file name still cannot be determined, patch will prompt the
          user for the file name to use.

     Additionally, 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 con-
     taining the patch.

Backup Files
     By default, the patched version is put in place of the original, with the
     original file backed up to the same name with the extension ".orig", or
     as specified by the -B, -V, or -z options.  The extension used for making
     backup files may also be specified in the SIMPLE_BACKUP_SUFFIX
     environment variable, which is overridden by the options above.

     If the backup file is a symbolic or hard link to the original file, patch
     creates a new backup file name by changing the first lowercase letter in
     the last component of the file's name into uppercase.  If there are no
     more lowercase letters in the name, it removes the first character from
     the name.  It repeats this process until it comes up with a backup file
     that does not already exist or is not linked to the original file.

     You may also specify where you want the output to go with the -o option;
     if that file already exists, it is backed up first.

Notes For Patch Senders
     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 will not let them apply patches out of order without some warning.

     Second, make sure you have specified the file names right, either in a
     context diff header, or with an "Index:" line.  If you are patching some-
     thing in a subdirectory, be sure to tell the patch user to specify a -p
     option 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 does not 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 sepa-
     rate files in case something goes haywire.

ENVIRONMENT
     POSIXLY_CORRECT        When set, patch behaves as if the --posix option
                            has been specified.
     SIMPLE_BACKUP_SUFFIX   Extension to use for backup file names instead of
                            ".orig".
     TMPDIR                 Directory to put temporary files in; default is
                            /tmp.
     PATCH_VERSION_CONTROL  Selects when numbered backup files are made.
     VERSION_CONTROL        Same as PATCH_VERSION_CONTROL.

FILES
     $TMPDIR/patch*  patch temporary files
     /dev/tty        used to read input when patch prompts the user

EXIT STATUS
     The patch utility exits with one of the following values:

           0       Successful completion.
           1       One or more lines were written to a reject file.
           >1      An error occurred.

     When applying a set of patches in a loop it behooves you to check this
     exit status so you do not apply a later patch to a partially patched
     file.

DIAGNOSTICS
     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.

SEE ALSO
     diff(1)

STANDARDS
     The patch utility is compliant with the specification (except as
     detailed above for the --posix option), though the presence of patch
     itself is optional.

     The flags [-BCEFfstVvxz] and [--posix] are extensions to that specifica-
     tion.

AUTHORS
     Larry Wall with many other contributors.

CAVEATS
     The patch utility 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.

     The utility 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.

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

     Check patch mode (-C) will fail if you try to check several patches in
     succession that build on each other.  The entire patch code would have to
     be restructured to keep temporary files around so that it can handle this
     situation.

     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 suc-
     ceeded to boot.

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

NOTES
       Portions of this document came from FreeBSD.