patch帮助

luther@gliethttp:~$ man patch

NAME
       patch - apply a diff file to an original

SYNOPSIS
       patch [options] [originalfile [patchfile]]

       but usually just

       patch -pnum
比如
1. 单个文件
luther@gliethttp:~$ diff -ruNa hello.c hello.new.c >gliethttp_hello.diff
luther@gliethttp:~$ patch -p0 hello.c < gliethttp_hello.diff
2. 目录
luther@gliethttp:~$ diff -ruNa test1 test1_new > gliethttp_test_folder.diff
luther@gliethttp:~/test1$ patch -p1 <../gliethttp_test_folder.diff


DESCRIPTION
       patch takes a patch file patchfile containing a difference listing produced by the 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 or --backup
       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 originalfile.

       Upon  startup,  patch  attempts to determine the type of the diff listing, unless overruled by a -c
       (--context), -e (--ed), -n (--normal), or -u (--unified) option.  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(1) editor via a pipe.

       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 patch, and it should work.  If
       the entire diff is indented by a consistent amount, or if a context diff contains lines  ending  in
       CRLF  or  is encapsulated one or more times by prepending "- " to lines starting with "-" as speci‐
       fied by Internet RFC 934, this is taken into account.  After removing indenting  or  encapsulation,
       lines beginning with # are ignored, as they are considered to be comments.

       With  context  diffs, and to a lesser extent with normal diffs, patch can detect when the line num‐
       bers 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, patch scans
       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 con‐
       text  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  puts  the  hunk out to a reject file, which normally is the name of the output file plus a .rej
       suffix, or # if .rej would generate a file name that is too long  (if  even  appending  the  single
       character  #  makes  the file name too long, then # 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.

       As  each hunk is completed, you are told if the hunk failed, and if so which line (in the new file)
       patch thought the hunk should go on.  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 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.  If the --verbose option is given, you
       are also told about hunks that match exactly.

       If no original file origfile is specified on the command line, patch tries to figure out  from  the
       leading garbage what the name of the file to edit is, using the following rules.

       First, patch takes an ordered list of candidate file names as follows:

        ?· If  the  header is that of a context diff, patch takes the old and new file names in the header.
          A name is ignored if it does not have enough slashes to satisfy the -pnum or --strip=num option.
          The name /dev/null is also ignored.

        ?· If  there  is an Index: line in the leading garbage and if either the old and new names are both
          absent or if patch is conforming to POSIX, patch takes the name in the Index: line.

        ?· For the purpose of the following rules, the candidate file names are considered  to  be  in  the
          order (old, new, index), regardless of the order that they appear in the header.

       Then patch selects a file name from the candidate list as follows:

        ?· If  some  of the named files exist, patch selects the first name if conforming to POSIX, and the
          best name otherwise.

        ?· If patch is not ignoring RCS, ClearCase,  Perforce,  and  SCCS  (see  the  -g num  or  --get=num
          option),  and  no  named  files  exist but an RCS, ClearCase, Perforce, or SCCS master is found,
          patch selects the first named file with an RCS, ClearCase, Perforce, or SCCS master.

        ?· If no named files exist, no RCS, ClearCase, Perforce, or SCCS master was found, some  names  are
          given,  patch  is not conforming to POSIX, and the patch appears to create a file, patch selects
          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, and patch selects that name.

       To  determine  the  best of a nonempty list of file names, 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.

       Additionally,  if  the leading garbage contains a Prereq: line, 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, patch asks for confirmation before proceeding.

       The upshot of all this is that you should be able to say, while in a news interface, something like
       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 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 list‐
       ing contains interesting things such as file names and revision level, as mentioned previously.

OPTIONS
       -b  or  --backup
          Make backup files.  That is, when patching a file, rename or copy the original instead of remov‐
          ing it.  When backing up a file that does not exist, an empty, unreadable backup file is created
          as  a placeholder to represent the nonexistent file.  See the -V or --version-control option for
          details about how backup file names are determined.

       --backup-if-mismatch
          Back up a file if the patch does not match the file exactly and if  backups  are  not  otherwise
          requested.  This is the default unless patch is conforming to POSIX.

       --no-backup-if-mismatch
          Do not back up a file if the patch does not match the file exactly and if backups are not other‐
          wise requested.  This is the default if patch is conforming to POSIX.

       -B pref  or  --prefix=pref
          Prefix pref to a file name when generating its simple  backup  file  name.   For  example,  with
          -B /junk/ the simple backup file name for src/patch/util.c is /junk/src/patch/util.c.

       --binary
          Read  and  write all files in binary mode, except for standard output and /dev/tty.  This option
          has no effect on POSIX-conforming systems.  On systems like DOS where this option makes  a  dif‐
          ference, the patch should be generated by diff -a --binary.

       -c  or  --context
          Interpret the patch file as a ordinary context diff.

       -d dir  or  --directory=dir
          Change to the directory dir immediately, before doing anything else.

       -D define  or  --ifdef=define
          Use the #ifdef ... #endif construct to mark changes, with define as the differentiating symbol.

       --dry-run
          Print the results of applying the patches without actually changing any files.

       -e  or  --ed
          Interpret the patch file as an ed script.

       -E  or  --remove-empty-files
          Remove output files that are empty after the patches have been applied.  Normally this option is
          unnecessary, since patch can examine the time stamps on the header to determine whether  a  file
          should  exist  after  patching.  However, if the input is not a context diff or if patch is con‐
          forming to POSIX, patch does not remove empty patched files unless this option is  given.   When
          patch removes a file, it also attempts to remove any empty ancestor directories.

       -f  or  --force
          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 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.

       -F num  or  --fuzz=num
          Set the maximum fuzz factor.  This option only applies to diffs that have  context,  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.

       -g num  or  --get=num
          This  option  controls  patch's  actions  when a file is under RCS or SCCS control, and does not
          exist or is read-only and matches the default version, or when a file is under ClearCase or Per‐
          force  control and does not exist.  If num is positive, patch gets (or checks out) the file from
          the revision control system; if zero, patch ignores RCS, ClearCase, Perforce, and SCCS and  does
          not  get  the  file;  and if negative, patch asks the user whether to get the file.  The default
          value of this option is given by the value of the PATCH_GET environment variable if it  is  set;
          if not, the default value is zero if patch is conforming to POSIX, negative otherwise.

       --help
          Print a summary of options and exit.

       -i patchfile  or  --input=patchfile
          Read the patch from patchfile.  If patchfile is -, read from standard input, the default.

       -l  or  --ignore-whitespace
          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.

       -n  or  --normal
          Interpret the patch file as a normal diff.

       -N  or  --forward
          Ignore patches that seem to be reversed or already applied.  See also -R.

       -o outfile  or  --output=outfile
          Send output to outfile instead of patching files in place.  Do not use this option if outfile is
          one of the files to be patched.

       -pnum  or  --strip=num
          Strip  the smallest prefix containing num 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

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

          setting -p0 gives the entire file name 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 option.

       --posix
          Conform more strictly to the POSIX standard, as follows.

           ?· Take the first existing file from the list (old, new, index) when intuiting file  names  from
             diff headers.

           ?· Do not remove files that are empty after patching.

           ?· Do not ask whether to get files from RCS, ClearCase, Perforce, or SCCS.

           ?· Require that all options precede the files in the command line.

           ?· Do not backup files when there is a mismatch.

       --quoting-style=word
          Use style word to quote output names.  The word should be one of the following:

          literal
                 Output names as-is.

          shell  Quote  names  for the shell if they contain shell metacharacters or would cause ambiguous
                 output.

          shell-always
                 Quote names for the shell, even if they would normally not require quoting.

          c      Quote names as for a C language string.

          escape Quote as with c except omit the surrounding double-quote characters.

          You can specify the default value of the --quoting-style option with  the  environment  variable
          QUOTING_STYLE.  If that environment variable is not set, the default value is shell.

       -r rejectfile  or  --reject-file=rejectfile
          Put rejects into rejectfile instead of the default .rej file.

       -U  or  --unified-reject-files
          Produce unified reject files. The default is to produce context type reject files.

       --global-reject-file=rejectfile
          Put  all  rejects  into  rejectfile instead of creating separate reject files for all files that
          have rejects. The rejectfile will contain headers that identify which file  each  reject  refers
          to. Note that the global reject file is created even if --dry-run is specified (while non-global
          reject files will only be created without --dry-run).

       -R  or  --reverse
          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.)  patch attempts to swap each hunk
          around before applying it.  Rejects come out in the swapped format.  The -R option does not work
          with  ed  diff scripts because there is too little information to reconstruct the reverse opera‐
          tion.

          If the first hunk of a patch fails, 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 -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.)

       -s  or  --silent  or  --quiet
          Work silently, unless an error occurs.

       -t  or  --batch
          Suppress questions like -f, but make some different assumptions: skip patches whose  headers  do
          not  contain  file names (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.

       -T  or  --set-time
          Set  the  modification  and access times of patched files from time stamps given in context diff
          headers, assuming that the context diff headers use local time.  This option is not recommended,
          because  patches  using  local  time  cannot  easily  be used by people in other time zones, and
          because local time stamps are ambiguous when local clocks move backwards during  daylight-saving
          time  adjustments.   Instead  of  using this option, generate patches with UTC and use the -Z or
          --set-utc option instead.

       -u  or  --unified
          Interpret the patch file as a unified context diff.

       -v  or  --version
          Print out patch's revision header and patch level, and exit.

       -V method  or  --version-control=method
          Use method to determine backup file names.  The method can  also  be  given  by  the  PATCH_VER??‐
          SION_CONTROL  (or,  if that's not set, the VERSION_CONTROL) environment variable, which is over‐
          ridden by this option.  The method does not affect whether backup files  are  made;  it  affects
          only the names of any backup files that are made.

          The value of method is like the GNU Emacs `version-control' variable; patch also recognizes syn‐
          onyms that are more descriptive.  The valid values for  method  are  (unique  abbreviations  are
          accepted):

          existing  or  nil
             Make numbered backups of files that already have them, otherwise simple backups.  This is the
             default.

          numbered  or  t
             Make numbered backups.  The numbered backup file name for F is F.~N~ where N is  the  version
             number.

          simple  or  never
             Make simple backups.  The -B or --prefix, -Y or --basename-prefix, and -z or --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 SIMPLE_BACKUP_SUFFIX environment variable if
             set, and is .orig otherwise.

          With numbered or simple backups, if the backup file name is too long, the  backup  suffix  ~  is
          used instead; if even appending ~ would make the name too long, then ~ replaces the last charac‐
          ter of the file name.

       --verbose
          Output extra information about the work being done.

       -x num  or  --debug=num
          Set internal debugging flags of interest only to patch patchers.

       -Y pref  or  --basename-prefix=pref
          Prefix pref to the basename of a file name when generating its simple  backup  file  name.   For
          example,    with    -Y .del/   the   simple   backup   file   name   for   src/patch/util.c   is
          src/patch/.del/util.c.

       -z suffix  or  --suffix=suffix
          Use suffix as the simple backup suffix.  For example, with -z - the simple backup file name  for
          src/patch/util.c  is  src/patch/util.c-.   The  backup  suffix may also be specified by the SIM??‐
          PLE_BACKUP_SUFFIX environment variable, which is overridden by this option.

       -Z  or  --set-utc
          Set the modification and access times of patched files from time stamps given  in  context  diff
          headers, assuming that the context diff headers use Coordinated Universal Time (UTC, often known
          as GMT).  Also see the -T or --set-time option.

          The -Z or --set-utc and -T or --set-time options normally refrain from setting a file's time  if
          the  file's  original time does not match the time given in the patch header, or if its contents
          do not match the patch exactly.  However, if the -f or --force option is given, the file time is
          set regardless.

          Due  to  the  limitations  of diff output format, these options cannot update the times of files
          whose contents have not changed.  Also, if you use these options, you should remove  (e.g.  with
          make clean) all files that depend on the patched files, so that later invocations of make do not
          get confused by the patched files' times.

ENVIRONMENT
       PATCH_GET
          This specifies whether patch gets missing or read-only files from RCS, ClearCase,  Perforce,  or
          SCCS by default; see the -g or --get option.

       POSIXLY_CORRECT
          If set, patch conforms more strictly to the POSIX standard by default: see the --posix option.

       QUOTING_STYLE
          Default value of the --quoting-style option.

       SIMPLE_BACKUP_SUFFIX
          Extension to use for simple backup file names instead of .orig.

       TMPDIR, TMP, TEMP
          Directory to put temporary files in; patch uses the first environment variable in this list that
          is set.  If none are set, the default is system-dependent; it is normally /tmp on Unix hosts.

       VERSION_CONTROL or PATCH_VERSION_CONTROL
          Selects version control style; see the -v or --version-control option.

FILES
       $TMPDIR/p*
          temporary files

       /dev/tty
          controlling terminal; used to get answers to questions asked of the user

SEE ALSO
       diff(1), ed(1)

       Marshall T. Rose and Einar A. Stefferud, Proposed Standard for Message Encapsulation, Internet  RFC
       934 (1985-01).

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

       Create  your  patch  systematically.  A good method is the command diff -Naur old new where old and
       new identify the old and new directories.  The names old and new should not  contain  any  slashes.
       The  diff  command's  headers  should have dates and times in Universal Time using traditional Unix
       format, so that patch recipients can use the -Z or --set-utc option.  Here is an  example  command,
       using Bourne shell syntax:

          LC_ALL=C TZ=UTC0 diff -Naur gcc-2.7 gcc-2.8

       Tell  your  recipients  how  to apply the patch by telling them which directory to cd to, and which
       patch options to use.  The option string -Np1 is recommended.  Test your procedure by pretending to
       be a recipient and applying your patch to a copy of the original files.

       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.

       You  can  create  a  file  by sending out a diff that compares /dev/null or an empty file dated the
       Epoch (1970-01-01 00:00:00 UTC) 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 context diff that compares the file to be deleted with an empty file dated the Epoch.
       The  file  will  be  removed unless patch is conforming to POSIX and the -E or --remove-empty-files
       option is not given.  An easy way to generate patches that create and remove files is  to  use  GNU
       diff's -N or --new-file option.

       If the recipient is supposed to use the -pN option, do not send output that looks like this:

          diff -Naur v2.0.29/prog/README prog/README
          --- v2.0.29/prog/README   Mon Mar 10 15:13:12 1997
          +++ prog/README   Mon Mar 17 14:58:22 1997

       because  the  two  file  names  have  different numbers of slashes, and different versions of patch
       interpret the file names differently.  To  avoid  confusion,  send  output  that  looks  like  this
       instead:

          diff -Naur v2.0.29/prog/README v2.0.30/prog/README
          --- v2.0.29/prog/README   Mon Mar 10 15:13:12 1997
          +++ v2.0.30/prog/README   Mon Mar 17 14:58:22 1997


       Avoid  sending  patches  that  compare backup file names like README.orig, since this might confuse
       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. old/README and new/README.

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

       Try not to have your patch modify derived files (e.g. the file configure where there is a line con??‐
       figure:  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, generate the diffs using UTC,  have
       the  recipients apply the patch with the -Z or --set-utc option, and have them remove any unpatched
       files that depend on patched files (e.g. with make clean).

       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.

DIAGNOSTICS
       Diagnostics generally indicate that patch couldn't parse your patch file.

       If  the  --verbose  option is given, 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.

       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.

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 own‐
       ership, 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.

       patch cannot tell if the line numbers are off in an 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 con‐
       text 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.

COMPATIBILITY ISSUES
       The  POSIX  standard specifies behavior that differs from patch's traditional behavior.  You should
       be aware of these differences if you must interoperate with patch versions 2.1 and  earlier,  which
       do not conform to POSIX.

        ?· In traditional patch, the -p option's operand was optional, and a bare -p was equivalent to -p0.
          The -p option now requires an operand, and -p 0 is now equivalent to -p0.  For maximum  compati‐
          bility, use options like -p0 and -p1.

          Also,  traditional  patch  simply counted slashes when stripping path prefixes; patch now counts
          pathname components.  That is, a sequence of one or more adjacent slashes now counts as a single
          slash.  For maximum portability, avoid sending patches containing // in file names.

        ?· In traditional patch, backups were enabled by default.  This behavior is now enabled with the -b
          or --backup option.

          Conversely, in POSIX patch, backups are never made, even when  there  is  a  mismatch.   In  GNU
          patch,  this  behavior  is  enabled with the --no-backup-if-mismatch option, or by conforming to
          POSIX with the --posix option or by setting the POSIXLY_CORRECT environment variable.

          The -b suffix option of traditional patch is equivalent  to  the  -b -z suffix  options  of  GNU
          patch.

        ?· Traditional  patch used a complicated (and incompletely documented) method to intuit the name of
          the file to be patched from the patch header.  This method did not conform to POSIX, and  had  a
          few  gotchas.   Now  patch  uses a different, equally complicated (but better documented) method
          that is optionally POSIX-conforming; we hope it has fewer gotchas.  The two methods are compati‐
          ble  if  the  file  names in the context diff header and the Index: line are all identical after
          prefix-stripping.  Your patch is normally compatible if each header's file names all contain the
          same number of slashes.

        ?· When  traditional  patch  asked  the user a question, it sent the question to standard error and
          looked for an answer from the first file in the following list that  was  a  terminal:  standard
          error,  standard  output,  /dev/tty,  and standard input.  Now patch sends questions to standard
          output and gets answers from /dev/tty.  Defaults for some answers  have  been  changed  so  that
          patch never goes into an infinite loop when using default answers.

        ?· Traditional  patch exited with a status value that counted the number of bad hunks, or with sta‐
          tus 1 if there was real trouble.  Now patch exits with status 1 if some hunks failed, or with  2
          if there was real trouble.

        ?· Limit yourself to the following options when sending instructions meant to be executed by anyone
          running GNU patch, traditional patch, or a patch that conforms to POSIX.  Spaces are significant
          in the following list, and operands are required.

             -c
             -d dir
             -D define
             -e
             -l
             -n
             -N
             -o outfile
             -pnum
             -R
             -r rejectfile

BUGS
       Please report bugs via email to .

       patch  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 inca‐
       pable 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 thinks it is a reversed patch, and offers to un-
       apply the patch.  This could be construed as a feature.

COPYING
       Copyright (C) 1984, 1985, 1986, 1988 Larry Wall.
       Copyright  (C)  1989,  1990, 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999, 2000, 2001, 2002
       Free Software Foundation, Inc.

       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.

       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.

       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.

AUTHORS
       Larry  Wall  wrote  the  original  version of patch.  Paul Eggert removed patch's arbitrary limits;
       added support for binary files, setting file times, and deleting files; and made it conform  better
       to  POSIX.   Other contributors include Wayne Davison, who added unidiff support, and David MacKen‐
       zie, who added configuration and backup support.



GNU                                             2002/05/25                                        PATCH(1)