DIFF(1)                                                                DIFF(1)


NAME
     diff - differential file and directory comparator

SYNOPSIS
     diff [-lrs] [-Sname] [-cefhn] [-xpat] [-biwt] dir1 dir2
     diff [-cefhn] [-biwt] file1 file2
     diff [-Dstring] [-biw] file1 file2
     GNU diff options:
     diff [-#] [-aBdHNpqTuv] [-Cnum] [-Fexp] [-Iexp] [-Llab]

DESCRIPTION
     If both arguments are directories, diff sorts the contents of the
     directories by name, and then runs the regular file diff algorithm
     (described below) on text files which are different.  Binary files which
     differ (unless -a is specified), common subdirectories, and files which
     appear in only one directory are listed.  Options when comparing
     directories are:

     -l   long output format; each text file diff is piped through pr(1) to
          paginate it, other differences are remembered and summarized after
          all text file differences are reported.

     -r   causes application of diff recursively to common subdirectories
          encountered.

     -s   causes diff to report files which are the same, which are otherwise
          not mentioned.

     -Sname
          starts a directory diff in the middle beginning with file name.

     -xpat
          adds pat to a list of regular expressions.  There can be several
          -xpat options.  Any directory or file (base) name matching one of
          the patterns is skipped.  The special characters available in the
          pattern pat are similar to those used in egrep(1).  For example, the
          following command will ignore all "hidden" files and all RCS
          archives:

                   diff -r '-x^\.' '-x,v$' foo bar

     When run on regular files, and when comparing text files which differ
     during directory comparison, diff tells what lines must be changed in the
     files to bring them into agreement.  Except in rare circumstances, diff
     finds a smallest sufficient set of file differences.  If neither file1
     nor file2 is a directory, then either may be given as `-', in which case
     the standard input is used.  If file1 is a directory, then a file in that
     directory whose file-name is the same as the file-name of file2 is used
     (and vice versa).  If both file1 and file2 are directories, then files
     (or files and directories with -r) with identical names are compared.


     There are several options for output format; the default output format
     contains lines of these forms:

          n1 a n3,n4
          n1,n2 d n3
          n1,n2 c n3,n4

     These lines resemble ed commands to convert file1 into file2.  The
     numbers after the letters pertain to file2.  In fact, by exchanging `a'
     for `d' and reading backward one may ascertain equally how to convert
     file2 into file1.  As in ed, identical pairs where n1 = n2 or n3 = n4 are
     abbreviated as a single number.

     Following each of these lines come all the lines that are affected in the
     first file flagged by `<', then all the lines that are affected in the
     second file flagged by `>'.

     Except for -b, -w, -i or -t which may be given with any of the others,
     the following options are mutually exclusive:

     -e       produces a script of a, c and d commands for the editor ed,
              which will recreate file2 from file1.  In connection with -e,
              the following shell program may help maintain multiple versions
              of a file.  Only an ancestral file ($1) and a chain of version-
              to-version ed scripts ($2,$3,...) made by diff need be on hand.
              A `latest version' appears on the standard output.

                      (shift; cat $*; echo '1,$p') | ed - $1

              Extra commands are added to the output when comparing
              directories with -e, so that the result is a sh(1) script for
              converting text files which are common to the two directories
              from their state in dir1 to their state in dir2.

     -f       produces a script similar to that of -e, not useful with ed, and
              in the opposite order.

     -n       produces a script similar to that of -e, but in the opposite
              order and with a count of changed lines on each insert or delete
              command.  This is the form used by the RCS commands for storing
              a revision change.

     -c       produces a diff with 3 lines of context.  Use the -Cnum option
              to specify the number of lines.  With -c the output format is
              modified slightly:  the output beginning with identification of
              the files involved and their creation dates and then each change
              is separated by a line with a dozen *'s.  The lines removed from
              file1 are marked with `- '; those added to file2 are marked `+
              '.  Lines which are changed from one file to the other are
              marked in both files with `! '.

              Changes which lie within <context> lines of each other are


              grouped together on output.

     -h       has no affect. It is present for compatibility with odiff(1).

     -Dstring causes diff to create a merged version of file1 and file2 on the
              standard output, with C preprocessor controls included so that a
              compilation of the result without defining string is equivalent
              to compiling file1, while defining string will yield file2.

     -b       causes trailing blanks (spaces and tabs) to be ignored, and
              other strings of blanks to compare equal.

     -w       is similar to -b but causes whitespace (blanks and tabs) to be
              totally ignored.  E.g., ``if ( a == b )'' will compare equal to
              ``if(a==b)''.

     -i       ignores the case of letters.  E.g., ``A'' will compare equal to
              ``a''.

     -t       will expand tabs in output lines.  Normal or -c output adds
              character(s) to the front of each line which may foul up the
              indentation of the original source lines and make the output
              listing difficult to interpret.  This option will preserve the
              original source's indentation.

GNU DIFF OPTIONS
     -#       Show # lines (an integer) of context.  This option does not
              specify an output format by itself; it has no effect unless it
              is combined with -c or -u.

     -a       Treat all files as text and compare them line-by-line, even if
              they do not appear to be text

     -B       Ignore changes that just insert or delete blank lines.

     -Cnum    Use the context output format, showing lines (an integer) lines
              of context

     -d       Change the algorithm to find a guaranteed minimal set of
              changes.  This makes diff slower (sometimes much slower).

     -Fexp    In context and unified format, for each hunk of differences,
              show some of the last preceding line that matches exp.

     -H       Use heuristics to speed handling of large files that have
              numerous scattered small changes

     -Iexp    Ignore changes that just insert or delete lines that match exp.

     -Llab    Use lab instead of the pathname in the context format and
              unified format headers


     -N       In directory comparison, if a file is found in only one
              directory, treat it as present but empty in the other directory.

     -p       Show which C function each change is in.

     -q       Report only whether the files differ, not the details of the
              differences.

     -T       Output a tab rather than a space before the text of a line in
              normal or context format.  This causes the alignment of tabs in
              the line to look normal.

     -u       Use the unified output format

     -v       Display the version string.

FILES
     /usr/bin/odiff    executable for odiff.
     /bin/diff         executable for GNU derived diff.
     /bin/pr           executed by the -l option.

SEE ALSO
     bdiff(1), cc(1), cmp(1), comm(1), odiff(1), diff3(1), ed(1),
     DIFFsource(5)

DIAGNOSTICS
     Exit status is 0 for no differences, 1 for some differences, 2 for
     trouble.

PERFORMANCE
     The GNU derived diff and older BSD derived odiff use different
     algorithms.  If the two files to be compared fit in main memory, then
     diff is faster, but if they don't fit, diff can cause severe paging.  The
     odiff algorithm uses a small, fixed amount of memory even on large files.
     The multiple pass algorithm of odiff is up to two or three times slower
     than diff, on files that fit in memory, but can be many times faster on
     files that don't fit.

     The diff command automatically invokes the odiff (searching for odiff
     along $PATH) on files that have a combined size larger than one-fourth of
     memory size, if no GNU specific diff options were specified.  If that
     fails, then diff silently does the work itself.

     You can force either variant: specifying -H always gets diff, specifying
     odiff always gets odiff.


                                                                        Page 4