Philip Gwyn
and 1 contributors


Text::Diff::Parser - Parse patch files containing unified and standard diffs


    use Text::Diff::Parser;

    # create the object
    my $parser = Text::Diff::Parser->new();

    # With options
    $parser = Text::Diff::Parser->new( Simplify=>1, # simplify the diff
                                       Strip=>2 );  # strip 2 directories

    # Create object.  Parse $file
    $parser = Text::Diff::Parser->new( $file );
    $parser = Text::Diff::Parser->new( File=>$file );

    # Create object.  Parse text
    my $parser = Text::Diff::Parser->new( $text );
    $parser = Text::Diff::Parser->new( Diff=>$text );

    # parse a file
    $parser->parse_file( $filename );

    # parse a string
    $parser->parse( $text );
    # Remove no-change lines.  Combine line substitutions

    # Find results
    foreach my $change ( $parser->changes ) {
        print "File1: ", $change->filename1;
        print "Line1: ", $change->line1;
        print "File2: ", $change->filename2;
        print "Line2: ", $change->line2;
        print "Type: ", $change->type;
        my $size = $change->size;
        foreach my $line ( 0..($size-1) ) {
            print "Line: ", $change->text( $line );

    # In scalar context, returns the number of changes
    my $n = $parser->changes;
    print "There are $n changes", 

    # Get the changes to a given file
    my @changes = $parser->changes( 'Makefile.PL' );

    # Get list of files changed by the diff
    my @files = $parser->files;


Text::Diff::Parser parses diff files and patches. It allows you to access the changes to a file in a standardized way, even if multiple patch formats are used.

A diff may be viewed a series of operations on a file, either adding, removing or modifying lines of one file (the from-file) to produce another file (the to-file). Diffs are generaly produced either by hand with diff, or by your version control system (cvs diff, svn diff, ...). Some diff formats, notably unified diffs, also contain null operations, that is lines that

Text::Diff::Parser currently parses unified diff format and standard diff format.

Unified diffs look like the following.

    --- Filename1 2006-04-12 18:47:22.000000000 -0400
    +++ Filename2 2006-04-12 19:21:16.000000000 -0400
    @@ -1,4 +1,6 @@

Standard diffs look like the following.

    diff something something.4
    < THREE
    > honk
    > honk
    > honk

The diff line isn't in fact part of the format but is necessary to find which files the chunks deal with. It is output by cvs diff and svn diff so that isn't a problem.



    $parser = Text::Diff::Parser->new;
    $parser = Text::Diff::Parser->new( $file );
    $parser = Text::Diff::Parser->new( $handle );
    $parser = Text::Diff::Parser->new( %params );
    $parser = Text::Diff::Parser->new( \%params );

Object constructor.


String that contains a diff. This diff will be parse before new returns.


File name or file handle that is parsed before new returns.


Simplifying a patch involves dropping all null-operations and converting and remove operation followed by an add operation (or an add followed by a remove) of the same size on the same lines into a modify operation.


Strip N leading directories from all filenames. Less then useful for standard diffs produced by cvs diff, because they don't contain directory information.


In a unified diff, various chunks are introduced with @@. By default, we trust these to reference the right line count. If you set this to 0, the lines will not be trust and a chunk must end with another @@ line or --- (which introduces a new file). Note that not trusting @@ means you can not parse a diff that removes a line that begins with --, because that also start with '---'.


If true, print copious details of what is going on.


    $parser->parse_file( $file );
    $parser->parse_file( $handle );

Read and parse the file or file handle specified. Will die if it fails, returns true on sucess. Contents of the file may then be accessed with changes and files.


    $parser->parse( $string );

Parses the diff present in $string. Will die if it fails, returns true on sucess. Contents of the file may then be accessed with changes and files.


    %files = $parser->files;

Fetch a list of all the files that were referenced in the patch. The keys are original files (from-file) and the values are the modified files (to-file).


    @changes = $parser->changes;
    $n = $parser->changes;
    @changes = $parser->changes( $file );
    $n = $parser->changes( $file );

Return all the operations (array context) or the number of operations in the patch file. If $file is specified, only returns changes to that file (from-file or to-file).

Elements of the returned array are change objects, as described in CHANGE METHODS below.



Simplifies the diff. Removes no-change lines. Combine line substitutions. Automatically called if you supply Simplify to ->new().


    my $file = $parser->source

Returns the filename of the last file that was parsed. Returns "user filehandle" if you supplied a file handle.


The changes method returns an array of objects that describe each operation. You may use the following methods to find out details of the operation.


Returns the type of operation, either 'ADD', 'REMOVE', 'MODIFY' or '' (null operation).


Filename of the from-file.


Filename of the to-file.


Line in the from-file the operation starts at.


Line in the to-file the operation starts at.


Number of lines affected by this operation.


    $line  = $ch->text( $N );
    @lines = $ch->text;

Fetch the text of the line $N if present or all lines affected by this operation. For '' (null) and 'REMOVE' operations, these are the lines present before the operation was done ('from-file'. For 'ADD' and 'MODIFY' operations, these are the lines present after the operation was done ('to-file'.


I'm not 100% sure of standard diff handling.

Missing support for context diffs.


Text::Diff, Arch, diff.


Philip Gwyn, <>


Copyright (C) 2006 by Philip Gwyn

This library is free software; you can redistribute it and/or modify it under the same terms as Perl itself, either Perl version 5.8.8 or, at your option, any later version of Perl 5 you may have available.