NAME

Arch::DiffParser - parse file's diff and perform some manipulations

SYNOPSIS

    use Arch::DiffParser;
    my $dp = Arch::DiffParser->new;
    # usable for "annotate" functionality
    my $changes = $dp->parse_file("f.diff")->changes;
    $dp->parse($diff_content);
    $dp->parse("--- f1.c\t2005-02-26\n+++ f2.c\t2005-02-28\n...");
    # prints "f1.c, f2.c"
    printf "%s, %s\n", $dp->filename1, $dp->filename2;
    # enclose lines in <span class="patch_{mod,orig,line,add,del}">
    my $html = $dp->markup_content;

DESCRIPTION

This class provides a limited functionality to parse a single file diff in unified format. Multiple diffs may be parsed sequentially. The parsed data is stored for the last diff, and is replaced on the following parse.

METHODS

The following class methods are available:

new, parse, parse_file, content, lines, filename1, filename2, mtime1, mtime2, hunks, changes.

new

Construct the "Arch::DiffParser" instanse.

parse diff_content

Parse the diff_content and store its parsed data.

parse_file diff_filename

Like parse, but read the diff_content from diff_filename.

diff_data

Return hashref containing certain parsed data. Die if called before any parse methods. The keys are: "lines", "filename1", "filename2", "mtime1", "mtime2", "hunks", "changes".

The value of "hunks" and "changes" is arrayref of arrayrefs with 5 elements: [ line-number-1, num-lines-1, line-number-2, num-lines-2, "lines"-index ].

A "hunk" describes a set of lines containing some combination of unmodified, deleted and added lines, a "change" describes an inter-hunk atom that only contains zero or more deleted lines and zero or more added lines.

These methods are just shortcuts for diff_data->{method}.

content [%args]

Return content of the last diff.

%args keys are "fileroot1" and "fileroot2"; if given, these will replace the subdirs "orig" and "mod" that arch usually uses in the filepaths.

markup_content [%args]

Like content, but every non-context line is enclosed into markup <span class="patch_name">line</span>, where name is one of "orig" (filename1), "mod" (filename2), "line" (hunk linenums), "add" (added), del (deleted).

Not implemented yet.

BUGS

No support for newlines in source file names yet.

AUTHORS