sourCEntral - mobile manpages




Perl::Tags − Generate (possibly exuberant) Ctags style tags for Perl sourcecode


        use Perl::Tags;
        my $naive_tagger = Perl::Tags::Naive−>new( max_level=>2 );
            files => ['', ''],

Recursively follows "use" and "require" statements, up to a maximum of "max_level".

The implemented tagger, "Perl::Tags::Naive" is a more-or-less straight ripoff, slightly updated, of the original pltags code, and is rather naive. It should be possible to subclass using something like "PPI" or "Text::Balanced", though be aware that this is alpha software and the internals are subject to change (so get in touch to let me know what you want to do and I’ll try to help).


    * Recursive, incremental tagging.
    * parses `use_ok`/`require_ok` line from Test::More


"Perl::Tags" is designed to be used with vim. My "~/.vim/ftplugin/perl.vim" contains the following:

    setlocal iskeyword+=:  " make tags with :: in them useful
    if ! exists("s:defined_functions")
    function s:init_tags()
        perl <<EOF
            use Perl::Tags;
            $naive_tagger = Perl::Tags::Naive−>new( max_level=>2 );
                # only go one level down by default
    " let vim do the tempfile cleanup and protection
    let s:tagsfile = tempname()
    function s:do_tags(filename)
        perl <<EOF
            my $filename = VIM::Eval('a:filename');
            $naive_tagger−>process(files => $filename, refresh=>1 );
            my $tagsfile=VIM::Eval('s:tagsfile');
            # of course, it may not even output, for example, if there's nothing new to process
            $naive_tagger−>output( outfile => $tagsfile );
    call s:init_tags() " only the first time
    let s:defined_functions = 1
    call s:do_tags(expand('%'))
    augroup perltags
    autocmd BufRead,BufWritePost *.pm,*.pl call s:do_tags(expand('%'))
    augroup END

Note the following:

You will need to have a vim with perl compiled in it. Debuntu packages this as "vim−perl". Alternatively you can compile from source (you’ll need Perl + the development headers "libperl−dev").

The "EOF" in the examples has to be at the beginning of the line (the verbatim text above has leading whitespace)


Perl::Tags is an abstract baseclass. Perl::Tags::Naive is provided and can be instantiated with "new".

    $naive_tagger = Perl::Tags::Naive−>new( max_level=>2 );

Accepts the following parameters

    max_level:    levels of "use" statements to descend into, default 2
    do_variables: tag variables?  default 1 (true)
    exts:         use the Exuberant extensions

A Perl::Tags object will stringify to a textual representation of a ctags file.

    print $tagger;

Delete all tags, but without touching the "order" seen, that way, if the tags are recreated, they will remain near the top of the "interestingness" tree

Save the file to disk if it has changed. (The private "{is_dirty}" attribute is used, as the tags object may be made up incrementally and recursively within your IDE .

Scan one or more Perl file for tags

        files => [ '',  '' ]
        files   => '',
        refresh => 1,

"queue", "popqueue"
Internal methods managing the processing

"process_item", "process_file"
Do the heavy lifting for "process" above.

The parsing is done by a number of lightweight objects (parsers) which look for subroutine references, variables, module inclusion etc. When they are successful, they call the "register" method in the main tags object.

Return the parses for this object. Abstract, see Perl::Tags::Naive below.


A naive implementation. That is to say, it’s based on the classic "" script distributed with Perl, which is by and large a better bet than the results produced by "ctags". But a "better" approach may be to integrate this with PPI .

See TodoTagger in the "t/" directory of the distribution for a fully working example (tested in <t/02_subclass.t>). You may want to reuse parsers in the ::Naive package, or use all of the existing parsers and add your own.

    package My::Tagger;
    use Perl::Tags;
    our @ISA = qw( Perl::Tags::Naive );
    sub get_parsers {
        my $self = shift;
        return (
            $self−>can('todo_line'),     # a new parser
            $self−>SUPER::get_parsers(), # all ::Naive's parsers
            # or maybe...
            $self−>can('variable'),      # one of ::Naive's parsers
    sub todo_line {
        # your new parser code here!
    sub package_line {
        # override one of ::Naive's parsers

Because ::Naive uses "can('parser')" instead of "\&parser", you can just override a particular parser by redefining in the subclass.

The following parsers are defined by this module.

A filter rather than a parser, removes whitespace and comments.


Tags definitions of "my", "our", and "local" variables.

Returns a Perl::Tags::Tag::Var if found


Parse a package declaration, returning a Perl::Tags::Tag::Package if found.


Parse the declaration of a subroutine, returning a Perl::Tags::Tag::Sub if found.


Parse a use constant directive


Parse a use, require, and also a use_ok line (from Test::More). Uses a dummy tag (Perl::Tags::Tag::Recurse to do so).


Parse label declaration


A superclass for tags

Returns a new tag object

"type", "modify_options"
Abstract methods

A tag stringifies to an appropriate line in a ctags file.

Allows tag to meddle with process when registered with the main tagger object. Return false if want to prevent registration (true normally).‘


"type": p

Sets static=0

Sets the package name


"type": v

        Make a tag for this variable unless we're told not to.  We
        assume that a variable is always static, unless it appears
        in a package before any sub.  (Not necessarily true, but
        it's ok for most purposes and Vim works fine even if it is
            − comments


"type": s

        Make a tag for this sub unless we're told not to.  We assume
        that a sub is static, unless it appears in a package.  (Not
        necessarily true, but it's ok for most purposes and Vim works
        fine even if it is incorrect)
            − pltags comments


"type": c


"type": l


"type": dummy

Recurse adding this new module to the queue.


Contributions are always welcome. The repo is in git:−tags

Please fork and make pull request. Maint bits available on request.

::PPI subclass

Ian Tegebo

patch to use File::Temp


patch to parse constant and label declarations


::Naive::Spiffy and ::Naive::Lib subclasses


prodding me to make repo public


::PPI fixes


Command line interface, applying patches


    osfameron (2006−2009) − osfameron AT cpan DOT org
                            and contributors, as above

For support, try emailing me or grabbing me on irc on

This was originally ripped off, as distributed with vim and available from <> Version 2.3, 28 February 2002 Written by Michael Schaap <pltags AT mscha DOT com>.

This is licensed under the same terms as Perl itself. (Or as Vim if you prefer).