Module::Info - Information about Perl modules

  use Module::Info;

  my $mod = Module::Info->new_from_file('Some/Module.pm');
  my $mod = Module::Info->new_from_module('Some::Module');
  my $mod = Module::Info->new_from_loaded('Some::Module');

  my @mods = Module::Info->all_installed('Some::Module');

  my $name    = $mod->name;
  my $version = $mod->version;
  my $dir     = $mod->inc_dir;
  my $file    = $mod->file;
  my $is_core = $mod->is_core;

  # Only available in perl 5.6.1 and up.
  # These do compile the module.
  my @packages = $mod->packages_inside;
  my @used     = $mod->modules_used;
  my @subs     = $mod->subroutines;
  my @isa      = $mod->superclasses;
  my @calls    = $mod->subroutines_called;

  # Check for constructs which make perl hard to predict.
  my @methods   = $mod->dynamic_method_calls;
  my @lines     = $mod->eval_string;    *UNIMPLEMENTED*
  my @lines     = $mod->gotos;          *UNIMPLEMENTED*
  my @controls  = $mod->exit_via_loop_control;      *UNIMPLEMENTED*
  my @unpredictables = $mod->has_unpredictables;    *UNIMPLEMENTED*

  # set/get Module::Info options
  $self->die_on_compilation_error(1);
  my $die_on_error = $mod->die_on_compilation_error;
  $self->safe(1);
  my $safe = $mod->safe;

Module::Info gives you information about Perl modules without actually loading the module. It actually isn't specific to modules and should work on any perl code.

There are a few ways to specify which module you want information for. They all return Module::Info objects.

new_from_file

  my $module = Module::Info->new_from_file('path/to/Some/Module.pm');

Given a file, it will interpret this as the module you want information about. You can also hand it a perl script. If the file doesn't exist or isn't readable it will return false.

new_from_module

  my $module = Module::Info->new_from_module('Some::Module');
  my $module = Module::Info->new_from_module('Some::Module', @INC);

Given a module name, CW@INC will be searched and the first module found used. This is the same module that would be loaded if you just say CWuse Some::Module. If you give your own CW@INC, that will be used to search instead.

new_from_loaded

  my $module = Module::Info->new_from_loaded('Some::Module');

Gets information about the currently loaded version of Some::Module. If it isn't loaded, returns false.

all_installed

  my @modules = Module::Info->all_installed('Some::Module');
  my @modules = Module::Info->all_installed('Some::Module', @INC);

Like new_from_module(), except all modules in CW@INC will be returned, in the order they are found. Thus CW$modules[0] is the one that would be loaded by CWuse Some::Module.

The following methods get their information without actually compiling the module.

name

  my $name = $module->name;
  $module->name($name);

Name of the module (ie. Some::Module). Module loaded using new_from_file() won't have this information in which case you can set it yourself.

version

  my $version = $module->version;

Divines the value of CW$VERSION. This uses the same method as ExtUtils::MakeMaker and all caveats therein apply.

inc_dir

  my $dir = $module->inc_dir;

Include directory in which this module was found. Module::Info objects created with new_from_file() won't have this info.

file

  my $file = $module->file;

The absolute path to this module.

is_core

  my $is_core = $module->is_core;

Checks if this module is the one distributed with Perl. NOTE This goes by what directory it's in. It's possible that the module has been altered or upgraded from CPAN since the original Perl installation.

WARNING! From here down reliability drops rapidly!

The following methods get their information by compiling the module and examining the opcode tree. The module will be compiled in a seperate process so as not to disturb the current program.

They will only work on 5.6.1 and up and requires the B::Utils module.

packages_inside

  my @packages = $module->packages_inside;

Looks for any explicit CWpackage declarations inside the module and returns a list. Useful for finding hidden classes and functionality (like Tie::StdHandle inside Tie::Handle). KNOWN BUG Currently doesn't spot package changes inside subroutines.

package_versions

  my %versions = $module->package_versions;

Returns a hash whose keys are the packages contained in the module (these are the same as what's returned by CWpackages_inside()), and whose values are the versions of those packages.

modules_used

  my @used = $module->modules_used;

Returns a list of all modules and files which may be CWuse'd or CWrequire'd by this module. NOTE These modules may be conditionally loaded, can't tell. Also can't find modules which might be used inside an CWeval.

modules_required

  my %required = $module->modules_required;

Returns a list of all modules and files which may be CWuse'd or CWrequire'd by this module, together with the minimum required version. The hash is keyed on the module/file name, the corrisponding value is an array reference containing the requied versions, or an empty array if no specific version was required. NOTE These modules may be conditionally loaded, can't tell. Also can't find modules which might be used inside an CWeval.

subroutines

  my %subs = $module->subroutines;

Returns a hash of all subroutines defined inside this module and some info about it. The key is the *full* name of the subroutine (ie. CW$subs{'Some::Module::foo'} rather than just CW$subs{'foo'}), value is a hash ref with information about the subroutine like so:

    start   => line number of the first statement in the subroutine
    end     => line number of the last statement in the subroutine

Note that the line numbers may not be entirely accurate and will change as perl's backend compiler improves. They typically correspond to the first and last run-time statements in a subroutine. For example:

    sub foo {
        package Wibble;
        $foo = "bar";
        return $foo;
    }

Taking CWsub foo { as line 1, Module::Info will report line 3 as the start and line 4 as the end. CWpackage Wibble; is a compile-time statement. Again, this will change as perl changes. Note this only catches simple CWsub foo {...} subroutine declarations. Anonymous, autoloaded or eval'd subroutines are not listed.

superclasses

  my @isa = $module->superclasses;

Returns the value of CW@ISA for this CW$module. Requires that CW$module->name be set to work. NOTE superclasses() is currently cheating. See CAVEATS below.

subroutines_called

  my @calls = $module->subroutines_called;

Finds all the methods and functions which are called inside the CW$module. Returns a list of hashes. Each hash represents a single function or method call and has the keys:

    line        line number where this call originated
    class       class called on if its a class method
    type        function, symbolic function, object method, 
                class method, dynamic object method or 
                dynamic class method.
                (NOTE  This format will probably change)
    name        name of the function/method called if not dynamic

Unpredictable constructs are things that make a Perl program hard to predict what its going to do without actually running it. There's nothing wrong with these constructs, but its nice to know where they are when maintaining a piece of code.

dynamic_method_calls

  my @methods = $module->dynamic_method_calls;

Returns a list of dynamic method calls (ie. CW$obj-$method()>) used by the CW$module. CW@methods has the same format as the return value of subroutines_called().

The following methods get/set specific option values for the Module::Info object.

die_on_compilation_error

  $module->die_on_compilation_error(0); # default
  $module->die_on_compilation_error(1);
  my $flag = $module->die_on_compilation_error;

Sets/gets the die on compilation error flag. When the flag is off (default), and a module fails to compile, Module::Info simply emits a watning and continues. When the flag is on and a module fails to compile, Module::Info CWdie()s with the same error message it would use in the warning.

safe

  $module->safe(0); # default
  $module->safe(1); # be safer
  my $flag = $module->safe;

Sets/gets the safe flag. When the flag is enabled all operations requiring module compilation are forbidden and the CWversion() method executes its code in a CWSafe compartment.

Michael G Schwern <schwern@pobox.com> with code from ExtUtils::MM_Unix, Module::InstalledVersion and lots of cargo-culting from B::Deparse.

Mattia Barbon <mbarbon@cpan.org> is the current maintainer.

This program is free software; you can redistribute it and/or modify it under the same terms as Perl itself.

Many thanks to Simon Cozens and Robin Houston for letting me chew their ears about B.

Code refs in CW@INC are currently ignored. If this bothers you submit a patch.

superclasses() is cheating and just loading the module in a seperate process and looking at CW@ISA. I don't think its worth the trouble to go through and parse the opcode tree as it still requires loading the module and running all the BEGIN blocks. Patches welcome.

I originally was going to call superclasses() isa() but then I remembered that would be bad.

All the methods that require loading are really inefficient as they're not caching anything. I'll worry about efficiency later.