PPIx::DocumentName - Utility to extract a name from a PPI Document
version 1.01
New API:
use PPIx::DocumentName 1.00 -api => 1; my $result = PPIx::DocumentName->extract( $ppi_document ); # say the "name" of the document say $result->name; # the result object can also be stringified into the name found: say "$result"; # the line number, column, filename etc. where the name was found my $location = $result->node->location;
Old API:
use PPIx::DocumentName; # assumes -api => 0 my $name = PPIx::DocumentName->extract( $ppi_document ); # say the "name" of the document say $name;
This module contains a few utilities for extracting a "name" out of an arbitrary Perl file.
Typically, this is the module name, in the form:
package Foo
However, it also supports extraction of an override statement in the form:
# PODNAME: OverrideName::Goes::Here
Which may be more applicable for documents that lack a package statement, or the package statement may be "wrong", but they still need the document parsed under the guise of having a name ( for purposes such as POD )
my $result = PPIx::Document->extract( $ppi_document);
This will first attempt to extract a name via the PODNAME: comment notation, and then fall back to using a package Package::Name statement.
$ppi_document is ideally a PPI::Document, but will be auto-up-cast if it is any of the parameters PPI::Document->new() understands.
The $result is the found name as a string under -api => 0 and a PPIx::DocumentName::Result object under -api => 1. If the name is not found, then it will be undef (with either API). Note that PPIx::DocumentName::Result is stringified to the found name, so in many circumstances the new API can be used in the same way as the old.
my $result = PPIx::DocumentName->extract_via_statement( $ppi_document );
This only extract package Package::Name statement based document names.
$ppi_document is ideally a PPI::Document, but will be auto-up-cast if it is any of the parameters PPI::Document->new() understands.
The $result is the found name as a string under -api => 0 and a PPIx::DocumentName::Result object under -api => 1. If the name is not found, then it will be undef (with either API).
my $result = PPIx::DocumentName->extract_via_comment( $ppi_document );
This will only extract PODNAME: comment based document names.
$ppi_document is ideally a PPI::Document, but will be auto-up-cast if it is any of the parameters PPI::Document->new() understands.
The $result is the found name as a string under -api => 0 and a PPIx::DocumentName::Result object under -api => 1. If the name is not found, then it will be undef (with either API).
The newer API (-api => 1) is packaged scoped in Perl 5.6 and 5.8. In newer Perls the API is block scoped as it should be. Because this can cause bugs if you are using an older version of Perl this module will complain loudly if you are using an older Perl with the newer API. If you don't like the warning, then either use the old API or upgrade to Perl 5.10+.
Under the older API (-api => 0; the default), extract_via_statement, unlike the other methods in this module, returns empty list instead of undef when it does find a name. When using the newer API (-api => 1), calls are consistent in scalar and list context. New code should therefore use the newer API.
Other things I could have called this
PPIx::PodName - But it isn't, because it doesn't extract from POD, only returns data that may be useful FOR PODPPIx::ModuleName - But it kinda isn't either, because its more generic than that and is tailored to extracting "a name" out of any PPI Document, and they're NOT all modules.Modules that are perceptibly similar to this ones tasks ( but are subtly different in important ways ) are as follows:
Module::Metadata - Module::Metadata does a bunch of things this module explicitly doesn't want or need to do, and it lacks a bunch of features this module needs.
Module::Metadata is predominantly concerned with extracting ALL name spaces and ALL versions from a module for the purposes of indexing and indexing related tasks. This also means it has a notion of "hideable" name spaces with the purpose of hiding them from CPAN.
Due to being core as well, it is not able to use PPI for its features, so the above concerns mean it is also mostly based on careful regex parsing, which can easily be false tripped on miscellaneous in document content.
Whereas PPIx::DocumentName only cares about the first name of a given class, and it cares much more about nested strings being ignored intentionally. It also has a motive to show names even for documents that won't be indexed ( And Module::Metadata has no short term plans on exposing hidden document names ).
PPIx::DocumentName also has special logic for the PODNAME: declaration, and may eventually support other mechanisms for extracting a name from "a document", which will be not in Module::Metadata's collection of desired use-cases.
Module::Extract::Namespaces - This is probably closer to PPIx::DocumentName's requirements, using PPI to extract content.
Most of Module::Extract::Namespaces's code seems to be glue for legacy versions of PPI and the remaining code is for loading modules from @INC ( Which we don't need ), or special casing IO ( Which is also not necessary, as this module assumes you're moderately acquainted with PPI and can do IO yourself )
Module::Extract::Namespaces also obliterates document comments, which of course stands in the way of our auxiliary requirements re PODNAME: declarations.
It will also not be flexible enough to support other name extraction features we may eventually add.
And like Module::Metadata, it also focuses on extracting many package declarations where this module prefers to extract only the first.
PPIx::DocumentName::Result - comes with this module, and contains the results of this module, when using the newer -api => 1 API.The bulk of this logic was extrapolated from Pod::Weaver::Section::Name and a related role, Pod::Weaver::Role::StringFromComment.
Thanks to RJBS for the initial implementation and DROLSKY for some of the improvement patches.
This software is copyright (c) 2015-2021 by Kent Fredric <kentfredric@gmail.com>.
This is free software; you can redistribute it and/or modify it under the same terms as the Perl 5 programming language system itself.