Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

Awesome Lists | Featured Topics | Projects

https://github.com/kentnl/path-scaninc

Emulate Perls internal handling of @INC
https://github.com/kentnl/path-scaninc

perl

Last synced: about 1 month ago
JSON representation

Emulate Perls internal handling of @INC

Awesome Lists containing this project

README

        

# NAME

Path::ScanINC - Emulate Perls internal handling of @INC.

# VERSION

version 1.000004

# SYNOPSIS

The Aim of this module is to fully implement everything Perl does with `@INC`, to be feature compatible with it, including
the behavior with regard to `sub refs` in `@INC`.

use Path::ScanINC;

# Normal usage.
my $inc = Path::ScanINC->new( );

# In case you need something that isn't @INC
# but works like it

my $inc = Path::ScanINC->new( inc => \@INC );

# Freeze the value of @INC at the time of object instantiation
# with regard to behaviour so later changes to @INC have no effect

my $inc = Path::ScanINC->new( immutable => 1 );

# Return the first file in @INC that matches.

my $file = $inc->first_file('Path', 'ScanINC.pm' );

# Find all possible versions of modules in @INC
my ( @files ) = $inc->all_files('Path', 'ScanINC.pm');

# Try to discover a File::ShareDir 'module' root.
my $dir = $inc->first_dir('auto','share','module');

# Should return the same as File::ShareDir::module_dir('Path::ScanINC')
# ( assuming such a directory existed, which there is presently no plans of )
my $dir = $inc->first_dir('auto','share','module','Path-ScanINC');

# Find All File::ShareDir roots in @INC
my ( @dirs ) = $inc->all_dirs('auto', 'share');

# DESCRIPTION

`Path::ScanINC` is a basic tool for traversing `@INC` in a `perl`-like manner, stepping over some common
pitfalls with using it. It also has the property of being able to capture `@INC` states to emulate a portable
isolated library resolver.

# REF SUPPORT IN @INC

This module has elemental support for discovery of results in `@INC` using `CODE`/`ARRAY`/`BLESSED` entries in
`@INC`. However, due to a limitation as to how `perl` itself implements this functionality, the best we can do at present
is simply return what the above are expected to return. This means if you have any of the above ref-types in `@INC`,
and one of those returns `a true value`, you'll get handed back an `ARRAY` reference instead of the file you were
expecting.

Fortunately, `@INC` barely ever has refs in it. But in the event you _need_ to work with refs in `@INC` and you
expect that those refs will return `true`, you have to pick one of two options, either :

- a. Write your code to work with the `array-ref` returned by the respective reference on a match
- b. Use the `all_` family of methods and try pretending that there are no `array-refs` in the list it returns.

Its possible in a future release we may have better choices how to handle this situation in future, but don't bet on it.

Given that the API as defined by Perl mandates `code-ref`'s return lists containing `file-handles` or iterative
`code-ref`'s , not actual files, the best I can foresee at this time we'd be able to do to make life easier for you is
creating a fake library somewhere in a `tempdir` and stuffing the result of the `code-ref`'s into files in that directory
prior to returning a path to the generated file.

( And it also tells me that they have to be "Real" file handles, not tied or blessed ones, so being able to ask a
`filehandle` what file it represents is equally slim.... if that is of course what you require )

For more details, see [`perldoc perlfunc` or `perldoc -f require` ](https://metacpan.org/pod/perlfunc#require).

# METHODS

## new

my $object = $class->new(
inc => [ 'x', 'y', 'z' , ],
immutable => 1 | undef
);

## immutable

if( $inc->immutable ) {
print "We're working with a snapshotted version of @INC";
}

## inc

for my $i ( $inc->inc ) {
say "Plain: $incer" if not ref $incer;
say "Callback: $incer" if ref $incer;
}

Returns a copy of the internal version of `@INC` it will be using.

If the object is `immutable`, then this method will continue to report the same value as c<@INC>, or will be updated
every time the original array reference passed during construction gets updated:

my $ref = [];
my $a = Path::ScanINC->new( inc => $ref );
my $b = Path::ScanINC->new( inc => $ref, immutable => 1 );

push @{$ref} , 'a';

is( [ $a->inc ]->[0] , 'a' , "non-immutable references keep tracking their original" );
isnt( [ $b->inc ]->[0] , 'a' , "immutable references are shallow-copied at construction" );

Do note of course that is a **SHALLOW** copy, so if you have multiple `@INC` copies sharing the same `array`/`bless`
references, changes to those references will be shared amongst all `@INC`'s .

## first\_file

if( defined ( my $file = $inc->first_file('Moose.pm') ) {
print "Yep, Moose seems to be available in \@INC , its at $file, but its not loaded (yet)\n";
}

This proves to be a handy little gem that replaces the oft used

if( try { require Moose ; 1 } ){
Yadayadayada
}

And adds the benefit of not needing to actually source the file to see if it exists or not.

#### **IMPORTANT**: PORTABILITIY

For best system portability, where possible, its suggested you specify paths as arrays
of strings, not slash-separated strings.

$inc->first_file('MooseX' , 'Declare.pm') # Good
$inc->first_file('MooseX/Declare.pm') # Bad.

This is for several reasons, all of which can be summarized as "Windows".

- `%INC` keys all use Unix notation.
- `@INC` callbacks expect Unix notation.
- `\` is a valid path part on Unix.
- On Win32, we have to use `\` Separation, not `/` for resolving physical files.

The sum of these means if you do this:

$inc->first_file('MooseX/Declare.pm')

On win32, it might just end up doing:

C:\some\path\here/MooseX/Declare.pm

Which may or may not work.

And additionally, if the above module is loaded, it will be loaded as

"MooseX/Declare.pm"

in `%INC`, not what you'd expect, `MooseX\Declare.pm`

## all\_files

Returns all matches in all `@INC` paths.

my $inc = Path::ScanINC->new();
push @INC, 'lib';
my ( @files ) = $inc->all_files('Something','Im','Working','On.pm');
pp(\@files );

# [
# '/something/........./lib/Something/Im/Working/On.pm',
# '/something/....../share/per5/lib/site_perl/5.15.9/Something/Im/Working/On.pm',
# ]

Chances are if you understand how this can be useful, you'll do so immediately.

Useful for debugging what module is being loaded, and possibly introspecting information about
multiple parallel installs of modules in `%ENV`, such as frequently the case with 'dual-life' modules.

perl -MPath::ScanINC -E 'my $scanner = Path::ScanINC->new(); say for $scanner->all_files(qw( Scalar Util.pm ))'
/usr/lib64/perl5/vendor_perl/5.12.4/x86_64-linux/Scalar/Util.pm
/usr/lib64/perl5/5.12.4/x86_64-linux/Scalar/Util.pm

Sort-of like ye' olde' `perldoc -l`, but more like `man -a`

I might even be tempted to make a sub-module to make one-liners easier like

perl -MPath::ScanINC::All=Scalar/Util.pm

**REMINDER**: If there are `REFS` in `@INC` that match, they'll return `array-ref`'s, not strings.

## first\_dir

Just like `first_file` except for locating directories.

## all\_dirs

Just like `all_dirs` except for locating directories.

# AUTHOR

Kent Fredric

# COPYRIGHT AND LICENSE

This software is copyright (c) 2017 by Kent Fredric .

This is free software; you can redistribute it and/or modify it under
the same terms as the Perl 5 programming language system itself.