NAME

File::Path::Tiny - recursive versions of mkdir() and rmdir() without as much overhead as File::Path

VERSION

This document describes File::Path::Tiny version 1.0

SYNOPSIS

    use File::Path::Tiny;

    if(!File::Path::Tiny::mk($path)) {
        die "Could not make path '$path': $!";
    }
    
    if(!File::Path::Tiny::rm($path)) {
        die "Could not remove path '$path': $!";
    }

DESCRIPTION

The goal here is simply to provide recursive versions of mkdir() and rmdir() with as little code and overhead as possible.

This module is in no way meant to derogate File::Path and is in no way an endorsement to go out and replace all use of File::Path with File::Path::Tiny.

File::Path is very good at what it does but there's simply a lot happening that we can do without much of the time.

Here are some things File::Path has/does that this module attempts to do without:

multiple interfaces
Backwards compatibility brings in a lot of code and logic that we don't need from here on out.
chdir()s
It forces a ton of chdir()s which could leave you somewhere you're not planning on being and requires much more overhead to do.

This module provides a way to disable that if you know it is safe to do so in your circumstance.
can croak not allowing you to detect and handle failure
Just let me handle errors how I want. Don't make my entire app die or have to wrap it in an eval

The exception here is the security checks can croak, which is what you want. See DIAGNOSTICS for more info.
A well intentioned output system
Just let me do the output how I want. (Nothing, As HTML, print to a filehandle, etc...)
A well intentioned and experimental (IE subject to change) error handling system.
Just keep it simple and detect failure via a boolean check and do what I want with the error. See "How can I make/remove multiple paths?"
According to its POD, removing a tree is apparently not safe unless you tell it to be with the ‘safe’ or 'keep_root' attributes.
Seems like that should just happen, I don't want to worry about accidentally removing / when I pass it /tmp

INTERFACE

Nothing in exported or exportable, that helps keep it '::Tiny'.

File::Path::Tiny::mk()

Takes a single path (like mkdir()) to recursively create and, optionally, a mask (like mkdir()) for all subsequent mkdir() calls.

Mask defaults to 0700 (also like mkdir())

Returns false if it could not be made, true otherwise (returns ‘2’ if it is -d already)

It is recursive in the sense that given “foo/bar/baz” as the path to create all 3 will be created if necessary.

File::Path::Tiny::rm()

Takes a single path (like rmdir()) to recursively empty and remove.

Returns false if it could not be emptied or removed, true otherwise (returns ‘2’ if it is !-d already). Also see DIAGNOSTICS.

It is recursive in the sense that given “/foo/bar/baz” as the path to remove it will recursively empty ‘baz’ and then remove it from /foo/bar.

Its parents, /, /foo, and /foo/bar are *not* touched.

By default it does chdir() for security reasons. If you are sure it is safe to do without that for the sake of a bit of speed you can pass a second true argument to skip that.

File::Path::Tiny::empty_dir()

Takes a single path to recursively empty but not remove.

Returns false when there is a problem. Also see DIAGNOSTICS.

By default it does chdir() for security reasons. If you are sure it is safe to do without that for the sake of a bit of speed you can pass a second true argument to skip that.

File::Path::Tiny::mk_parent()

Like mk() but recursively creates the parent. e.g. given “foo/bar/baz.txt” creates foo/bar.

From Cwd

It uses these internally so, for convenience, these are exposed in case you want to use them also.

File::Path::Tiny::cwd()

Comes directly from Cwd’s cwd().

File::Path::Tiny::chdir()

Comes directly from Cwd’s chdir().

DIAGNOSTICS

If the functions ever return false, $! will have been set either explicitly or by the mkdir(), rmdir(), unlink(), or opendir() that ultimately returned false.

"directory %s changed: expected dev=%d ino=$d, actual dev=%d ino=%d, aborting": empty_dir() and rm() throw this if any of the directories being operated on change during the operation.

MISC

How can I make/remove multiple paths?

For simplicity sake && ::Tiny status this module must be very very very simple.

That said it is also very simple to do multiple paths:

    for my $path (@paths) {
        File::Path::Tiny::::mk($path) or _my_handle_failed_mk($path, $!);    
    }
    
    for my $path (@paths) {
        File::Path::Tiny::::rm($path) or _my_handle_failed_rm($path, $!);    
    }

That also lets you keep going or short circuit it or handle errors however you like:

     PATH:
     for my $path qw(foo/bar bar/rat) {
         if (!File::Path::Tiny::mk($path)) {
             print "problem unlinking '$path': $!\n";
             last PATH;
         }
         else {
             print "No error encountered with '$path'\n" 
         }
     }

About the '::Tiny' status

See Acme::Tiny for information on the ::Tiny suffix.

  #3 is almost there (< 1/5th +/-), a bit more trimming and I think we'll have it!
  #8 is N/A since part of the "sub set" is to do single paths like their non-recursive core counterparts and there are so many ways to invoke it with different consequences 

   [ -- RSS Memory -- ]
    Baseline perl 1168
    File::Path 1808 (+640)
    File::Path::Tiny 1288 (+120)

Even though "time" isn't really a ::Tiny factor, it does improve loading a bit:

   [ -- time -- ]
    Baseline perl
      real      0m0.007s
      user      0m0.002s
      sys       0m0.004s
    
    File::Path
      real      0m0.017s
      user      0m0.011s
      sys       0m0.005s

    File::Path::Tiny
      real      0m0.007s
      user      0m0.003s
      sys       0m0.004s

As time allows and more tests are added I'll try to include more comprehensive benchmark results.

How do I make sure the path is safe to create or remove?

Of course the answer depends on what you mean by "safe".

This module makes no assumptions on interpreting the "safeness" of a path, just like mkdir() and rmdir().

Also like mkdir() and rmdir() typically you'll find that filesystem permissions are a pretty reliable tool (of course if the code will be run as root you would want to setuid first...)

You might use Cwd::abs_path() to sanitize a path before sending it to be made or removed.

Even after that it might not be "safe" so you'll need to discern what your particular definition of "safe" is and take appropriate action.

DEPENDENCIES

File::Spec of course but its only loaded if needed

INCOMPATIBILITIES

None reported.

BUGS AND FEATURES

Please report any bugs or feature requests (and a pull request for bonus points) through the issue tracker at <https://github.com/drmuey/p5-File-Path-Tiny/issues>.

AUTHOR

Daniel Muey "<http://drmuey.com/cpan_contact.pl>"

LICENCE AND COPYRIGHT

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

DISCLAIMER OF WARRANTY

BECAUSE THIS SOFTWARE IS LICENSED FREE OF CHARGE, THERE IS NO WARRANTY FOR THE SOFTWARE, TO THE EXTENT PERMITTED BY APPLICABLE LAW. EXCEPT WHEN OTHERWISE STATED IN WRITING THE COPYRIGHT HOLDERS AND/OR OTHER PARTIES PROVIDE THE SOFTWARE "AS IS" WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESSED OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. THE ENTIRE RISK AS TO THE QUALITY AND PERFORMANCE OF THE SOFTWARE IS WITH YOU. SHOULD THE SOFTWARE PROVE DEFECTIVE, YOU ASSUME THE COST OF ALL NECESSARY SERVICING, REPAIR, OR CORRECTION.

IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MAY MODIFY AND/OR REDISTRIBUTE THE SOFTWARE AS PERMITTED BY THE ABOVE LICENCE, BE LIABLE TO YOU FOR DAMAGES, INCLUDING ANY GENERAL, SPECIAL, INCIDENTAL, OR CONSEQUENTIAL DAMAGES ARISING OUT OF THE USE OR INABILITY TO USE THE SOFTWARE (INCLUDING BUT NOT LIMITED TO LOSS OF DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY YOU OR THIRD PARTIES OR A FAILURE OF THE SOFTWARE TO OPERATE WITH ANY OTHER SOFTWARE), EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES.