[pm-h] File::Find and system

[Lanny Ripple]

First to Russell --

perldoc is your friend.  `perldoc File::Find` will give you the 
documentation on File::Find and  has several examples which pretty much 
cover everything you are looking for.

The File::Find module, part of the standard perl install, should put 
`find2perl` in your path so you can execute the following

   find2perl . -type f -exec dos2unix

which will give you a perl script doing what you want and ready for 
changes to implement your search-and-replace processing.  Change "." to 
your target directory if you want to hard-code that.  If not then you'll 
need to be in your target directory when you run the script.


Mark --

You can golf your readdir down.  See `perldoc -f readdir`.

   Enjoy,
   -ljr

On 09/20/2011 05:16 PM, Mark Allen wrote:
> Find::File is definitely more cross platform and tested for edge cases, but something like this works pretty well too.
>
> == SNIP ==
>
> use strict;
> use warnings;
>
> my $base = $1 || $ENV{'PWD'};
>
> my @dirs = ( $base );
>
> foreach my $dir ( @dirs ) {
>        opendir DIR, $dir;
>        while ( my $file = readdir DIR ) {
>                if ( -d "$dir/$file" ) {
>                    next if ( $file eq "." or $file eq ".." );
>                    push @dirs, "$dir/$file";
>                    next;
>                }
>
>                # do stuff, like maybe
>                print "Now processing $dir/$file...";
>                `/usr/bin/dos2unix $dir/$file`;
>                print "done\n";
>        }
>        closedir DIR;
> }
>
> == SNIP END ==
>
> Mark
>
>
>
> ________________________________
> From: Russell L. Harris<rlharris at oplink.net>
> To: houston at pm.org
> Sent: Tuesday, September 20, 2011 4:42 PM
> Subject: [pm-h] File::Find and system
>
> Running Linux, I need to execute various system utilities (including
> "dos2unix") on multiple files in multiple directories.  The file
> structure is similar to the following:
>
>      foo/jan/01.txt
>      foo/jan/02.txt
>      foo/jan/03.txt
>      ...
>      foo/feb/01.txt
>      foo/feb/02.txt
>      foo/feb/03.txt
>      ...
>      foo/dec/01.txt
>      foo/dec/02.txt
>      foo/dec/03.txt
>      ...
>      bar/jan/01.txt
>      bar/jan/02.txt
>      bar/jan/03.txt
>      ...
>      bar/feb/01.txt
>      bar/feb/02.txt
>      bar/feb/03.txt
>      ...
>      bar/dec/01.txt
>      bar/dec/02.txt
>      bar/dec/03.txt
>      ...
>      foobar/jan/01.txt
>      foobar/jan/02.txt
>      foobar/jan/03.txt
>      ...
>      foobar/feb/01.txt
>      foobar/feb/02.txt
>      foobar/feb/03.txt
>      ...
>      foobar/dec/01.txt
>      foobar/dec/02.txt
>      foobar/dec/03.txt
>      ...
>
> Regrettably, not all Linux utilities have a recursive option, and I do
> not wish to take the time to re-write and debug functions which
> already are available as a standard utilities.
>
> A very tedious approach would be to "cd" to "foo/jan/" and run
> "dos2unix *.txt", then "cd" to "foo/feb/" and run "dos2unix *.txt",
> etc.
>
> I know that a Perl script can automate the process.  I just discovered
> the Perl "File::Find" module and the Perl "system" function, and now I
> am perusing the O'Reilly Perl books, trying to understand how to
> combine the two into a script.
>
> Afterward, I need to do involved search-and-replace processing on
> these files which cannot be handled with system utilities.  I
> previously have used Perl scripts for similar tasks, but never on a
> multi-level directory.
>
> So, learning how to run system utilities with "File::Find" appears to
> me to be the logical first step.
>
> RLH
> _______________________________________________
> Houston mailing list
> Houston at pm.org
> http://mail.pm.org/mailman/listinfo/houston
> Website: http://houston.pm.org/
>
>
>
> _______________________________________________
> Houston mailing list
> Houston at pm.org
> http://mail.pm.org/mailman/listinfo/houston
> Website: http://houston.pm.org/