Regular expression parse pathname

unixguy has asked for the wisdom of the Perl Monks concerning the following question:

Replies are listed 'Best First'.
Re: Regular expression parse pathname by toolic (Bishop) on Jan 29, 2010 at 01:13 UTC
The key is to accumulate your string as you loop through the directories. Not a one-liner, but this split solution gets the job done... `use strict; use warnings; my $s = '/home/code/work'; my @ds = split m{/}, $s; shift @ds; my $accum; for (@ds) { $accum .= "/$_"; print "$accum "; } print "\n";` [download] Update: or, adapting your regex attempt: `my $s = '/home/code/work'; my $acc; while ($s =~ m#([^/]+)#g) { $acc .= "/$1"; print "$acc " }` [download]	[reply] [d/l] [select]
Re: Regular expression parse pathname by jwkrahn (Abbot) on Jan 29, 2010 at 02:49 UTC
`$ perl -le' use File::Basename; my $name = "/home/code/work"; print join " ", reverse map { my $d = dirname my $x = $name; $name = $ +d; $x } 1 .. $name =~ y!/!!; ' /home /home/code /home/code/work` [download]	[reply] [d/l]
Re: Regular expression parse pathname by rubasov (Friar) on Jan 29, 2010 at 02:38 UTC
Try this: `echo "/home/code/work" \|\ perl -nle 'chomp; print $p .= $& while m{ \G (?: [^/]* /+ \| [^/]+ \z ) + }xg'` [download] The `m/\G.../g` match can be called several times, and each time it will start at the position where last time it left off. However be aware, that this one-liner produces a little different tree than your example, but probably you want to represent the real file system tree and this result is a better representation of it.	[reply] [d/l] [select]
Re: Regular expression parse pathname by aquarium (Curate) on Jan 29, 2010 at 02:47 UTC
Just a little comment about the original regex. There's no need to escape special characters inside a character class match = character matches inside square brackets. Even no need to escape a forward slash inside square brackets, as regex engines know they're inside a character class and won't confuse with a regex delimiter (normally forward slash.) So if the regex was originally intended for unix only paths, no need for the backslash..unless you also want to catch dos style path delimiter. the hardest line to type correctly is: stty erase ^H	[reply]
Re: Regular expression parse pathname by JavaFan (Canon) on Jan 29, 2010 at 10:47 UTC
`use 5.010; my $_ = "/home/code/work"; my @paths; again: unshift @paths, $_; s!/[^/]*$!! and length $_ and goto again; say "@paths"; __END__ /home /home/code /home/code/work` [download]	[reply] [d/l]
Re: Regular expression parse pathname by shawnhcorey (Friar) on Jan 29, 2010 at 14:42 UTC
You should ``use File::Spec;`` when dealing with paths. `File::Spec` is a standard module and should have been installed with Perl. For a complete list of all the standard modules and pragmatics, see ``perldoc perlmodlib``.	[reply] [d/l] [select]
Re: Regular expression parse pathname by hbm (Hermit) on Jan 29, 2010 at 19:46 UTC
This is as tight as I can currently squeeze it: echo "/home/code/work"\|\ perl -pe's\|(/[^/]+)\|$`$1 \|g' [download] Update: My previous, longer solutions: perl -pe's\|(/[^/]+)\|$`$1$"\|g' ### $" - DUH! perl -l40p057e'$_=$;.$_;$;="$_/"' perl -ne'print$1,$"while m\|($&/[^/]+)\|' perl -e'$_=<>;print$1,$"while m\|($&/[^/]+)\|' perl -e'@_=<>=~m\|(/[^/]+)\|g;print@_[0..$_],$"for(0..$#_)' [download]	[reply] [d/l] [select]