[Za-pm] getting out of <STDIN> loop

[Anne Wainwright]

Oskar, lotsa thanks, 

comment below at end.

On Mon, 10 Aug 2009 02:44:54 +0100
Oskar Pearson <oskar at deckle.co.za> wrote:

> Hi
> 
> 
> > Hi, all,
> >
> > (Am in fact referring to Exercise 2 Chapter 2 of
> > Schwartz et al INTERMEDIATE PERL.)
> >
> > First, while realising that the official solution is a masterpiece
> > of condensed perl code, I don't understand why the first
> > line has the (1) ).
> >
> > -----------------------------------------------------------
> > while(1) {
> >    print "Enter a regular expression to match filenames> ";
> >    chomp(my $regex = <STDIN>);
> >    last unless (defined $regex && length $regex);
> >    print map {"    $_\n"} grep {eval{/$regex/}} glob(".* *");
> > }
> > -----------------------------------------------------------
> 
> The while loop continues indefinitely - the value is always true, so  
> it'll loop until the "last" gets hit.
> 
> There are two reasons that happens - the "end of file" (pressing  
> control-D will do that), or if the length of the input is 0 (when  
> someone inputs a blank line - assuming it's "chomped"). That's what  
> this line means:
> >    last unless (defined $regex && length $regex);
> 
> 
> 
> 
> > My own more primitive solution fails to get out of the
> > input loop, albeit with simpler criteria. My code is:
> >
> > -------------------------------------------------------
> > opendir THISDIR, "." or die "serious braindamage: $!";
> > @allfiles = readdir THISDIR;
> > closedir THISDIR;
> >
> > while (defined($regexpr = <STDIN>)) {		# 1
> >        chomp $regexpr;
> >        foreach $file(@allfiles) {        	# 2
> > 	        if ($file =~ m/$regexpr/) {     # 3
> > 		    print "$file\n";
> > 	        }                               # -3
> >        }                                   	# -2
> > }						# -1
> > -------------------------------------------------------
> > Essentially on no regex input it outputs all the files as if ".+"
> > had been input.
> 
> First hint is that you should use "perl -w" and "use strict;" -
> you'll need to declare your variables though.
> 
> > I thought the line marked #1 was a dead ringer for getting out of
> > the input loop. (p.72 Chapter 6 of Schwartz et al LEARNING PERL).
> > If I take
> > "&& length $regex" out of their code then that also fails to exit. I
> > don't understand  :(
> 
> Pseudocode for their version:
> 
> Loop forever (unless we hit the "last below, of course") {
> 	print something
> 	read something from stderr
> 	Unless I got a valid input value, and that input value has a
> valid length, then jump out of this loop
> 	Do the regex stuff
> }
> 
> 
> 
> 
> 
> > opendir THISDIR, "." or die "serious braindamage: $!";
> > @allfiles = readdir THISDIR;
> > closedir THISDIR;
> >
> > while (defined($regexpr = <STDIN>)) {		# 1
> >        chomp $regexpr;
> >        foreach $file(@allfiles) {        	# 2
> > 	        if ($file =~ m/$regexpr/) {     # 3
> > 		    print "$file\n";
> > 	        }                               # -3
> >        }                                   	# -2
> > }						# -1
> 
> 
> 
> I'd prefer to change this as follows. I'm using slightly different  
> structures here.
> 
> #!/usr/bin/perl -w
> use strict;
> opendir THISDIR, "." or die "serious braindamage: $!";
> my @allfiles = readdir THISDIR;
> closedir THISDIR;
> 
> while (my $regexpr = <STDIN>) { 			# If this is
> undefined here, the "while" is a while of an undef value, and the
> while will terminate chomp $regexpr;
>         last unless length($regexpr); 			#If I
> got a blank line, jump out
>         foreach my $file (sort(@allfiles)) {		# Sort
> for consistency if ($file =~ m/$regexpr/i) {
> # guessing you probably want /i heere

all my filenames are in small print  ;)
>              print "$file\n";
>            }
>         }
> }
> 
> Hope that helps!

Yes, bigly! Very clear and exactly how I would have written it if I
could have  ;)

Along with the analysis of the why things did what they did from
Francois I can now move on.  

Anne
> 
> Oskar