#!/usr/bin/env perl -w

#Perl Homework 3

#prompt the user for an RCS Id
print "Please enter the id of an RCS user\n";
chomp($rcs = <STDIN>);

($dir) = glob("~$rcs");
if ($dir eq "~$rcs"){ #glob() failed
    die "No such user $rcs!\n";
}

$dir .= "/public_html";

#open given directory and read files in it.
opendir DIR, $dir or die "Cannot open directory $dir: $!\n";
@files = grep /.htm(l?)$/i, readdir DIR;

#go through all the files in the directory
foreach $file (@files){
    
    #open the file if we can, print the name, and open a local copy
    if (!open FILE, "$dir/$file"){
	print "Could not open file $file: $!\n";
	next;
    }
    print "-" x 20, "\n";
    print "Links in $file:\n";
    print "-" x 20, "\n";
    open OFILE, ">$file" or die "Cannot create file in current directory: $!\n";
    
    #read the file into one big scalar
    {
	local $/;
	undef $/;
	$one_line = <FILE>;
    }
    close FILE;

    #find all <a href=...> tags
    @matches = (
		$one_line =~ m{
	        <a		    #the opening <a tag
		(.*?)		    #the link address
		                    #(really, everything in <a> tag -
			            #we'll get rid of other attributes later)
		>		    #ends at a >
		(.*?)		    #the link reference
		<\s*/a\s*>	    #closing </a> tag
		}xigs);		    #allow whitespace, case insensitive,...
				    #global, treat as single line

    #Go through each of the matches
    for ($i=0; $i<@matches; $i+=2){ #increment by 2 - only look at address
	#get rid of other <a> attributes
	#search for href= (with any spaces), then match and save all 
	#of the following non-whitespace characters
	if ($matches[$i] =~ /href\s*=\s*(\S+)/i){
	    $matches[$i] = $1;
	} else { #no href= found
	    next;
	}
	
	#get rid of quotes
	$matches[$i] =~ s/^["']|["']$//g;
	
	#change newlines to spaces, to make more readable
	$matches[$i] =~ s/\n/ /g;
	$matches[$i+1] =~ s/\n/ /g;
	
	#print link and reference
	print "\t$matches[$i] ==> $matches[$i+1]\n";		
    }
    
    #make substituions, print to local copy, close local copy
    #Below is the most compact way I could think of to do this - but it
    #won't quite work correctly, as it wouldn't replace the phrase
    #"Rennselaer University Polytechnic Institute", which should be replaced
    #following a strict interpretation of the Homework Description
    #$one_line =~ s/Rensselaer(\s+University)?(?!\s+Polytechnic\s+Institute)/RPI/gi;
    #so instead, we do this:
    $one_line =~ s/Rensselaer\s+University/RPI/gi;
    $one_line =~ s/Rensselaer(?!\s+Polytechnic\s+Institute)/RPI/gi;
    print OFILE $one_line;
    close OFILE;
    
}