Re^2: trouble with regular expressions

[programmer.perl]

in reply to Re: trouble with regular expressions
in thread trouble with regular expressions

Hi 2teez,

my last solution was (anyway it puts a js comment "//" and next codes are invalid, but here were an unwanted bits in HD memory)

#!/usr/bin/perl -w
use 5.010;
use strict;
use warnings;

open (FINDIT, "find /home/auu/Documents/js -name '*.js' -type f -print
+ |") || die "Couldn't execute find!\n";

while (my $filename = <FINDIT>) {
open (TABLETKA, "+<$filename") || die "Can't open a file $filename: $!
+\n";
my @lines = <TABLETKA>;
foreach (@lines) {
print "$filename is cleaned\n" if s{(\;do.*)(</iframe>'\);)}{\; //eof}
+;
}
seek(TABLETKA,0,0) || die;
print TABLETKA @lines;
close (TABLETKA);
}

this is (above) is my final code, now I'm thinking how to integrate your code to my code... do you have any idea? ))...

[Lotus1]

When you overwrite the original file with less text it doesn't clear out the original text, that is where the extra stuff after //eof comes from. Try writing to a new file to see this for yourself.

[programmer.perl]

Yes, you are right, I'm happy that we found the bug... Now, but I have to save to file itself, I can't (musn't) write to another file...

[ysth]

You can't do what you are trying to do with the file...unless you write out exactly the same number of characters as were there before. Usually instead you open a new file, write out all the lines (whether changed or unchanged) to it, then rename it to replace the existing file. For a quick script operating on only small files, you could use Tie::File to automate most of this for you.

[Lotus1]

It isn't a safe thing to do but you could close the file after reading it into your array. Then re-open it for output open $fh, ">", "file.txt" or similar which will clobber (clear the contents) your file then write to it. The danger is that a minor mistake in your program could wipe out or mangle the original file. You could write to a test file until you finish debugging.