http://qs1969.pair.com?node_id=58701


in reply to Multiline Regex

Asumming I've understood your question correctly, something like the following?
#!/usr/bin/perl -w use strict; my $data = join '', <DATA>; $data =~ s|(?=^\d{1,3}\.\d{1,3}\.\d{1,3}\.\d{1,3}\nOS Type: unknown$)| +\n|gm; print $data; exit(1); __DATA__ 10.1.1.1 bogus info 10.1.1.2 this could be anything 10.1.1.3 OS Type: unknown 10.1.1.4 filler information 10.1.1.5 OS Type: unknown

Replies are listed 'Best First'.
Re: Re: Multiline Regex
by InfiniteSilence (Curate) on Feb 16, 2001 at 01:18 UTC
    Although the right solution is to use multiline mode, your code doesn't do anything as the test on the regex reveals. See here:
    #!/usr/bin/perl -w use strict; my $data = join '', <DATA>; if($data =~ s|(?=^\d{1,3}\.\d{1,3}\.\d{1,3}\.\d{1,3}\nOS Type: unknown +$)|\n|gm) { print "yes, it worked!"; } #print $data; exit(1); __DATA__ 10.1.1.1 bogus info 10.1.1.2 this could be anything 10.1.1.3 OS Type: unknown 10.1.1.4 filler information 10.1.1.5 OS Type: unknown
    The above code returns nothing, meaning that the regex never does the substitution. From a quick rereading of the algorithm, it looks like he is trying to do the following:
    1. Read through the file
    2. find lines that start with IP addresses
    3. ASK if the next line following it has some specific text
    4. write an extra space BETWEEN the IP address line and the specific text line
    I may be wrong, as the spec. was pretty poorly written. A quick command line example that does this is as follows:
    perl -e "$q = qq(the brown dog\n is hailing a cab); if (($q=~m/\n(.*)/ +mg) && ($1 eq q( is hailing a cab))) { print qq(\n$1;)} else {print ' +no luck buddy.';}"

    Celebrate Intellectual Diversity

      Actually, the regex "mostly works".

      The problem with this regex is that the DATA does not have an immediate newline following it (there is a space). If the space is removed - or a simple (.+)? is added between the IP match and the newline, the first regex works fine. I'll leave someone else to do the benchmarks :-)

      That's of course, if the original poster wants the new line before the line with the IP
        Ack, sorry, nasty artefact of cut and paste :-/
        Well spotted and ++ to that man :-)