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

Hi monkers

have a small problem which looks like
http://www.perlmonks.com/index.pl?node_id=366431
i want to make a replacement inline, so a replace in replace so here is the code 

$content =~ s/(\w+)(: *)([^\n]*)\n<xxxxyyxxxx>\n/$1.":".$2.encode_base
+64($3,'')."\n"/ges;

[download]

for now it doesn't work if $3 contains multiples lines (aka '\n')

first case, rewrite the rule to make replacement on multiple lines. In general i have few success with this.

seconde case, before this code, i decode the code. so there i can replace '\n' with '%%NEWLINE%%' for example and in the above line reverse.
s/%%NEWLINE%%/\n/ is ok, but how can i make it in the previous line ?


thanks
note: code use with perl/redhat enterprise or activeperl 5.8/winxp

Replies are listed 'Best First'.
Re: inline replacement in existing regex
by tall_man (Parson) on Nov 30, 2004 at 20:59 UTC
    Your question is very cryptic. Please include a bit more detail, such as some example input and desired output.
[reply]
Re: inline replacement in existing regex
by duff (Parson) on Nov 30, 2004 at 21:28 UTC

    It sounds like you're asking how to substitute newlines in the middle of that substitution. The reason you're asking is that

    it doesn't work if $3 contains multiples lines
    I think this premise is wrong. What do you mean that it "doesn't work"? I just tried a small test case (assuming that the encode_base64 in your code is the one from MIME::Base64) and it works just fine.

    duff

[reply]
[d/l]
      if using
      $content =~ s/(\w+)(: *)(.*)\n<xxxxyyxxxx>\n/$1.":".$2.encode_base64($
+3,'')."\n"/ges;

      [download]
      it works but it takes the biggest text segment, not the smallest one which i want.
      to answer to tall_man, i'm trying to clean the ldif output from mozilla to get standard ldap to include in a directory.
      so i get a file with multiple 
      cn: dn=toto
sn: XXX
description: YYY

      [download]

      some of them base64 encoded, i have to decode them, for cn, add a good suffix, remove others, reformat ones, ...
      some of theses base64 fields contains multiples lines and at the end of the conversion (reencoding base64) my file has:
      cn: dn=toto
sn: XXX
description: YYY
yyyy
<xxxxyyxxxx>

cn: dn=toto2
sn: XXX2
description: YYY2

cn: dn=toto3
sn: XXX3
description: YYY3
tttt
uuuuuu
<xxxxyyxxxx>

      [download]

      $content =~ s/(\w+)(: *)(.*)\n<xxxxyyxxxx>\n/$1.":".$2.encode_base64($
+3,'')."\n"/ges;

      [download]

      => encode the largest matching
      $content =~ s/(\w+)(: *)([^\n]*)\n<xxxxyyxxxx>\n/$1.":".$2.encode_base
+64($3,'')."\n"/ges;

      [download]

      => doesn't match field with more than 1 lines. i hope have been clearer.

      thanks
[reply]
[d/l]
[select]

        It sounds like you just want to modify the greediness of *: 

        $content =~ s/(\w+)(: *)(.*?)\n<xxxxyyxxxx>\n/$1.":".$2.encode_base64(
+$3,'')."\n"/ges;

        [download]
        The ? following the * will cause it to be non-greedy and thus match at the earliest possible point in the string rather than the latest.

        duff

[reply]
[d/l]
[select]

Back to Seekers of Perl Wisdom