replacing text file separators in regex

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

This relates to the question I asked yesterday at Net::FTP - syntax with username containing / - since it turns out it's a separate yet related issue than what I thought it was, here goes:

The sub-routine I'm still having a problem with is:

sub ParseURL
{
  my($url)= shift;
  my($protocol, $site, $login, $password);

  $url=~ /^(\w+):\/*([^\/]+)\/(.*)$/; # chunk the URL components
  $protocol= uc($1);        # preserve matches in real variables
  $site= $2;
  $path= $3;

  if ($site=~ /\@/)
  {                # check for the optional login field
    ($login, $site)= split("\@", $site);
    ($login, $password)=    # check for the optional password field
      split(":", $login);

    return($protocol, $site, $login, $password, $path);
  }
  else
  {                # no account information
    return($protocol, $site, "", "", $path);
  }
}
[download]

Basically my ftp username contains a / (XX/yyyy), and XX is getting parsed as the hostname of the ftp site.

As was suggested, I think the fastest way to fix this is to change the item separator in the text file to a comma or something, rather than a /.

The regex I'm having trouble with is:

$url=~ /^(\w+):\/*([^\/]+)\/(.*)$/;
[download]

The way I'm reading this is: between /^(\w+) is $protocol, between \/*([^\/]+) is $site, and \/(.*)$ is $path.

So, based on the above, if I use a comma separator for the $site portion, the line should look like:

$url=~ /^(\w+):\/*([^\/]+),(.*)$/;
[download]

and the appropriate line in my text file looks like: ftp://XX/yyyy:password@123.456.789.012,/ However, when I then run the script, I get several "uninitalized value" errors, one for the pattern match ($site=~ /\@/), one for a line trying to use $host, the other for a line trying to use $path.

So, it seems I'm missing something (hopefully) fairly obvious, but I can't seem to figure it out.

Any help appreciated.

Thanks,

Glenn

Comment on replacing text file separators in regex Select or Download Code

Replies are listed 'Best First'.
Re: replacing text file separators in regex by ehdonhon (Curate) on Jul 05, 2002 at 20:11 UTC
How about this? `$url=~ m\|^(\w+):/((.\@)?[^/]+)/(.)$\|; $protocol= uc($1); $site= $2; $path= $4;` [download] That would accept everything up till the last '@' as part of the site regardless of any '/' characters that it might contain. UPDATE:* Corrected the '@' that needed escaped	[reply] [d/l]
Re: replacing text file separators in regex by perlkid (Novice) on Jul 06, 2002 at 04:31 UTC
I would think, the following RE works with your case. #Given that a comma separator is used between site and path #$2 may or may not contain / or @ $url=~ /^(\w+):\/\/(.),(.)$/ ; I assumed that your data looks like one of the following cases ftp://XX/yyy:password@123.123.123.123,/ ftp://site,path ftp://site, For the "uninitailized value" warning use the defined fcn, for example: if (defined($site =~ /\@/)) { .. } else { .. } -perlkid	[reply]
Re: replacing text file separators in regex by fglock (Vicar) on Jul 05, 2002 at 20:16 UTC
`[^\/]` means "not a slash". You might try something like "find the at-sign, then the slash": `# untested $url=~ /^(\w+):\/(.?\@[^\/]+)\/(.*)$/;` [download]	[reply] [d/l] [select]
Re: Re: replacing text file separators in regex by ehdonhon (Curate) on Jul 05, 2002 at 21:19 UTC
The at-sign might not exist at all. If that is the case, your regexp will not match.	[reply]


No such thing as a small change
	PerlMonks