It seems to work as expected and, as far as I can tell, as you wish for me:

use strict;

my $text = <<TEXT;
<p>Before new line.
After new line.</p>
TEXT

print "Before: |$text|\n";

$text=~s#<p>(.+?)<\/p>#"<p>".&enter($1)."<\/p>"#gsie;

print "After: |$text|\n";


sub enter
{
my $new;

($new)=$1;
$new=~s/\n//gsi;
return $new;
}
[download]

Prints:

Before: |<p>Before new line.
After new line.</p>
|
After: |<p>Before new line.After new line.</p>
|
[download]

If that is not what you want then perhaps you should provide your own sample and expected output?

That aside, there are a few things worth mentioning:

• it is generally considered bad practice to call subs using & these days
• you pass a parameter to the sub, but don't use it (see below)
• you use $1 as a "global" to the sub which is bad in itself, but also there is no need! (see below)
• hand parsing HTML is a bad, mad and dangerous thing to do. Use HTML::TreeBuilder or one of the other modules designed for the purpose
• (minor) you use # to delimit the parts of the substitution, but quote the / - / is not special if not the delimiter so doesn't need to be quoted

The sub would be better written:

sub enter
{
my $new = shift;

$new=~s/\n\s*/ /gi;
return $new;
}
[download]

Note the alteration to the regex so that a space is inserted in place of a new line, but that leading white space on the line is removed:

After: |<p>Before new line. After new line.</p>
|
[download]