Can someone suggest a solution or tell me what I'm doing wrong?

The exit needs to be inside the if.

For a more efficient way to do what you're doing without foreach, see my tutorial Building Regex Alternations Dynamically.

Like https://metacpan.org/pod/Regexp::Profanity::US

So you don't need to implement your own solution.

More general you might need a generic spam word solution, e.g. "Nigerian prince" is not really a profanity as such... ;)

On another note:

I'm not sure if returning a detailed error message is a good approach against spammers, they might use it to train against your filter.

So at least consider a JS based feedback, to make it harder to be spotted by bots.

#!/usr/bin/perl

use strict;
use warnings;

my @BADWORDS = ('hell');
$a =  ' H e l l o   w o r l d ! ';
$a =~ tr|a-zA-Z||cd;
$a = lc($a);
print $a;
[download]

my $fail=0; my @ErrMsg;
my @forbidden=('porn','sex');

if (map {$Message =~ /$_/i} @forbidden){
  $fail=1; # or $fail++;
}

if ($fail){ # or ($fail > 2)
  push @ErrMsg,"Bad Language Not Permitted";
}
[download]

This problem space is pretty difficult to get right even in naïve cases.

ack -i sex /usr/share/dict/words | nl
1 ambosexous
# ...
45 misexplain
# ...
199 Wessexman
[download]

Quite a few false positives there and it will pass through these and their variations or the even more common emoji replacements for the topic I won’t include.

s3x & pr0n
s̸͚͍̺̺͇̝̥͎͕̃e̸̲͆̅̈́̈̈́̆͆̀͆͘͜x̶̧̟̭̘̿̂̈́͝ ̷̨͇͓͖̅̄͑̓̕͝a̵̰̜̣͎͔͛͗̆̋̓n̸͍̳͍̤̊́͌͌͗͂͘d̵͇̱̣̈́̀͒͆̐͠͝ ̷̡͓̥̙͕̦͙̬̠͒̐ͅp̶̡̖̫̱̞̙͔͑̊͛͌̈́͘ô̷̫̰̫̹̘̖̪͙͌͋͗̓̓̂̿r̴͎͇͖̲̦̤͕̉n̸̘̝͇͌̂͒̈́͝
รєא คภ๔ ק๏гภ
ˢᵉˣ ᵃⁿᵈ ᵖᵒʳⁿ
𝚜𝚎𝚡 𝚊𝚗𝚍 𝚙𝚘𝚛𝚗
uɹod puɐ xǝs
🅂🄴🅇 🄰🄽🄳 🄿🄾🅁🄽
🎀 𝓈𝑒𝓍 𝒶𝓃𝒹 𝓅🍑𝓇𝓃 🎀
𝘀𝗲𝘅 𝗮𝗻𝗱 𝗽𝗼𝗿𝗻

Update: I ran those through Text::Unidecode to see if it would untangle them because it’s surprisingly good at that, usually. Not here. Best conversion on any of them was s[?]x [?]n[?] [?][?]rn. The rest were nowhere close.

Okay, this is silly and I probably could have found a better way to spend the last couple hours but I was curious and surprised by Text::Unidecode’s lack of efficacy here so… (has to be <pre/> instead of <code/> tags because of content.)

Posting it because I do these things so rarely I’ll lose the memory of how to do it if I don’t. :P

#!/usr/bin/env perl
use 5.14.0; # Not sure this is earliest/best.
use utf8;
use strictures;
use open ":std", ":encoding(utf8)";
use Unicode::UCD "charinfo";
use charnames ":full";
use YAML; # For introspection if you're curious.

my @sex_and_porn = <DATA>;

for my $sap ( @sex_and_porn )
{
    chomp $sap;
    my @string;
    for my $chr ( split "", $sap )
    {
        next unless $chr =~ /\A[[:print:]]+\z/;
        my $info = charinfo(ord($chr));
        # print Dump($info);
        my $name = $info->{name};
        next if $name =~ /COMBINING/;
        my $replacement;
        if ( $name =~ /\b(SMALL|CAPITAL)?(?: LETTER)?(?: TURNED| SHARP)? ([A-Z])\b/ )
        {
            no warnings "uninitialized";
            my $case = $1 eq "SMALL" ? "SMALL" : "CAPITAL";
            $replacement = "LATIN $case LETTER $2";
        }

        my $final = charnames::string_vianame( $replacement || $name );

        push @string, $final =~ /(?:\b[A-Z]\b| )/i ?
            $final : "\x{0}";
    }
    s/\x{0}+/[?]/g for @string;
    printf " %15s -> %s\n\n", $sap, join "", @string;
}

__DATA__
s3x & pr0n
s̸͚͍̺̺͇̝̥͎͕̃e̸̲͆̅̈́̈̈́̆͆̀͆͘͜x̶̧̟̭̘̿̂̈́͝ ̷̨͇͓͖̅̄͑̓̕͝a̵̰̜̣͎͔͛͗̆̋̓n̸͍̳͍̤̊́͌͌͗͂͘d̵͇̱̣̈́̀͒͆̐͠͝ ̷̡͓̥̙͕̦͙̬̠͒̐ͅp̶̡̖̫̱̞̙͔͑̊͛͌̈́͘ô̷̫̰̫̹̘̖̪͙͌͋͗̓̓̂̿r̴͎͇͖̲̦̤͕̉n̸̘̝͇͌̂͒̈́͝
รєא คภ๔ ק๏гภ
ˢᵉˣ ᵃⁿᵈ ᵖᵒʳⁿ
𝚜𝚎𝚡 𝚊𝚗𝚍 𝚙𝚘𝚛𝚗
uɹod puɐ xǝs
🅂🄴🅇 🄰🄽🄳 🄿🄾🅁🄽
🎀 𝓈𝑒𝓍 𝒶𝓃𝒹 𝓅🍑𝓇𝓃 🎀
𝘀𝗲𝘅 𝗮𝗻𝗱 𝗽𝗼𝗿𝗻 
ßêx and pørñ
ẞӬ𐊜 AƝḊ ⓅỚℝꞐ

Output–

      s3x & pr0n -> s[?]x [?] pr[?]n

 s̸͚͍̺̺͇̝̥͎͕̃e̸̲͆̅̈́̈̈́̆͆̀͆͘͜x̶̧̟̭̘̿̂̈́͝ ̷̨͇͓͖̅̄͑̓̕͝a̵̰̜̣͎͔͛͗̆̋̓n̸͍̳͍̤̊́͌͌͗͂͘d̵͇̱̣̈́̀͒͆̐͠͝ ̷̡͓̥̙͕̦͙̬̠͒̐ͅp̶̡̖̫̱̞̙͔͑̊͛͌̈́͘ô̷̫̰̫̹̘̖̪͙͌͋͗̓̓̂̿r̴͎͇͖̲̦̤͕̉n̸̘̝͇͌̂͒̈́͝ -> sex and porn

    รєא คภ๔ ק๏гภ -> [?][?][?] [?][?][?] [?][?][?][?]

    ˢᵉˣ ᵃⁿᵈ ᵖᵒʳⁿ -> sex and porn

    𝚜𝚎𝚡 𝚊𝚗𝚍 𝚙𝚘𝚛𝚗 -> sex and porn

    uɹod puɐ xǝs -> urod pua xes

    🅂🄴🅇 🄰🄽🄳 🄿🄾🅁🄽 -> SEX AND PORN

 🎀 𝓈𝑒𝓍 𝒶𝓃𝒹 𝓅🍑𝓇𝓃 🎀 -> [?] sex and p[?]rn [?]

   𝘀𝗲𝘅 𝗮𝗻𝗱 𝗽𝗼𝗿𝗻  -> sex and porn 

    ßêx and pørñ -> sex and porn

    ẞӬ𐊜 AƝḊ ⓅỚℝꞐ -> SEX AND PORN

And it gets worse if you have to support multiple languages. If the company is in North America, they can probably get away with supporting 3 different languages (spanish, french and english).

If the company in the EU, there are 24 official languages at last count: Bulgarian, Croatian, Czech, Danish, Dutch, English, Estonian, Finnish, French, German, Greek, Hungarian, Irish, Italian, Latvian, Lithuanian, Maltese, Polish, Portuguese, Romanian, Slovak, Slovenian, Spanish and Swedish.

You might get away with supporting fewer languages in some private sectors, but if you are running a government system (or a system that wants to receive funding from the government), you might have to support all of those. And as anyone can imagine, a "bad" word in one of those languages might be a standard word with a different meaning in another language.

Plus, even for countries speaking the same language, there might be big cultural differences. For U.S. citizens, a "tea party" means throwing valuable cargo into the harbor because they endorse "taxation without representation" but want to keep all the taxes for their own government(*). For Brits, it means "5 p.m."

(*) Puerto Rico

PerlMonks XP is useless? Not anymore: XPD - Do more with your PerlMonks XP

> This problem space is pretty difficult to get right even in naïve cases.

I told this story before, but for some time my old faculty blocked emails from the University of Sussex ...

(those filthy people ;-)

Cheers Rolf
_{(addicted to the Perl Programming Language :)
Wikisyntax for the Monastery}