Fixed MIXED_CHARSET spam filter check

crates/common/src/scripts/functions/mod.rs

@@ -51,7 +51,7 @@ pub fn register_functions_trusted() -> FunctionMap {
         .with_function("winnow", fn_winnow)
         .with_function("has_zwsp", fn_has_zwsp)
         .with_function("has_obscured", fn_has_obscured)
-        .with_function("is_single_script", fn_is_single_script)
+        .with_function("is_mixed_charset", fn_is_mixed_charset)
         .with_function("puny_decode", fn_puny_decode)
         .with_function("unicode_skeleton", fn_unicode_skeleton)
         .with_function("cure_text", fn_cure_text)

crates/common/src/scripts/functions/unicode.rs

@@ -5,7 +5,8 @@
  */
 
 use sieve::{runtime::Variable, Context};
-use unicode_security::MixedScript;
+
+use crate::scripts::IsMixedCharset;
 
 pub fn fn_is_ascii<'x>(_: &'x Context<'x>, v: Vec<Variable>) -> Variable {
     match &v[0] {
@@ -80,12 +81,12 @@ pub fn fn_unicode_skeleton<'x>(_: &'x Context<'x>, v: Vec<Variable>) -> Variable
         .into()
 }
 
-pub fn fn_is_single_script<'x>(_: &'x Context<'x>, v: Vec<Variable>) -> Variable {
+pub fn fn_is_mixed_charset<'x>(_: &'x Context<'x>, v: Vec<Variable>) -> Variable {
     let text = v[0].to_string();
     if !text.is_empty() {
-        text.as_ref().is_single_script()
+        text.as_ref().is_mixed_charset()
     } else {
-        true
+        false
     }
     .into()
 }

crates/common/src/scripts/mod.rs

@@ -8,6 +8,7 @@ use std::sync::Arc;
 
 use sieve::{runtime::Variable, Envelope};
 use store::Value;
+use unicode_security::mixed_script::AugmentedScriptSet;
 
 use crate::IntoString;
 
@@ -56,3 +57,21 @@ pub fn to_store_value(value: &Variable) -> Value<'static> {
         v => Value::Text(v.to_string().into_owned().into()),
     }
 }
+
+pub trait IsMixedCharset {
+    fn is_mixed_charset(&self) -> bool;
+}
+
+impl<T: AsRef<str>> IsMixedCharset for T {
+    fn is_mixed_charset(&self) -> bool {
+        let mut set: Option<AugmentedScriptSet> = None;
+
+        for ch in self.as_ref().chars() {
+            if !ch.is_ascii() {
+                set.get_or_insert_default().intersect_with(ch.into());
+            }
+        }
+
+        set.map_or(false, |set| set.is_empty())
+    }
+}

crates/spam-filter/src/analysis/mime.rs

@@ -7,12 +7,14 @@
 use std::{collections::HashSet, future::Future, vec};
 
 use common::{
-    scripts::functions::{array::cosine_similarity, unicode::CharUtils},
+    scripts::{
+        functions::{array::cosine_similarity, unicode::CharUtils},
+        IsMixedCharset,
+    },
     Server,
 };
 use mail_parser::{HeaderName, MimeHeaders, PartType};
 use nlp::tokenizers::types::TokenType;
-use unicode_security::MixedScript;
 
 use crate::{SpamFilterContext, TextPart};
 
@@ -304,8 +306,8 @@ impl SpamFilterAnalyzeMime for Server {
                                 || ctx.input.message.html_body.contains(&part_id)
                         })
                         .map_or(false, |p| match p {
-                            TextPart::Plain { text_body, .. } => !text_body.is_single_script(),
-                            TextPart::Html { text_body, .. } => !text_body.is_single_script(),
+                            TextPart::Plain { text_body, .. } => text_body.is_mixed_charset(),
+                            TextPart::Html { text_body, .. } => text_body.is_mixed_charset(),
                             TextPart::None => false,
                         })
                     {

crates/spam-filter/src/analysis/url.rs

@@ -10,11 +10,11 @@ use std::{borrow::Cow, future::Future, time::Duration};
 
 use common::config::spamfilter::{Element, IpResolver, Location};
 use common::scripts::functions::unicode::CharUtils;
+use common::scripts::IsMixedCharset;
 use common::Server;
 use hyper::{header::LOCATION, Uri};
 use nlp::tokenizers::types::TokenType;
 use reqwest::redirect::Policy;
-use unicode_security::MixedScript;
 
 use crate::modules::dnsbl::check_dnsbl;
 use crate::modules::expression::StringResolver;
@@ -244,7 +244,7 @@ impl SpamFilterAnalyzeUrl for Server {
                         }
                     }
 
-                    if !host.fqdn.is_single_script() {
+                    if host.fqdn.is_mixed_charset() {
                         ctx.result.add_tag("MIXED_CHARSET_URL");
                     }
                 }

tests/resources/smtp/antispam/combined.test

@@ -1021,8 +1021,8 @@ dmarc.result pass
 dmarc.policy reject
 remote_ip 173.224.123.255
 tls.version TLS1_2
-expect_header X-Spam-Result: DMARC_POLICY_ALLOW (-0.50), DKIM_ALLOW (-0.20), SPF_ALLOW (-0.20), ARC_NA (0.00), DKIM_SIGNED (0.00), FROM_EQ_ENV_FROM (0.00), FROM_HAS_DN (0.00), HAS_EXTERNAL_IMG (0.00), HAS_REPLYTO (0.00), HAS_X_PRIO_THREE (0.00), HTML_SHORT_1 (0.00), RCPT_COUNT_ONE (0.00), REPLYTO_DN_EQ_FROM_DN (0.00), REPLYTO_DOM_EQ_FROM_DOM (0.00), TO_DN_ALL (0.00), TO_EQ_FROM (0.00), RCVD_COUNT_ZERO (0.10), RCVD_NO_TLS_LAST (0.10), HELO_NORES_A_OR_MX (0.30), MID_RHS_NOT_FQDN (0.50), UNPARSABLE_URL (0.50), FROMHOST_NORES_A_OR_MX (1.50), DIRECT_TO_MX (2.00), FORGED_RECIPIENTS (2.00), SUBJ_ALL_CAPS (3.00)
-expect_header X-Spam-Status: Yes, score=9.10
+expect_header X-Spam-Result: DMARC_POLICY_ALLOW (-0.50), DKIM_ALLOW (-0.20), SPF_ALLOW (-0.20), ARC_NA (0.00), DKIM_SIGNED (0.00), FROM_EQ_ENV_FROM (0.00), FROM_HAS_DN (0.00), HAS_EXTERNAL_IMG (0.00), HAS_REPLYTO (0.00), HAS_X_PRIO_THREE (0.00), HTML_SHORT_1 (0.00), RCPT_COUNT_ONE (0.00), REPLYTO_DN_EQ_FROM_DN (0.00), REPLYTO_DOM_EQ_FROM_DOM (0.00), TO_DN_ALL (0.00), TO_EQ_FROM (0.00), RCVD_COUNT_ZERO (0.10), RCVD_NO_TLS_LAST (0.10), HELO_NORES_A_OR_MX (0.30), MID_RHS_NOT_FQDN (0.50), UNPARSABLE_URL (0.50), DATE_IN_PAST (1.00), FROMHOST_NORES_A_OR_MX (1.50), DIRECT_TO_MX (2.00), FORGED_RECIPIENTS (2.00), SUBJ_ALL_CAPS (3.00)
+expect_header X-Spam-Status: Yes, score=10.10
 
 Return-Path: <[email protected]>
 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; s=default; d=landeray.com;

tests/resources/smtp/antispam/url.test

@@ -26,7 +26,7 @@ expect MIXED_CHARSET_URL
 
 Subject: test
 
-my site is https://www.xn--80ak6aa92e.com/
+my site is https://www.xn--1ca81o6aa92e.com/
 <!-- NEXT TEST -->
 expect UNPARSABLE_URL