Enhance isEmail validation (#250)

This enhances the current isEmail validation by using a consistent regex
that will be used across protovalidate implementations.

Author: smaye81

build.gradle.kts

@@ -272,7 +272,6 @@ dependencies {
     implementation(libs.cel.core)
     implementation(libs.guava)
     implementation(libs.ipaddress)
-    implementation(libs.jakarta.mail.api)
 
     buf("build.buf:buf:${libs.versions.buf.get()}:${osdetector.classifier}@exe")
 

conformance/expected-failures.yaml

@@ -107,46 +107,6 @@ custom_constraints:
   #ERROR: <input>:1:1: expression of type 'int' cannot be range of a comprehension (must be list, map, or dynamic)
   # | this.all(e, e == 1)
   # | ^
-library/is_email:
-  - invalid/non_ascii
-  #       input: [type.googleapis.com/buf.validate.conformance.cases.IsEmail]:{val:"µ@example.com"}
-  #        want: validation error (1 violation)
-  #          1. constraint_id: "library.is_email"
-  #         got: valid
-  - invalid/quoted-string/a
-  #       input: [type.googleapis.com/buf.validate.conformance.cases.IsEmail]:{val:"\"foo bar\"@example.com"}
-  #        want: validation error (1 violation)
-  #          1. constraint_id: "library.is_email"
-  #         got: valid
-  - invalid/quoted-string/b
-  #       input: [type.googleapis.com/buf.validate.conformance.cases.IsEmail]:{val:"\"foo..bar\"@example.com"}
-  #        want: validation error (1 violation)
-  #          1. constraint_id: "library.is_email"
-  #         got: valid
-  - valid/empty_atext
-  #       input: [type.googleapis.com/buf.validate.conformance.cases.IsEmail]:{val:"[email protected]"}
-  #        want: valid
-  #         got: validation error (1 violation)
-  #          1. constraint_id: "library.is_email"
-  #             message: ""
-  - valid/exhaust_atext
-  #       input: [type.googleapis.com/buf.validate.conformance.cases.IsEmail]:{val:"0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ!#$%&'*+-/=?^_`{|}[email protected]"}
-  #        want: valid
-  #         got: validation error (1 violation)
-  #          1. constraint_id: "library.is_email"
-  #             message: ""
-  - valid/label_all_digits
-  #       input: [type.googleapis.com/buf.validate.conformance.cases.IsEmail]:{val:"[email protected]"}
-  #        want: valid
-  #         got: validation error (1 violation)
-  #          1. constraint_id: "library.is_email"
-  #             message: ""
-  - valid/multiple_empty_atext
-  #       input: [type.googleapis.com/buf.validate.conformance.cases.IsEmail]:{val:"[email protected]"}
-  #        want: valid
-  #         got: validation error (1 violation)
-  #          1. constraint_id: "library.is_email"
-  #             message: ""
 library/is_host_and_port:
   - port_required/false/invalid/ipv6_zone-id_too_short
   #       input: [type.googleapis.com/buf.validate.conformance.cases.IsHostAndPort]:{val:"[::1%]"}

gradle/libs.versions.toml

@@ -20,7 +20,6 @@ cel-core = { module = "org.projectnessie.cel:cel-core" }
 errorprone = { module = "com.google.errorprone:error_prone_core", version = "2.37.0" }
 guava = { module = "com.google.guava:guava", version = "33.4.0-jre" }
 ipaddress = { module = "com.github.seancfoley:ipaddress", version.ref = "ipaddress" }
-jakarta-mail-api = { module = "jakarta.mail:jakarta.mail-api", version = "2.1.3" }
 junit-bom = { module = "org.junit:junit-bom", version.ref = "junit" }
 maven-plugin = { module = "com.vanniktech:gradle-maven-publish-plugin", version.ref = "maven-publish" }
 nullaway = { module = "com.uber.nullaway:nullaway", version = "0.12.4" }

src/main/java/build/buf/protovalidate/CustomOverload.java

@@ -20,15 +20,14 @@
 import com.google.common.primitives.Bytes;
 import inet.ipaddr.IPAddress;
 import inet.ipaddr.IPAddressString;
-import jakarta.mail.internet.AddressException;
-import jakarta.mail.internet.InternetAddress;
 import java.net.Inet4Address;
 import java.net.Inet6Address;
 import java.net.InetAddress;
 import java.net.URI;
 import java.net.URISyntaxException;
 import java.util.HashSet;
 import java.util.Set;
+import java.util.regex.Pattern;
 import org.projectnessie.cel.common.types.BoolT;
 import org.projectnessie.cel.common.types.Err;
 import org.projectnessie.cel.common.types.IntT;
@@ -58,6 +57,11 @@ final class CustomOverload {
   private static final String OVERLOAD_IS_INF = "isInf";
   private static final String OVERLOAD_IS_HOST_AND_PORT = "isHostAndPort";
 
+  // See https://html.spec.whatwg.org/multipage/input.html#valid-e-mail-address
+  private static final Pattern EMAIL_REGEX =
+      Pattern.compile(
+          "^[a-zA-Z0-9.!#$%&'*+/=?^_`{|}~-]+@[a-zA-Z0-9](?:[a-zA-Z0-9-]{0,61}[a-zA-Z0-9])?(?:\\.[a-zA-Z0-9](?:[a-zA-Z0-9-]{0,61}[a-zA-Z0-9])?)*$");
+
   /**
    * Create custom function overload list.
    *
@@ -514,27 +518,17 @@ private static Val uniqueList(Lister list) {
   }
 
   /**
-   * Validates if the input string is a valid email address.
+   * validateEmail returns true if addr is a valid email address.
+   *
+   * <p>This regex conforms to the definition for a valid email address from the HTML standard. Note
+   * that this standard willfully deviates from RFC 5322, which allows many unexpected forms of
+   * email addresses and will easily match a typographical error.
    *
    * @param addr The input string to validate as an email address.
    * @return {@code true} if the input string is a valid email address, {@code false} otherwise.
    */
   private static boolean validateEmail(String addr) {
-    try {
-      InternetAddress emailAddr = new InternetAddress(addr);
-      emailAddr.validate();
-      if (addr.contains("<") || !emailAddr.getAddress().equals(addr)) {
-        return false;
-      }
-      addr = emailAddr.getAddress();
-      if (addr.length() > 254) {
-        return false;
-      }
-      String[] parts = addr.split("@", 2);
-      return parts[0].length() < 64 && validateHostname(parts[1]);
-    } catch (AddressException ex) {
-      return false;
-    }
+    return EMAIL_REGEX.matcher(addr).matches();
   }
 
   /**