feat: add setting to block disposable emails

- Add a new setting `EMAIL_DOMAIN_BLOCK_DISPOSABLE` that will append a list of
  domains that are known for being used by temporary or disposable email
  services.

- Add a utility to automatically download and format the list of domains from
  the disposable-email-domains project on github.

  (https://github.com/disposable-email-domains/disposable-email-domains)
  license: CC0 1.0 Universal (CC0 1.0) [Public Domain]

  from README:
  """
  This repo contains a list of disposable and temporary email address domains often used to register dummy users in order to spam or abuse some services.

  We cannot guarantee all of these can still be considered disposable but we do basic checking so chances are they were disposable at one point in time.
  """

modules/setting/service.go

@@ -5,6 +5,7 @@ package setting
 
 import (
 	"regexp"
+	"slices"
 	"strings"
 	"time"
 
@@ -37,6 +38,7 @@ var Service = struct {
 	RegisterManualConfirm                   bool
 	EmailDomainAllowList                    []glob.Glob
 	EmailDomainBlockList                    []glob.Glob
+	EmailDomainBlockDisposable              bool
 	DisableRegistration                     bool
 	AllowOnlyInternalRegistration           bool
 	AllowOnlyExternalRegistration           bool
@@ -156,6 +158,22 @@ func loadServiceFrom(rootCfg ConfigProvider) {
 	}
 	Service.EmailDomainAllowList = CompileEmailGlobList(sec, "EMAIL_DOMAIN_WHITELIST", "EMAIL_DOMAIN_ALLOWLIST")
 	Service.EmailDomainBlockList = CompileEmailGlobList(sec, "EMAIL_DOMAIN_BLOCKLIST")
+	Service.EmailDomainBlockDisposable = sec.Key("EMAIL_DOMAIN_BLOCK_DISPOSABLE").MustBool(false)
+	if Service.EmailDomainBlockDisposable {
+		toAdd := make([]glob.Glob, 0, len(DisposableEmailDomains()))
+		for _, domain := range DisposableEmailDomains() {
+			domain = strings.ToLower(domain)
+			// Only add domains that aren't blocked yet.
+			if !slices.ContainsFunc(Service.EmailDomainBlockList, func(g glob.Glob) bool { return g.Match(domain) }) {
+				if g, err := glob.Compile(domain); err != nil {
+					log.Error("Error in disposable domain %s: %v", domain, err)
+				} else {
+					toAdd = append(toAdd, g)
+				}
+			}
+		}
+		Service.EmailDomainBlockList = append(Service.EmailDomainBlockList, toAdd...)
+	}
 	Service.ShowRegistrationButton = sec.Key("SHOW_REGISTRATION_BUTTON").MustBool(!(Service.DisableRegistration || Service.AllowOnlyExternalRegistration))
 	Service.ShowMilestonesDashboardPage = sec.Key("SHOW_MILESTONES_DASHBOARD_PAGE").MustBool(true)
 	Service.RequireSignInView = sec.Key("REQUIRE_SIGNIN_VIEW").MustBool()

modules/setting/service_test.go

@@ -4,6 +4,9 @@
 package setting
 
 import (
+	"fmt"
+	"sort"
+	"strings"
 	"testing"
 
 	"code.gitea.io/gitea/modules/structs"
@@ -11,8 +14,18 @@ import (
 	"github.com/gobwas/glob"
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
+	"golang.org/x/net/publicsuffix"
 )
 
+func match(globs []glob.Glob, s string) bool {
+	for _, g := range globs {
+		if g.Match(s) {
+			return true
+		}
+	}
+	return false
+}
+
 func TestLoadServices(t *testing.T) {
 	oldService := Service
 	defer func() {
@@ -28,15 +41,6 @@ EMAIL_DOMAIN_BLOCKLIST = d3, *.b
 	require.NoError(t, err)
 	loadServiceFrom(cfg)
 
-	match := func(globs []glob.Glob, s string) bool {
-		for _, g := range globs {
-			if g.Match(s) {
-				return true
-			}
-		}
-		return false
-	}
-
 	assert.True(t, match(Service.EmailDomainAllowList, "d1"))
 	assert.True(t, match(Service.EmailDomainAllowList, "foo.w"))
 	assert.True(t, match(Service.EmailDomainAllowList, "d2"))
@@ -48,6 +52,121 @@ EMAIL_DOMAIN_BLOCKLIST = d3, *.b
 	assert.False(t, match(Service.EmailDomainBlockList, "d1"))
 }
 
+func TestLoadServiceBlockDisposable(t *testing.T) {
+	oldService := Service
+	defer func() {
+		Service = oldService
+	}()
+
+	cfg, err := NewConfigProviderFromData(`
+[service]
+EMAIL_DOMAIN_BLOCK_DISPOSABLE = true
+`)
+
+	require.NoError(t, err)
+	loadServiceFrom(cfg)
+
+	for _, domain := range DisposableEmailDomains() {
+		require.True(t, match(Service.EmailDomainBlockList, domain))
+	}
+
+	require.Len(t, Service.EmailDomainBlockList, len(DisposableEmailDomains()))
+
+	knownGood := [...]string{
+		"aol.com",
+		"gmx.com",
+		"mail.com",
+		"zoho.com",
+		"proton.me",
+		"gmail.com",
+		"yahoo.com",
+		"icloud.com",
+		"outlook.com",
+		"protonmail.com",
+	}
+
+	for _, domain := range knownGood {
+		require.False(t, match(Service.EmailDomainBlockList, domain))
+	}
+}
+
+func TestLoadServiceBlockDisposableWithExistingGlobs(t *testing.T) {
+	oldService := Service
+	defer func() {
+		Service = oldService
+	}()
+
+	tldCounts := make(map[string]int)
+	for _, domain := range DisposableEmailDomains() {
+		tld, _ := publicsuffix.PublicSuffix(domain)
+		tldCounts[tld]++
+	}
+
+	type tldkv struct {
+		Tld   string
+		Count int
+	}
+
+	sortedTldCounts := make([]tldkv, 0)
+	for tld, count := range tldCounts {
+		sortedTldCounts = append(sortedTldCounts, tldkv{tld, count})
+	}
+
+	sort.Slice(sortedTldCounts, func(i, j int) bool {
+		return sortedTldCounts[i].Count > sortedTldCounts[j].Count
+	})
+	require.GreaterOrEqual(t, len(sortedTldCounts), 2)
+
+	blockString := fmt.Sprintf("*.%s,*.%s", sortedTldCounts[0].Tld, sortedTldCounts[1].Tld)
+
+	cfg, err := NewConfigProviderFromData(fmt.Sprintf(`
+[service]
+EMAIL_DOMAIN_BLOCKLIST = %s
+EMAIL_DOMAIN_BLOCK_DISPOSABLE = true
+`, blockString))
+
+	require.NoError(t, err)
+	loadServiceFrom(cfg)
+
+	for _, domain := range DisposableEmailDomains() {
+		require.True(t, match(Service.EmailDomainBlockList, domain))
+	}
+
+	redundant := 0
+	for _, val := range DisposableEmailDomains() {
+		if strings.HasSuffix(val, sortedTldCounts[0].Tld) ||
+			strings.HasSuffix(val, sortedTldCounts[1].Tld) {
+			redundant++
+		}
+	}
+
+	expected := len(DisposableEmailDomains()) - redundant + 2
+	require.Len(t, Service.EmailDomainBlockList, expected)
+}
+
+func TestLoadServiceBlockDisposableWithComplementGlobs(t *testing.T) {
+	oldService := Service
+	defer func() {
+		Service = oldService
+	}()
+
+	cfg, err := NewConfigProviderFromData(`
+[service]
+EMAIL_DOMAIN_BLOCKLIST = *.random
+EMAIL_DOMAIN_BLOCK_DISPOSABLE = true
+`)
+
+	require.NoError(t, err)
+	loadServiceFrom(cfg)
+
+	for _, domain := range DisposableEmailDomains() {
+		require.True(t, match(Service.EmailDomainBlockList, domain))
+	}
+
+	expected := len(DisposableEmailDomains()) + 1
+	require.Len(t, Service.EmailDomainBlockList, expected)
+}
+
 func TestLoadServiceVisibilityModes(t *testing.T) {
 	oldService := Service
 	defer func() {