Pull request: all: add idna handling, imp domain validation

Updates #2915.

Squashed commit of the following:

commit b907324426c87ee7334edbd61e43c44444ad27a9
Author: Ainar Garipov <A.Garipov@AdGuard.COM>
Date:   Wed Apr 7 16:26:41 2021 +0300

    all: imp docs, upd

commit c022f75cac006e077095cad283fea0a91d3a0eea
Author: Ainar Garipov <A.Garipov@AdGuard.COM>
Date:   Wed Apr 7 15:51:30 2021 +0300

    all: add idna handling, imp domain validation

internal/aghnet/addr.go

@@ -3,8 +3,10 @@ package aghnet
 import (
 	"fmt"
 	"net"
+	"strings"
 
 	"github.com/AdguardTeam/AdGuardHome/internal/agherr"
+	"golang.org/x/net/idna"
 )
 
 // ValidateHardwareAddress returns an error if hwa is not a valid EUI-48,
@@ -21,3 +23,79 @@ func ValidateHardwareAddress(hwa net.HardwareAddr) (err error) {
 		return fmt.Errorf("bad len: %d", l)
 	}
 }
+
+// maxDomainLabelLen is the maximum allowed length of a domain name label
+// according to RFC 1035.
+const maxDomainLabelLen = 63
+
+// maxDomainNameLen is the maximum allowed length of a full domain name
+// according to RFC 1035.
+//
+// See https://stackoverflow.com/a/32294443/1892060.
+const maxDomainNameLen = 253
+
+const invalidCharMsg = "invalid char %q at index %d in %q"
+
+// isValidHostFirstRune returns true if r is a valid first rune for a hostname
+// label.
+func isValidHostFirstRune(r rune) (ok bool) {
+	return (r >= 'a' && r <= 'z') ||
+		(r >= 'A' && r <= 'Z') ||
+		(r >= '0' && r <= '9')
+}
+
+// isValidHostRune returns true if r is a valid rune for a hostname label.
+func isValidHostRune(r rune) (ok bool) {
+	return r == '-' || isValidHostFirstRune(r)
+}
+
+// ValidateDomainNameLabel returns an error if label is not a valid label of
+// a domain name.
+func ValidateDomainNameLabel(label string) (err error) {
+	if len(label) > maxDomainLabelLen {
+		return fmt.Errorf("%q is too long, max: %d", label, maxDomainLabelLen)
+	} else if len(label) == 0 {
+		return agherr.Error("label is empty")
+	}
+
+	if r := label[0]; !isValidHostFirstRune(rune(r)) {
+		return fmt.Errorf(invalidCharMsg, r, 0, label)
+	}
+
+	for i, r := range label[1:] {
+		if !isValidHostRune(r) {
+			return fmt.Errorf(invalidCharMsg, r, i+1, label)
+		}
+	}
+
+	return nil
+}
+
+// ValidateDomainName validates the domain name in accordance to RFC 952, RFC
+// 1035, and with RFC-1123's inclusion of digits at the start of the host.  It
+// doesn't validate against two or more hyphens to allow punycode and
+// internationalized domains.
+//
+// TODO(a.garipov): After making sure that this works correctly, port this into
+// module golibs.
+func ValidateDomainName(name string) (err error) {
+	name, err = idna.ToASCII(name)
+	if err != nil {
+		return err
+	}
+
+	l := len(name)
+	if l == 0 || l > maxDomainNameLen {
+		return fmt.Errorf("%q is too long, max: %d", name, maxDomainNameLen)
+	}
+
+	labels := strings.Split(name, ".")
+	for i, l := range labels {
+		err = ValidateDomainNameLabel(l)
+		if err != nil {
+			return fmt.Errorf("invalid domain name label at index %d: %w", i, err)
+		}
+	}
+
+	return nil
+}

internal/aghnet/addr_test.go

@@ -2,6 +2,7 @@ package aghnet
 
 import (
 	"net"
+	"strings"
 	"testing"
 
 	"github.com/stretchr/testify/assert"
@@ -50,6 +51,81 @@ func TestValidateHardwareAddress(t *testing.T) {
 				assert.NoError(t, err)
 			} else {
 				require.Error(t, err)
+
+				assert.Equal(t, tc.wantErrMsg, err.Error())
+			}
+		})
+	}
+}
+
+func repeatStr(b *strings.Builder, s string, n int) {
+	for i := 0; i < n; i++ {
+		_, _ = b.WriteString(s)
+	}
+}
+
+func TestValidateDomainName(t *testing.T) {
+	b := &strings.Builder{}
+	repeatStr(b, "a", 255)
+	longDomainName := b.String()
+
+	b.Reset()
+	repeatStr(b, "a", 64)
+	longLabel := b.String()
+
+	_, _ = b.WriteString(".com")
+	longLabelDomainName := b.String()
+
+	testCases := []struct {
+		name       string
+		in         string
+		wantErrMsg string
+	}{{
+		name:       "success",
+		in:         "example.com",
+		wantErrMsg: "",
+	}, {
+		name:       "success_idna",
+		in:         "пример.рф",
+		wantErrMsg: "",
+	}, {
+		name: "bad_symbol",
+		in:   "!!!",
+		wantErrMsg: `invalid domain name label at index 0: ` +
+			`invalid char '!' at index 0 in "!!!"`,
+	}, {
+		name:       "bad_length",
+		in:         longDomainName,
+		wantErrMsg: `"` + longDomainName + `" is too long, max: 253`,
+	}, {
+		name: "bad_label_length",
+		in:   longLabelDomainName,
+		wantErrMsg: `invalid domain name label at index 0: "` + longLabel +
+			`" is too long, max: 63`,
+	}, {
+		name:       "bad_label_empty",
+		in:         "example..com",
+		wantErrMsg: `invalid domain name label at index 1: label is empty`,
+	}, {
+		name: "bad_label_first_symbol",
+		in:   "example.-aa.com",
+		wantErrMsg: `invalid domain name label at index 1:` +
+			` invalid char '-' at index 0 in "-aa"`,
+	}, {
+		name: "bad_label_symbol",
+		in:   "example.a!!!.com",
+		wantErrMsg: `invalid domain name label at index 1:` +
+			` invalid char '!' at index 1 in "a!!!"`,
+	}}
+
+	for _, tc := range testCases {
+		t.Run(tc.name, func(t *testing.T) {
+			err := ValidateDomainName(tc.in)
+			if tc.wantErrMsg == "" {
+				assert.NoError(t, err)
+			} else {
+				require.Error(t, err)
+
 				assert.Equal(t, tc.wantErrMsg, err.Error())
 			}
 		})