idna: allow specifying CheckHyphens and CheckJoiners
This copies CL 258837 from x/text over to this module.
Fixes golang/go#41732.
Change-Id: I41fc11c9a13faacb42504827063b60f7648426d2
Reviewed-on: https://go-review.googlesource.com/c/net/+/273586
Run-TryBot: Tobias Klauser <tobias.klauser@gmail.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Nigel Tao <nigeltao@golang.org>
Trust: Nigel Tao <nigeltao@golang.org>
Trust: Cherry Zhang <cherryyz@google.com>
diff --git a/idna/idna10.0.0.go b/idna/idna10.0.0.go
index 7e69ee1..5208ba6 100644
--- a/idna/idna10.0.0.go
+++ b/idna/idna10.0.0.go
@@ -67,15 +67,14 @@
// VerifyDNSLength sets whether a Profile should fail if any of the IDN parts
// are longer than allowed by the RFC.
+//
+// This option corresponds to the VerifyDnsLength flag in UTS #46.
func VerifyDNSLength(verify bool) Option {
return func(o *options) { o.verifyDNSLength = verify }
}
// RemoveLeadingDots removes leading label separators. Leading runes that map to
// dots, such as U+3002 IDEOGRAPHIC FULL STOP, are removed as well.
-//
-// This is the behavior suggested by the UTS #46 and is adopted by some
-// browsers.
func RemoveLeadingDots(remove bool) Option {
return func(o *options) { o.removeLeadingDots = remove }
}
@@ -83,6 +82,8 @@
// ValidateLabels sets whether to check the mandatory label validation criteria
// as defined in Section 5.4 of RFC 5891. This includes testing for correct use
// of hyphens ('-'), normalization, validity of runes, and the context rules.
+// In particular, ValidateLabels also sets the CheckHyphens and CheckJoiners flags
+// in UTS #46.
func ValidateLabels(enable bool) Option {
return func(o *options) {
// Don't override existing mappings, but set one that at least checks
@@ -91,25 +92,48 @@
o.mapping = normalize
}
o.trie = trie
- o.validateLabels = enable
- o.fromPuny = validateFromPunycode
+ o.checkJoiners = enable
+ o.checkHyphens = enable
+ if enable {
+ o.fromPuny = validateFromPunycode
+ } else {
+ o.fromPuny = nil
+ }
+ }
+}
+
+// CheckHyphens sets whether to check for correct use of hyphens ('-') in
+// labels. Most web browsers do not have this option set, since labels such as
+// "r3---sn-apo3qvuoxuxbt-j5pe" are in common use.
+//
+// This option corresponds to the CheckHyphens flag in UTS #46.
+func CheckHyphens(enable bool) Option {
+ return func(o *options) { o.checkHyphens = enable }
+}
+
+// CheckJoiners sets whether to check the ContextJ rules as defined in Appendix
+// A of RFC 5892, concerning the use of joiner runes.
+//
+// This option corresponds to the CheckJoiners flag in UTS #46.
+func CheckJoiners(enable bool) Option {
+ return func(o *options) {
+ o.trie = trie
+ o.checkJoiners = enable
}
}
// StrictDomainName limits the set of permissible ASCII characters to those
// allowed in domain names as defined in RFC 1034 (A-Z, a-z, 0-9 and the
-// hyphen). This is set by default for MapForLookup and ValidateForRegistration.
+// hyphen). This is set by default for MapForLookup and ValidateForRegistration,
+// but is only useful if ValidateLabels is set.
//
// This option is useful, for instance, for browsers that allow characters
// outside this range, for example a '_' (U+005F LOW LINE). See
-// http://www.rfc-editor.org/std/std3.txt for more details This option
-// corresponds to the UseSTD3ASCIIRules option in UTS #46.
+// http://www.rfc-editor.org/std/std3.txt for more details.
+//
+// This option corresponds to the UseSTD3ASCIIRules flag in UTS #46.
func StrictDomainName(use bool) Option {
- return func(o *options) {
- o.trie = trie
- o.useSTD3Rules = use
- o.fromPuny = validateFromPunycode
- }
+ return func(o *options) { o.useSTD3Rules = use }
}
// NOTE: the following options pull in tables. The tables should not be linked
@@ -117,6 +141,8 @@
// BidiRule enables the Bidi rule as defined in RFC 5893. Any application
// that relies on proper validation of labels should include this rule.
+//
+// This option corresponds to the CheckBidi flag in UTS #46.
func BidiRule() Option {
return func(o *options) { o.bidirule = bidirule.ValidString }
}
@@ -152,7 +178,8 @@
type options struct {
transitional bool
useSTD3Rules bool
- validateLabels bool
+ checkHyphens bool
+ checkJoiners bool
verifyDNSLength bool
removeLeadingDots bool
@@ -225,8 +252,11 @@
if p.useSTD3Rules {
s += ":UseSTD3Rules"
}
- if p.validateLabels {
- s += ":ValidateLabels"
+ if p.checkHyphens {
+ s += ":CheckHyphens"
+ }
+ if p.checkJoiners {
+ s += ":CheckJoiners"
}
if p.verifyDNSLength {
s += ":VerifyDNSLength"
@@ -254,26 +284,29 @@
punycode = &Profile{}
lookup = &Profile{options{
- transitional: true,
- useSTD3Rules: true,
- validateLabels: true,
- trie: trie,
- fromPuny: validateFromPunycode,
- mapping: validateAndMap,
- bidirule: bidirule.ValidString,
+ transitional: true,
+ useSTD3Rules: true,
+ checkHyphens: true,
+ checkJoiners: true,
+ trie: trie,
+ fromPuny: validateFromPunycode,
+ mapping: validateAndMap,
+ bidirule: bidirule.ValidString,
}}
display = &Profile{options{
- useSTD3Rules: true,
- validateLabels: true,
- trie: trie,
- fromPuny: validateFromPunycode,
- mapping: validateAndMap,
- bidirule: bidirule.ValidString,
+ useSTD3Rules: true,
+ checkHyphens: true,
+ checkJoiners: true,
+ trie: trie,
+ fromPuny: validateFromPunycode,
+ mapping: validateAndMap,
+ bidirule: bidirule.ValidString,
}}
registration = &Profile{options{
useSTD3Rules: true,
- validateLabels: true,
verifyDNSLength: true,
+ checkHyphens: true,
+ checkJoiners: true,
trie: trie,
fromPuny: validateFromPunycode,
mapping: validateRegistration,
@@ -340,7 +373,7 @@
}
isBidi = isBidi || bidirule.DirectionString(u) != bidi.LeftToRight
labels.set(u)
- if err == nil && p.validateLabels {
+ if err == nil && p.fromPuny != nil {
err = p.fromPuny(p, u)
}
if err == nil {
@@ -681,16 +714,18 @@
}
return nil
}
- if !p.validateLabels {
+ if p.checkHyphens {
+ if len(s) > 4 && s[2] == '-' && s[3] == '-' {
+ return &labelError{s, "V2"}
+ }
+ if s[0] == '-' || s[len(s)-1] == '-' {
+ return &labelError{s, "V3"}
+ }
+ }
+ if !p.checkJoiners {
return nil
}
- trie := p.trie // p.validateLabels is only set if trie is set.
- if len(s) > 4 && s[2] == '-' && s[3] == '-' {
- return &labelError{s, "V2"}
- }
- if s[0] == '-' || s[len(s)-1] == '-' {
- return &labelError{s, "V3"}
- }
+ trie := p.trie // p.checkJoiners is only set if trie is set.
// TODO: merge the use of this in the trie.
v, sz := trie.lookupString(s)
x := info(v)
diff --git a/idna/idna9.0.0.go b/idna/idna9.0.0.go
index 7c74563..55f718f 100644
--- a/idna/idna9.0.0.go
+++ b/idna/idna9.0.0.go
@@ -66,15 +66,14 @@
// VerifyDNSLength sets whether a Profile should fail if any of the IDN parts
// are longer than allowed by the RFC.
+//
+// This option corresponds to the VerifyDnsLength flag in UTS #46.
func VerifyDNSLength(verify bool) Option {
return func(o *options) { o.verifyDNSLength = verify }
}
// RemoveLeadingDots removes leading label separators. Leading runes that map to
// dots, such as U+3002 IDEOGRAPHIC FULL STOP, are removed as well.
-//
-// This is the behavior suggested by the UTS #46 and is adopted by some
-// browsers.
func RemoveLeadingDots(remove bool) Option {
return func(o *options) { o.removeLeadingDots = remove }
}
@@ -82,6 +81,8 @@
// ValidateLabels sets whether to check the mandatory label validation criteria
// as defined in Section 5.4 of RFC 5891. This includes testing for correct use
// of hyphens ('-'), normalization, validity of runes, and the context rules.
+// In particular, ValidateLabels also sets the CheckHyphens and CheckJoiners flags
+// in UTS #46.
func ValidateLabels(enable bool) Option {
return func(o *options) {
// Don't override existing mappings, but set one that at least checks
@@ -90,25 +91,48 @@
o.mapping = normalize
}
o.trie = trie
- o.validateLabels = enable
- o.fromPuny = validateFromPunycode
+ o.checkJoiners = enable
+ o.checkHyphens = enable
+ if enable {
+ o.fromPuny = validateFromPunycode
+ } else {
+ o.fromPuny = nil
+ }
+ }
+}
+
+// CheckHyphens sets whether to check for correct use of hyphens ('-') in
+// labels. Most web browsers do not have this option set, since labels such as
+// "r3---sn-apo3qvuoxuxbt-j5pe" are in common use.
+//
+// This option corresponds to the CheckHyphens flag in UTS #46.
+func CheckHyphens(enable bool) Option {
+ return func(o *options) { o.checkHyphens = enable }
+}
+
+// CheckJoiners sets whether to check the ContextJ rules as defined in Appendix
+// A of RFC 5892, concerning the use of joiner runes.
+//
+// This option corresponds to the CheckJoiners flag in UTS #46.
+func CheckJoiners(enable bool) Option {
+ return func(o *options) {
+ o.trie = trie
+ o.checkJoiners = enable
}
}
// StrictDomainName limits the set of permissable ASCII characters to those
// allowed in domain names as defined in RFC 1034 (A-Z, a-z, 0-9 and the
-// hyphen). This is set by default for MapForLookup and ValidateForRegistration.
+// hyphen). This is set by default for MapForLookup and ValidateForRegistration,
+// but is only useful if ValidateLabels is set.
//
// This option is useful, for instance, for browsers that allow characters
// outside this range, for example a '_' (U+005F LOW LINE). See
-// http://www.rfc-editor.org/std/std3.txt for more details This option
-// corresponds to the UseSTD3ASCIIRules option in UTS #46.
+// http://www.rfc-editor.org/std/std3.txt for more details.
+//
+// This option corresponds to the UseSTD3ASCIIRules flag in UTS #46.
func StrictDomainName(use bool) Option {
- return func(o *options) {
- o.trie = trie
- o.useSTD3Rules = use
- o.fromPuny = validateFromPunycode
- }
+ return func(o *options) { o.useSTD3Rules = use }
}
// NOTE: the following options pull in tables. The tables should not be linked
@@ -116,6 +140,8 @@
// BidiRule enables the Bidi rule as defined in RFC 5893. Any application
// that relies on proper validation of labels should include this rule.
+//
+// This option corresponds to the CheckBidi flag in UTS #46.
func BidiRule() Option {
return func(o *options) { o.bidirule = bidirule.ValidString }
}
@@ -152,7 +178,8 @@
type options struct {
transitional bool
useSTD3Rules bool
- validateLabels bool
+ checkHyphens bool
+ checkJoiners bool
verifyDNSLength bool
removeLeadingDots bool
@@ -225,8 +252,11 @@
if p.useSTD3Rules {
s += ":UseSTD3Rules"
}
- if p.validateLabels {
- s += ":ValidateLabels"
+ if p.checkHyphens {
+ s += ":CheckHyphens"
+ }
+ if p.checkJoiners {
+ s += ":CheckJoiners"
}
if p.verifyDNSLength {
s += ":VerifyDNSLength"
@@ -255,9 +285,10 @@
punycode = &Profile{}
lookup = &Profile{options{
transitional: true,
- useSTD3Rules: true,
- validateLabels: true,
removeLeadingDots: true,
+ useSTD3Rules: true,
+ checkHyphens: true,
+ checkJoiners: true,
trie: trie,
fromPuny: validateFromPunycode,
mapping: validateAndMap,
@@ -265,8 +296,9 @@
}}
display = &Profile{options{
useSTD3Rules: true,
- validateLabels: true,
removeLeadingDots: true,
+ checkHyphens: true,
+ checkJoiners: true,
trie: trie,
fromPuny: validateFromPunycode,
mapping: validateAndMap,
@@ -274,8 +306,9 @@
}}
registration = &Profile{options{
useSTD3Rules: true,
- validateLabels: true,
verifyDNSLength: true,
+ checkHyphens: true,
+ checkJoiners: true,
trie: trie,
fromPuny: validateFromPunycode,
mapping: validateRegistration,
@@ -339,7 +372,7 @@
continue
}
labels.set(u)
- if err == nil && p.validateLabels {
+ if err == nil && p.fromPuny != nil {
err = p.fromPuny(p, u)
}
if err == nil {
@@ -629,16 +662,18 @@
if p.bidirule != nil && !p.bidirule(s) {
return &labelError{s, "B"}
}
- if !p.validateLabels {
+ if p.checkHyphens {
+ if len(s) > 4 && s[2] == '-' && s[3] == '-' {
+ return &labelError{s, "V2"}
+ }
+ if s[0] == '-' || s[len(s)-1] == '-' {
+ return &labelError{s, "V3"}
+ }
+ }
+ if !p.checkJoiners {
return nil
}
- trie := p.trie // p.validateLabels is only set if trie is set.
- if len(s) > 4 && s[2] == '-' && s[3] == '-' {
- return &labelError{s, "V2"}
- }
- if s[0] == '-' || s[len(s)-1] == '-' {
- return &labelError{s, "V3"}
- }
+ trie := p.trie // p.checkJoiners is only set if trie is set.
// TODO: merge the use of this in the trie.
v, sz := trie.lookupString(s)
x := info(v)
