jason/atom
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases 14 Wiki Activity

Change URI to IRI after finding validation pattern

Browse Source
This commit is contained in:
Jason Streifling 2024-10-16 17:33:25 +02:00
parent b76e529ca3
commit 082e71a698
10 changed files with 41 additions and 32 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

35
atom.go
View File

@ -2,7 +2,6 @@ package atomfeed
import ( import (
"mime" "mime"
"net/url"
"regexp" "regexp"
"strings" "strings"
) )
@ -11,21 +10,31 @@ type (
EmailAddress string EmailAddress string
LanguageTag string LanguageTag string
MediaType string MediaType string
URI string // URI string
IRI string
) )
func isValidURL(uri URI) bool { // func isValidURL(uri URI) bool {
_, err := url.ParseRequestURI(string(uri)) // _, err := url.ParseRequestURI(string(uri))
return err == nil // return err == nil
} // }
//
// func isValidURN(uri URI) bool {
// pattern := `\A(?i:urn:(?!urn:)(?<nid>[a-z0-9][a-z0-9-]{1,31}):(?<nss>(?:[-a-z0-9()+,.:=@;$_!*'&~\/]|%[0-9a-f]{2})+)(?:\?\+(?<rcomponent>.*?))?(?:\?=(?<qcomponent>.*?))?(?:#(?<fcomponent>.*?))?)\z`
// return regexp.MustCompile(pattern).MatchString(string(uri))
// }
//
// // isValidURI checks whether an URI is valid or not.
// func isValidURI(uri URI) bool {
// return isValidURL(uri) || isValidURN(uri)
// }
func isValidURN(uri URI) bool { // isValidIRI checks whether an IRI is valid or not.
pattern := `\A(?i:urn:(?!urn:)(?<nid>[a-z0-9][a-z0-9-]{1,31}):(?<nss>(?:[-a-z0-9()+,.:=@;$_!*'&~\/]|%[0-9a-f]{2})+)(?:\?\+(?<rcomponent>.*?))?(?:\?=(?<qcomponent>.*?))?(?:#(?<fcomponent>.*?))?)\z` // The used pattern stems from
return regexp.MustCompile(pattern).MatchString(string(uri)) // https://www.w3.org/2011/04/XMLSchema/TypeLibrary-IRI-RFC3987.xsd
} func isValidIRI(iri IRI) bool {
pattern := `((([A-Za-z])[A-Za-z0-9+\-\.]*):((//(((([A-Za-z0-9\-\._~ -퟿豈-﷏ﷰ-￯𐀀-🿽𠀀-𯿽𰀀-𿿽񀀀-񏿽񐀀-񟿽񠀀-񯿽񰀀-񿿽򀀀-򏿽򐀀-򟿽򠀀-򯿽򰀀-򿿽󀀀-󏿽󐀀-󟿽󡀀-󯿽!$&'()*+,;=:]|(%[0-9A-Fa-f][0-9A-Fa-f]))*@))?((\[((((([0-9A-Fa-f]{0,4}:)){6}(([0-9A-Fa-f]{0,4}:[0-9A-Fa-f]{0,4})|(([0-9]|([1-9][0-9])|(1([0-9]){2})|(2[0-4][0-9])|(25[0-5]))\.([0-9]|([1-9][0-9])|(1([0-9]){2})|(2[0-4][0-9])|(25[0-5]))\.([0-9]|([1-9][0-9])|(1([0-9]){2})|(2[0-4][0-9])|(25[0-5]))\.([0-9]|([1-9][0-9])|(1([0-9]){2})|(2[0-4][0-9])|(25[0-5])))))|(::(([0-9A-Fa-f]{0,4}:)){5}(([0-9A-Fa-f]{0,4}:[0-9A-Fa-f]{0,4})|(([0-9]|([1-9][0-9])|(1([0-9]){2})|(2[0-4][0-9])|(25[0-5]))\.([0-9]|([1-9][0-9])|(1([0-9]){2})|(2[0-4][0-9])|(25[0-5]))\.([0-9]|([1-9][0-9])|(1([0-9]){2})|(2[0-4][0-9])|(25[0-5]))\.([0-9]|([1-9][0-9])|(1([0-9]){2})|(2[0-4][0-9])|(25[0-5])))))|(([0-9A-Fa-f]{0,4})?::(([0-9A-Fa-f]{0,4}:)){4}(([0-9A-Fa-f]{0,4}:[0-9A-Fa-f]{0,4})|(([0-9]|([1-9][0-9])|(1([0-9]){2})|(2[0-4][0-9])|(25[0-5]))\.([0-9]|([1-9][0-9])|(1([0-9]){2})|(2[0-4][0-9])|(25[0-5]))\.([0-9]|([1-9][0-9])|(1([0-9]){2})|(2[0-4][0-9])|(25[0-5]))\.([0-9]|([1-9][0-9])|(1([0-9]){2})|(2[0-4][0-9])|(25[0-5])))))|((((([0-9A-Fa-f]{0,4}:))?[0-9A-Fa-f]{0,4}))?::(([0-9A-Fa-f]{0,4}:)){3}(([0-9A-Fa-f]{0,4}:[0-9A-Fa-f]{0,4})|(([0-9]|([1-9][0-9])|(1([0-9]){2})|(2[0-4][0-9])|(25[0-5]))\.([0-9]|([1-9][0-9])|(1([0-9]){2})|(2[0-4][0-9])|(25[0-5]))\.([0-9]|([1-9][0-9])|(1([0-9]){2})|(2[0-4][0-9])|(25[0-5]))\.([0-9]|([1-9][0-9])|(1([0-9]){2})|(2[0-4][0-9])|(25[0-5])))))|((((([0-9A-Fa-f]{0,4}:)){0,2}[0-9A-Fa-f]{0,4}))?::(([0-9A-Fa-f]{0,4}:)){2}(([0-9A-Fa-f]{0,4}:[0-9A-Fa-f]{0,4})|(([0-9]|([1-9][0-9])|(1([0-9]){2})|(2[0-4][0-9])|(25[0-5]))\.([0-9]|([1-9][0-9])|(1([0-9]){2})|(2[0-4][0-9])|(25[0-5]))\.([0-9]|([1-9][0-9])|(1([0-9]){2})|(2[0-4][0-9])|(25[0-5]))\.([0-9]|([1-9][0-9])|(1([0-9]){2})|(2[0-4][0-9])|(25[0-5])))))|((((([0-9A-Fa-f]{0,4}:)){0,3}[0-9A-Fa-f]{0,4}))?::[0-9A-Fa-f]{0,4}:(([0-9A-Fa-f]{0,4}:[0-9A-Fa-f]{0,4})|(([0-9]|([1-9][0-9])|(1([0-9]){2})|(2[0-4][0-9])|(25[0-5]))\.([0-9]|([1-9][0-9])|(1([0-9]){2})|(2[0-4][0-9])|(25[0-5]))\.([0-9]|([1-9][0-9])|(1([0-9]){2})|(2[0-4][0-9])|(25[0-5]))\.([0-9]|([1-9][0-9])|(1([0-9]){2})|(2[0-4][0-9])|(25[0-5])))))|((((([0-9A-Fa-f]{0,4}:)){0,4}[0-9A-Fa-f]{0,4}))?::(([0-9A-Fa-f]{0,4}:[0-9A-Fa-f]{0,4})|(([0-9]|([1-9][0-9])|(1([0-9]){2})|(2[0-4][0-9])|(25[0-5]))\.([0-9]|([1-9][0-9])|(1([0-9]){2})|(2[0-4][0-9])|(25[0-5]))\.([0-9]|([1-9][0-9])|(1([0-9]){2})|(2[0-4][0-9])|(25[0-5]))\.([0-9]|([1-9][0-9])|(1([0-9]){2})|(2[0-4][0-9])|(25[0-5])))))|((((([0-9A-Fa-f]{0,4}:)){0,5}[0-9A-Fa-f]{0,4}))?::[0-9A-Fa-f]{0,4})|((((([0-9A-Fa-f]{0,4}:)){0,6}[0-9A-Fa-f]{0,4}))?::))|(v[0-9A-Fa-f]+\.[A-Za-z0-9\-\._~!$&'()*+,;=:]+))\])|(([0-9]|([1-9][0-9])|(1([0-9]){2})|(2[0-4][0-9])|(25[0-5]))\.([0-9]|([1-9][0-9])|(1([0-9]){2})|(2[0-4][0-9])|(25[0-5]))\.([0-9]|([1-9][0-9])|(1([0-9]){2})|(2[0-4][0-9])|(25[0-5]))\.([0-9]|([1-9][0-9])|(1([0-9]){2})|(2[0-4][0-9])|(25[0-5])))|(([A-Za-z0-9\-\._~ -퟿豈-﷏ﷰ-￯𐀀-🿽𠀀-𯿽𰀀-𿿽񀀀-񏿽񐀀-񟿽񠀀-񯿽񰀀-񿿽򀀀-򏿽򐀀-򟿽򠀀-򯿽򰀀-򿿽󀀀-󏿽󐀀-󟿽󡀀-󯿽]|(%[0-9A-Fa-f][0-9A-Fa-f])|[!$&'()*+,;=]))*)((:[0-9]*))?)((/(([A-Za-z0-9\-\._~ -퟿豈-﷏ﷰ-￯𐀀-🿽𠀀-𯿽𰀀-𿿽񀀀-񏿽񐀀-񟿽񠀀-񯿽񰀀-񿿽򀀀-򏿽򐀀-򟿽򠀀-򯿽򰀀-򿿽󀀀-󏿽󐀀-󟿽󡀀-󯿽]|(%[0-9A-Fa-f][0-9A-Fa-f])|[!$&'()*+,;=:@]))*))*)|(/(((([A-Za-z0-9\-\._~ -퟿豈-﷏ﷰ-￯𐀀-🿽𠀀-𯿽𰀀-𿿽񀀀-񏿽񐀀-񟿽񠀀-񯿽񰀀-񿿽򀀀-򏿽򐀀-򟿽򠀀-򯿽򰀀-򿿽󀀀-󏿽󐀀-󟿽󡀀-󯿽]|(%[0-9A-Fa-f][0-9A-Fa-f])|[!$&'()*+,;=:@]))+((/(([A-Za-z0-9\-\._~ -퟿豈-﷏ﷰ-￯𐀀-🿽𠀀-𯿽𰀀-𿿽񀀀-񏿽񐀀-񟿽񠀀-񯿽񰀀-񿿽򀀀-򏿽򐀀-򟿽򠀀-򯿽򰀀-򿿽󀀀-󏿽󐀀-󟿽󡀀-󯿽]|(%[0-9A-Fa-f][0-9A-Fa-f])|[!$&'()*+,;=:@]))*))*))?)|((([A-Za-z0-9\-\._~ -퟿豈-﷏ﷰ-￯𐀀-🿽𠀀-𯿽𰀀-𿿽񀀀-񏿽񐀀-񟿽񠀀-񯿽񰀀-񿿽򀀀-򏿽򐀀-򟿽򠀀-򯿽򰀀-򿿽󀀀-󏿽󐀀-󟿽󡀀-󯿽]|(%[0-9A-Fa-f][0-9A-Fa-f])|[!$&'()*+,;=:@]))+((/(([A-Za-z0-9\-\._~ -퟿豈-﷏ﷰ-￯𐀀-🿽𠀀-𯿽𰀀-𿿽񀀀-񏿽񐀀-񟿽񠀀-񯿽񰀀-񿿽򀀀-򏿽򐀀-򟿽򠀀-򯿽򰀀-򿿽󀀀-󏿽󐀀-󟿽󡀀-󯿽]|(%[0-9A-Fa-f][0-9A-Fa-f])|[!$&'()*+,;=:@]))*))*)|)((\?(([A-Za-z0-9\-\._~ -퟿豈-﷏ﷰ-￯𐀀-🿽𠀀-𯿽𰀀-𿿽񀀀-񏿽񐀀-񟿽񠀀-񯿽񰀀-񿿽򀀀-򏿽򐀀-򟿽򠀀-򯿽򰀀-򿿽󀀀-󏿽󐀀-󟿽󡀀-󯿽]|(%[0-9A-Fa-f][0-9A-Fa-f])|[!$&'()*+,;=:@])|[-󰀀-󿿽􀀀-􏿽/?])*))?((#((([A-Za-z0-9\-\._~ -퟿豈-﷏ﷰ-￯𐀀-🿽𠀀-𯿽𰀀-𿿽񀀀-񏿽񐀀-񟿽񠀀-񯿽񰀀-񿿽򀀀-򏿽򐀀-򟿽򠀀-򯿽򰀀-򿿽󀀀-󏿽󐀀-󟿽󡀀-󯿽]|(%[0-9A-Fa-f][0-9A-Fa-f])|[!$&'()*+,;=:@])|/|\?))*))?)`
func isValidURI(uri URI) bool { return regexp.MustCompile(pattern).MatchString(string(iri))
return isValidURL(uri) || isValidURN(uri)
} }
func isCorrectlyEscaped(text string) bool { func isCorrectlyEscaped(text string) bool {

4
category.go
View File

@ -10,7 +10,7 @@ type Category struct {
*CommonAttributes *CommonAttributes
Content Content `xml:"content"` // undefinedContent in RFC4287 Content Content `xml:"content"` // undefinedContent in RFC4287
Term string `xml:"term,attr"` Term string `xml:"term,attr"`
Scheme URI `xml:"scheme,attr,omitempty"` Scheme IRI `xml:"scheme,attr,omitempty"`
Label string `xml:"label,attr,omitempty"` Label string `xml:"label,attr,omitempty"`
} }
@ -33,7 +33,7 @@ func (c *Category) Check() error {
} }
if c.Scheme != "" { if c.Scheme != "" {
if !isValidURI(c.Scheme) { if !isValidIRI(c.Scheme) {
return fmt.Errorf("scheme attribute %v of category not correctly formatted", c.Scheme) return fmt.Errorf("scheme attribute %v of category not correctly formatted", c.Scheme)
} }
} }

2
commonAttributes.go
View File

@ -3,7 +3,7 @@ package atomfeed
import "fmt" import "fmt"
type CommonAttributes struct { type CommonAttributes struct {
Base URI `xml:"base,attr,omitempty"` Base IRI `xml:"base,attr,omitempty"`
Lang LanguageTag `xml:"lang,attr,omitempty"` Lang LanguageTag `xml:"lang,attr,omitempty"`
UndefinedAttributes []*ExtensionAttribute `xml:",any"` UndefinedAttributes []*ExtensionAttribute `xml:",any"`
} }

4
generator.go
View File

@ -8,7 +8,7 @@ import (
type Generator struct { type Generator struct {
*CommonAttributes *CommonAttributes
URI URI `xml:"uri,attr,omitempty"` URI IRI `xml:"uri,attr,omitempty"`
Version string `xml:"version,attr,omitempty"` Version string `xml:"version,attr,omitempty"`
Text string `xml:"text"` Text string `xml:"text"`
} }
@ -19,7 +19,7 @@ func NewGenerator(text string) *Generator {
func (g *Generator) Check() error { func (g *Generator) Check() error {
if g.URI != "" { if g.URI != "" {
if !isValidURI(g.URI) { if !isValidIRI(g.URI) {
return fmt.Errorf("uri attribute %v of generator not correctly formatted", g.URI) return fmt.Errorf("uri attribute %v of generator not correctly formatted", g.URI)
} }
} }

6
icon.go
View File

@ -7,18 +7,18 @@ import (
type Icon struct { type Icon struct {
*CommonAttributes *CommonAttributes
URI URI `xml:"uri"` URI IRI `xml:"uri"`
} }
func NewIcon(uri string) *Icon { func NewIcon(uri string) *Icon {
return &Icon{URI: URI(uri)} return &Icon{URI: IRI(uri)}
} }
func (i *Icon) Check() error { func (i *Icon) Check() error {
if i.URI == "" { if i.URI == "" {
return errors.New("uri element of icon empty") return errors.New("uri element of icon empty")
} else { } else {
if !isValidURI(i.URI) { if !isValidIRI(i.URI) {
return fmt.Errorf("uri attribute %v of icon not correctly formatted", i.URI) return fmt.Errorf("uri attribute %v of icon not correctly formatted", i.URI)
} }
} }

6
id.go
View File

@ -9,18 +9,18 @@ import (
type ID struct { type ID struct {
*CommonAttributes *CommonAttributes
URI URI `xml:"uri"` URI IRI `xml:"uri"`
} }
func NewID() *ID { func NewID() *ID {
return &ID{URI: URI(fmt.Sprint("urn:uuid:", uuid.New()))} return &ID{URI: IRI(fmt.Sprint("urn:uuid:", uuid.New()))}
} }
func (i *ID) Check() error { func (i *ID) Check() error {
if i.URI == "" { if i.URI == "" {
return errors.New("uri element of id empty") return errors.New("uri element of id empty")
} else { } else {
if !isValidURI(i.URI) { if !isValidIRI(i.URI) {
return fmt.Errorf("uri element %v of id not correctly formatted", i.URI) return fmt.Errorf("uri element %v of id not correctly formatted", i.URI)
} }
} }

4
link.go
View File

@ -9,7 +9,7 @@ type Link struct {
*CommonAttributes *CommonAttributes
Title Text `xml:"title,attr,omitempty"` Title Text `xml:"title,attr,omitempty"`
Content Content `xml:"content"` // undefinedContent in RFC4287 Content Content `xml:"content"` // undefinedContent in RFC4287
Href URI `xml:"href,attr"` Href IRI `xml:"href,attr"`
Rel string `xml:"rel,attr,omitempty"` Rel string `xml:"rel,attr,omitempty"`
Type MediaType `xml:"type,attr,omitempty"` Type MediaType `xml:"type,attr,omitempty"`
HrefLang LanguageTag `xml:"hreflang,attr,omitempty"` HrefLang LanguageTag `xml:"hreflang,attr,omitempty"`
@ -22,7 +22,7 @@ func NewLink(href string) (*Link, error) {
return nil, fmt.Errorf("error creating content element: %v", err) return nil, fmt.Errorf("error creating content element: %v", err)
} }
return &Link{Href: URI(href), Content: content}, nil return &Link{Href: IRI(href), Content: content}, nil
} }
func (l *Link) Check() error { func (l *Link) Check() error {

2
logo.go
View File

@ -4,7 +4,7 @@ import "errors"
type Logo struct { type Logo struct {
*CommonAttributes *CommonAttributes
URI URI `xml:"uri"` URI IRI `xml:"uri"`
} }
func (l *Logo) Check() error { func (l *Logo) Check() error {

6
outOfLineContent.go
View File

@ -10,7 +10,7 @@ import (
type OutOfLineContent struct { type OutOfLineContent struct {
*CommonAttributes *CommonAttributes
Type MediaType `xml:"type,attr,omitempty"` Type MediaType `xml:"type,attr,omitempty"`
SRC URI `xml:"src,attr"` SRC IRI `xml:"src,attr"`
} }
func newOutOfLineContent(mediaType string, content any) (*OutOfLineContent, error) { func newOutOfLineContent(mediaType string, content any) (*OutOfLineContent, error) {
@ -22,11 +22,11 @@ func newOutOfLineContent(mediaType string, content any) (*OutOfLineContent, erro
return nil, fmt.Errorf("content type %T incompatible with out of line content", content) return nil, fmt.Errorf("content type %T incompatible with out of line content", content)
} }
if !isValidURI(content.(URI)) { if !isValidIRI(content.(IRI)) {
return nil, errors.New("content not a valid uri") return nil, errors.New("content not a valid uri")
} }
return &OutOfLineContent{Type: MediaType(mediaType), SRC: content.(URI)}, nil return &OutOfLineContent{Type: MediaType(mediaType), SRC: content.(IRI)}, nil
} }
func (o *OutOfLineContent) isContent() bool { return true } func (o *OutOfLineContent) isContent() bool { return true }

4
person.go
View File

@ -10,7 +10,7 @@ import (
type Person struct { type Person struct {
*CommonAttributes *CommonAttributes
Name string `xml:"name"` Name string `xml:"name"`
URI URI `xml:"uri,omitempty"` URI IRI `xml:"uri,omitempty"`
Email EmailAddress `xml:"email,omitempty"` Email EmailAddress `xml:"email,omitempty"`
Extensions []*ExtensionElement `xml:",any,omitempty"` Extensions []*ExtensionElement `xml:",any,omitempty"`
} }
@ -29,7 +29,7 @@ func (p *Person) Check() error {
} }
if p.URI != "" { if p.URI != "" {
if !isValidURI(p.URI) { if !isValidIRI(p.URI) {
return fmt.Errorf("uri element of person %v not correctly formatted", p.Name) return fmt.Errorf("uri element of person %v not correctly formatted", p.Name)
} }
} }