From 4b54582deed916c7f716ea01a7cf44112e6945c7 Mon Sep 17 00:00:00 2001
From: yuin <yuin@inforno.net>
Date: Thu, 2 Jan 2020 01:36:46 +0900
Subject: [PATCH] Fixes #83, Adds options for Linkify

---
 README.md                   |  70 +++++++++++---
 extension/_test/linkify.txt |   8 ++
 extension/linkify.go        | 179 +++++++++++++++++++++++++++++++++---
 extension/linkify_test.go   |  79 ++++++++++++++++
 4 files changed, 308 insertions(+), 28 deletions(-)

diff --git a/README.md b/README.md
index 9e61344..0bde9c5 100644
--- a/README.md
+++ b/README.md
@@ -64,8 +64,8 @@ Import packages:
 
 ```
 import (
-	"bytes"
-	"github.com/yuin/goldmark"
+    "bytes"
+    "github.com/yuin/goldmark"
 )
 ```
 
@@ -105,11 +105,11 @@ Custom parser and renderer
 --------------------------
 ```go
 import (
-	"bytes"
-	"github.com/yuin/goldmark"
-	"github.com/yuin/goldmark/extension"
-	"github.com/yuin/goldmark/parser"
-	"github.com/yuin/goldmark/renderer/html"
+    "bytes"
+    "github.com/yuin/goldmark"
+    "github.com/yuin/goldmark/extension"
+    "github.com/yuin/goldmark/parser"
+    "github.com/yuin/goldmark/renderer/html"
 )
 
 md := goldmark.New(
@@ -215,14 +215,53 @@ You can overwrite the substitutions by `extensions.WithTypographicSubstitutions`
 
 ```go
 markdown := goldmark.New(
-	goldmark.WithExtensions(
-		extension.NewTypographer(
-			extension.WithTypographicSubstitutions(extension.TypographicSubstitutions{
-				extension.LeftSingleQuote:  []byte("&sbquo;"),
-				extension.RightSingleQuote: nil, // nil disables a substitution
-			}),
-		),
-	),
+    goldmark.WithExtensions(
+        extension.NewTypographer(
+            extension.WithTypographicSubstitutions(extension.TypographicSubstitutions{
+                extension.LeftSingleQuote:  []byte("&sbquo;"),
+                extension.RightSingleQuote: nil, // nil disables a substitution
+            }),
+        ),
+    ),
+)
+```
+
+### Linkify extension
+Linkify extension implements [Autolinks(extension)](https://github.github.com/gfm/#autolinks-extension-)
+defined in [GitHub Flavored Markdown Spec](https://github.github.com/gfm/).
+
+Since spec does not define details about URL, there are many ambiguous cases.
+
+You can overwrite autolinking patterns by options.
+
+| Functional option | Type | Description |
+| ----------------- | ---- | ----------- |
+| `extension.WithLinkifyAllowedProtocols` | `[][]byte` | List of allowed protocols such as `[][]byte{ []byte("http:") }` |
+| `extension.WithLinkifyURLRegexp` | `*regexp.Regexp` | Regexp that defines URL including protocols |
+| `extension.WithLinkifyWWWRegexp` | `*regexp.Regexp` | Regexp that defines URL starting with `www.`. This pattern corresponds to [the extended www autolink](https://github.github.com/gfm/#extended-www-autolink) |
+| `extension.WithLinkifyEmailRegexp` | `*regexp.Regexp` | Regexp that defines email address` |
+
+Example: using [xurls](https://github.com/mvdan/xurls)
+
+```go
+import "mvdan.cc/xurls/v2"
+
+markdown := goldmark.New(
+    goldmark.WithRendererOptions(
+        html.WithXHTML(),
+        html.WithUnsafe(),
+    ),
+    goldmark.WithExtensions(
+        extension.NewLinkify(
+            extension.WithLinkifyAllowedProtocols([][]byte{
+                []byte("http:"),
+                []byte("https:"),
+            }),
+            extension.WithLinkifyURLRegexp(
+                xurls.Strict(),
+            ),
+        ),
+    ),
 )
 ```
 
@@ -317,6 +356,7 @@ AST nodes do not have concrete text. AST nodes have segment information of the d
 
 `text.Segment` has 3 attributes: `Start`, `End`, `Padding` .
 
+(TBC)
 
 **TODO**
 
diff --git a/extension/_test/linkify.txt b/extension/_test/linkify.txt
index 669ab72..1e75aeb 100644
--- a/extension/_test/linkify.txt
+++ b/extension/_test/linkify.txt
@@ -153,3 +153,11 @@ This is a `git@github.com:vim/vim`
 //- - - - - - - - -//
 <p>This is a <code>git@github.com:vim/vim</code></p>
 //= = = = = = = = = = = = = = = = = = = = = = = =//
+
+
+16
+//- - - - - - - - -//
+https://nic.college
+//- - - - - - - - -//
+<p><a href="https://nic.college">https://nic.college</a></p>
+//= = = = = = = = = = = = = = = = = = = = = = = =//
diff --git a/extension/linkify.go b/extension/linkify.go
index 0a584e8..565c6bd 100644
--- a/extension/linkify.go
+++ b/extension/linkify.go
@@ -2,27 +2,153 @@ package extension
 
 import (
 	"bytes"
+	"regexp"
+
 	"github.com/yuin/goldmark"
 	"github.com/yuin/goldmark/ast"
 	"github.com/yuin/goldmark/parser"
 	"github.com/yuin/goldmark/text"
 	"github.com/yuin/goldmark/util"
-	"regexp"
 )
 
-var wwwURLRegxp = regexp.MustCompile(`^www\.[-a-zA-Z0-9@:%._\+~#=]{2,256}\.[a-z]{2,6}((?:/|[#?])[-a-zA-Z0-9@:%_\+.~#!?&//=\(\);,'">\^{}\[\]` + "`" + `]*)?`)
+var wwwURLRegxp = regexp.MustCompile(`^www\.[-a-zA-Z0-9@:%._\+~#=]{2,256}\.[a-z]+(?:(?:/|[#?])[-a-zA-Z0-9@:%_\+.~#!?&//=\(\);,'">\^{}\[\]` + "`" + `]*)?`)
 
-var urlRegexp = regexp.MustCompile(`^(?:http|https|ftp):\/\/(?:www\.)?[-a-zA-Z0-9@:%._\+~#=]{2,256}\.[a-z]{2,6}((?:/|[#?])[-a-zA-Z0-9@:%_+.~#$!?&//=\(\);,'">\^{}\[\]` + "`" + `]*)?`)
+var urlRegexp = regexp.MustCompile(`^(?:http|https|ftp):\/\/(?:www\.)?[-a-zA-Z0-9@:%._\+~#=]{2,256}\.[a-z]+(?:(?:/|[#?])[-a-zA-Z0-9@:%_+.~#$!?&//=\(\);,'">\^{}\[\]` + "`" + `]*)?`)
 
-type linkifyParser struct {
+// An LinkifyConfig struct is a data structure that holds configuration of the
+// Linkify extension.
+type LinkifyConfig struct {
+	AllowedProtocols [][]byte
+	URLRegexp        *regexp.Regexp
+	WWWRegexp        *regexp.Regexp
+	EmailRegexp      *regexp.Regexp
 }
 
-var defaultLinkifyParser = &linkifyParser{}
+const optLinkifyAllowedProtocols parser.OptionName = "LinkifyAllowedProtocols"
+const optLinkifyURLRegexp parser.OptionName = "LinkifyURLRegexp"
+const optLinkifyWWWRegexp parser.OptionName = "LinkifyWWWRegexp"
+const optLinkifyEmailRegexp parser.OptionName = "LinkifyEmailRegexp"
+
+// SetOption implements SetOptioner.
+func (c *LinkifyConfig) SetOption(name parser.OptionName, value interface{}) {
+	switch name {
+	case optLinkifyAllowedProtocols:
+		c.AllowedProtocols = value.([][]byte)
+	case optLinkifyURLRegexp:
+		c.URLRegexp = value.(*regexp.Regexp)
+	case optLinkifyWWWRegexp:
+		c.WWWRegexp = value.(*regexp.Regexp)
+	case optLinkifyEmailRegexp:
+		c.EmailRegexp = value.(*regexp.Regexp)
+	}
+}
+
+// A LinkifyOption interface sets options for the LinkifyOption.
+type LinkifyOption interface {
+	parser.Option
+	SetLinkifyOption(*LinkifyConfig)
+}
+
+type withLinkifyAllowedProtocols struct {
+	value [][]byte
+}
+
+func (o *withLinkifyAllowedProtocols) SetParserOption(c *parser.Config) {
+	c.Options[optLinkifyAllowedProtocols] = o.value
+}
+
+func (o *withLinkifyAllowedProtocols) SetLinkifyOption(p *LinkifyConfig) {
+	p.AllowedProtocols = o.value
+}
+
+// WithLinkifyAllowedProtocols is a functional otpion that specify allowed
+// protocols in autolinks. Each protocol must end with ':' like
+// 'http:' .
+func WithLinkifyAllowedProtocols(value [][]byte) LinkifyOption {
+	return &withLinkifyAllowedProtocols{
+		value: value,
+	}
+}
+
+type withLinkifyURLRegexp struct {
+	value *regexp.Regexp
+}
+
+func (o *withLinkifyURLRegexp) SetParserOption(c *parser.Config) {
+	c.Options[optLinkifyURLRegexp] = o.value
+}
+
+func (o *withLinkifyURLRegexp) SetLinkifyOption(p *LinkifyConfig) {
+	p.URLRegexp = o.value
+}
+
+// WithLinkifyURLRegexp is a functional otpion that specify
+// a pattern of the URL including a protocol.
+func WithLinkifyURLRegexp(value *regexp.Regexp) LinkifyOption {
+	return &withLinkifyURLRegexp{
+		value: value,
+	}
+}
+
+// WithLinkifyWWWRegexp is a functional otpion that specify
+// a pattern of the URL without a protocol.
+// This pattern must start with 'www.' .
+type withLinkifyWWWRegexp struct {
+	value *regexp.Regexp
+}
+
+func (o *withLinkifyWWWRegexp) SetParserOption(c *parser.Config) {
+	c.Options[optLinkifyWWWRegexp] = o.value
+}
+
+func (o *withLinkifyWWWRegexp) SetLinkifyOption(p *LinkifyConfig) {
+	p.WWWRegexp = o.value
+}
+
+func WithLinkifyWWWRegexp(value *regexp.Regexp) LinkifyOption {
+	return &withLinkifyWWWRegexp{
+		value: value,
+	}
+}
+
+// WithLinkifyWWWRegexp is a functional otpion that specify
+// a pattern of the email address.
+type withLinkifyEmailRegexp struct {
+	value *regexp.Regexp
+}
+
+func (o *withLinkifyEmailRegexp) SetParserOption(c *parser.Config) {
+	c.Options[optLinkifyEmailRegexp] = o.value
+}
+
+func (o *withLinkifyEmailRegexp) SetLinkifyOption(p *LinkifyConfig) {
+	p.EmailRegexp = o.value
+}
+
+func WithLinkifyEmailRegexp(value *regexp.Regexp) LinkifyOption {
+	return &withLinkifyEmailRegexp{
+		value: value,
+	}
+}
+
+type linkifyParser struct {
+	LinkifyConfig
+}
 
 // NewLinkifyParser return a new InlineParser can parse
 // text that seems like a URL.
-func NewLinkifyParser() parser.InlineParser {
-	return defaultLinkifyParser
+func NewLinkifyParser(opts ...LinkifyOption) parser.InlineParser {
+	p := &linkifyParser{
+		LinkifyConfig: LinkifyConfig{
+			AllowedProtocols: nil,
+			URLRegexp:        urlRegexp,
+			WWWRegexp:        wwwURLRegxp,
+		},
+	}
+	for _, o := range opts {
+		o.SetLinkifyOption(&p.LinkifyConfig)
+	}
+	return p
 }
 
 func (s *linkifyParser) Trigger() []byte {
@@ -53,14 +179,26 @@ func (s *linkifyParser) Parse(parent ast.Node, block text.Reader, pc parser.Cont
 	var m []int
 	var protocol []byte
 	var typ ast.AutoLinkType = ast.AutoLinkURL
-	if bytes.HasPrefix(line, protoHTTP) || bytes.HasPrefix(line, protoHTTPS) || bytes.HasPrefix(line, protoFTP) {
-		m = urlRegexp.FindSubmatchIndex(line)
+	if s.LinkifyConfig.AllowedProtocols == nil {
+		if bytes.HasPrefix(line, protoHTTP) || bytes.HasPrefix(line, protoHTTPS) || bytes.HasPrefix(line, protoFTP) {
+			m = s.LinkifyConfig.URLRegexp.FindSubmatchIndex(line)
+		}
+	} else {
+		for _, prefix := range s.LinkifyConfig.AllowedProtocols {
+			if bytes.HasPrefix(line, prefix) {
+				m = s.LinkifyConfig.URLRegexp.FindSubmatchIndex(line)
+				break
+			}
+		}
 	}
 	if m == nil && bytes.HasPrefix(line, domainWWW) {
-		m = wwwURLRegxp.FindSubmatchIndex(line)
+		m = s.LinkifyConfig.WWWRegexp.FindSubmatchIndex(line)
 		protocol = []byte("http")
 	}
-	if m != nil {
+	if m != nil && m[0] != 0 {
+		m = nil
+	}
+	if m != nil && m[0] == 0 {
 		lastChar := line[m[1]-1]
 		if lastChar == '.' {
 			m[1]--
@@ -96,7 +234,15 @@ func (s *linkifyParser) Parse(parent ast.Node, block text.Reader, pc parser.Cont
 			return nil
 		}
 		typ = ast.AutoLinkEmail
-		stop := util.FindEmailIndex(line)
+		stop := -1
+		if s.LinkifyConfig.EmailRegexp == nil {
+			stop = util.FindEmailIndex(line)
+		} else {
+			m := s.LinkifyConfig.EmailRegexp.FindSubmatchIndex(line)
+			if m != nil && m[0] == 0 {
+				stop = m[1]
+			}
+		}
 		if stop < 0 {
 			return nil
 		}
@@ -136,15 +282,22 @@ func (s *linkifyParser) CloseBlock(parent ast.Node, pc parser.Context) {
 }
 
 type linkify struct {
+	options []LinkifyOption
 }
 
 // Linkify is an extension that allow you to parse text that seems like a URL.
 var Linkify = &linkify{}
 
+func NewLinkify(opts ...LinkifyOption) goldmark.Extender {
+	return &linkify{
+		options: opts,
+	}
+}
+
 func (e *linkify) Extend(m goldmark.Markdown) {
 	m.Parser().AddOptions(
 		parser.WithInlineParsers(
-			util.Prioritized(NewLinkifyParser(), 999),
+			util.Prioritized(NewLinkifyParser(e.options...), 999),
 		),
 	)
 }
diff --git a/extension/linkify_test.go b/extension/linkify_test.go
index 1abefad..18d1331 100644
--- a/extension/linkify_test.go
+++ b/extension/linkify_test.go
@@ -1,6 +1,7 @@
 package extension
 
 import (
+	"regexp"
 	"testing"
 
 	"github.com/yuin/goldmark"
@@ -19,3 +20,81 @@ func TestLinkify(t *testing.T) {
 	)
 	testutil.DoTestCaseFile(markdown, "_test/linkify.txt", t)
 }
+
+func TestLinkifyWithAllowedProtocols(t *testing.T) {
+	markdown := goldmark.New(
+		goldmark.WithRendererOptions(
+			html.WithXHTML(),
+			html.WithUnsafe(),
+		),
+		goldmark.WithExtensions(
+			NewLinkify(
+				WithLinkifyAllowedProtocols([][]byte{
+					[]byte("ssh:"),
+				}),
+				WithLinkifyURLRegexp(
+					regexp.MustCompile(`\w+://[^\s]+`),
+				),
+			),
+		),
+	)
+	testutil.DoTestCase(
+		markdown,
+		testutil.MarkdownTestCase{
+			No:       1,
+			Markdown: `hoge ssh://user@hoge.com. http://example.com/`,
+			Expected: `<p>hoge <a href="ssh://user@hoge.com">ssh://user@hoge.com</a>. http://example.com/</p>`,
+		},
+		t,
+	)
+}
+
+func TestLinkifyWithWWWRegexp(t *testing.T) {
+	markdown := goldmark.New(
+		goldmark.WithRendererOptions(
+			html.WithXHTML(),
+			html.WithUnsafe(),
+		),
+		goldmark.WithExtensions(
+			NewLinkify(
+				WithLinkifyWWWRegexp(
+					regexp.MustCompile(`www\.example\.com`),
+				),
+			),
+		),
+	)
+	testutil.DoTestCase(
+		markdown,
+		testutil.MarkdownTestCase{
+			No:       1,
+			Markdown: `www.google.com www.example.com`,
+			Expected: `<p>www.google.com <a href="http://www.example.com">www.example.com</a></p>`,
+		},
+		t,
+	)
+}
+
+func TestLinkifyWithEmailRegexp(t *testing.T) {
+	markdown := goldmark.New(
+		goldmark.WithRendererOptions(
+			html.WithXHTML(),
+			html.WithUnsafe(),
+		),
+		goldmark.WithExtensions(
+			NewLinkify(
+				WithLinkifyEmailRegexp(
+					regexp.MustCompile(`user@example\.com`),
+				),
+			),
+		),
+	)
+	testutil.DoTestCase(
+		markdown,
+		testutil.MarkdownTestCase{
+			No:       1,
+			Markdown: `hoge@example.com user@example.com`,
+			Expected: `<p>hoge@example.com <a href="mailto:user@example.com">user@example.com</a></p>`,
+		},
+		t,
+	)
+}