Add link metadata fetching package

vendor/github.com/golang-commonmark/markdown/linkify/charset.go

@@ -0,0 +1,50 @@
+// Copyright 2015 The Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+package linkify
+
+import "unicode"
+
+var (
+	trigger    [256]bool
+	unreserved [256]bool
+	subdelims  [256]bool
+	emailcs    [256]bool
+	basicPunct [256]bool
+)
+
+func init() {
+	for _, b := range "-._~" {
+		unreserved[b] = true
+	}
+	for _, b := range "!$&'()*+,;=" {
+		subdelims[b] = true
+	}
+	for _, b := range "abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789!#$%&'*+/=?^_`{|}~-" {
+		emailcs[b] = true
+	}
+	for _, b := range ".,?!;:" {
+		basicPunct[b] = true
+	}
+}
+
+func isAllowedInEmail(r rune) bool {
+	return r < 0x7f && emailcs[r]
+}
+
+func isLetterOrDigit(r rune) bool {
+	return unicode.In(r, unicode.Letter, unicode.Digit)
+}
+
+func isPunctOrSpaceOrControl(r rune) bool {
+	return r == '<' || r == '>' || unicode.In(r, unicode.Punct, unicode.Space, unicode.Cc)
+}
+
+func isUnreserved(r rune) bool {
+	return (r < 0x7f && unreserved[r]) || isLetterOrDigit(r)
+}
+
+func isSubDelimiter(r rune) bool {
+	return r < 0x7f && subdelims[r]
+}

vendor/github.com/golang-commonmark/markdown/linkify/email.go

@@ -0,0 +1,93 @@
+// Copyright 2015 The Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+package linkify
+
+import (
+	"unicode"
+	"unicode/utf8"
+)
+
+func findEmailStart(s string, start int) (_ int, _ bool) {
+	end := start
+	allowDot := false
+	for end >= 0 {
+		b := s[end]
+		switch {
+		case emailcs[b]:
+			allowDot = true
+		case b == '.':
+			if !allowDot {
+				return
+			}
+			allowDot = false
+		default:
+			if end == start {
+				return
+			}
+			if s[end+1] == '.' {
+				return
+			}
+			r, _ := utf8.DecodeLastRuneInString(s[:end+1])
+			if r == utf8.RuneError {
+				return
+			}
+			if !unicode.IsSpace(r) {
+				return
+			}
+			return end + 1, true
+		}
+		end--
+	}
+	if end < start && s[end+1] == '.' {
+		return
+	}
+	return end + 1, true
+}
+
+func findEmailEnd(s string, start int) (_ int, _ bool) {
+	end := start
+	allowDot := false
+loop:
+	for end < len(s) {
+		b := s[end]
+		switch {
+		case emailcs[b]:
+			allowDot = true
+		case b == '.':
+			if !allowDot {
+				return
+			}
+			allowDot = false
+		case b == '@':
+			break loop
+		default:
+			return
+		}
+		end++
+	}
+	if end >= len(s)-5 {
+		return
+	}
+	if end > start && s[end-1] == '.' {
+		return
+	}
+
+	var dot int
+	var ok bool
+	end, dot, ok = findHostnameEnd(s, end+1)
+	if !ok || dot == -1 {
+		return
+	}
+
+	if dot+5 <= len(s) && s[dot+1:dot+5] == "xn--" {
+		return end, true
+	}
+
+	if length := match(s[dot+1:]); dot+length+1 != end {
+		return
+	}
+
+	return end, true
+}

vendor/github.com/golang-commonmark/markdown/linkify/linkify.go

@@ -0,0 +1,463 @@
+// Copyright 2015 The Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+// Package linkify provides a way to find links in plain text.
+package linkify
+
+import (
+	"unicode/utf8"
+
+	"github.com/golang-commonmark/markdown/byteutil"
+)
+
+// Link represents a link found in a string with a schema and a position in the string.
+type Link struct {
+	Scheme     string
+	Start, End int
+}
+
+func max(a, b int) int {
+	if a >= b {
+		return a
+	}
+	return b
+}
+
+// Links returns links found in s.
+func Links(s string) (links []Link) {
+	for i := 0; i < len(s)-2; i++ {
+		switch s[i] {
+		case '.': // IP address or domain name
+			if i == 0 {
+				continue // . at the start of a line
+			}
+			if length := match(s[i+1:]); length > 0 {
+				pos := i + 1 + length
+				switch s[pos-1] {
+				case '.': // IP address
+					if pos >= len(s) {
+						continue // . at the end of line
+					}
+					if !byteutil.IsDigit(s[i-1]) {
+						i = pos
+						continue // . should be preceded by a digit
+					}
+					if !byteutil.IsDigit(s[pos]) {
+						i = pos
+						continue // . should be followed by a digit
+					}
+
+					// find the start of the IP address
+					j := i - 2
+					m := max(0, j-3)
+					for j >= m && byteutil.IsDigit(s[j]) {
+						j--
+					}
+					if i-2-j > 2 {
+						i = pos + 1
+						continue // at most 3 digits
+					}
+					start := 0
+					if j >= 0 {
+						r, rlen := utf8.DecodeLastRuneInString(s[:j+1])
+						if !isPunctOrSpaceOrControl(r) {
+							i = pos + 1
+							continue
+						}
+						switch r {
+						case '.', ':', '/', '\\', '-', '_':
+							i = pos + 1
+							continue
+						}
+						start = j + 2 - rlen
+					}
+
+					length, ok := skipIPv4(s[start:])
+					if !ok {
+						i = pos + 1
+						continue
+					}
+					end := start + length
+					if end == len(s) {
+						links = append(links, Link{
+							Scheme: "",
+							Start:  start,
+							End:    end,
+						})
+						return
+					}
+
+					r, _ := utf8.DecodeRuneInString(s[end:])
+					if !isPunctOrSpaceOrControl(r) {
+						continue
+					}
+
+					end = skipPort(s, end)
+					end = skipPath(s, end)
+					end = skipQuery(s, end)
+					end = skipFragment(s, end)
+					end = unskipPunct(s, end)
+
+					if end < len(s) {
+						r, _ = utf8.DecodeRuneInString(s[end:])
+						if !isPunctOrSpaceOrControl(r) || r == '%' {
+							continue
+						}
+					}
+
+					links = append(links, Link{
+						Scheme: "",
+						Start:  start,
+						End:    end,
+					})
+					i = end
+
+				default: // domain name
+					r, _ := utf8.DecodeLastRuneInString(s[:i])
+					if !isLetterOrDigit(r) {
+						continue // should be preceded by a letter or a digit
+					}
+
+					if pos == len(s) {
+						start, ok := findHostnameStart(s, i)
+						if !ok {
+							continue
+						}
+						links = append(links, Link{
+							Scheme: "",
+							Start:  start,
+							End:    pos,
+						})
+						return
+					}
+
+					if s[i+1:pos] != "xn--" {
+						r, _ = utf8.DecodeRuneInString(s[pos:])
+						if isLetterOrDigit(r) {
+							continue // should not be followed by a letter or a digit
+						}
+					}
+
+					end, dot, ok := findHostnameEnd(s, pos)
+					if !ok {
+						continue
+					}
+					dot = max(dot, i)
+
+					if !(dot+5 <= len(s) && s[dot+1:dot+5] == "xn--") {
+						if length := match(s[dot+1:]); dot+length+1 != end {
+							continue
+						}
+					}
+
+					start, ok := findHostnameStart(s, i)
+					if !ok {
+						continue
+					}
+
+					end = skipPort(s, end)
+					end = skipPath(s, end)
+					end = skipQuery(s, end)
+					end = skipFragment(s, end)
+					end = unskipPunct(s, end)
+
+					if end < len(s) {
+						r, _ = utf8.DecodeRuneInString(s[end:])
+						if !isPunctOrSpaceOrControl(r) || r == '%' {
+							continue // should be followed by punctuation or space
+						}
+					}
+
+					links = append(links, Link{
+						Scheme: "",
+						Start:  start,
+						End:    end,
+					})
+					i = end
+				}
+			}
+
+		case '/': // schema-less link
+			if s[i+1] != '/' {
+				continue
+			}
+
+			if i > 0 {
+				if s[i-1] == ':' {
+					i++
+					continue // should not be preceded by a colon
+				}
+				r, _ := utf8.DecodeLastRuneInString(s[:i])
+				if !isPunctOrSpaceOrControl(r) {
+					i++
+					continue // should be preceded by punctuation or space
+				}
+			}
+
+			r, _ := utf8.DecodeRuneInString(s[i+2:])
+			if !isLetterOrDigit(r) {
+				i++
+				continue // should be followed by a letter or a digit
+			}
+
+			start := i
+			end, dot, ok := findHostnameEnd(s, i+2)
+			if !ok {
+				continue
+			}
+			if s[i+2:end] != "localhost" {
+				if dot == -1 {
+					continue // no dot
+				}
+				if length, ok := skipIPv4(s[i+2:]); !ok || i+2+length != end {
+					if length := match(s[dot+1:]); dot+length+1 != end {
+						continue
+					}
+				}
+			}
+
+			end = skipPort(s, end)
+			end = skipPath(s, end)
+			end = skipQuery(s, end)
+			end = skipFragment(s, end)
+			end = unskipPunct(s, end)
+
+			if end < len(s) {
+				r, _ = utf8.DecodeRuneInString(s[end:])
+				if !isPunctOrSpaceOrControl(r) || r == '%' {
+					continue // should be followed by punctuation or space
+				}
+			}
+
+			links = append(links, Link{
+				Scheme: "//",
+				Start:  start,
+				End:    end,
+			})
+			i = end
+
+		case ':': // http, https, ftp, mailto or localhost
+			if i < 3 { // at least ftp:
+				continue
+			}
+
+			if i >= 9 && s[i-1] == 't' && s[i-9:i] == "localhost" {
+				j := i - 9
+				if !byteutil.IsDigit(s[j+10]) {
+					continue
+				}
+				if j > 0 {
+					r, _ := utf8.DecodeLastRuneInString(s[:j])
+					if !isPunctOrSpaceOrControl(r) {
+						i++
+						continue // should be preceded by punctuation or space
+					}
+				}
+
+				start := j
+				pos := j + 9
+				end := skipPort(s, pos)
+				if end == pos {
+					continue // invalid port
+				}
+				end = skipPath(s, end)
+				end = skipQuery(s, end)
+				end = skipFragment(s, end)
+				end = unskipPunct(s, end)
+
+				if end < len(s) {
+					r, _ := utf8.DecodeRuneInString(s[end:])
+					if !isPunctOrSpaceOrControl(r) || r == '%' {
+						i++
+						continue // should be followed by punctuation or space
+					}
+				}
+
+				links = append(links, Link{
+					Scheme: "",
+					Start:  start,
+					End:    end,
+				})
+				i = end
+
+				break
+			}
+
+			j := i - 1
+			var start int
+			var schema string
+
+			switch byteutil.ByteToLower(s[j]) {
+			case 'o': // mailto
+				if j < 5 {
+					continue // too short for mailto
+				}
+				if len(s)-j < 8 {
+					continue // insufficient length after
+				}
+				if byteutil.ToLower(s[j-5:j+2]) != "mailto:" {
+					continue
+				}
+				r, _ := utf8.DecodeLastRuneInString(s[:j-5])
+				if isLetterOrDigit(r) {
+					continue // should not be preceded by a letter or a digit
+				}
+				r, _ = utf8.DecodeRuneInString(s[j+2:])
+				if !isAllowedInEmail(r) {
+					continue // should be followed by a valid e-mail character
+				}
+
+				start = j - 5
+				end, ok := findEmailEnd(s, j+2)
+				if !ok {
+					continue
+				}
+
+				links = append(links, Link{
+					Scheme: "mailto:",
+					Start:  start,
+					End:    end,
+				})
+				i = end
+				continue // continue processing
+
+			case 'p': // http or ftp
+				if len(s)-j < 8 {
+					continue // insufficient length after
+				}
+				switch byteutil.ByteToLower(s[j-2]) {
+				case 'f':
+					if byteutil.ToLower(s[j-2:j+4]) != "ftp://" {
+						continue
+					}
+					start = j - 2
+					schema = "ftp:"
+				case 't':
+					if j < 3 {
+						continue
+					}
+					if byteutil.ToLower(s[j-3:j+4]) != "http://" {
+						continue
+					}
+					start = j - 3
+					schema = "http:"
+				default:
+					continue
+				}
+
+			case 's': // https
+				if j < 4 {
+					continue // too short for https
+				}
+				if len(s)-j < 8 {
+					continue // insufficient length after
+				}
+				start = j - 4
+				if byteutil.ToLower(s[start:j+4]) != "https://" {
+					continue
+				}
+				schema = "https:"
+
+			default:
+				continue
+			}
+
+			// http, https or ftp
+
+			if start > 0 {
+				r, _ := utf8.DecodeLastRuneInString(s[:start])
+				if !isPunctOrSpaceOrControl(r) {
+					continue // should be preceded by punctuation or space
+				}
+			}
+
+			r, _ := utf8.DecodeRuneInString(s[j+4:])
+			if !isLetterOrDigit(r) {
+				continue // should be followed by a letter or a digit
+			}
+
+			end, dot, ok := findHostnameEnd(s, j+4)
+			if !ok {
+				continue
+			}
+			if s[j+4:end] != "localhost" {
+				if dot == -1 {
+					continue // no dot
+				}
+				if length, ok := skipIPv4(s[j+4:]); !ok || j+4+length != end {
+					if !(dot+5 <= len(s) && s[dot+1:dot+5] == "xn--") {
+						if length := match(s[dot+1:]); dot+length+1 != end {
+							continue
+						}
+					}
+				}
+			}
+
+			end = skipPort(s, end)
+			end = skipPath(s, end)
+			end = skipQuery(s, end)
+			end = skipFragment(s, end)
+			end = unskipPunct(s, end)
+
+			if end < len(s) {
+				r, _ = utf8.DecodeRuneInString(s[end:])
+				if !isPunctOrSpaceOrControl(r) || r == '%' {
+					continue // should be followed by punctuation or space
+				}
+			}
+
+			links = append(links, Link{
+				Scheme: schema,
+				Start:  start,
+				End:    end,
+			})
+			i = end
+
+		case '@': // schema-less e-mail
+			if i == 0 {
+				continue // @ at the start of a line
+			}
+
+			if len(s)-i < 5 {
+				continue // insufficient length after
+			}
+
+			r, _ := utf8.DecodeLastRuneInString(s[:i])
+			if !isAllowedInEmail(r) {
+				continue // should be preceded by a valid e-mail character
+			}
+
+			r, _ = utf8.DecodeRuneInString(s[i+1:])
+			if !isLetterOrDigit(r) {
+				continue // should be followed by a letter or a digit
+			}
+
+			start, ok := findEmailStart(s, i-1)
+			if !ok {
+				continue
+			}
+
+			end, dot, ok := findHostnameEnd(s, i+1)
+			if !ok {
+				continue
+			}
+			if dot == -1 {
+				continue // no dot
+			}
+			if !(dot+5 <= len(s) && s[dot+1:dot+5] == "xn--") {
+				if length := match(s[dot+1:]); dot+length+1 != end {
+					continue
+				}
+			}
+
+			links = append(links, Link{
+				Scheme: "mailto:",
+				Start:  start,
+				End:    end,
+			})
+			i = end
+		}
+	}
+	return
+}

vendor/github.com/golang-commonmark/markdown/linkify/url.go

@@ -0,0 +1,412 @@
+// Copyright 2015 The Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+package linkify
+
+import (
+	"unicode/utf8"
+
+	"github.com/golang-commonmark/markdown/byteutil"
+)
+
+func atoi3(s string, start int) (int, bool) {
+	n := 0
+	var i int
+	for i = start; i < len(s) && byteutil.IsDigit(s[i]); i++ {
+		n = n*10 + int(s[i]-'0')
+		if n > 255 {
+			return 0, false
+		}
+	}
+	if i == start {
+		return 0, false
+	}
+	return i, true
+}
+
+func skipIPv4(s string) (_ int, _ bool) {
+	j := 0
+	for i := 0; i < 4; i++ {
+		if j >= len(s) {
+			return
+		}
+		if i > 0 {
+			if s[j] != '.' {
+				return
+			}
+			j++
+		}
+		if n, ok := atoi3(s, j); !ok {
+			return
+		} else {
+			j = n
+		}
+	}
+	return j, true
+}
+
+func atoi5(s string, start int) (int, bool) {
+	n := 0
+	var i int
+	for i = start; i < len(s) && byteutil.IsDigit(s[i]); i++ {
+		n = n*10 + int(s[i]-'0')
+		if n > 65535 {
+			return 0, false
+		}
+	}
+	if i == start || n == 0 {
+		return 0, false
+	}
+	return i, true
+}
+
+func skipPort(s string, start int) int {
+	if start >= len(s) || s[start] != ':' {
+		return start
+	}
+	end, ok := atoi5(s, start+1)
+	if !ok {
+		return start
+	}
+	return end
+}
+
+func skipPath(s string, start int) int {
+	if start >= len(s) || s[start] != '/' {
+		return start // skip empty path
+	}
+	var stack []rune
+	var notClosedIndex int
+	var nHyphen int
+	end := start + 1
+loop:
+	for end < len(s) {
+		r, rlen := utf8.DecodeRuneInString(s[end:])
+		if r == utf8.RuneError {
+			nHyphen = 0
+			break
+		}
+
+		switch {
+		case isUnreserved(r):
+			if r == '-' {
+				nHyphen++
+				if nHyphen > 1 {
+					break loop
+				}
+			} else {
+				nHyphen = 0
+			}
+		case isSubDelimiter(r) || r == '[' || r == ']':
+			nHyphen = 0
+			switch r {
+			case '[', '(':
+				if len(stack) == 0 {
+					notClosedIndex = end
+				}
+				stack = append(stack, r)
+			case ']', ')':
+				opening := '['
+				if r == ')' {
+					opening = '('
+				}
+				if len(stack) == 0 || stack[len(stack)-1] != opening {
+					break loop
+				}
+				stack = stack[:len(stack)-1]
+			}
+		case r == '/' || r == ':' || r == '@':
+			nHyphen = 0
+		case r == '%':
+			nHyphen = 0
+			if end+2 >= len(s) {
+				break loop
+			}
+			if !(byteutil.IsHexDigit(s[end+1]) &&
+				byteutil.IsHexDigit(s[end+2])) {
+				break loop
+			}
+			end += 2
+		default:
+			nHyphen = 0
+			if r != ' ' || len(stack) == 0 {
+				break loop
+			}
+		}
+		end += rlen
+	}
+	if len(stack) > 0 {
+		return notClosedIndex
+	}
+	if nHyphen > 0 {
+		return end - nHyphen + 1
+	}
+	return end
+}
+
+func skipQuery(s string, start int) int {
+	if start >= len(s) || s[start] != '?' {
+		return start
+	}
+	var stack []rune
+	var notClosedIndex int
+	var nHyphen int
+	end := start + 1
+loop:
+	for end < len(s) {
+		r, rlen := utf8.DecodeRuneInString(s[end:])
+		if r == utf8.RuneError {
+			nHyphen = 0
+			break
+		}
+
+		switch {
+		case isUnreserved(r):
+			if r == '-' {
+				nHyphen++
+				if nHyphen > 1 {
+					break loop
+				}
+			} else {
+				nHyphen = 0
+			}
+		case isSubDelimiter(r) || r == '[' || r == ']':
+			nHyphen = 0
+			switch r {
+			case '[', '(':
+				if len(stack) == 0 {
+					notClosedIndex = end
+				}
+				stack = append(stack, r)
+			case ']', ')':
+				opening := '['
+				if r == ')' {
+					opening = '('
+				}
+				if len(stack) == 0 || stack[len(stack)-1] != opening {
+					break loop
+				}
+				stack = stack[:len(stack)-1]
+			}
+		case r == '?' || r == '/' || r == ':' || r == '@':
+			nHyphen = 0
+		case r == '%':
+			nHyphen = 0
+			if end+2 >= len(s) {
+				break loop
+			}
+			if !(byteutil.IsHexDigit(s[end+1]) &&
+				byteutil.IsHexDigit(s[end+2])) {
+				break loop
+			}
+			end += 2
+		default:
+			nHyphen = 0
+			if r != ' ' || len(stack) == 0 {
+				break loop
+			}
+		}
+		end += rlen
+	}
+	if len(stack) > 0 {
+		return notClosedIndex
+	}
+	if nHyphen > 0 {
+		return end - nHyphen + 1
+	}
+	return end
+}
+
+func skipFragment(s string, start int) int {
+	if start >= len(s) || s[start] != '#' {
+		return start
+	}
+	var stack []rune
+	var notClosedIndex int
+	var nHyphen int
+	end := start + 1
+loop:
+	for end < len(s) {
+		r, rlen := utf8.DecodeRuneInString(s[end:])
+		if r == utf8.RuneError {
+			nHyphen = 0
+			break
+		}
+
+		switch {
+		case isUnreserved(r):
+			if r == '-' {
+				nHyphen++
+				if nHyphen > 1 {
+					break loop
+				}
+			} else {
+				nHyphen = 0
+			}
+		case isSubDelimiter(r) || r == '[' || r == ']':
+			nHyphen = 0
+			switch r {
+			case '[', '(':
+				if len(stack) == 0 {
+					notClosedIndex = end
+				}
+				stack = append(stack, r)
+			case ']', ')':
+				opening := '['
+				if r == ')' {
+					opening = '('
+				}
+				if len(stack) == 0 || stack[len(stack)-1] != opening {
+					break loop
+				}
+				stack = stack[:len(stack)-1]
+			}
+		case r == '?' || r == '/' || r == ':' || r == '@':
+			nHyphen = 0
+		case r == '%':
+			nHyphen = 0
+			if end+2 >= len(s) {
+				break loop
+			}
+			if !(byteutil.IsHexDigit(s[end+1]) &&
+				byteutil.IsHexDigit(s[end+2])) {
+				break loop
+			}
+			end += 2
+		default:
+			nHyphen = 0
+			if r != ' ' || len(stack) == 0 {
+				break loop
+			}
+		}
+		end += rlen
+	}
+	if len(stack) > 0 {
+		return notClosedIndex
+	}
+	if nHyphen > 0 {
+		return end - nHyphen + 1
+	}
+	return end
+}
+
+func unskipPunct(s string, start int) int {
+	end := start - 1
+	if end < 0 || end >= len(s) || !basicPunct[s[end]] {
+		return start
+	}
+	return end
+}
+
+func findHostnameStart(s string, start int) (_ int, _ bool) {
+	end := start
+	lastDot := true
+	nHyphen := 0
+loop:
+	for end > 0 {
+		r, rlen := utf8.DecodeLastRuneInString(s[:end])
+		if r == utf8.RuneError {
+			return
+		}
+
+		switch {
+		case isLetterOrDigit(r):
+			lastDot = false
+			nHyphen = 0
+		case r == '.':
+			if nHyphen > 0 {
+				return
+			}
+			lastDot = true
+		case r == '-':
+			if end == start {
+				return
+			}
+			if lastDot {
+				return
+			}
+			nHyphen++
+			if nHyphen == 3 {
+				return
+			}
+		case r == ':' || r == '/' || r == '\\' || r == '_':
+			return
+		case isPunctOrSpaceOrControl(r):
+			break loop
+		default:
+			return
+		}
+		end -= rlen
+	}
+	if lastDot || nHyphen > 0 {
+		return
+	}
+	return end, true
+}
+
+func findHostnameEnd(s string, start int) (_ int, _ int, _ bool) {
+	end := start
+	lastDot := false
+	lastDotPos := -1
+	nHyphen := 0
+loop:
+	for end < len(s) {
+		r, rlen := utf8.DecodeRuneInString(s[end:])
+		if r == utf8.RuneError {
+			return
+		}
+
+		switch {
+		case isLetterOrDigit(r):
+			lastDot = false
+			nHyphen = 0
+		case r == '.':
+			if nHyphen > 0 {
+				return
+			}
+			if lastDot {
+				break loop
+			}
+			lastDot = true
+			lastDotPos = end
+			nHyphen = 0
+		case r == '-':
+			lastDot = false
+			if end == start {
+				return
+			}
+			if lastDot {
+				return
+			}
+			nHyphen++
+			if nHyphen == 3 {
+				break loop
+			}
+		case r == '\\' || r == '_':
+			return
+		case isPunctOrSpaceOrControl(r):
+			break loop
+		default:
+			return
+		}
+		end += rlen
+	}
+
+	if nHyphen > 0 {
+		end -= nHyphen
+	} else if lastDot {
+		if s[end-1] == '.' {
+			end--
+		}
+		lastDotPos = end - 1
+		for lastDotPos >= start && s[lastDotPos] != '.' {
+			lastDotPos--
+		}
+		if lastDotPos < start {
+			lastDotPos = -1
+		}
+	}
+
+	return end, lastDotPos, true
+}