[pull] master from PaulSonOfLars:master

common.go

@@ -22,7 +22,7 @@ func validStart(pos int, input []rune) bool {
 }
 
 func validEnd(pos int, input []rune) bool {
-	// First char is not a valid end char.
+	// First char is not a valid end char; we do NOT allow empty entities.
 	// If the end char has a space before it, its not valid either.
 	if pos == 0 || unicode.IsSpace(input[pos-1]) {
 		return false

commonV2.go

@@ -148,7 +148,7 @@ func isClosingTag(in []rune, pos int) bool {
 	return false
 }
 
-func getClosingTag(in []rune, tag string) (int, int) {
+func getClosingTag(in []rune, openingTag string, closingTag string) (int, int) {
 	offset := 0
 	subtags := 0
 	for offset < len(in) {
@@ -164,9 +164,9 @@ func getClosingTag(in []rune, tag string) (int, int) {
 		}
 
 		closingTagIdx := openingTagIdx + 2 + c
-		if string(in[openingTagIdx+1:closingTagIdx]) == tag { // found a nested tag, this is annoying
+		if string(in[openingTagIdx+1:closingTagIdx]) == openingTag { // found a nested tag, this is annoying
 			subtags++
-		} else if isClosingTag(in, openingTagIdx) && string(in[openingTagIdx+2:closingTagIdx]) == tag {
+		} else if isClosingTag(in, openingTagIdx) && string(in[openingTagIdx+2:closingTagIdx]) == closingTag {
 			if subtags == 0 {
 				return openingTagIdx, closingTagIdx
 			}

md2htmlV2.go

@@ -4,6 +4,7 @@ import (
 	"html"
 	"sort"
 	"strings"
+	"unicode"
 )
 
 var defaultConverterV2 = ConverterV2{
@@ -42,20 +43,23 @@ func MD2HTMLButtonsV2(in string) (string, []ButtonV2) {
 }
 
 var chars = map[string]string{
-	"`":   "code",
-	"```": "pre",
-	"_":   "i",
-	"*":   "b",
-	"~":   "s",
-	"__":  "u",
-	"|":   "", // this is a placeholder for || to work
-	"||":  "span class=\"tg-spoiler\"",
-	"!":   "", // for emoji
-	"[":   "", // for links
-	"]":   "", // for links/emoji
-	"(":   "", // for links/emoji
-	")":   "", // for links/emoji
-	"\\":  "", // for escapes
+	"`":    "code",
+	"```":  "pre",
+	"_":    "i",
+	"*":    "b",
+	"~":    "s",
+	"__":   "u",
+	"|":    "", // this is a placeholder for || to work
+	"||":   "span class=\"tg-spoiler\"",
+	"!":    "", // for emoji
+	"![":   "", // for emoji
+	"[":    "", // for links
+	"]":    "", // for links/emoji
+	"(":    "", // for links/emoji
+	")":    "", // for links/emoji
+	"\\":   "", // for escapes
+	"&":    "", // for blockquotes
+	">": "blockquote",
 }
 
 var AllMarkdownV2Chars = func() []rune {
@@ -86,6 +90,53 @@ var skipStarts = map[rune]bool{
 	'[': true, // links
 }
 
+func getItem(in []rune, i int) (string, int, bool) {
+	c := in[i]
+	if _, ok := chars[string(c)]; !ok {
+		return "", 0, false
+	}
+
+	if !validStart(i, in) && !skipStarts[c] {
+		if c == '\\' && i+1 < len(in) {
+			escaped := string(in[i+1])
+			if _, ok := chars[escaped]; ok {
+				return escaped, 1, false
+			}
+		}
+		return "", 0, false
+	}
+
+	item := string(c)
+	if c == '|' &&
+		i+1 < len(in) && in[i+1] == '|' {
+		return "||", 1, true
+
+	} else if c == '_' &&
+		i+1 < len(in) && in[i+1] == '_' { // support __
+		return "__", 1, true
+
+	} else if c == '`' &&
+		i+2 < len(in) && in[i+1] == '`' && in[i+2] == '`' { // support ```
+		return "```", 2, true
+
+	} else if c == '&' &&
+		i+3 < len(in) && in[i+1] == 'g' && in[i+2] == 't' && in[i+3] == ';' &&
+		validBlockQuoteStart(in, i) {
+		return "&gt;", 3, true
+
+	} else if c == '*' &&
+		i+5 < len(in) && in[i+1] == '*' && in[i+2] == '&' && in[i+3] == 'g' && in[i+4] == 't' && in[i+5] == ';' &&
+		// We force support for **> to allow for people to separate quotes/expandable quote blocks with **
+		validBlockQuoteStart(in, i) {
+		return "**&gt;", 5, true
+
+	} else if c == '!' && i+1 < len(in) && in[i+1] == '[' {
+		return "![", 1, true
+	}
+
+	return item, 0, true
+}
+
 // TODO: add support for a map-like check of which items cannot be included.
 //
 //	Eg: `code` cannot be italic/bold/underline/strikethrough
@@ -96,49 +147,20 @@ func (cv ConverterV2) md2html(in []rune, enableButtons bool) (string, []ButtonV2
 	out := strings.Builder{}
 
 	for i := 0; i < len(in); i++ {
-		c := in[i]
-		if _, ok := chars[string(c)]; !ok {
-			out.WriteRune(c)
-			continue
-		}
-
-		if !validStart(i, in) && !skipStarts[c] {
-			if c == '\\' && i+1 < len(in) {
-				if _, ok := chars[string(in[i+1])]; ok {
-					out.WriteRune(in[i+1])
-					i++
-					continue
-				}
+		item, offset, ok := getItem(in, i)
+		if !ok {
+			if item == "" {
+				item = string(in[i])
 			}
-			out.WriteRune(c)
+			out.WriteString(item)
+			i += offset
 			continue
 		}
+		i += offset
 
-		switch c {
-		case '`', '*', '~', '_', '|': // '||', '__', and '```' are included here too
-			item := string(c)
-			if c == '|' { // support ||
-				// if single |, ignore. We only care about double ||
-				if i+1 >= len(in) || in[i+1] != '|' {
-					out.WriteRune(c)
-					continue
-				}
-
-				item = "||"
-				i++
-			} else if c == '_' && i+1 < len(in) && in[i+1] == '_' { // support __
-				item = "__"
-				i++
-			} else if c == '`' && i+2 < len(in) && in[i+1] == '`' && in[i+2] == '`' { // support ```
-				item = "```"
-				i += 2
-			}
-
-			if i+1 >= len(in) {
-				out.WriteString(item)
-				continue
-			}
-
+		switch item {
+		// All cases where start and closing tags are the same.
+		case "`", "*", "~", "_", "```", "||", "__":
 			idx := getValidEnd(in[i+1:], item)
 			if idx < 0 {
 				// not found; write and move on.
@@ -174,29 +196,44 @@ func (cv ConverterV2) md2html(in []rune, enableButtons bool) (string, []ButtonV2
 			nestedT, nestedB := cv.md2html(in[nStart:nEnd], enableButtons)
 			return out.String() + "<" + chars[item] + ">" + nestedT + "</" + closeSpans(chars[item]) + ">" + followT, append(nestedB, followB...)
 
-		case '!':
-			if len(in) <= i+1 || in[i+1] != '[' {
-				out.WriteRune(c)
+		case "&gt;", "**&gt;":
+			nStart := i + 1
+			for nStart < len(in) && unicode.IsSpace(in[nStart]) {
+				nStart++
+			}
+
+			if nStart >= len(in) {
+				out.WriteString(item)
 				continue
 			}
 
-			ok, text, content, newEnd := getLinkContents(in[i+1:], true)
+			nEnd, contents, expandable := getBlockQuoteEnd(in, nStart)
+			nestedT, nestedB := cv.md2html(contents, enableButtons)
+			followT, followB := cv.md2html(in[nEnd:], enableButtons)
+
+			if expandable {
+				return out.String() + "<blockquote expandable>" + strings.TrimSpace(nestedT) + "</blockquote>" + followT, append(nestedB, followB...)
+			}
+			return out.String() + "<blockquote>" + strings.TrimSpace(nestedT) + "</blockquote>" + followT, append(nestedB, followB...)
+
+		case "![":
+			ok, text, content, newEnd := getLinkContents(in[i:], true)
 			if !ok {
-				out.WriteRune(c)
+				out.WriteString(item)
 				continue
 			}
-			end := i + 1 + newEnd
+			end := i + newEnd
 
 			content = strings.TrimPrefix(content, "tg://emoji?id=")
 
 			nestedT, nestedB := cv.md2html(text, enableButtons)
 			followT, followB := cv.md2html(in[end:], enableButtons)
 			return out.String() + `<tg-emoji emoji-id="` + content + `">` + nestedT + "</tg-emoji>" + followT, append(nestedB, followB...)
 
-		case '[':
+		case "[":
 			ok, text, content, newEnd := getLinkContents(in[i:], false)
 			if !ok {
-				out.WriteRune(c)
+				out.WriteString(item)
 				continue
 			}
 			end := i + newEnd
@@ -227,24 +264,78 @@ func (cv ConverterV2) md2html(in []rune, enableButtons bool) (string, []ButtonV2
 			nestedT, nestedB := cv.md2html(text, enableButtons)
 			return out.String() + `<a href="` + content + `">` + nestedT + "</a>" + followT, append(nestedB, followB...)
 
-		case ']', '(', ')':
-			out.WriteRune(c)
-
-		case '\\':
+		case "\\":
 			if i+1 < len(in) {
 				if _, ok := chars[string(in[i+1])]; ok {
 					out.WriteRune(in[i+1])
 					i++
 					continue
 				}
 			}
-			out.WriteRune(c)
+			out.WriteString(item)
+
+		default:
+			out.WriteString(item)
 		}
 	}
 
 	return out.String(), nil
 }
 
+func getBlockQuoteEnd(in []rune, nStart int) (int, []rune, bool) {
+	var contents []rune // We store all the contents, minus the > characters, so we avoid double-html tags
+	lineStart := true
+	for j := nStart; j < len(in); j++ {
+		if lineStart && in[j] == ' ' {
+			// Skip space chars at start of lines
+			continue
+		}
+
+		lineStart = in[j] == '\n'
+		contents = append(contents, in[j])
+
+		// Keep skipping until we get a newline
+		if in[j] != '\n' {
+			continue
+		}
+
+		if isExpandableEnd(in, j) {
+			// Extra -1 to include newline
+			return j, contents[:len(contents)-3], true
+		}
+
+		if j+4 < len(in) && in[j+1] == '&' && in[j+2] == 'g' && in[j+3] == 't' && in[j+4] == ';' {
+			j = j + 4 // skip '>' symbol for the next blockquote start
+			continue
+		}
+		return j, contents, false
+	}
+
+	if isExpandableEnd(in, len(in)) {
+		return len(in), contents[:len(contents)-2], true
+	}
+
+	return len(in), contents, false
+}
+
+func isExpandableEnd(in []rune, j int) bool {
+	return j-2 >= 0 && in[j-1] == '|' && in[j-2] == '|'
+}
+
+func validBlockQuoteStart(in []rune, i int) bool {
+	for j := i - 1; j >= 0; j-- {
+		if !unicode.IsSpace(in[j]) {
+			return false
+		}
+		if in[j] == '\n' {
+			return true
+		}
+	}
+
+	// Start of message; must be valid.
+	return true
+}
+
 func EscapeMarkdownV2(r []rune) string {
 	out := strings.Builder{}
 	for i, x := range r {