package html import ( "bytes" "fmt" "strconv" "github.com/yuin/goldmark/ast" "github.com/yuin/goldmark/renderer" "github.com/yuin/goldmark/util" ) // A Config struct has configurations for the HTML based renderers. type Config struct { Writer Writer HardWraps bool XHTML bool Unsafe bool } // NewConfig returns a new Config with defaults. func NewConfig() Config { return Config{ Writer: DefaultWriter, HardWraps: false, XHTML: false, Unsafe: false, } } // SetOption implements renderer.NodeRenderer.SetOption. func (c *Config) SetOption(name renderer.OptionName, value interface{}) { switch name { case optHardWraps: c.HardWraps = value.(bool) case optXHTML: c.XHTML = value.(bool) case optUnsafe: c.Unsafe = value.(bool) case optTextWriter: c.Writer = value.(Writer) } } // An Option interface sets options for HTML based renderers. type Option interface { SetHTMLOption(*Config) } // TextWriter is an option name used in WithWriter. const optTextWriter renderer.OptionName = "Writer" type withWriter struct { value Writer } func (o *withWriter) SetConfig(c *renderer.Config) { c.Options[optTextWriter] = o.value } func (o *withWriter) SetHTMLOption(c *Config) { c.Writer = o.value } // WithWriter is a functional option that allow you to set the given writer to // the renderer. func WithWriter(writer Writer) interface { renderer.Option Option } { return &withWriter{writer} } // HardWraps is an option name used in WithHardWraps. const optHardWraps renderer.OptionName = "HardWraps" type withHardWraps struct { } func (o *withHardWraps) SetConfig(c *renderer.Config) { c.Options[optHardWraps] = true } func (o *withHardWraps) SetHTMLOption(c *Config) { c.HardWraps = true } // WithHardWraps is a functional option that indicates whether softline breaks // should be rendered as '
'. func WithHardWraps() interface { renderer.Option Option } { return &withHardWraps{} } // XHTML is an option name used in WithXHTML. const optXHTML renderer.OptionName = "XHTML" type withXHTML struct { } func (o *withXHTML) SetConfig(c *renderer.Config) { c.Options[optXHTML] = true } func (o *withXHTML) SetHTMLOption(c *Config) { c.XHTML = true } // WithXHTML is a functional option indicates that nodes should be rendered in // xhtml instead of HTML5. func WithXHTML() interface { Option renderer.Option } { return &withXHTML{} } // Unsafe is an option name used in WithUnsafe. const optUnsafe renderer.OptionName = "Unsafe" type withUnsafe struct { } func (o *withUnsafe) SetConfig(c *renderer.Config) { c.Options[optUnsafe] = true } func (o *withUnsafe) SetHTMLOption(c *Config) { c.Unsafe = true } // WithUnsafe is a functional option that renders dangerous contents // (raw htmls and potentially dangerous links) as it is. func WithUnsafe() interface { renderer.Option Option } { return &withUnsafe{} } // A Renderer struct is an implementation of renderer.NodeRenderer that renders // nodes as (X)HTML. type Renderer struct { Config } // NewRenderer returns a new Renderer with given options. func NewRenderer(opts ...Option) renderer.NodeRenderer { r := &Renderer{ Config: NewConfig(), } for _, opt := range opts { opt.SetHTMLOption(&r.Config) } return r } // RegisterFuncs implements NodeRenderer.RegisterFuncs . func (r *Renderer) RegisterFuncs(reg renderer.NodeRendererFuncRegisterer) { // blocks reg.Register(ast.KindDocument, r.renderDocument) reg.Register(ast.KindHeading, r.renderHeading) reg.Register(ast.KindBlockquote, r.renderBlockquote) reg.Register(ast.KindCodeBlock, r.renderCodeBlock) reg.Register(ast.KindFencedCodeBlock, r.renderFencedCodeBlock) reg.Register(ast.KindHTMLBlock, r.renderHTMLBlock) reg.Register(ast.KindList, r.renderList) reg.Register(ast.KindListItem, r.renderListItem) reg.Register(ast.KindParagraph, r.renderParagraph) reg.Register(ast.KindTextBlock, r.renderTextBlock) reg.Register(ast.KindThematicBreak, r.renderThematicBreak) // inlines reg.Register(ast.KindAutoLink, r.renderAutoLink) reg.Register(ast.KindCodeSpan, r.renderCodeSpan) reg.Register(ast.KindEmphasis, r.renderEmphasis) reg.Register(ast.KindImage, r.renderImage) reg.Register(ast.KindLink, r.renderLink) reg.Register(ast.KindRawHTML, r.renderRawHTML) reg.Register(ast.KindText, r.renderText) reg.Register(ast.KindString, r.renderString) } func (r *Renderer) writeLines(w util.BufWriter, source []byte, n ast.Node) { l := n.Lines().Len() for i := 0; i < l; i++ { line := n.Lines().At(i) r.Writer.RawWrite(w, line.Value(source)) } } // GlobalAttributeFilter defines attribute names which any elements can have. var GlobalAttributeFilter = util.NewBytesFilter( []byte("accesskey"), []byte("autocapitalize"), []byte("autofocus"), []byte("class"), []byte("contenteditable"), []byte("dir"), []byte("draggable"), []byte("enterkeyhint"), []byte("hidden"), []byte("id"), []byte("inert"), []byte("inputmode"), []byte("is"), []byte("itemid"), []byte("itemprop"), []byte("itemref"), []byte("itemscope"), []byte("itemtype"), []byte("lang"), []byte("part"), []byte("slot"), []byte("spellcheck"), []byte("style"), []byte("tabindex"), []byte("title"), []byte("translate"), ) func (r *Renderer) renderDocument(w util.BufWriter, source []byte, node ast.Node, entering bool) (ast.WalkStatus, error) { // nothing to do return ast.WalkContinue, nil } // HeadingAttributeFilter defines attribute names which heading elements can have var HeadingAttributeFilter = GlobalAttributeFilter func (r *Renderer) renderHeading(w util.BufWriter, source []byte, node ast.Node, entering bool) (ast.WalkStatus, error) { n := node.(*ast.Heading) if entering { _, _ = w.WriteString("') } else { _, _ = w.WriteString("\n") } return ast.WalkContinue, nil } // BlockquoteAttributeFilter defines attribute names which blockquote elements can have var BlockquoteAttributeFilter = GlobalAttributeFilter.Extend( []byte("cite"), ) func (r *Renderer) renderBlockquote(w util.BufWriter, source []byte, n ast.Node, entering bool) (ast.WalkStatus, error) { if entering { if n.Attributes() != nil { _, _ = w.WriteString("') } else { _, _ = w.WriteString("
\n") } } else { _, _ = w.WriteString("
\n") } return ast.WalkContinue, nil } func (r *Renderer) renderCodeBlock(w util.BufWriter, source []byte, n ast.Node, entering bool) (ast.WalkStatus, error) { if entering { _, _ = w.WriteString("
")
		r.writeLines(w, source, n)
	} else {
		_, _ = w.WriteString("
\n") } return ast.WalkContinue, nil } func (r *Renderer) renderFencedCodeBlock(w util.BufWriter, source []byte, node ast.Node, entering bool) (ast.WalkStatus, error) { n := node.(*ast.FencedCodeBlock) if entering { _, _ = w.WriteString("
')
		r.writeLines(w, source, n)
	} else {
		_, _ = w.WriteString("
\n") } return ast.WalkContinue, nil } func (r *Renderer) renderHTMLBlock(w util.BufWriter, source []byte, node ast.Node, entering bool) (ast.WalkStatus, error) { n := node.(*ast.HTMLBlock) if entering { if r.Unsafe { l := n.Lines().Len() for i := 0; i < l; i++ { line := n.Lines().At(i) r.Writer.SecureWrite(w, line.Value(source)) } } else { _, _ = w.WriteString("\n") } } else { if n.HasClosure() { if r.Unsafe { closure := n.ClosureLine r.Writer.SecureWrite(w, closure.Value(source)) } else { _, _ = w.WriteString("\n") } } } return ast.WalkContinue, nil } // ListAttributeFilter defines attribute names which list elements can have. var ListAttributeFilter = GlobalAttributeFilter.Extend( []byte("start"), []byte("reversed"), []byte("type"), ) func (r *Renderer) renderList(w util.BufWriter, source []byte, node ast.Node, entering bool) (ast.WalkStatus, error) { n := node.(*ast.List) tag := "ul" if n.IsOrdered() { tag = "ol" } if entering { _ = w.WriteByte('<') _, _ = w.WriteString(tag) if n.IsOrdered() && n.Start != 1 { fmt.Fprintf(w, " start=\"%d\"", n.Start) } if n.Attributes() != nil { RenderAttributes(w, n, ListAttributeFilter) } _, _ = w.WriteString(">\n") } else { _, _ = w.WriteString("\n") } return ast.WalkContinue, nil } // ListItemAttributeFilter defines attribute names which list item elements can have. var ListItemAttributeFilter = GlobalAttributeFilter.Extend( []byte("value"), ) func (r *Renderer) renderListItem(w util.BufWriter, source []byte, n ast.Node, entering bool) (ast.WalkStatus, error) { if entering { if n.Attributes() != nil { _, _ = w.WriteString("') } else { _, _ = w.WriteString("
  • ") } fc := n.FirstChild() if fc != nil { if _, ok := fc.(*ast.TextBlock); !ok { _ = w.WriteByte('\n') } } } else { _, _ = w.WriteString("
  • \n") } return ast.WalkContinue, nil } // ParagraphAttributeFilter defines attribute names which paragraph elements can have. var ParagraphAttributeFilter = GlobalAttributeFilter func (r *Renderer) renderParagraph(w util.BufWriter, source []byte, n ast.Node, entering bool) (ast.WalkStatus, error) { if entering { if n.Attributes() != nil { _, _ = w.WriteString("') } else { _, _ = w.WriteString("

    ") } } else { _, _ = w.WriteString("

    \n") } return ast.WalkContinue, nil } func (r *Renderer) renderTextBlock(w util.BufWriter, source []byte, n ast.Node, entering bool) (ast.WalkStatus, error) { if !entering { if _, ok := n.NextSibling().(ast.Node); ok && n.FirstChild() != nil { _ = w.WriteByte('\n') } } return ast.WalkContinue, nil } // ThematicAttributeFilter defines attribute names which hr elements can have. var ThematicAttributeFilter = GlobalAttributeFilter.Extend( []byte("align"), // [Deprecated] []byte("color"), // [Not Standardized] []byte("noshade"), // [Deprecated] []byte("size"), // [Deprecated] []byte("width"), // [Deprecated] ) func (r *Renderer) renderThematicBreak(w util.BufWriter, source []byte, n ast.Node, entering bool) (ast.WalkStatus, error) { if !entering { return ast.WalkContinue, nil } _, _ = w.WriteString("\n") } else { _, _ = w.WriteString(">\n") } return ast.WalkContinue, nil } // LinkAttributeFilter defines attribute names which link elements can have. var LinkAttributeFilter = GlobalAttributeFilter.Extend( []byte("download"), // []byte("href"), []byte("hreflang"), []byte("media"), []byte("ping"), []byte("referrerpolicy"), []byte("rel"), []byte("shape"), []byte("target"), ) func (r *Renderer) renderAutoLink(w util.BufWriter, source []byte, node ast.Node, entering bool) (ast.WalkStatus, error) { n := node.(*ast.AutoLink) if !entering { return ast.WalkContinue, nil } _, _ = w.WriteString(`') } else { _, _ = w.WriteString(`">`) } _, _ = w.Write(util.EscapeHTML(label)) _, _ = w.WriteString(``) return ast.WalkContinue, nil } // CodeAttributeFilter defines attribute names which code elements can have. var CodeAttributeFilter = GlobalAttributeFilter func (r *Renderer) renderCodeSpan(w util.BufWriter, source []byte, n ast.Node, entering bool) (ast.WalkStatus, error) { if entering { if n.Attributes() != nil { _, _ = w.WriteString("') } else { _, _ = w.WriteString("") } for c := n.FirstChild(); c != nil; c = c.NextSibling() { segment := c.(*ast.Text).Segment value := segment.Value(source) if bytes.HasSuffix(value, []byte("\n")) { r.Writer.RawWrite(w, value[:len(value)-1]) r.Writer.RawWrite(w, []byte(" ")) } else { r.Writer.RawWrite(w, value) } } return ast.WalkSkipChildren, nil } _, _ = w.WriteString("") return ast.WalkContinue, nil } // EmphasisAttributeFilter defines attribute names which emphasis elements can have. var EmphasisAttributeFilter = GlobalAttributeFilter func (r *Renderer) renderEmphasis(w util.BufWriter, source []byte, node ast.Node, entering bool) (ast.WalkStatus, error) { n := node.(*ast.Emphasis) tag := "em" if n.Level == 2 { tag = "strong" } if entering { _ = w.WriteByte('<') _, _ = w.WriteString(tag) if n.Attributes() != nil { RenderAttributes(w, n, EmphasisAttributeFilter) } _ = w.WriteByte('>') } else { _, _ = w.WriteString("') } return ast.WalkContinue, nil } func (r *Renderer) renderLink(w util.BufWriter, source []byte, node ast.Node, entering bool) (ast.WalkStatus, error) { n := node.(*ast.Link) if entering { _, _ = w.WriteString("') } else { _, _ = w.WriteString("") } return ast.WalkContinue, nil } // ImageAttributeFilter defines attribute names which image elements can have. var ImageAttributeFilter = GlobalAttributeFilter.Extend( []byte("align"), []byte("border"), []byte("crossorigin"), []byte("decoding"), []byte("height"), []byte("importance"), []byte("intrinsicsize"), []byte("ismap"), []byte("loading"), []byte("referrerpolicy"), []byte("sizes"), []byte("srcset"), []byte("usemap"), []byte("width"), ) func (r *Renderer) renderImage(w util.BufWriter, source []byte, node ast.Node, entering bool) (ast.WalkStatus, error) { if !entering { return ast.WalkContinue, nil } n := node.(*ast.Image) _, _ = w.WriteString("`)
	_, _ = w.Write(util.EscapeHTML(n.Text(source)))
	_ = w.WriteByte('") } else { _, _ = w.WriteString(">") } return ast.WalkSkipChildren, nil } func (r *Renderer) renderRawHTML(w util.BufWriter, source []byte, node ast.Node, entering bool) (ast.WalkStatus, error) { if !entering { return ast.WalkSkipChildren, nil } if r.Unsafe { n := node.(*ast.RawHTML) l := n.Segments.Len() for i := 0; i < l; i++ { segment := n.Segments.At(i) _, _ = w.Write(segment.Value(source)) } return ast.WalkSkipChildren, nil } _, _ = w.WriteString("") return ast.WalkSkipChildren, nil } func (r *Renderer) renderText(w util.BufWriter, source []byte, node ast.Node, entering bool) (ast.WalkStatus, error) { if !entering { return ast.WalkContinue, nil } n := node.(*ast.Text) segment := n.Segment if n.IsRaw() { r.Writer.RawWrite(w, segment.Value(source)) } else { r.Writer.Write(w, segment.Value(source)) if n.HardLineBreak() || (n.SoftLineBreak() && r.HardWraps) { if r.XHTML { _, _ = w.WriteString("
    \n") } else { _, _ = w.WriteString("
    \n") } } else if n.SoftLineBreak() { _ = w.WriteByte('\n') } } return ast.WalkContinue, nil } func (r *Renderer) renderString(w util.BufWriter, source []byte, node ast.Node, entering bool) (ast.WalkStatus, error) { if !entering { return ast.WalkContinue, nil } n := node.(*ast.String) if n.IsCode() { _, _ = w.Write(n.Value) } else { if n.IsRaw() { r.Writer.RawWrite(w, n.Value) } else { r.Writer.Write(w, n.Value) } } return ast.WalkContinue, nil } var dataPrefix = []byte("data-") // RenderAttributes renders given node's attributes. // You can specify attribute names to render by the filter. // If filter is nil, RenderAttributes renders all attributes. func RenderAttributes(w util.BufWriter, node ast.Node, filter util.BytesFilter) { for _, attr := range node.Attributes() { if filter != nil && !filter.Contains(attr.Name) { if !bytes.HasPrefix(attr.Name, dataPrefix) { continue } } _, _ = w.WriteString(" ") _, _ = w.Write(attr.Name) _, _ = w.WriteString(`="`) // TODO: convert numeric values to strings _, _ = w.Write(util.EscapeHTML(attr.Value.([]byte))) _ = w.WriteByte('"') } } // A Writer interface writes textual contents to a writer. type Writer interface { // Write writes the given source to writer with resolving references and unescaping // backslash escaped characters. Write(writer util.BufWriter, source []byte) // RawWrite writes the given source to writer without resolving references and // unescaping backslash escaped characters. RawWrite(writer util.BufWriter, source []byte) // SecureWrite writes the given source to writer with replacing insecure characters. SecureWrite(writer util.BufWriter, source []byte) } var replacementCharacter = []byte("\ufffd") type defaultWriter struct { } func escapeRune(writer util.BufWriter, r rune) { if r < 256 { v := util.EscapeHTMLByte(byte(r)) if v != nil { _, _ = writer.Write(v) return } } _, _ = writer.WriteRune(util.ToValidRune(r)) } func (d *defaultWriter) SecureWrite(writer util.BufWriter, source []byte) { n := 0 l := len(source) for i := 0; i < l; i++ { if source[i] == '\u0000' { _, _ = writer.Write(source[i-n : i]) n = 0 _, _ = writer.Write(replacementCharacter) continue } n++ } if n != 0 { _, _ = writer.Write(source[l-n:]) } } func (d *defaultWriter) RawWrite(writer util.BufWriter, source []byte) { n := 0 l := len(source) for i := 0; i < l; i++ { v := util.EscapeHTMLByte(source[i]) if v != nil { _, _ = writer.Write(source[i-n : i]) n = 0 _, _ = writer.Write(v) continue } n++ } if n != 0 { _, _ = writer.Write(source[l-n:]) } } func (d *defaultWriter) Write(writer util.BufWriter, source []byte) { escaped := false var ok bool limit := len(source) n := 0 for i := 0; i < limit; i++ { c := source[i] if escaped { if util.IsPunct(c) { d.RawWrite(writer, source[n:i-1]) n = i escaped = false continue } } if c == '\x00' { d.RawWrite(writer, source[n:i]) d.RawWrite(writer, replacementCharacter) n = i + 1 escaped = false continue } if c == '&' { pos := i next := i + 1 if next < limit && source[next] == '#' { nnext := next + 1 if nnext < limit { nc := source[nnext] // code point like #x22; if nnext < limit && nc == 'x' || nc == 'X' { start := nnext + 1 i, ok = util.ReadWhile(source, [2]int{start, limit}, util.IsHexDecimal) if ok && i < limit && source[i] == ';' && i-start < 7 { v, _ := strconv.ParseUint(util.BytesToReadOnlyString(source[start:i]), 16, 32) d.RawWrite(writer, source[n:pos]) n = i + 1 escapeRune(writer, rune(v)) continue } // code point like #1234; } else if nc >= '0' && nc <= '9' { start := nnext i, ok = util.ReadWhile(source, [2]int{start, limit}, util.IsNumeric) if ok && i < limit && i-start < 8 && source[i] == ';' { v, _ := strconv.ParseUint(util.BytesToReadOnlyString(source[start:i]), 10, 32) d.RawWrite(writer, source[n:pos]) n = i + 1 escapeRune(writer, rune(v)) continue } } } } else { start := next i, ok = util.ReadWhile(source, [2]int{start, limit}, util.IsAlphaNumeric) // entity reference if ok && i < limit && source[i] == ';' { name := util.BytesToReadOnlyString(source[start:i]) entity, ok := util.LookUpHTML5EntityByName(name) if ok { d.RawWrite(writer, source[n:pos]) n = i + 1 d.RawWrite(writer, entity.Characters) continue } } } i = next - 1 } if c == '\\' { escaped = true continue } escaped = false } d.RawWrite(writer, source[n:]) } // DefaultWriter is a default implementation of the Writer. var DefaultWriter = &defaultWriter{} var bDataImage = []byte("data:image/") var bPng = []byte("png;") var bGif = []byte("gif;") var bJpeg = []byte("jpeg;") var bWebp = []byte("webp;") var bSvg = []byte("svg;") var bJs = []byte("javascript:") var bVb = []byte("vbscript:") var bFile = []byte("file:") var bData = []byte("data:") // IsDangerousURL returns true if the given url seems a potentially dangerous url, // otherwise false. func IsDangerousURL(url []byte) bool { if bytes.HasPrefix(url, bDataImage) && len(url) >= 11 { v := url[11:] if bytes.HasPrefix(v, bPng) || bytes.HasPrefix(v, bGif) || bytes.HasPrefix(v, bJpeg) || bytes.HasPrefix(v, bWebp) || bytes.HasPrefix(v, bSvg) { return false } return true } return bytes.HasPrefix(url, bJs) || bytes.HasPrefix(url, bVb) || bytes.HasPrefix(url, bFile) || bytes.HasPrefix(url, bData) }