1
0
mirror of https://github.com/danog/blackfriday.git synced 2024-11-30 04:29:13 +01:00
blackfriday/html.go

1447 lines
34 KiB
Go
Raw Normal View History

2011-05-29 05:17:53 +02:00
//
2011-06-28 04:11:32 +02:00
// Blackfriday Markdown Processor
// Available at http://github.com/russross/blackfriday
//
// Copyright © 2011 Russ Ross <russ@russross.com>.
// Distributed under the Simplified BSD License.
2011-06-28 04:11:32 +02:00
// See README.md for details.
2011-05-29 05:17:53 +02:00
//
//
//
// HTML rendering backend
//
//
package blackfriday
import (
"bytes"
"fmt"
"html"
"io"
"regexp"
2011-05-29 05:17:53 +02:00
"strconv"
"strings"
2011-05-29 05:17:53 +02:00
)
type HTMLFlags int
// HTML renderer configuration options.
2011-05-29 05:17:53 +02:00
const (
HTMLFlagsNone HTMLFlags = 0
SkipHTML HTMLFlags = 1 << iota // Skip preformatted HTML blocks
SkipStyle // Skip embedded <style> elements
SkipImages // Skip embedded images
SkipLinks // Skip all links
Safelink // Only link to trusted protocols
NofollowLinks // Only link with rel="nofollow"
NoreferrerLinks // Only link with rel="noreferrer"
HrefTargetBlank // Add a blank target
CompletePage // Generate a complete HTML page
UseXHTML // Generate XHTML output instead of HTML
FootnoteReturnLinks // Generate a link at the end of a footnote to return to the source
TagName = "[A-Za-z][A-Za-z0-9-]*"
AttributeName = "[a-zA-Z_:][a-zA-Z0-9:._-]*"
UnquotedValue = "[^\"'=<>`\\x00-\\x20]+"
SingleQuotedValue = "'[^']*'"
DoubleQuotedValue = "\"[^\"]*\""
AttributeValue = "(?:" + UnquotedValue + "|" + SingleQuotedValue + "|" + DoubleQuotedValue + ")"
AttributeValueSpec = "(?:" + "\\s*=" + "\\s*" + AttributeValue + ")"
Attribute = "(?:" + "\\s+" + AttributeName + AttributeValueSpec + "?)"
OpenTag = "<" + TagName + Attribute + "*" + "\\s*/?>"
CloseTag = "</" + TagName + "\\s*[>]"
HTMLComment = "<!---->|<!--(?:-?[^>-])(?:-?[^-])*-->"
ProcessingInstruction = "[<][?].*?[?][>]"
Declaration = "<![A-Z]+" + "\\s+[^>]*>"
CDATA = "<!\\[CDATA\\[[\\s\\S]*?\\]\\]>"
HTMLTag = "(?:" + OpenTag + "|" + CloseTag + "|" + HTMLComment + "|" +
ProcessingInstruction + "|" + Declaration + "|" + CDATA + ")"
2011-05-29 05:17:53 +02:00
)
var (
// TODO: improve this regexp to catch all possible entities:
htmlEntity = regexp.MustCompile(`&[a-z]{2,5};`)
reHtmlTag = regexp.MustCompile("(?i)^" + HTMLTag)
)
2016-04-01 14:37:21 +02:00
type HTMLRendererParameters struct {
// Prepend this text to each relative URL.
AbsolutePrefix string
2014-05-29 06:52:45 +02:00
// Add this text to each footnote anchor, to ensure uniqueness.
FootnoteAnchorPrefix string
// Show this text inside the <a> tag for a footnote return link, if the
// HTML_FOOTNOTE_RETURN_LINKS flag is enabled. If blank, the string
// <sup>[return]</sup> is used.
FootnoteReturnLinkContents string
// If set, add this text to the front of each Header ID, to ensure
// uniqueness.
HeaderIDPrefix string
// If set, add this text to the back of each Header ID, to ensure uniqueness.
HeaderIDSuffix string
Title string // Document title (used if CompletePage is set)
CSS string // Optional CSS file URL (used if CompletePage is set)
Flags HTMLFlags // Flags allow customizing this renderer's behavior
Extensions Extensions // Extensions give Smartypants and HTML renderer access to Blackfriday's global extensions
}
// HTML is a type that implements the Renderer interface for HTML output.
2011-07-07 19:56:45 +02:00
//
// Do not create this directly, instead use the NewHTMLRenderer function.
type HTML struct {
HTMLRendererParameters
closeTag string // how to end singleton tags: either " />" or ">"
// table of contents data
2011-06-29 18:36:56 +02:00
tocMarker int
headerCount int
currentLevel int
toc bytes.Buffer
// Track header IDs to prevent ID collision in a single generation.
headerIDs map[string]int
2016-04-01 14:37:21 +02:00
w HTMLWriter
lastOutputLen int
disableTags int
2011-05-29 05:17:53 +02:00
}
2011-06-29 19:13:17 +02:00
const (
xhtmlClose = " />"
htmlClose = ">"
2011-06-29 19:13:17 +02:00
)
2011-05-29 05:17:53 +02:00
2016-04-01 14:37:21 +02:00
type HTMLWriter struct {
bytes.Buffer
}
// Writes out a newline if the output is not pristine. Used at the beginning of
// every rendering func
2016-04-01 14:37:21 +02:00
func (w *HTMLWriter) Newline() {
2016-03-30 14:56:53 +02:00
w.WriteByte('\n')
}
// NewHTMLRenderer creates and configures an HTML object, which
// satisfies the Renderer interface.
func NewHTMLRenderer(params HTMLRendererParameters) Renderer {
2011-05-29 05:17:53 +02:00
// configure the rendering engine
closeTag := htmlClose
if params.Flags&UseXHTML != 0 {
closeTag = xhtmlClose
2011-05-29 05:17:53 +02:00
}
if params.FootnoteReturnLinkContents == "" {
params.FootnoteReturnLinkContents = `<sup>[return]</sup>`
}
2016-04-01 14:37:21 +02:00
var writer HTMLWriter
return &HTML{
HTMLRendererParameters: params,
2011-05-29 05:17:53 +02:00
closeTag: closeTag,
headerIDs: make(map[string]int),
w: writer,
2011-05-29 05:17:53 +02:00
}
}
// Using if statements is a bit faster than a switch statement. As the compiler
// improves, this should be unnecessary this is only worthwhile because
// attrEscape is the single largest CPU user in normal use.
// Also tried using map, but that gave a ~3x slowdown.
func escapeSingleChar(char byte) (string, bool) {
if char == '"' {
return "&quot;", true
}
if char == '&' {
return "&amp;", true
}
if char == '<' {
return "&lt;", true
}
if char == '>' {
return "&gt;", true
}
return "", false
}
func (r *HTML) attrEscape(src []byte) {
org := 0
for i, ch := range src {
if entity, ok := escapeSingleChar(ch); ok {
if i > org {
// copy all the normal characters since the last escape
r.w.Write(src[org:i])
}
org = i + 1
r.w.WriteString(entity)
2011-05-29 05:17:53 +02:00
}
}
if org < len(src) {
r.w.Write(src[org:])
}
2011-05-29 05:17:53 +02:00
}
func attrEscape2(src []byte) []byte {
unesc := []byte(html.UnescapeString(string(src)))
esc1 := []byte(html.EscapeString(string(unesc)))
esc2 := bytes.Replace(esc1, []byte("&#34;"), []byte("&quot;"), -1)
return bytes.Replace(esc2, []byte("&#39;"), []byte{'\''}, -1)
}
func (r *HTML) entityEscapeWithSkip(src []byte, skipRanges [][]int) {
end := 0
for _, rang := range skipRanges {
r.attrEscape(src[end:rang[0]])
r.w.Write(src[rang[0]:rang[1]])
end = rang[1]
}
r.attrEscape(src[end:])
}
func (r *HTML) TitleBlock(text []byte) {
2014-08-02 04:54:21 +02:00
text = bytes.TrimPrefix(text, []byte("% "))
text = bytes.Replace(text, []byte("\n% "), []byte("\n"), -1)
r.w.WriteString("<h1 class=\"title\">")
r.w.Write(text)
r.w.WriteString("\n</h1>")
2014-08-02 04:54:21 +02:00
}
func (r *HTML) BeginHeader(level int, id string) {
r.w.Newline()
2011-05-29 05:17:53 +02:00
if id == "" && r.Extensions&TOC != 0 {
id = fmt.Sprintf("toc_%d", r.headerCount)
}
if id != "" {
id = r.ensureUniqueHeaderID(id)
if r.HeaderIDPrefix != "" {
id = r.HeaderIDPrefix + id
}
if r.HeaderIDSuffix != "" {
id = id + r.HeaderIDSuffix
}
r.w.WriteString(fmt.Sprintf("<h%d id=\"%s\">", level, id))
2011-05-29 05:17:53 +02:00
} else {
r.w.WriteString(fmt.Sprintf("<h%d>", level))
2011-05-29 05:17:53 +02:00
}
}
func (r *HTML) EndHeader(level int, id string, header []byte) {
// are we building a table of contents?
if r.Extensions&TOC != 0 {
r.TocHeaderWithAnchor(header, level, id)
}
r.w.WriteString(fmt.Sprintf("</h%d>\n", level))
2011-05-29 05:17:53 +02:00
}
func (r *HTML) BlockHtml(text []byte) {
if r.Flags&SkipHTML != 0 {
return
}
r.w.Newline()
r.w.Write(text)
r.w.WriteByte('\n')
2011-05-29 05:17:53 +02:00
}
func (r *HTML) HRule() {
r.w.Newline()
r.w.WriteString("<hr")
r.w.WriteString(r.closeTag)
r.w.WriteByte('\n')
2011-05-29 05:17:53 +02:00
}
func (r *HTML) BlockCode(text []byte, lang string) {
r.w.Newline()
2011-05-29 05:17:53 +02:00
// parse out the language names/classes
count := 0
for _, elt := range strings.Fields(lang) {
if elt[0] == '.' {
elt = elt[1:]
2011-05-29 05:17:53 +02:00
}
if len(elt) == 0 {
continue
}
if count == 0 {
r.w.WriteString("<pre><code class=\"language-")
} else {
r.w.WriteByte(' ')
}
r.attrEscape([]byte(elt))
count++
2011-05-29 05:17:53 +02:00
}
if count == 0 {
r.w.WriteString("<pre><code>")
} else {
r.w.WriteString("\">")
2011-05-29 05:17:53 +02:00
}
r.attrEscape(text)
r.w.WriteString("</code></pre>\n")
2011-05-29 05:17:53 +02:00
}
func (r *HTML) BlockQuote(text []byte) {
r.w.Newline()
r.w.WriteString("<blockquote>\n")
r.w.Write(text)
r.w.WriteString("</blockquote>\n")
2011-05-29 05:17:53 +02:00
}
func (r *HTML) Table(header []byte, body []byte, columnData []CellAlignFlags) {
r.w.Newline()
r.w.WriteString("<table>\n<thead>\n")
r.w.Write(header)
r.w.WriteString("</thead>\n\n<tbody>\n")
r.w.Write(body)
r.w.WriteString("</tbody>\n</table>\n")
2011-05-29 05:17:53 +02:00
}
func (r *HTML) TableRow(text []byte) {
r.w.Newline()
r.w.WriteString("<tr>\n")
r.w.Write(text)
r.w.WriteString("\n</tr>\n")
}
func leadingNewline(out *bytes.Buffer) {
if out.Len() > 0 {
out.WriteByte('\n')
}
2011-05-29 05:17:53 +02:00
}
func (r *HTML) TableHeaderCell(out *bytes.Buffer, text []byte, align CellAlignFlags) {
leadingNewline(out)
switch align {
case TableAlignmentLeft:
out.WriteString("<th align=\"left\">")
case TableAlignmentRight:
out.WriteString("<th align=\"right\">")
case TableAlignmentCenter:
out.WriteString("<th align=\"center\">")
default:
out.WriteString("<th>")
}
out.Write(text)
out.WriteString("</th>")
}
func (r *HTML) TableCell(out *bytes.Buffer, text []byte, align CellAlignFlags) {
leadingNewline(out)
2011-05-29 05:17:53 +02:00
switch align {
case TableAlignmentLeft:
out.WriteString("<td align=\"left\">")
case TableAlignmentRight:
out.WriteString("<td align=\"right\">")
case TableAlignmentCenter:
out.WriteString("<td align=\"center\">")
2011-05-29 05:17:53 +02:00
default:
out.WriteString("<td>")
2011-05-29 05:17:53 +02:00
}
out.Write(text)
out.WriteString("</td>")
2011-05-29 05:17:53 +02:00
}
func (r *HTML) BeginFootnotes() {
r.w.WriteString("<div class=\"footnotes\">\n")
r.HRule()
r.BeginList(ListTypeOrdered)
}
func (r *HTML) EndFootnotes() {
r.EndList(ListTypeOrdered)
r.w.WriteString("</div>\n")
}
func (r *HTML) FootnoteItem(name, text []byte, flags ListType) {
if flags&ListItemContainsBlock != 0 || flags&ListItemBeginningOfList != 0 {
r.w.Newline()
}
slug := slugify(name)
r.w.WriteString(`<li id="`)
r.w.WriteString(`fn:`)
r.w.WriteString(r.FootnoteAnchorPrefix)
r.w.Write(slug)
r.w.WriteString(`">`)
r.w.Write(text)
if r.Flags&FootnoteReturnLinks != 0 {
r.w.WriteString(` <a class="footnote-return" href="#`)
r.w.WriteString(`fnref:`)
r.w.WriteString(r.FootnoteAnchorPrefix)
r.w.Write(slug)
r.w.WriteString(`">`)
r.w.WriteString(r.FootnoteReturnLinkContents)
r.w.WriteString(`</a>`)
}
r.w.WriteString("</li>\n")
}
func (r *HTML) BeginList(flags ListType) {
r.w.Newline()
2011-06-25 23:02:46 +02:00
if flags&ListTypeDefinition != 0 {
r.w.WriteString("<dl>")
} else if flags&ListTypeOrdered != 0 {
r.w.WriteString("<ol>")
2011-05-29 05:17:53 +02:00
} else {
r.w.WriteString("<ul>")
2011-05-29 05:17:53 +02:00
}
}
func (r *HTML) EndList(flags ListType) {
if flags&ListTypeDefinition != 0 {
r.w.WriteString("</dl>\n")
} else if flags&ListTypeOrdered != 0 {
r.w.WriteString("</ol>\n")
2011-05-29 05:17:53 +02:00
} else {
r.w.WriteString("</ul>\n")
2011-05-29 05:17:53 +02:00
}
}
func (r *HTML) ListItem(text []byte, flags ListType) {
if (flags&ListItemContainsBlock != 0 && flags&ListTypeDefinition == 0) ||
flags&ListItemBeginningOfList != 0 {
r.w.Newline()
}
if flags&ListTypeTerm != 0 {
r.w.WriteString("<dt>")
} else if flags&ListTypeDefinition != 0 {
r.w.WriteString("<dd>")
2015-05-29 13:30:49 +02:00
} else {
r.w.WriteString("<li>")
2015-05-29 13:30:49 +02:00
}
r.w.Write(text)
if flags&ListTypeTerm != 0 {
r.w.WriteString("</dt>\n")
} else if flags&ListTypeDefinition != 0 {
r.w.WriteString("</dd>\n")
2015-05-29 13:30:49 +02:00
} else {
r.w.WriteString("</li>\n")
2015-05-29 13:30:49 +02:00
}
2011-05-29 05:17:53 +02:00
}
func (r *HTML) BeginParagraph() {
r.w.Newline()
r.w.WriteString("<p>")
}
func (r *HTML) EndParagraph() {
r.w.WriteString("</p>\n")
2011-05-29 05:17:53 +02:00
}
func (r *HTML) AutoLink(link []byte, kind LinkType) {
skipRanges := htmlEntity.FindAllIndex(link, -1)
if r.Flags&Safelink != 0 && !isSafeLink(link) && kind != LinkTypeEmail {
// mark it but don't link it if it is not a safe link: no smartypants
r.w.WriteString("<tt>")
r.entityEscapeWithSkip(link, skipRanges)
r.w.WriteString("</tt>")
2011-06-29 21:00:54 +02:00
return
2011-05-29 05:17:53 +02:00
}
r.w.WriteString("<a href=\"")
if kind == LinkTypeEmail {
r.w.WriteString("mailto:")
} else {
r.maybeWriteAbsolutePrefix(link)
2011-05-29 05:17:53 +02:00
}
r.entityEscapeWithSkip(link, skipRanges)
2014-03-21 03:52:46 +01:00
2015-03-15 00:46:32 +01:00
var relAttrs []string
if r.Flags&NofollowLinks != 0 && !isRelativeLink(link) {
2015-03-15 00:46:32 +01:00
relAttrs = append(relAttrs, "nofollow")
2014-03-21 03:52:46 +01:00
}
if r.Flags&NoreferrerLinks != 0 && !isRelativeLink(link) {
2015-03-15 00:46:32 +01:00
relAttrs = append(relAttrs, "noreferrer")
}
if len(relAttrs) > 0 {
r.w.WriteString(fmt.Sprintf("\" rel=\"%s", strings.Join(relAttrs, " ")))
2015-03-15 00:46:32 +01:00
}
2014-03-21 03:52:46 +01:00
// blank target only add to external link
if r.Flags&HrefTargetBlank != 0 && !isRelativeLink(link) {
r.w.WriteString("\" target=\"_blank")
2014-03-21 03:52:46 +01:00
}
r.w.WriteString("\">")
2011-05-29 05:17:53 +02:00
// Pretty print: if we get an email address as
// an actual URI, e.g. `mailto:foo@bar.com`, we don't
// want to print the `mailto:` prefix
2011-05-31 19:49:49 +02:00
switch {
case bytes.HasPrefix(link, []byte("mailto://")):
r.attrEscape(link[len("mailto://"):])
2011-05-31 19:49:49 +02:00
case bytes.HasPrefix(link, []byte("mailto:")):
r.attrEscape(link[len("mailto:"):])
default:
r.entityEscapeWithSkip(link, skipRanges)
2011-05-29 05:17:53 +02:00
}
r.w.WriteString("</a>")
2011-05-29 05:17:53 +02:00
}
func (r *HTML) CodeSpan(text []byte) {
r.w.WriteString("<code>")
r.attrEscape(text)
r.w.WriteString("</code>")
2011-05-29 05:17:53 +02:00
}
func (r *HTML) DoubleEmphasis(text []byte) {
r.w.WriteString("<strong>")
r.w.Write(text)
r.w.WriteString("</strong>")
2011-05-29 05:17:53 +02:00
}
func (r *HTML) Emphasis(text []byte) {
2011-05-29 05:17:53 +02:00
if len(text) == 0 {
2011-06-29 21:00:54 +02:00
return
2011-05-29 05:17:53 +02:00
}
r.w.WriteString("<em>")
r.w.Write(text)
r.w.WriteString("</em>")
2011-05-29 05:17:53 +02:00
}
func (r *HTML) maybeWriteAbsolutePrefix(link []byte) {
if r.AbsolutePrefix != "" && isRelativeLink(link) && link[0] != '.' {
r.w.WriteString(r.AbsolutePrefix)
if link[0] != '/' {
r.w.WriteByte('/')
}
}
}
func (r *HTML) Image(link []byte, title []byte, alt []byte) {
if r.Flags&SkipImages != 0 {
2011-06-29 21:00:54 +02:00
return
}
r.w.WriteString("<img src=\"")
r.maybeWriteAbsolutePrefix(link)
r.attrEscape(link)
r.w.WriteString("\" alt=\"")
2011-05-29 05:17:53 +02:00
if len(alt) > 0 {
r.attrEscape(alt)
2011-05-29 05:17:53 +02:00
}
if len(title) > 0 {
r.w.WriteString("\" title=\"")
r.attrEscape(title)
2011-05-29 05:17:53 +02:00
}
r.w.WriteByte('"')
r.w.WriteString(r.closeTag)
2011-05-29 05:17:53 +02:00
}
func (r *HTML) LineBreak() {
r.w.WriteString("<br")
r.w.WriteString(r.closeTag)
r.w.WriteByte('\n')
2011-05-29 05:17:53 +02:00
}
func (r *HTML) Link(link []byte, title []byte, content []byte) {
if r.Flags&SkipLinks != 0 {
// write the link text out but don't link it, just mark it with typewriter font
r.w.WriteString("<tt>")
r.attrEscape(content)
r.w.WriteString("</tt>")
2011-06-29 21:00:54 +02:00
return
}
2011-05-29 05:17:53 +02:00
if r.Flags&Safelink != 0 && !isSafeLink(link) {
// write the link text out but don't link it, just mark it with typewriter font
r.w.WriteString("<tt>")
r.attrEscape(content)
r.w.WriteString("</tt>")
2011-06-29 21:00:54 +02:00
return
2011-05-29 05:17:53 +02:00
}
r.w.WriteString("<a href=\"")
r.maybeWriteAbsolutePrefix(link)
r.attrEscape(link)
2011-05-29 05:17:53 +02:00
if len(title) > 0 {
r.w.WriteString("\" title=\"")
r.attrEscape(title)
2011-05-29 05:17:53 +02:00
}
2015-03-15 00:46:32 +01:00
var relAttrs []string
if r.Flags&NofollowLinks != 0 && !isRelativeLink(link) {
2015-03-15 00:46:32 +01:00
relAttrs = append(relAttrs, "nofollow")
2014-02-25 15:21:57 +01:00
}
if r.Flags&NoreferrerLinks != 0 && !isRelativeLink(link) {
2015-03-15 00:46:32 +01:00
relAttrs = append(relAttrs, "noreferrer")
}
if len(relAttrs) > 0 {
r.w.WriteString(fmt.Sprintf("\" rel=\"%s", strings.Join(relAttrs, " ")))
2015-03-15 00:46:32 +01:00
}
2014-03-21 03:52:46 +01:00
// blank target only add to external link
if r.Flags&HrefTargetBlank != 0 && !isRelativeLink(link) {
r.w.WriteString("\" target=\"_blank")
2014-03-21 03:52:46 +01:00
}
r.w.WriteString("\">")
r.w.Write(content)
r.w.WriteString("</a>")
2011-06-29 21:00:54 +02:00
return
2011-05-29 05:17:53 +02:00
}
func (r *HTML) RawHtmlTag(text []byte) {
if r.Flags&SkipHTML != 0 {
2011-06-29 21:00:54 +02:00
return
2011-05-29 05:17:53 +02:00
}
if r.Flags&SkipStyle != 0 && isHtmlTag(text, "style") {
2011-06-29 21:00:54 +02:00
return
2011-05-29 05:17:53 +02:00
}
if r.Flags&SkipLinks != 0 && isHtmlTag(text, "a") {
2011-06-29 21:00:54 +02:00
return
2011-05-29 05:17:53 +02:00
}
if r.Flags&SkipImages != 0 && isHtmlTag(text, "img") {
2011-06-29 21:00:54 +02:00
return
2011-05-29 05:17:53 +02:00
}
r.w.Write(text)
2011-05-29 05:17:53 +02:00
}
func (r *HTML) TripleEmphasis(text []byte) {
r.w.WriteString("<strong><em>")
r.w.Write(text)
r.w.WriteString("</em></strong>")
2011-05-29 05:17:53 +02:00
}
func (r *HTML) StrikeThrough(text []byte) {
r.w.WriteString("<del>")
r.w.Write(text)
r.w.WriteString("</del>")
2011-05-29 05:17:53 +02:00
}
func (r *HTML) FootnoteRef(ref []byte, id int) {
slug := slugify(ref)
r.w.WriteString(`<sup class="footnote-ref" id="`)
r.w.WriteString(`fnref:`)
r.w.WriteString(r.FootnoteAnchorPrefix)
r.w.Write(slug)
r.w.WriteString(`"><a rel="footnote" href="#`)
r.w.WriteString(`fn:`)
r.w.WriteString(r.FootnoteAnchorPrefix)
r.w.Write(slug)
r.w.WriteString(`">`)
r.w.WriteString(strconv.Itoa(id))
r.w.WriteString(`</a></sup>`)
}
func (r *HTML) Entity(entity []byte) {
r.w.Write(entity)
}
func (r *HTML) NormalText(text []byte) {
if r.Extensions&Smartypants != 0 {
r.Smartypants(text)
} else {
r.attrEscape(text)
}
2011-05-29 05:17:53 +02:00
}
func (r *HTML) Smartypants(text []byte) {
r.w.Write(NewSmartypantsRenderer(r.Extensions).Process(text))
2011-06-29 19:13:17 +02:00
}
func (r *HTML) DocumentHeader() {
if r.Flags&CompletePage == 0 {
return
}
ending := ""
if r.Flags&UseXHTML != 0 {
r.w.WriteString("<!DOCTYPE html PUBLIC \"-//W3C//DTD XHTML 1.0 Transitional//EN\" ")
r.w.WriteString("\"http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd\">\n")
r.w.WriteString("<html xmlns=\"http://www.w3.org/1999/xhtml\">\n")
ending = " /"
} else {
r.w.WriteString("<!DOCTYPE html>\n")
r.w.WriteString("<html>\n")
}
r.w.WriteString("<head>\n")
r.w.WriteString(" <title>")
r.NormalText([]byte(r.Title))
r.w.WriteString("</title>\n")
r.w.WriteString(" <meta name=\"GENERATOR\" content=\"Blackfriday Markdown Processor v")
r.w.WriteString(VERSION)
r.w.WriteString("\"")
r.w.WriteString(ending)
r.w.WriteString(">\n")
r.w.WriteString(" <meta charset=\"utf-8\"")
r.w.WriteString(ending)
r.w.WriteString(">\n")
if r.CSS != "" {
r.w.WriteString(" <link rel=\"stylesheet\" type=\"text/css\" href=\"")
r.attrEscape([]byte(r.CSS))
r.w.WriteString("\"")
r.w.WriteString(ending)
r.w.WriteString(">\n")
}
r.w.WriteString("</head>\n")
r.w.WriteString("<body>\n")
2011-06-29 18:36:56 +02:00
r.tocMarker = r.w.Len() // XXX
}
func (r *HTML) DocumentFooter() {
2011-06-29 18:36:56 +02:00
// finalize and insert the table of contents
if r.Extensions&TOC != 0 {
r.TocFinalize()
2011-06-29 18:36:56 +02:00
// now we have to insert the table of contents into the document
var temp bytes.Buffer
// start by making a copy of everything after the document header
temp.Write(r.w.Bytes()[r.tocMarker:])
2011-06-29 18:36:56 +02:00
// now clear the copied material from the main output buffer
r.w.Truncate(r.tocMarker)
2011-06-29 18:36:56 +02:00
// corner case spacing issue
if r.Flags&CompletePage != 0 {
r.w.WriteByte('\n')
}
2011-06-29 18:36:56 +02:00
// insert the table of contents
r.w.WriteString("<nav>\n")
r.w.Write(r.toc.Bytes())
r.w.WriteString("</nav>\n")
2011-06-29 18:36:56 +02:00
// corner case spacing issue
if r.Flags&CompletePage == 0 && r.Extensions&OmitContents == 0 {
r.w.WriteByte('\n')
}
2011-06-29 18:36:56 +02:00
// write out everything that came after it
if r.Extensions&OmitContents == 0 {
r.w.Write(temp.Bytes())
2011-06-29 18:36:56 +02:00
}
}
if r.Flags&CompletePage != 0 {
r.w.WriteString("\n</body>\n")
r.w.WriteString("</html>\n")
2011-05-29 05:17:53 +02:00
}
2011-05-29 05:17:53 +02:00
}
func (r *HTML) TocHeaderWithAnchor(text []byte, level int, anchor string) {
for level > r.currentLevel {
2011-06-29 18:36:56 +02:00
switch {
case bytes.HasSuffix(r.toc.Bytes(), []byte("</li>\n")):
// this sublist can nest underneath a header
size := r.toc.Len()
r.toc.Truncate(size - len("</li>\n"))
2011-06-29 18:36:56 +02:00
case r.currentLevel > 0:
r.toc.WriteString("<li>")
2011-06-29 18:36:56 +02:00
}
if r.toc.Len() > 0 {
r.toc.WriteByte('\n')
}
r.toc.WriteString("<ul>\n")
r.currentLevel++
2011-06-29 18:36:56 +02:00
}
for level < r.currentLevel {
r.toc.WriteString("</ul>")
if r.currentLevel > 1 {
r.toc.WriteString("</li>\n")
2011-06-29 18:36:56 +02:00
}
r.currentLevel--
2011-06-29 18:36:56 +02:00
}
r.toc.WriteString("<li><a href=\"#")
if anchor != "" {
r.toc.WriteString(anchor)
} else {
r.toc.WriteString("toc_")
r.toc.WriteString(strconv.Itoa(r.headerCount))
}
r.toc.WriteString("\">")
r.headerCount++
2011-06-29 18:36:56 +02:00
r.toc.Write(text)
2011-06-29 18:36:56 +02:00
r.toc.WriteString("</a></li>\n")
2011-06-29 18:36:56 +02:00
}
func (r *HTML) TocHeader(text []byte, level int) {
r.TocHeaderWithAnchor(text, level, "")
}
func (r *HTML) TocFinalize() {
for r.currentLevel > 1 {
r.toc.WriteString("</ul></li>\n")
r.currentLevel--
2011-05-29 05:17:53 +02:00
}
if r.currentLevel > 0 {
r.toc.WriteString("</ul>\n")
2011-05-29 05:17:53 +02:00
}
}
func isHtmlTag(tag []byte, tagname string) bool {
2013-04-18 02:15:47 +02:00
found, _ := findHtmlTagPos(tag, tagname)
return found
}
2014-01-21 23:45:43 +01:00
// Look for a character, but ignore it when it's in any kind of quotes, it
// might be JavaScript
func skipUntilCharIgnoreQuotes(html []byte, start int, char byte) int {
inSingleQuote := false
inDoubleQuote := false
inGraveQuote := false
i := start
for i < len(html) {
switch {
case html[i] == char && !inSingleQuote && !inDoubleQuote && !inGraveQuote:
return i
case html[i] == '\'':
inSingleQuote = !inSingleQuote
case html[i] == '"':
inDoubleQuote = !inDoubleQuote
case html[i] == '`':
inGraveQuote = !inGraveQuote
}
i++
}
return start
}
2013-04-18 02:15:47 +02:00
func findHtmlTagPos(tag []byte, tagname string) (bool, int) {
2011-05-29 05:17:53 +02:00
i := 0
if i < len(tag) && tag[0] != '<' {
2013-04-18 02:15:47 +02:00
return false, -1
2011-05-29 05:17:53 +02:00
}
i++
2013-04-13 21:26:29 +02:00
i = skipSpace(tag, i)
2011-05-29 05:17:53 +02:00
if i < len(tag) && tag[i] == '/' {
i++
}
2013-04-13 21:26:29 +02:00
i = skipSpace(tag, i)
j := 0
2011-06-29 00:02:12 +02:00
for ; i < len(tag); i, j = i+1, j+1 {
if j >= len(tagname) {
2011-05-29 05:17:53 +02:00
break
}
2013-04-13 21:34:37 +02:00
if strings.ToLower(string(tag[i]))[0] != tagname[j] {
2013-04-18 02:15:47 +02:00
return false, -1
2011-05-29 05:17:53 +02:00
}
}
if i == len(tag) {
2013-04-18 02:15:47 +02:00
return false, -1
}
2014-01-21 23:45:43 +01:00
rightAngle := skipUntilCharIgnoreQuotes(tag, i, '>')
if rightAngle >= i {
2014-01-21 23:45:43 +01:00
return true, rightAngle
2011-05-29 05:17:53 +02:00
}
2013-04-18 02:15:47 +02:00
return false, -1
2011-05-29 05:17:53 +02:00
}
func skipUntilChar(text []byte, start int, char byte) int {
i := start
for i < len(text) && text[i] != char {
i++
}
return i
}
2013-04-13 21:26:29 +02:00
func skipSpace(tag []byte, i int) int {
for i < len(tag) && isspace(tag[i]) {
i++
}
return i
}
func skipChar(data []byte, start int, char byte) int {
i := start
for i < len(data) && data[i] == char {
i++
}
return i
}
2014-03-21 03:52:46 +01:00
func isRelativeLink(link []byte) (yes bool) {
// a tag begin with '#'
if link[0] == '#' {
2015-04-11 17:06:30 +02:00
return true
2014-03-21 03:52:46 +01:00
}
// link begin with '/' but not '//', the second maybe a protocol relative link
if len(link) >= 2 && link[0] == '/' && link[1] != '/' {
2015-04-11 17:06:30 +02:00
return true
2014-03-21 03:52:46 +01:00
}
// only the root '/'
if len(link) == 1 && link[0] == '/' {
2015-04-11 17:06:30 +02:00
return true
2014-03-21 03:52:46 +01:00
}
// current directory : begin with "./"
2015-04-11 17:06:30 +02:00
if bytes.HasPrefix(link, []byte("./")) {
return true
}
// parent directory : begin with "../"
2015-04-11 17:06:30 +02:00
if bytes.HasPrefix(link, []byte("../")) {
return true
}
2015-04-11 17:06:30 +02:00
return false
2014-03-21 03:52:46 +01:00
}
func (r *HTML) ensureUniqueHeaderID(id string) string {
for count, found := r.headerIDs[id]; found; count, found = r.headerIDs[id] {
tmp := fmt.Sprintf("%s-%d", id, count+1)
if _, tmpFound := r.headerIDs[tmp]; !tmpFound {
r.headerIDs[id] = count + 1
id = tmp
} else {
id = id + "-1"
}
}
if _, found := r.headerIDs[id]; !found {
r.headerIDs[id] = 0
}
return id
}
func (r *HTML) addAbsPrefix(link []byte) []byte {
if r.AbsolutePrefix != "" && isRelativeLink(link) && link[0] != '.' {
newDest := r.AbsolutePrefix
if link[0] != '/' {
newDest += "/"
}
newDest += string(link)
return []byte(newDest)
}
return link
}
func appendLinkAttrs(attrs []string, flags HTMLFlags, link []byte) []string {
if isRelativeLink(link) {
return attrs
}
val := []string{}
if flags&NofollowLinks != 0 {
val = append(val, "nofollow")
}
if flags&NoreferrerLinks != 0 {
val = append(val, "noreferrer")
}
if flags&HrefTargetBlank != 0 {
attrs = append(attrs, "target=\"_blank\"")
}
if len(val) == 0 {
return attrs
}
attr := fmt.Sprintf("rel=%q", strings.Join(val, " "))
return append(attrs, attr)
}
func isMailto(link []byte) bool {
return bytes.HasPrefix(link, []byte("mailto:"))
}
2016-04-04 13:08:35 +02:00
func needSkipLink(flags HTMLFlags, dest []byte) bool {
if flags&SkipLinks != 0 {
return true
}
return flags&Safelink != 0 && !isSafeLink(dest) && !isMailto(dest)
}
func isSmartypantable(node *Node) bool {
pt := node.Parent.Type
return pt != Link && pt != CodeBlock && pt != Code
}
func appendLanguageAttr(attrs []string, info []byte) []string {
infoWords := bytes.Split(info, []byte("\t "))
if len(infoWords) > 0 && len(infoWords[0]) > 0 {
attrs = append(attrs, fmt.Sprintf("class=\"language-%s\"", infoWords[0]))
}
return attrs
}
func tag(name string, attrs []string, selfClosing bool) []byte {
result := "<" + name
if attrs != nil && len(attrs) > 0 {
result += " " + strings.Join(attrs, " ")
}
if selfClosing {
result += " /"
}
return []byte(result + ">")
}
func footnoteRef(prefix string, node *Node) []byte {
urlFrag := prefix + string(slugify(node.Destination))
anchor := fmt.Sprintf(`<a rel="footnote" href="#fn:%s">%d</a>`, urlFrag, node.NoteID)
return []byte(fmt.Sprintf(`<sup class="footnote-ref" id="fnref:%s">%s</sup>`, urlFrag, anchor))
}
func footnoteItem(prefix string, slug []byte) []byte {
return []byte(fmt.Sprintf(`<li id="fn:%s%s">`, prefix, slug))
}
func footnoteReturnLink(prefix, returnLink string, slug []byte) []byte {
const format = ` <a class="footnote-return" href="#fnref:%s%s">%s</a>`
return []byte(fmt.Sprintf(format, prefix, slug, returnLink))
}
func itemOpenCR(node *Node) bool {
if node.Prev == nil {
return false
}
ld := node.Parent.ListData
return !ld.Tight && ld.ListFlags&ListTypeDefinition == 0
}
func skipParagraphTags(node *Node) bool {
grandparent := node.Parent.Parent
if grandparent == nil || grandparent.Type != List {
return false
}
tightOrTerm := grandparent.Tight || node.Parent.ListFlags&ListTypeTerm != 0
return grandparent.Type == List && tightOrTerm
}
func cellAlignment(align CellAlignFlags) string {
switch align {
case TableAlignmentLeft:
return "left"
case TableAlignmentRight:
return "right"
case TableAlignmentCenter:
return "center"
default:
return ""
}
}
func esc(text []byte, preserveEntities bool) []byte {
return attrEscape2(text)
}
func escCode(text []byte, preserveEntities bool) []byte {
e1 := []byte(html.EscapeString(string(text)))
e2 := bytes.Replace(e1, []byte("&#34;"), []byte("&quot;"), -1)
return bytes.Replace(e2, []byte("&#39;"), []byte{'\''}, -1)
}
func (r *HTML) out(w io.Writer, text []byte) {
if r.disableTags > 0 {
w.Write(reHtmlTag.ReplaceAll(text, []byte{}))
} else {
w.Write(text)
}
r.lastOutputLen = len(text)
}
func (r *HTML) cr(w io.Writer) {
if r.lastOutputLen > 0 {
r.out(w, []byte{'\n'})
}
}
func (r *HTML) RenderNode(w io.Writer, node *Node, entering bool) WalkStatus {
attrs := []string{}
switch node.Type {
case Text:
r.out(w, node.Literal)
break
case Softbreak:
r.out(w, []byte("\n"))
// TODO: make it configurable via out(renderer.softbreak)
case Hardbreak:
r.out(w, tag("br", nil, true))
r.cr(w)
case Emph:
if entering {
r.out(w, tag("em", nil, false))
} else {
r.out(w, tag("/em", nil, false))
}
break
case Strong:
if entering {
r.out(w, tag("strong", nil, false))
} else {
r.out(w, tag("/strong", nil, false))
}
break
case Del:
if entering {
r.out(w, tag("del", nil, false))
} else {
r.out(w, tag("/del", nil, false))
}
case HTMLSpan:
if r.Flags&SkipHTML != 0 {
2016-04-05 11:37:02 +02:00
break
}
if r.Flags&SkipStyle != 0 && isHtmlTag(node.Literal, "style") {
break
}
//if options.safe {
// out(w, "<!-- raw HTML omitted -->")
//} else {
r.out(w, node.Literal)
//}
case Link:
// mark it but don't link it if it is not a safe link: no smartypants
dest := node.LinkData.Destination
if needSkipLink(r.Flags, dest) {
if entering {
r.out(w, tag("tt", nil, false))
} else {
r.out(w, tag("/tt", nil, false))
}
} else {
if entering {
dest = r.addAbsPrefix(dest)
//if (!(options.safe && potentiallyUnsafe(node.destination))) {
attrs = append(attrs, fmt.Sprintf("href=%q", esc(dest, true)))
//}
if node.NoteID != 0 {
r.out(w, footnoteRef(r.FootnoteAnchorPrefix, node))
break
}
attrs = appendLinkAttrs(attrs, r.Flags, dest)
if len(node.LinkData.Title) > 0 {
attrs = append(attrs, fmt.Sprintf("title=%q", esc(node.LinkData.Title, true)))
}
r.out(w, tag("a", attrs, false))
} else {
if node.NoteID != 0 {
break
}
r.out(w, tag("/a", nil, false))
}
}
case Image:
if r.Flags&SkipImages != 0 {
2016-04-05 10:12:30 +02:00
return SkipChildren
}
if entering {
dest := node.LinkData.Destination
dest = r.addAbsPrefix(dest)
if r.disableTags == 0 {
//if options.safe && potentiallyUnsafe(dest) {
//out(w, `<img src="" alt="`)
//} else {
r.out(w, []byte(fmt.Sprintf(`<img src="%s" alt="`, esc(dest, true))))
//}
}
r.disableTags++
} else {
r.disableTags--
if r.disableTags == 0 {
if node.LinkData.Title != nil {
r.out(w, []byte(`" title="`))
r.out(w, esc(node.LinkData.Title, true))
}
r.out(w, []byte(`" />`))
}
}
case Code:
r.out(w, tag("code", nil, false))
r.out(w, escCode(node.Literal, false))
r.out(w, tag("/code", nil, false))
case Document:
break
case Paragraph:
if skipParagraphTags(node) {
break
}
if entering {
// TODO: untangle this clusterfuck about when the newlines need
// to be added and when not.
if node.Prev != nil {
t := node.Prev.Type
if t == HTMLBlock || t == List || t == Paragraph || t == Header || t == CodeBlock || t == BlockQuote || t == HorizontalRule {
r.cr(w)
}
}
if node.Parent.Type == BlockQuote && node.Prev == nil {
r.cr(w)
}
r.out(w, tag("p", attrs, false))
} else {
r.out(w, tag("/p", attrs, false))
if !(node.Parent.Type == Item && node.Next == nil) {
r.cr(w)
}
}
break
case BlockQuote:
if entering {
r.cr(w)
r.out(w, tag("blockquote", attrs, false))
} else {
r.out(w, tag("/blockquote", nil, false))
r.cr(w)
}
break
case HTMLBlock:
if r.Flags&SkipHTML != 0 {
2016-04-05 11:37:02 +02:00
break
}
r.cr(w)
r.out(w, node.Literal)
r.cr(w)
case Header:
tagname := fmt.Sprintf("h%d", node.Level)
if entering {
if node.IsTitleblock {
attrs = append(attrs, `class="title"`)
}
if node.HeaderID != "" {
id := r.ensureUniqueHeaderID(node.HeaderID)
if r.HeaderIDPrefix != "" {
id = r.HeaderIDPrefix + id
}
if r.HeaderIDSuffix != "" {
id = id + r.HeaderIDSuffix
}
attrs = append(attrs, fmt.Sprintf(`id="%s"`, id))
}
r.cr(w)
r.out(w, tag(tagname, attrs, false))
} else {
r.out(w, tag("/"+tagname, nil, false))
if !(node.Parent.Type == Item && node.Next == nil) {
r.cr(w)
}
}
break
case HorizontalRule:
r.cr(w)
r.out(w, tag("hr", attrs, r.Flags&UseXHTML != 0))
r.cr(w)
break
case List:
tagName := "ul"
if node.ListFlags&ListTypeOrdered != 0 {
tagName = "ol"
}
if node.ListFlags&ListTypeDefinition != 0 {
tagName = "dl"
}
if entering {
// var start = node.listStart;
// if (start !== null && start !== 1) {
// attrs.push(['start', start.toString()]);
// }
r.cr(w)
if node.Parent.Type == Item && node.Parent.Parent.Tight {
r.cr(w)
}
r.out(w, tag(tagName, attrs, false))
r.cr(w)
} else {
r.out(w, tag("/"+tagName, nil, false))
//cr(w)
//if node.parent.Type != Item {
// cr(w)
//}
if node.Parent.Type == Item && node.Next != nil {
r.cr(w)
}
if node.Parent.Type == Document || node.Parent.Type == BlockQuote {
r.cr(w)
}
}
case Item:
tagName := "li"
if node.ListFlags&ListTypeDefinition != 0 {
tagName = "dd"
}
if node.ListFlags&ListTypeTerm != 0 {
tagName = "dt"
}
if entering {
if itemOpenCR(node) {
r.cr(w)
}
if node.ListData.RefLink != nil {
slug := slugify(node.ListData.RefLink)
r.out(w, footnoteItem(r.FootnoteAnchorPrefix, slug))
break
}
r.out(w, tag(tagName, nil, false))
} else {
if node.ListData.RefLink != nil {
slug := slugify(node.ListData.RefLink)
if r.Flags&FootnoteReturnLinks != 0 {
r.out(w, footnoteReturnLink(r.FootnoteAnchorPrefix, r.FootnoteReturnLinkContents, slug))
}
}
r.out(w, tag("/"+tagName, nil, false))
r.cr(w)
}
case CodeBlock:
attrs = appendLanguageAttr(attrs, node.Info)
r.cr(w)
r.out(w, tag("pre", nil, false))
r.out(w, tag("code", attrs, false))
r.out(w, escCode(node.Literal, false))
r.out(w, tag("/code", nil, false))
r.out(w, tag("/pre", nil, false))
if node.Parent.Type != Item {
r.cr(w)
}
case Table:
if entering {
r.cr(w)
r.out(w, tag("table", nil, false))
} else {
r.out(w, tag("/table", nil, false))
r.cr(w)
}
case TableCell:
tagName := "td"
if node.IsHeader {
tagName = "th"
}
if entering {
align := cellAlignment(node.Align)
if align != "" {
attrs = append(attrs, fmt.Sprintf(`align="%s"`, align))
}
if node.Prev == nil {
r.cr(w)
}
r.out(w, tag(tagName, attrs, false))
} else {
r.out(w, tag("/"+tagName, nil, false))
r.cr(w)
}
case TableHead:
if entering {
r.cr(w)
r.out(w, tag("thead", nil, false))
} else {
r.out(w, tag("/thead", nil, false))
r.cr(w)
}
case TableBody:
if entering {
r.cr(w)
r.out(w, tag("tbody", nil, false))
// XXX: this is to adhere to a rather silly test. Should fix test.
if node.FirstChild == nil {
r.cr(w)
}
} else {
r.out(w, tag("/tbody", nil, false))
r.cr(w)
}
case TableRow:
if entering {
r.cr(w)
r.out(w, tag("tr", nil, false))
} else {
r.out(w, tag("/tr", nil, false))
r.cr(w)
}
default:
panic("Unknown node type " + node.Type.String())
}
return GoToNext
}
func (r *HTML) writeDocumentHeader(w *bytes.Buffer, sr *SPRenderer) {
if r.Flags&CompletePage == 0 {
return
}
ending := ""
if r.Flags&UseXHTML != 0 {
w.WriteString("<!DOCTYPE html PUBLIC \"-//W3C//DTD XHTML 1.0 Transitional//EN\" ")
w.WriteString("\"http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd\">\n")
w.WriteString("<html xmlns=\"http://www.w3.org/1999/xhtml\">\n")
ending = " /"
} else {
w.WriteString("<!DOCTYPE html>\n")
w.WriteString("<html>\n")
}
w.WriteString("<head>\n")
w.WriteString(" <title>")
if r.Extensions&Smartypants != 0 {
w.Write(sr.Process([]byte(r.Title)))
} else {
w.Write(esc([]byte(r.Title), false))
}
w.WriteString("</title>\n")
w.WriteString(" <meta name=\"GENERATOR\" content=\"Blackfriday Markdown Processor v")
w.WriteString(VERSION)
w.WriteString("\"")
w.WriteString(ending)
w.WriteString(">\n")
w.WriteString(" <meta charset=\"utf-8\"")
w.WriteString(ending)
w.WriteString(">\n")
if r.CSS != "" {
w.WriteString(" <link rel=\"stylesheet\" type=\"text/css\" href=\"")
r.attrEscape([]byte(r.CSS))
w.WriteString("\"")
w.WriteString(ending)
w.WriteString(">\n")
}
w.WriteString("</head>\n")
w.WriteString("<body>\n\n")
}
func (r *HTML) writeDocumentFooter(w *bytes.Buffer) {
if r.Flags&CompletePage == 0 {
return
}
w.WriteString("\n</body>\n</html>\n")
}
func (r *HTML) Render(ast *Node) []byte {
//println("render_Blackfriday")
//dump(ast)
// Run Smartypants if it's enabled or simply escape text if not
sr := NewSmartypantsRenderer(r.Extensions)
ast.Walk(func(node *Node, entering bool) WalkStatus {
if node.Type == Text {
if r.Extensions&Smartypants != 0 {
node.Literal = sr.Process(node.Literal)
} else {
node.Literal = esc(node.Literal, false)
}
}
return GoToNext
})
var buff bytes.Buffer
r.writeDocumentHeader(&buff, sr)
ast.Walk(func(node *Node, entering bool) WalkStatus {
return r.RenderNode(&buff, node, entering)
})
r.writeDocumentFooter(&buff)
return buff.Bytes()
}