blackfriday/html.go

635 lines
13 KiB
Go
Raw Normal View History

2011-05-29 11:17:53 +08:00
//
2011-06-28 10:11:32 +08:00
// Blackfriday Markdown Processor
// Available at http://github.com/russross/blackfriday
//
// Copyright © 2011 Russ Ross <russ@russross.com>.
// Licensed under the Simplified BSD License.
// See README.md for details.
2011-05-29 11:17:53 +08:00
//
//
//
// HTML rendering backend
//
//
package blackfriday
import (
"bytes"
"fmt"
"strconv"
)
const (
HTML_SKIP_HTML = 1 << iota
HTML_SKIP_STYLE
HTML_SKIP_IMAGES
HTML_SKIP_LINKS
HTML_SAFELINK
HTML_TOC
HTML_GITHUB_BLOCKCODE
HTML_USE_XHTML
HTML_USE_SMARTYPANTS
HTML_SMARTYPANTS_FRACTIONS
HTML_SMARTYPANTS_LATEX_DASHES
)
type htmlOptions struct {
2011-06-01 01:49:49 +08:00
flags int
closeTag string // how to end singleton tags: either " />\n" or ">\n"
tocData struct {
headerCount int
currentLevel int
2011-05-29 11:17:53 +08:00
}
smartypants *SmartypantsRenderer
}
var xhtmlClose = " />\n"
var htmlClose = ">\n"
2011-05-29 11:17:53 +08:00
func HtmlRenderer(flags int) *Renderer {
// configure the rendering engine
r := new(Renderer)
if flags&HTML_GITHUB_BLOCKCODE == 0 {
r.BlockCode = htmlBlockCode
2011-05-29 11:17:53 +08:00
} else {
r.BlockCode = htmlBlockCodeGithub
2011-05-29 11:17:53 +08:00
}
r.BlockQuote = htmlBlockQuote
2011-05-29 11:17:53 +08:00
if flags&HTML_SKIP_HTML == 0 {
r.BlockHtml = htmlRawBlock
}
r.Header = htmlHeader
r.HRule = htmlHRule
r.List = htmlList
r.ListItem = htmlListItem
r.Paragraph = htmlParagraph
r.Table = htmlTable
r.TableRow = htmlTableRow
r.TableCell = htmlTableCell
r.AutoLink = htmlAutoLink
r.CodeSpan = htmlCodeSpan
r.DoubleEmphasis = htmlDoubleEmphasis
r.Emphasis = htmlEmphasis
2011-05-29 11:17:53 +08:00
if flags&HTML_SKIP_IMAGES == 0 {
r.Image = htmlImage
2011-05-29 11:17:53 +08:00
}
r.LineBreak = htmlLineBreak
2011-05-29 11:17:53 +08:00
if flags&HTML_SKIP_LINKS == 0 {
r.Link = htmlLink
2011-05-29 11:17:53 +08:00
}
r.RawHtmlTag = htmlRawTag
r.TripleEmphasis = htmlTripleEmphasis
r.StrikeThrough = htmlStrikeThrough
2011-05-29 11:17:53 +08:00
var cb *SmartypantsRenderer
if flags&HTML_USE_SMARTYPANTS == 0 {
r.NormalText = htmlNormalText
2011-05-29 11:17:53 +08:00
} else {
cb = Smartypants(flags)
r.NormalText = htmlSmartypants
2011-05-29 11:17:53 +08:00
}
closeTag := htmlClose
2011-05-29 11:17:53 +08:00
if flags&HTML_USE_XHTML != 0 {
closeTag = xhtmlClose
2011-05-29 11:17:53 +08:00
}
r.Opaque = &htmlOptions{flags: flags, closeTag: closeTag, smartypants: cb}
2011-05-29 11:17:53 +08:00
return r
}
func HtmlTocRenderer(flags int) *Renderer {
// configure the rendering engine
r := new(Renderer)
r.Header = htmlTocHeader
2011-05-29 11:17:53 +08:00
r.CodeSpan = htmlCodeSpan
r.DoubleEmphasis = htmlDoubleEmphasis
r.Emphasis = htmlEmphasis
r.TripleEmphasis = htmlTripleEmphasis
r.StrikeThrough = htmlStrikeThrough
2011-05-29 11:17:53 +08:00
r.DocumentFooter = htmlTocFinalize
2011-05-29 11:17:53 +08:00
closeTag := ">\n"
2011-05-29 11:17:53 +08:00
if flags&HTML_USE_XHTML != 0 {
closeTag = " />\n"
2011-05-29 11:17:53 +08:00
}
r.Opaque = &htmlOptions{flags: flags | HTML_TOC, closeTag: closeTag}
2011-05-29 11:17:53 +08:00
return r
}
func attrEscape(out *bytes.Buffer, src []byte) {
org := 0
for i, ch := range src {
2011-06-26 05:02:46 +08:00
// 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
if ch == '"' {
if i > org {
// copy all the normal characters since the last escape
out.Write(src[org:i])
}
org = i + 1
2011-06-26 05:02:46 +08:00
out.WriteString("&quot;")
continue
}
if ch == '&' {
if i > org {
out.Write(src[org:i])
}
org = i + 1
2011-06-26 05:02:46 +08:00
out.WriteString("&amp;")
continue
}
if ch == '<' {
if i > org {
out.Write(src[org:i])
}
org = i + 1
2011-06-26 05:02:46 +08:00
out.WriteString("&lt;")
continue
}
if ch == '>' {
if i > org {
out.Write(src[org:i])
}
org = i + 1
2011-06-26 05:02:46 +08:00
out.WriteString("&gt;")
continue
2011-05-29 11:17:53 +08:00
}
}
if org < len(src) {
out.Write(src[org:])
}
2011-05-29 11:17:53 +08:00
}
func htmlHeader(out *bytes.Buffer, text func() bool, level int, opaque interface{}) {
2011-05-29 11:17:53 +08:00
options := opaque.(*htmlOptions)
marker := out.Len()
2011-05-29 11:17:53 +08:00
if marker > 0 {
out.WriteByte('\n')
2011-05-29 11:17:53 +08:00
}
if options.flags&HTML_TOC != 0 {
out.WriteString(fmt.Sprintf("<h%d id=\"toc_%d\">", level, options.tocData.headerCount))
options.tocData.headerCount++
2011-05-29 11:17:53 +08:00
} else {
out.WriteString(fmt.Sprintf("<h%d>", level))
2011-05-29 11:17:53 +08:00
}
if !text() {
out.Truncate(marker)
return
}
out.WriteString(fmt.Sprintf("</h%d>\n", level))
2011-05-29 11:17:53 +08:00
}
func htmlRawBlock(out *bytes.Buffer, text []byte, opaque interface{}) {
2011-05-29 11:17:53 +08:00
sz := len(text)
for sz > 0 && text[sz-1] == '\n' {
sz--
}
org := 0
for org < sz && text[org] == '\n' {
org++
}
if org >= sz {
return
}
if out.Len() > 0 {
out.WriteByte('\n')
2011-05-29 11:17:53 +08:00
}
out.Write(text[org:sz])
out.WriteByte('\n')
2011-05-29 11:17:53 +08:00
}
func htmlHRule(out *bytes.Buffer, opaque interface{}) {
2011-05-29 11:17:53 +08:00
options := opaque.(*htmlOptions)
if out.Len() > 0 {
out.WriteByte('\n')
2011-05-29 11:17:53 +08:00
}
out.WriteString("<hr")
out.WriteString(options.closeTag)
2011-05-29 11:17:53 +08:00
}
func htmlBlockCode(out *bytes.Buffer, text []byte, lang string, opaque interface{}) {
if out.Len() > 0 {
out.WriteByte('\n')
2011-05-29 11:17:53 +08:00
}
if lang != "" {
out.WriteString("<pre><code class=\"")
2011-05-29 11:17:53 +08:00
for i, cls := 0, 0; i < len(lang); i, cls = i+1, cls+1 {
for i < len(lang) && isspace(lang[i]) {
i++
}
if i < len(lang) {
org := i
for i < len(lang) && !isspace(lang[i]) {
i++
}
if lang[org] == '.' {
org++
}
if cls > 0 {
out.WriteByte(' ')
2011-05-29 11:17:53 +08:00
}
attrEscape(out, []byte(lang[org:]))
2011-05-29 11:17:53 +08:00
}
}
out.WriteString("\">")
2011-05-29 11:17:53 +08:00
} else {
out.WriteString("<pre><code>")
2011-05-29 11:17:53 +08:00
}
if len(text) > 0 {
attrEscape(out, text)
2011-05-29 11:17:53 +08:00
}
out.WriteString("</code></pre>\n")
2011-05-29 11:17:53 +08:00
}
/*
* GitHub style code block:
*
* <pre lang="LANG"><code>
* ...
* </pre></code>
*
* Unlike other parsers, we store the language identifier in the <pre>,
* and don't let the user generate custom classes.
*
* The language identifier in the <pre> block gets postprocessed and all
* the code inside gets syntax highlighted with Pygments. This is much safer
* than letting the user specify a CSS class for highlighting.
*
* Note that we only generate HTML for the first specifier.
* E.g.
* ~~~~ {.python .numbered} => <pre lang="python"><code>
*/
func htmlBlockCodeGithub(out *bytes.Buffer, text []byte, lang string, opaque interface{}) {
if out.Len() > 0 {
out.WriteByte('\n')
2011-05-29 11:17:53 +08:00
}
if len(lang) > 0 {
out.WriteString("<pre lang=\"")
2011-05-29 11:17:53 +08:00
i := 0
for i < len(lang) && !isspace(lang[i]) {
i++
}
if lang[0] == '.' {
attrEscape(out, []byte(lang[1:i]))
2011-05-29 11:17:53 +08:00
} else {
attrEscape(out, []byte(lang[:i]))
2011-05-29 11:17:53 +08:00
}
out.WriteString("\"><code>")
2011-05-29 11:17:53 +08:00
} else {
out.WriteString("<pre><code>")
2011-05-29 11:17:53 +08:00
}
if len(text) > 0 {
attrEscape(out, text)
2011-05-29 11:17:53 +08:00
}
out.WriteString("</code></pre>\n")
2011-05-29 11:17:53 +08:00
}
func htmlBlockQuote(out *bytes.Buffer, text []byte, opaque interface{}) {
out.WriteString("<blockquote>\n")
out.Write(text)
out.WriteString("</blockquote>")
2011-05-29 11:17:53 +08:00
}
func htmlTable(out *bytes.Buffer, header []byte, body []byte, columnData []int, opaque interface{}) {
if out.Len() > 0 {
out.WriteByte('\n')
2011-05-29 11:17:53 +08:00
}
out.WriteString("<table><thead>\n")
out.Write(header)
out.WriteString("\n</thead><tbody>\n")
out.Write(body)
out.WriteString("\n</tbody></table>")
2011-05-29 11:17:53 +08:00
}
func htmlTableRow(out *bytes.Buffer, text []byte, opaque interface{}) {
if out.Len() > 0 {
out.WriteByte('\n')
2011-05-29 11:17:53 +08:00
}
out.WriteString("<tr>\n")
out.Write(text)
out.WriteString("\n</tr>")
2011-05-29 11:17:53 +08:00
}
func htmlTableCell(out *bytes.Buffer, text []byte, align int, opaque interface{}) {
if out.Len() > 0 {
out.WriteByte('\n')
2011-05-29 11:17:53 +08:00
}
switch align {
case TABLE_ALIGNMENT_LEFT:
out.WriteString("<td align=\"left\">")
2011-05-29 11:17:53 +08:00
case TABLE_ALIGNMENT_RIGHT:
out.WriteString("<td align=\"right\">")
2011-05-29 11:17:53 +08:00
case TABLE_ALIGNMENT_CENTER:
out.WriteString("<td align=\"center\">")
2011-05-29 11:17:53 +08:00
default:
out.WriteString("<td>")
2011-05-29 11:17:53 +08:00
}
out.Write(text)
out.WriteString("</td>")
2011-05-29 11:17:53 +08:00
}
2011-06-26 05:02:46 +08:00
func htmlList(out *bytes.Buffer, text func() bool, flags int, opaque interface{}) {
marker := out.Len()
if marker > 0 {
out.WriteByte('\n')
2011-05-29 11:17:53 +08:00
}
if flags&LIST_TYPE_ORDERED != 0 {
out.WriteString("<ol>\n")
2011-05-29 11:17:53 +08:00
} else {
out.WriteString("<ul>\n")
2011-05-29 11:17:53 +08:00
}
2011-06-26 05:02:46 +08:00
if !text() {
out.Truncate(marker)
return
}
2011-05-29 11:17:53 +08:00
if flags&LIST_TYPE_ORDERED != 0 {
out.WriteString("</ol>\n")
2011-05-29 11:17:53 +08:00
} else {
out.WriteString("</ul>\n")
2011-05-29 11:17:53 +08:00
}
}
func htmlListItem(out *bytes.Buffer, text []byte, flags int, opaque interface{}) {
out.WriteString("<li>")
2011-05-29 11:17:53 +08:00
size := len(text)
for size > 0 && text[size-1] == '\n' {
size--
}
out.Write(text[:size])
out.WriteString("</li>\n")
2011-05-29 11:17:53 +08:00
}
func htmlParagraph(out *bytes.Buffer, text func() bool, opaque interface{}) {
marker := out.Len()
if marker > 0 {
out.WriteByte('\n')
2011-05-29 11:17:53 +08:00
}
out.WriteString("<p>")
if !text() {
out.Truncate(marker)
return
}
out.WriteString("</p>\n")
2011-05-29 11:17:53 +08:00
}
func htmlAutoLink(out *bytes.Buffer, link []byte, kind int, opaque interface{}) int {
2011-05-29 11:17:53 +08:00
options := opaque.(*htmlOptions)
if len(link) == 0 {
return 0
}
if options.flags&HTML_SAFELINK != 0 && !isSafeLink(link) && kind != LINK_TYPE_EMAIL {
2011-05-29 11:17:53 +08:00
return 0
}
out.WriteString("<a href=\"")
2011-05-29 11:17:53 +08:00
if kind == LINK_TYPE_EMAIL {
out.WriteString("mailto:")
2011-05-29 11:17:53 +08:00
}
attrEscape(out, link)
out.WriteString("\">")
2011-05-29 11:17:53 +08:00
/*
* 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
*/
2011-06-01 01:49:49 +08:00
switch {
case bytes.HasPrefix(link, []byte("mailto://")):
attrEscape(out, link[9:])
2011-06-01 01:49:49 +08:00
case bytes.HasPrefix(link, []byte("mailto:")):
attrEscape(out, link[7:])
default:
attrEscape(out, link)
2011-05-29 11:17:53 +08:00
}
out.WriteString("</a>")
2011-05-29 11:17:53 +08:00
return 1
}
func htmlCodeSpan(out *bytes.Buffer, text []byte, opaque interface{}) int {
out.WriteString("<code>")
attrEscape(out, text)
out.WriteString("</code>")
2011-05-29 11:17:53 +08:00
return 1
}
func htmlDoubleEmphasis(out *bytes.Buffer, text []byte, opaque interface{}) int {
2011-05-29 11:17:53 +08:00
if len(text) == 0 {
return 0
}
out.WriteString("<strong>")
out.Write(text)
out.WriteString("</strong>")
2011-05-29 11:17:53 +08:00
return 1
}
func htmlEmphasis(out *bytes.Buffer, text []byte, opaque interface{}) int {
2011-05-29 11:17:53 +08:00
if len(text) == 0 {
return 0
}
out.WriteString("<em>")
out.Write(text)
out.WriteString("</em>")
2011-05-29 11:17:53 +08:00
return 1
}
func htmlImage(out *bytes.Buffer, link []byte, title []byte, alt []byte, opaque interface{}) int {
2011-05-29 11:17:53 +08:00
options := opaque.(*htmlOptions)
if len(link) == 0 {
return 0
}
out.WriteString("<img src=\"")
attrEscape(out, link)
out.WriteString("\" alt=\"")
2011-05-29 11:17:53 +08:00
if len(alt) > 0 {
attrEscape(out, alt)
2011-05-29 11:17:53 +08:00
}
if len(title) > 0 {
out.WriteString("\" title=\"")
attrEscape(out, title)
2011-05-29 11:17:53 +08:00
}
out.WriteByte('"')
out.WriteString(options.closeTag)
2011-05-29 11:17:53 +08:00
return 1
}
func htmlLineBreak(out *bytes.Buffer, opaque interface{}) int {
2011-05-29 11:17:53 +08:00
options := opaque.(*htmlOptions)
out.WriteString("<br")
out.WriteString(options.closeTag)
2011-05-29 11:17:53 +08:00
return 1
}
func htmlLink(out *bytes.Buffer, link []byte, title []byte, content []byte, opaque interface{}) int {
2011-05-29 11:17:53 +08:00
options := opaque.(*htmlOptions)
if options.flags&HTML_SAFELINK != 0 && !isSafeLink(link) {
2011-05-29 11:17:53 +08:00
return 0
}
out.WriteString("<a href=\"")
attrEscape(out, link)
2011-05-29 11:17:53 +08:00
if len(title) > 0 {
out.WriteString("\" title=\"")
attrEscape(out, title)
2011-05-29 11:17:53 +08:00
}
out.WriteString("\">")
out.Write(content)
out.WriteString("</a>")
2011-05-29 11:17:53 +08:00
return 1
}
func htmlRawTag(out *bytes.Buffer, text []byte, opaque interface{}) int {
2011-05-29 11:17:53 +08:00
options := opaque.(*htmlOptions)
if options.flags&HTML_SKIP_HTML != 0 {
2011-05-29 11:17:53 +08:00
return 1
}
if options.flags&HTML_SKIP_STYLE != 0 && isHtmlTag(text, "style") {
2011-05-29 11:17:53 +08:00
return 1
}
if options.flags&HTML_SKIP_LINKS != 0 && isHtmlTag(text, "a") {
2011-05-29 11:17:53 +08:00
return 1
}
if options.flags&HTML_SKIP_IMAGES != 0 && isHtmlTag(text, "img") {
2011-05-29 11:17:53 +08:00
return 1
}
out.Write(text)
2011-05-29 11:17:53 +08:00
return 1
}
func htmlTripleEmphasis(out *bytes.Buffer, text []byte, opaque interface{}) int {
2011-05-29 11:17:53 +08:00
if len(text) == 0 {
return 0
}
out.WriteString("<strong><em>")
out.Write(text)
out.WriteString("</em></strong>")
2011-05-29 11:17:53 +08:00
return 1
}
func htmlStrikeThrough(out *bytes.Buffer, text []byte, opaque interface{}) int {
2011-05-29 11:17:53 +08:00
if len(text) == 0 {
return 0
}
out.WriteString("<del>")
out.Write(text)
out.WriteString("</del>")
2011-05-29 11:17:53 +08:00
return 1
}
func htmlNormalText(out *bytes.Buffer, text []byte, opaque interface{}) {
attrEscape(out, text)
2011-05-29 11:17:53 +08:00
}
func htmlTocHeader(out *bytes.Buffer, text func() bool, level int, opaque interface{}) {
2011-05-29 11:17:53 +08:00
options := opaque.(*htmlOptions)
marker := out.Len()
for level > options.tocData.currentLevel {
if options.tocData.currentLevel > 0 {
out.WriteString("<li>")
2011-05-29 11:17:53 +08:00
}
out.WriteString("<ul>\n")
options.tocData.currentLevel++
2011-05-29 11:17:53 +08:00
}
for level < options.tocData.currentLevel {
out.WriteString("</ul>")
if options.tocData.currentLevel > 1 {
out.WriteString("</li>\n")
2011-05-29 11:17:53 +08:00
}
options.tocData.currentLevel--
2011-05-29 11:17:53 +08:00
}
out.WriteString("<li><a href=\"#toc_")
out.WriteString(strconv.Itoa(options.tocData.headerCount))
out.WriteString("\">")
options.tocData.headerCount++
2011-05-29 11:17:53 +08:00
if !text() {
out.Truncate(marker)
return
2011-05-29 11:17:53 +08:00
}
out.WriteString("</a></li>\n")
2011-05-29 11:17:53 +08:00
}
func htmlTocFinalize(out *bytes.Buffer, opaque interface{}) {
2011-05-29 11:17:53 +08:00
options := opaque.(*htmlOptions)
for options.tocData.currentLevel > 1 {
out.WriteString("</ul></li>\n")
options.tocData.currentLevel--
2011-05-29 11:17:53 +08:00
}
if options.tocData.currentLevel > 0 {
out.WriteString("</ul>\n")
2011-05-29 11:17:53 +08:00
}
}
func isHtmlTag(tag []byte, tagname string) bool {
2011-05-29 11:17:53 +08:00
i := 0
if i < len(tag) && tag[0] != '<' {
return false
}
i++
for i < len(tag) && isspace(tag[i]) {
i++
}
if i < len(tag) && tag[i] == '/' {
i++
}
for i < len(tag) && isspace(tag[i]) {
i++
}
tag_i := i
for ; i < len(tag); i, tag_i = i+1, tag_i+1 {
if tag_i >= len(tagname) {
break
}
if tag[i] != tagname[tag_i] {
return false
}
}
if i == len(tag) {
return false
}
return isspace(tag[i]) || tag[i] == '>'
}