all repos — grayfriday @ a2fda5e98fedf99810df6fb0f5877988795a3733

blackfriday fork with a few changes

html.go (view raw)

  1//
  2// Blackfriday Markdown Processor
  3// Available at http://github.com/russross/blackfriday
  4//
  5// Copyright © 2011 Russ Ross <russ@russross.com>.
  6// Distributed under the Simplified BSD License.
  7// See README.md for details.
  8//
  9
 10//
 11//
 12// HTML rendering backend
 13//
 14//
 15
 16package blackfriday
 17
 18import (
 19	"bytes"
 20	"fmt"
 21	"strconv"
 22	"strings"
 23)
 24
 25// Html renderer configuration options.
 26const (
 27	HTML_SKIP_HTML                = 1 << iota // skip preformatted HTML blocks
 28	HTML_SKIP_STYLE                           // skip embedded <style> elements
 29	HTML_SKIP_IMAGES                          // skip embedded images
 30	HTML_SKIP_LINKS                           // skip all links
 31	HTML_SAFELINK                             // only link to trusted protocols
 32	HTML_TOC                                  // generate a table of contents
 33	HTML_OMIT_CONTENTS                        // skip the main contents (for a standalone table of contents)
 34	HTML_COMPLETE_PAGE                        // generate a complete HTML page
 35	HTML_GITHUB_BLOCKCODE                     // use github fenced code rendering rules
 36	HTML_USE_XHTML                            // generate XHTML output instead of HTML
 37	HTML_USE_SMARTYPANTS                      // enable smart punctuation substitutions
 38	HTML_SMARTYPANTS_FRACTIONS                // enable smart fractions (with HTML_USE_SMARTYPANTS)
 39	HTML_SMARTYPANTS_LATEX_DASHES             // enable LaTeX-style dashes (with HTML_USE_SMARTYPANTS)
 40)
 41
 42// Html is a type that implements the Renderer interface for HTML output.
 43//
 44// Do not create this directly, instead use the HtmlRenderer function.
 45type Html struct {
 46	flags    int    // HTML_* options
 47	closeTag string // how to end singleton tags: either " />\n" or ">\n"
 48	title    string // document title
 49	css      string // optional css file url (used with HTML_COMPLETE_PAGE)
 50
 51	// table of contents data
 52	tocMarker    int
 53	headerCount  int
 54	currentLevel int
 55	toc          *bytes.Buffer
 56
 57	smartypants *smartypantsRenderer
 58}
 59
 60const (
 61	xhtmlClose = " />\n"
 62	htmlClose  = ">\n"
 63)
 64
 65// HtmlRenderer creates and configures an Html object, which
 66// satisfies the Renderer interface.
 67//
 68// flags is a set of HTML_* options ORed together.
 69// title is the title of the document, and css is a URL for the document's
 70// stylesheet.
 71// title and css are only used when HTML_COMPLETE_PAGE is selected.
 72func HtmlRenderer(flags int, title string, css string) Renderer {
 73	// configure the rendering engine
 74	closeTag := htmlClose
 75	if flags&HTML_USE_XHTML != 0 {
 76		closeTag = xhtmlClose
 77	}
 78
 79	return &Html{
 80		flags:    flags,
 81		closeTag: closeTag,
 82		title:    title,
 83		css:      css,
 84
 85		headerCount:  0,
 86		currentLevel: 0,
 87		toc:          new(bytes.Buffer),
 88
 89		smartypants: smartypants(flags),
 90	}
 91}
 92
 93func attrEscape(out *bytes.Buffer, src []byte) {
 94	org := 0
 95	for i, ch := range src {
 96		// using if statements is a bit faster than a switch statement.
 97		// as the compiler improves, this should be unnecessary
 98		// this is only worthwhile because attrEscape is the single
 99		// largest CPU user in normal use
100		if ch == '"' {
101			if i > org {
102				// copy all the normal characters since the last escape
103				out.Write(src[org:i])
104			}
105			org = i + 1
106			out.WriteString("&quot;")
107			continue
108		}
109		if ch == '&' {
110			if i > org {
111				out.Write(src[org:i])
112			}
113			org = i + 1
114			out.WriteString("&amp;")
115			continue
116		}
117		if ch == '<' {
118			if i > org {
119				out.Write(src[org:i])
120			}
121			org = i + 1
122			out.WriteString("&lt;")
123			continue
124		}
125		if ch == '>' {
126			if i > org {
127				out.Write(src[org:i])
128			}
129			org = i + 1
130			out.WriteString("&gt;")
131			continue
132		}
133	}
134	if org < len(src) {
135		out.Write(src[org:])
136	}
137}
138
139func (options *Html) Header(out *bytes.Buffer, text func() bool, level int) {
140	marker := out.Len()
141	doubleSpace(out)
142
143	if options.flags&HTML_TOC != 0 {
144		// headerCount is incremented in htmlTocHeader
145		out.WriteString(fmt.Sprintf("<h%d id=\"toc_%d\">", level, options.headerCount))
146	} else {
147		out.WriteString(fmt.Sprintf("<h%d>", level))
148	}
149
150	tocMarker := out.Len()
151	if !text() {
152		out.Truncate(marker)
153		return
154	}
155
156	// are we building a table of contents?
157	if options.flags&HTML_TOC != 0 {
158		options.TocHeader(out.Bytes()[tocMarker:], level)
159	}
160
161	out.WriteString(fmt.Sprintf("</h%d>\n", level))
162}
163
164func (options *Html) BlockHtml(out *bytes.Buffer, text []byte) {
165	if options.flags&HTML_SKIP_HTML != 0 {
166		return
167	}
168
169	doubleSpace(out)
170	out.Write(text)
171	out.WriteByte('\n')
172}
173
174func (options *Html) HRule(out *bytes.Buffer) {
175	doubleSpace(out)
176	out.WriteString("<hr")
177	out.WriteString(options.closeTag)
178}
179
180func (options *Html) BlockCode(out *bytes.Buffer, text []byte, lang string) {
181	if options.flags&HTML_GITHUB_BLOCKCODE != 0 {
182		options.BlockCodeGithub(out, text, lang)
183	} else {
184		options.BlockCodeNormal(out, text, lang)
185	}
186}
187
188func (options *Html) BlockCodeNormal(out *bytes.Buffer, text []byte, lang string) {
189	doubleSpace(out)
190
191	// parse out the language names/classes
192	count := 0
193	for _, elt := range strings.Fields(lang) {
194		if elt[0] == '.' {
195			elt = elt[1:]
196		}
197		if len(elt) == 0 {
198			continue
199		}
200		if count == 0 {
201			out.WriteString("<pre><code class=\"")
202		} else {
203			out.WriteByte(' ')
204		}
205		attrEscape(out, []byte(elt))
206		count++
207	}
208
209	if count == 0 {
210		out.WriteString("<pre><code>")
211	} else {
212		out.WriteString("\">")
213	}
214
215	attrEscape(out, text)
216	out.WriteString("</code></pre>\n")
217}
218
219// GitHub style code block:
220//
221//              <pre lang="LANG"><code>
222//              ...
223//              </code></pre>
224//
225// Unlike other parsers, we store the language identifier in the <pre>,
226// and don't let the user generate custom classes.
227//
228// The language identifier in the <pre> block gets postprocessed and all
229// the code inside gets syntax highlighted with Pygments. This is much safer
230// than letting the user specify a CSS class for highlighting.
231//
232// Note that we only generate HTML for the first specifier.
233// E.g.
234//              ~~~~ {.python .numbered}        =>      <pre lang="python"><code>
235func (options *Html) BlockCodeGithub(out *bytes.Buffer, text []byte, lang string) {
236	doubleSpace(out)
237
238	// parse out the language name
239	count := 0
240	for _, elt := range strings.Fields(lang) {
241		if elt[0] == '.' {
242			elt = elt[1:]
243		}
244		if len(elt) == 0 {
245			continue
246		}
247		out.WriteString("<pre lang=\"")
248		attrEscape(out, []byte(elt))
249		out.WriteString("\"><code>")
250		count++
251		break
252	}
253
254	if count == 0 {
255		out.WriteString("<pre><code>")
256	}
257
258	attrEscape(out, text)
259	out.WriteString("</code></pre>\n")
260}
261
262func (options *Html) BlockQuote(out *bytes.Buffer, text []byte) {
263	doubleSpace(out)
264	out.WriteString("<blockquote>\n")
265	out.Write(text)
266	out.WriteString("</blockquote>\n")
267}
268
269func (options *Html) Table(out *bytes.Buffer, header []byte, body []byte, columnData []int) {
270	doubleSpace(out)
271	out.WriteString("<table>\n<thead>\n")
272	out.Write(header)
273	out.WriteString("</thead>\n\n<tbody>\n")
274	out.Write(body)
275	out.WriteString("</tbody>\n</table>\n")
276}
277
278func (options *Html) TableRow(out *bytes.Buffer, text []byte) {
279	doubleSpace(out)
280	out.WriteString("<tr>\n")
281	out.Write(text)
282	out.WriteString("\n</tr>\n")
283}
284
285func (options *Html) TableCell(out *bytes.Buffer, text []byte, align int) {
286	doubleSpace(out)
287	switch align {
288	case TABLE_ALIGNMENT_LEFT:
289		out.WriteString("<td align=\"left\">")
290	case TABLE_ALIGNMENT_RIGHT:
291		out.WriteString("<td align=\"right\">")
292	case TABLE_ALIGNMENT_CENTER:
293		out.WriteString("<td align=\"center\">")
294	default:
295		out.WriteString("<td>")
296	}
297
298	out.Write(text)
299	out.WriteString("</td>")
300}
301
302func (options *Html) List(out *bytes.Buffer, text func() bool, flags int) {
303	marker := out.Len()
304	doubleSpace(out)
305
306	if flags&LIST_TYPE_ORDERED != 0 {
307		out.WriteString("<ol>")
308	} else {
309		out.WriteString("<ul>")
310	}
311	if !text() {
312		out.Truncate(marker)
313		return
314	}
315	if flags&LIST_TYPE_ORDERED != 0 {
316		out.WriteString("</ol>\n")
317	} else {
318		out.WriteString("</ul>\n")
319	}
320}
321
322func (options *Html) ListItem(out *bytes.Buffer, text []byte, flags int) {
323	if flags&LIST_ITEM_CONTAINS_BLOCK != 0 || flags&LIST_ITEM_BEGINNING_OF_LIST != 0 {
324		doubleSpace(out)
325	}
326	out.WriteString("<li>")
327	out.Write(text)
328	out.WriteString("</li>\n")
329}
330
331func (options *Html) Paragraph(out *bytes.Buffer, text func() bool) {
332	marker := out.Len()
333	doubleSpace(out)
334
335	out.WriteString("<p>")
336	if !text() {
337		out.Truncate(marker)
338		return
339	}
340	out.WriteString("</p>\n")
341}
342
343func (options *Html) AutoLink(out *bytes.Buffer, link []byte, kind int) {
344	if options.flags&HTML_SAFELINK != 0 && !isSafeLink(link) && kind != LINK_TYPE_EMAIL {
345		// mark it but don't link it if it is not a safe link: no smartypants
346		out.WriteString("<tt>")
347		attrEscape(out, link)
348		out.WriteString("</tt>")
349		return
350	}
351
352	out.WriteString("<a href=\"")
353	if kind == LINK_TYPE_EMAIL {
354		out.WriteString("mailto:")
355	}
356	attrEscape(out, link)
357	out.WriteString("\">")
358
359	// Pretty print: if we get an email address as
360	// an actual URI, e.g. `mailto:foo@bar.com`, we don't
361	// want to print the `mailto:` prefix
362	switch {
363	case bytes.HasPrefix(link, []byte("mailto://")):
364		attrEscape(out, link[len("mailto://"):])
365	case bytes.HasPrefix(link, []byte("mailto:")):
366		attrEscape(out, link[len("mailto:"):])
367	default:
368		attrEscape(out, link)
369	}
370
371	out.WriteString("</a>")
372}
373
374func (options *Html) CodeSpan(out *bytes.Buffer, text []byte) {
375	out.WriteString("<code>")
376	attrEscape(out, text)
377	out.WriteString("</code>")
378}
379
380func (options *Html) DoubleEmphasis(out *bytes.Buffer, text []byte) {
381	out.WriteString("<strong>")
382	out.Write(text)
383	out.WriteString("</strong>")
384}
385
386func (options *Html) Emphasis(out *bytes.Buffer, text []byte) {
387	if len(text) == 0 {
388		return
389	}
390	out.WriteString("<em>")
391	out.Write(text)
392	out.WriteString("</em>")
393}
394
395func (options *Html) Image(out *bytes.Buffer, link []byte, title []byte, alt []byte) {
396	if options.flags&HTML_SKIP_IMAGES != 0 {
397		return
398	}
399
400	out.WriteString("<img src=\"")
401	attrEscape(out, link)
402	out.WriteString("\" alt=\"")
403	if len(alt) > 0 {
404		attrEscape(out, alt)
405	}
406	if len(title) > 0 {
407		out.WriteString("\" title=\"")
408		attrEscape(out, title)
409	}
410
411	out.WriteByte('"')
412	out.WriteString(options.closeTag)
413	return
414}
415
416func (options *Html) LineBreak(out *bytes.Buffer) {
417	out.WriteString("<br")
418	out.WriteString(options.closeTag)
419}
420
421func (options *Html) Link(out *bytes.Buffer, link []byte, title []byte, content []byte) {
422	if options.flags&HTML_SKIP_LINKS != 0 {
423		// write the link text out but don't link it, just mark it with typewriter font
424		out.WriteString("<tt>")
425		attrEscape(out, content)
426		out.WriteString("</tt>")
427		return
428	}
429
430	if options.flags&HTML_SAFELINK != 0 && !isSafeLink(link) {
431		// write the link text out but don't link it, just mark it with typewriter font
432		out.WriteString("<tt>")
433		attrEscape(out, content)
434		out.WriteString("</tt>")
435		return
436	}
437
438	out.WriteString("<a href=\"")
439	attrEscape(out, link)
440	if len(title) > 0 {
441		out.WriteString("\" title=\"")
442		attrEscape(out, title)
443	}
444	out.WriteString("\">")
445	out.Write(content)
446	out.WriteString("</a>")
447	return
448}
449
450func (options *Html) RawHtmlTag(out *bytes.Buffer, text []byte) {
451	if options.flags&HTML_SKIP_HTML != 0 {
452		return
453	}
454	if options.flags&HTML_SKIP_STYLE != 0 && isHtmlTag(text, "style") {
455		return
456	}
457	if options.flags&HTML_SKIP_LINKS != 0 && isHtmlTag(text, "a") {
458		return
459	}
460	if options.flags&HTML_SKIP_IMAGES != 0 && isHtmlTag(text, "img") {
461		return
462	}
463	out.Write(text)
464}
465
466func (options *Html) TripleEmphasis(out *bytes.Buffer, text []byte) {
467	out.WriteString("<strong><em>")
468	out.Write(text)
469	out.WriteString("</em></strong>")
470}
471
472func (options *Html) StrikeThrough(out *bytes.Buffer, text []byte) {
473	out.WriteString("<del>")
474	out.Write(text)
475	out.WriteString("</del>")
476}
477
478func (options *Html) Entity(out *bytes.Buffer, entity []byte) {
479	out.Write(entity)
480}
481
482func (options *Html) NormalText(out *bytes.Buffer, text []byte) {
483	if options.flags&HTML_USE_SMARTYPANTS != 0 {
484		options.Smartypants(out, text)
485	} else {
486		attrEscape(out, text)
487	}
488}
489
490func (options *Html) Smartypants(out *bytes.Buffer, text []byte) {
491	smrt := smartypantsData{false, false}
492
493	// first do normal entity escaping
494	var escaped bytes.Buffer
495	attrEscape(&escaped, text)
496	text = escaped.Bytes()
497
498	mark := 0
499	for i := 0; i < len(text); i++ {
500		if action := options.smartypants[text[i]]; action != nil {
501			if i > mark {
502				out.Write(text[mark:i])
503			}
504
505			previousChar := byte(0)
506			if i > 0 {
507				previousChar = text[i-1]
508			}
509			i += action(out, &smrt, previousChar, text[i:])
510			mark = i + 1
511		}
512	}
513
514	if mark < len(text) {
515		out.Write(text[mark:])
516	}
517}
518
519func (options *Html) DocumentHeader(out *bytes.Buffer) {
520	if options.flags&HTML_COMPLETE_PAGE == 0 {
521		return
522	}
523
524	ending := ""
525	if options.flags&HTML_USE_XHTML != 0 {
526		out.WriteString("<!DOCTYPE html PUBLIC \"-//W3C//DTD XHTML 1.0 Transitional//EN\" ")
527		out.WriteString("\"http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd\">\n")
528		out.WriteString("<html xmlns=\"http://www.w3.org/1999/xhtml\">\n")
529		ending = " /"
530	} else {
531		out.WriteString("<!DOCTYPE html>\n")
532		out.WriteString("<html>\n")
533	}
534	out.WriteString("<head>\n")
535	out.WriteString("  <title>")
536	options.NormalText(out, []byte(options.title))
537	out.WriteString("</title>\n")
538	out.WriteString("  <meta name=\"GENERATOR\" content=\"Blackfriday Markdown Processor v")
539	out.WriteString(VERSION)
540	out.WriteString("\"")
541	out.WriteString(ending)
542	out.WriteString(">\n")
543	out.WriteString("  <meta charset=\"utf-8\"")
544	out.WriteString(ending)
545	out.WriteString(">\n")
546	if options.css != "" {
547		out.WriteString("  <link rel=\"stylesheet\" type=\"text/css\" href=\"")
548		attrEscape(out, []byte(options.css))
549		out.WriteString("\"")
550		out.WriteString(ending)
551		out.WriteString(">\n")
552	}
553	out.WriteString("</head>\n")
554	out.WriteString("<body>\n")
555
556	options.tocMarker = out.Len()
557}
558
559func (options *Html) DocumentFooter(out *bytes.Buffer) {
560	// finalize and insert the table of contents
561	if options.flags&HTML_TOC != 0 {
562		options.TocFinalize()
563
564		// now we have to insert the table of contents into the document
565		var temp bytes.Buffer
566
567		// start by making a copy of everything after the document header
568		temp.Write(out.Bytes()[options.tocMarker:])
569
570		// now clear the copied material from the main output buffer
571		out.Truncate(options.tocMarker)
572
573		// corner case spacing issue
574		if options.flags&HTML_COMPLETE_PAGE != 0 {
575			out.WriteByte('\n')
576		}
577
578		// insert the table of contents
579		out.WriteString("<nav>\n")
580		out.Write(options.toc.Bytes())
581		out.WriteString("</nav>\n")
582
583		// corner case spacing issue
584		if options.flags&HTML_COMPLETE_PAGE == 0 && options.flags&HTML_OMIT_CONTENTS == 0 {
585			out.WriteByte('\n')
586		}
587
588		// write out everything that came after it
589		if options.flags&HTML_OMIT_CONTENTS == 0 {
590			out.Write(temp.Bytes())
591		}
592	}
593
594	if options.flags&HTML_COMPLETE_PAGE != 0 {
595		out.WriteString("\n</body>\n")
596		out.WriteString("</html>\n")
597	}
598
599}
600
601func (options *Html) TocHeader(text []byte, level int) {
602	for level > options.currentLevel {
603		switch {
604		case bytes.HasSuffix(options.toc.Bytes(), []byte("</li>\n")):
605			// this sublist can nest underneath a header
606			size := options.toc.Len()
607			options.toc.Truncate(size - len("</li>\n"))
608
609		case options.currentLevel > 0:
610			options.toc.WriteString("<li>")
611		}
612		if options.toc.Len() > 0 {
613			options.toc.WriteByte('\n')
614		}
615		options.toc.WriteString("<ul>\n")
616		options.currentLevel++
617	}
618
619	for level < options.currentLevel {
620		options.toc.WriteString("</ul>")
621		if options.currentLevel > 1 {
622			options.toc.WriteString("</li>\n")
623		}
624		options.currentLevel--
625	}
626
627	options.toc.WriteString("<li><a href=\"#toc_")
628	options.toc.WriteString(strconv.Itoa(options.headerCount))
629	options.toc.WriteString("\">")
630	options.headerCount++
631
632	options.toc.Write(text)
633
634	options.toc.WriteString("</a></li>\n")
635}
636
637func (options *Html) TocFinalize() {
638	for options.currentLevel > 1 {
639		options.toc.WriteString("</ul></li>\n")
640		options.currentLevel--
641	}
642
643	if options.currentLevel > 0 {
644		options.toc.WriteString("</ul>\n")
645	}
646}
647
648func isHtmlTag(tag []byte, tagname string) bool {
649	i := 0
650	if i < len(tag) && tag[0] != '<' {
651		return false
652	}
653	i++
654	i = skipSpace(tag, i)
655
656	if i < len(tag) && tag[i] == '/' {
657		i++
658	}
659
660	i = skipSpace(tag, i)
661	j := 0
662	for ; i < len(tag); i, j = i+1, j+1 {
663		if j >= len(tagname) {
664			break
665		}
666
667		if tag[i] != tagname[j] {
668			return false
669		}
670	}
671
672	if i == len(tag) {
673		return false
674	}
675
676	return isspace(tag[i]) || tag[i] == '>'
677}
678
679func skipSpace(tag []byte, i int) int {
680	for i < len(tag) && isspace(tag[i]) {
681		i++
682	}
683	return i
684}
685
686func doubleSpace(out *bytes.Buffer) {
687	if out.Len() > 0 {
688		out.WriteByte('\n')
689	}
690}