all repos — grayfriday @ 77aeb0ca37d425a404d87eb4d10cb3e33400c3d0

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	"regexp"
 22	"strconv"
 23	"strings"
 24)
 25
 26// Html renderer configuration options.
 27const (
 28	HTML_SKIP_HTML                = 1 << iota // skip preformatted HTML blocks
 29	HTML_SKIP_STYLE                           // skip embedded <style> elements
 30	HTML_SKIP_IMAGES                          // skip embedded images
 31	HTML_SKIP_LINKS                           // skip all links
 32	HTML_SAFELINK                             // only link to trusted protocols
 33	HTML_NOFOLLOW_LINKS                       // only link with rel="nofollow"
 34	HTML_HREF_TARGET_BLANK                    // add a blank target
 35	HTML_TOC                                  // generate a table of contents
 36	HTML_OMIT_CONTENTS                        // skip the main contents (for a standalone table of contents)
 37	HTML_COMPLETE_PAGE                        // generate a complete HTML page
 38	HTML_USE_XHTML                            // generate XHTML output instead of HTML
 39	HTML_USE_SMARTYPANTS                      // enable smart punctuation substitutions
 40	HTML_SMARTYPANTS_FRACTIONS                // enable smart fractions (with HTML_USE_SMARTYPANTS)
 41	HTML_SMARTYPANTS_LATEX_DASHES             // enable LaTeX-style dashes (with HTML_USE_SMARTYPANTS)
 42	HTML_FOOTNOTE_RETURN_LINKS                // generate a link at the end of a footnote to return to the source
 43)
 44
 45var (
 46	alignments = []string{
 47		"left",
 48		"right",
 49		"center",
 50	}
 51
 52	// TODO: improve this regexp to catch all possible entities:
 53	htmlEntity = regexp.MustCompile(`&[a-z]{2,5};`)
 54)
 55
 56type HtmlRendererParameters struct {
 57	// Prepend this text to each relative URL.
 58	AbsolutePrefix string
 59	// Add this text to each footnote anchor, to ensure uniqueness.
 60	FootnoteAnchorPrefix string
 61	// Show this text inside the <a> tag for a footnote return link, if the
 62	// HTML_FOOTNOTE_RETURN_LINKS flag is enabled. If blank, the string
 63	// <sup>[return]</sup> is used.
 64	FootnoteReturnLinkContents string
 65	// If set, add this text to the front of each Header ID, to ensure
 66	// uniqueness.
 67	HeaderIDPrefix string
 68	// If set, add this text to the back of each Header ID, to ensure uniqueness.
 69	HeaderIDSuffix string
 70}
 71
 72// Html is a type that implements the Renderer interface for HTML output.
 73//
 74// Do not create this directly, instead use the HtmlRenderer function.
 75type Html struct {
 76	flags    int    // HTML_* options
 77	closeTag string // how to end singleton tags: either " />\n" or ">\n"
 78	title    string // document title
 79	css      string // optional css file url (used with HTML_COMPLETE_PAGE)
 80
 81	parameters HtmlRendererParameters
 82
 83	// table of contents data
 84	tocMarker    int
 85	headerCount  int
 86	currentLevel int
 87	toc          *bytes.Buffer
 88
 89	// Track header IDs to prevent ID collision in a single generation.
 90	headerIDs map[string]int
 91
 92	smartypants *smartypantsRenderer
 93}
 94
 95const (
 96	xhtmlClose = " />\n"
 97	htmlClose  = ">\n"
 98)
 99
100// HtmlRenderer creates and configures an Html object, which
101// satisfies the Renderer interface.
102//
103// flags is a set of HTML_* options ORed together.
104// title is the title of the document, and css is a URL for the document's
105// stylesheet.
106// title and css are only used when HTML_COMPLETE_PAGE is selected.
107func HtmlRenderer(flags int, title string, css string) Renderer {
108	return HtmlRendererWithParameters(flags, title, css, HtmlRendererParameters{})
109}
110
111func HtmlRendererWithParameters(flags int, title string,
112	css string, renderParameters HtmlRendererParameters) Renderer {
113	// configure the rendering engine
114	closeTag := htmlClose
115	if flags&HTML_USE_XHTML != 0 {
116		closeTag = xhtmlClose
117	}
118
119	if renderParameters.FootnoteReturnLinkContents == "" {
120		renderParameters.FootnoteReturnLinkContents = `<sup>[return]</sup>`
121	}
122
123	return &Html{
124		flags:      flags,
125		closeTag:   closeTag,
126		title:      title,
127		css:        css,
128		parameters: renderParameters,
129
130		headerCount:  0,
131		currentLevel: 0,
132		toc:          new(bytes.Buffer),
133
134		headerIDs: make(map[string]int),
135
136		smartypants: smartypants(flags),
137	}
138}
139
140// Using if statements is a bit faster than a switch statement. As the compiler
141// improves, this should be unnecessary this is only worthwhile because
142// attrEscape is the single largest CPU user in normal use.
143// Also tried using map, but that gave a ~3x slowdown.
144func escapeSingleChar(char byte) (string, bool) {
145	if char == '"' {
146		return "&quot;", true
147	}
148	if char == '&' {
149		return "&amp;", true
150	}
151	if char == '<' {
152		return "&lt;", true
153	}
154	if char == '>' {
155		return "&gt;", true
156	}
157	return "", false
158}
159
160func attrEscape(out *bytes.Buffer, src []byte) {
161	org := 0
162	for i, ch := range src {
163		if entity, ok := escapeSingleChar(ch); ok {
164			if i > org {
165				// copy all the normal characters since the last escape
166				out.Write(src[org:i])
167			}
168			org = i + 1
169			out.WriteString(entity)
170		}
171	}
172	if org < len(src) {
173		out.Write(src[org:])
174	}
175}
176
177func entityEscapeWithSkip(out *bytes.Buffer, src []byte, skipRanges [][]int) {
178	end := 0
179	for _, rang := range skipRanges {
180		attrEscape(out, src[end:rang[0]])
181		out.Write(src[rang[0]:rang[1]])
182		end = rang[1]
183	}
184	attrEscape(out, src[end:])
185}
186
187func (options *Html) GetFlags() int {
188	return options.flags
189}
190
191func (options *Html) TitleBlock(out *bytes.Buffer, text []byte) {
192	text = bytes.TrimPrefix(text, []byte("% "))
193	text = bytes.Replace(text, []byte("\n% "), []byte("\n"), -1)
194	out.WriteString("<h1 class=\"title\">")
195	out.Write(text)
196	out.WriteString("\n</h1>")
197}
198
199func (options *Html) Header(out *bytes.Buffer, text func() bool, level int, id string) {
200	marker := out.Len()
201	doubleSpace(out)
202
203	if id == "" && options.flags&HTML_TOC != 0 {
204		id = fmt.Sprintf("toc_%d", options.headerCount)
205	}
206
207	if id != "" {
208		id = options.ensureUniqueHeaderID(id)
209
210		if options.parameters.HeaderIDPrefix != "" {
211			id = options.parameters.HeaderIDPrefix + id
212		}
213
214		if options.parameters.HeaderIDSuffix != "" {
215			id = id + options.parameters.HeaderIDSuffix
216		}
217
218		out.WriteString(fmt.Sprintf("<h%d id=\"%s\">", level, id))
219	} else {
220		out.WriteString(fmt.Sprintf("<h%d>", level))
221	}
222
223	tocMarker := out.Len()
224	if !text() {
225		out.Truncate(marker)
226		return
227	}
228
229	// are we building a table of contents?
230	if options.flags&HTML_TOC != 0 {
231		options.TocHeaderWithAnchor(out.Bytes()[tocMarker:], level, id)
232	}
233
234	out.WriteString(fmt.Sprintf("</h%d>\n", level))
235}
236
237func (options *Html) BlockHtml(out *bytes.Buffer, text []byte) {
238	if options.flags&HTML_SKIP_HTML != 0 {
239		return
240	}
241
242	doubleSpace(out)
243	out.Write(text)
244	out.WriteByte('\n')
245}
246
247func (options *Html) HRule(out *bytes.Buffer) {
248	doubleSpace(out)
249	out.WriteString("<hr")
250	out.WriteString(options.closeTag)
251}
252
253func (options *Html) BlockCode(out *bytes.Buffer, text []byte, lang string) {
254	doubleSpace(out)
255
256	// parse out the language names/classes
257	count := 0
258	for _, elt := range strings.Fields(lang) {
259		if elt[0] == '.' {
260			elt = elt[1:]
261		}
262		if len(elt) == 0 {
263			continue
264		}
265		if count == 0 {
266			out.WriteString("<pre><code class=\"language-")
267		} else {
268			out.WriteByte(' ')
269		}
270		attrEscape(out, []byte(elt))
271		count++
272	}
273
274	if count == 0 {
275		out.WriteString("<pre><code>")
276	} else {
277		out.WriteString("\">")
278	}
279
280	attrEscape(out, text)
281	out.WriteString("</code></pre>\n")
282}
283
284func (options *Html) BlockQuote(out *bytes.Buffer, text []byte) {
285	doubleSpace(out)
286	out.WriteString("<blockquote>\n")
287	out.Write(text)
288	out.WriteString("</blockquote>\n")
289}
290
291func (options *Html) Table(out *bytes.Buffer, header []byte, body []byte, columnData []int) {
292	doubleSpace(out)
293	out.WriteString("<table>\n<thead>\n")
294	out.Write(header)
295	out.WriteString("</thead>\n\n<tbody>\n")
296	out.Write(body)
297	out.WriteString("</tbody>\n</table>\n")
298}
299
300func (options *Html) TableRow(out *bytes.Buffer, text []byte) {
301	doubleSpace(out)
302	out.WriteString("<tr>\n")
303	out.Write(text)
304	out.WriteString("\n</tr>\n")
305}
306
307func (options *Html) TableHeaderCell(out *bytes.Buffer, text []byte, align int) {
308	doubleSpace(out)
309	switch align {
310	case TABLE_ALIGNMENT_LEFT:
311		out.WriteString("<th align=\"left\">")
312	case TABLE_ALIGNMENT_RIGHT:
313		out.WriteString("<th align=\"right\">")
314	case TABLE_ALIGNMENT_CENTER:
315		out.WriteString("<th align=\"center\">")
316	default:
317		out.WriteString("<th>")
318	}
319
320	out.Write(text)
321	out.WriteString("</th>")
322}
323
324func (options *Html) TableCell(out *bytes.Buffer, text []byte, align int) {
325	doubleSpace(out)
326	switch align {
327	case TABLE_ALIGNMENT_LEFT:
328		out.WriteString("<td align=\"left\">")
329	case TABLE_ALIGNMENT_RIGHT:
330		out.WriteString("<td align=\"right\">")
331	case TABLE_ALIGNMENT_CENTER:
332		out.WriteString("<td align=\"center\">")
333	default:
334		out.WriteString("<td>")
335	}
336
337	out.Write(text)
338	out.WriteString("</td>")
339}
340
341func (options *Html) Footnotes(out *bytes.Buffer, text func() bool) {
342	out.WriteString("<div class=\"footnotes\">\n")
343	options.HRule(out)
344	options.List(out, text, LIST_TYPE_ORDERED)
345	out.WriteString("</div>\n")
346}
347
348func (options *Html) FootnoteItem(out *bytes.Buffer, name, text []byte, flags int) {
349	if flags&LIST_ITEM_CONTAINS_BLOCK != 0 || flags&LIST_ITEM_BEGINNING_OF_LIST != 0 {
350		doubleSpace(out)
351	}
352	slug := slugify(name)
353	out.WriteString(`<li id="`)
354	out.WriteString(`fn:`)
355	out.WriteString(options.parameters.FootnoteAnchorPrefix)
356	out.Write(slug)
357	out.WriteString(`">`)
358	out.Write(text)
359	if options.flags&HTML_FOOTNOTE_RETURN_LINKS != 0 {
360		out.WriteString(` <a class="footnote-return" href="#`)
361		out.WriteString(`fnref:`)
362		out.WriteString(options.parameters.FootnoteAnchorPrefix)
363		out.Write(slug)
364		out.WriteString(`">`)
365		out.WriteString(options.parameters.FootnoteReturnLinkContents)
366		out.WriteString(`</a>`)
367	}
368	out.WriteString("</li>\n")
369}
370
371func (options *Html) List(out *bytes.Buffer, text func() bool, flags int) {
372	marker := out.Len()
373	doubleSpace(out)
374
375	if flags&LIST_TYPE_ORDERED != 0 {
376		out.WriteString("<ol>")
377	} else {
378		out.WriteString("<ul>")
379	}
380	if !text() {
381		out.Truncate(marker)
382		return
383	}
384	if flags&LIST_TYPE_ORDERED != 0 {
385		out.WriteString("</ol>\n")
386	} else {
387		out.WriteString("</ul>\n")
388	}
389}
390
391func (options *Html) ListItem(out *bytes.Buffer, text []byte, flags int) {
392	if flags&LIST_ITEM_CONTAINS_BLOCK != 0 || flags&LIST_ITEM_BEGINNING_OF_LIST != 0 {
393		doubleSpace(out)
394	}
395	out.WriteString("<li>")
396	out.Write(text)
397	out.WriteString("</li>\n")
398}
399
400func (options *Html) Paragraph(out *bytes.Buffer, text func() bool) {
401	marker := out.Len()
402	doubleSpace(out)
403
404	out.WriteString("<p>")
405	if !text() {
406		out.Truncate(marker)
407		return
408	}
409	out.WriteString("</p>\n")
410}
411
412func (options *Html) AutoLink(out *bytes.Buffer, link []byte, kind int) {
413	skipRanges := htmlEntity.FindAllIndex(link, -1)
414	if options.flags&HTML_SAFELINK != 0 && !isSafeLink(link) && kind != LINK_TYPE_EMAIL {
415		// mark it but don't link it if it is not a safe link: no smartypants
416		out.WriteString("<tt>")
417		entityEscapeWithSkip(out, link, skipRanges)
418		out.WriteString("</tt>")
419		return
420	}
421
422	out.WriteString("<a href=\"")
423	if kind == LINK_TYPE_EMAIL {
424		out.WriteString("mailto:")
425	} else {
426		options.maybeWriteAbsolutePrefix(out, link)
427	}
428
429	entityEscapeWithSkip(out, link, skipRanges)
430
431	if options.flags&HTML_NOFOLLOW_LINKS != 0 && !isRelativeLink(link) {
432		out.WriteString("\" rel=\"nofollow")
433	}
434	// blank target only add to external link
435	if options.flags&HTML_HREF_TARGET_BLANK != 0 && !isRelativeLink(link) {
436		out.WriteString("\" target=\"_blank")
437	}
438
439	out.WriteString("\">")
440
441	// Pretty print: if we get an email address as
442	// an actual URI, e.g. `mailto:foo@bar.com`, we don't
443	// want to print the `mailto:` prefix
444	switch {
445	case bytes.HasPrefix(link, []byte("mailto://")):
446		attrEscape(out, link[len("mailto://"):])
447	case bytes.HasPrefix(link, []byte("mailto:")):
448		attrEscape(out, link[len("mailto:"):])
449	default:
450		entityEscapeWithSkip(out, link, skipRanges)
451	}
452
453	out.WriteString("</a>")
454}
455
456func (options *Html) CodeSpan(out *bytes.Buffer, text []byte) {
457	out.WriteString("<code>")
458	attrEscape(out, text)
459	out.WriteString("</code>")
460}
461
462func (options *Html) DoubleEmphasis(out *bytes.Buffer, text []byte) {
463	out.WriteString("<strong>")
464	out.Write(text)
465	out.WriteString("</strong>")
466}
467
468func (options *Html) Emphasis(out *bytes.Buffer, text []byte) {
469	if len(text) == 0 {
470		return
471	}
472	out.WriteString("<em>")
473	out.Write(text)
474	out.WriteString("</em>")
475}
476
477func (options *Html) maybeWriteAbsolutePrefix(out *bytes.Buffer, link []byte) {
478	if options.parameters.AbsolutePrefix != "" && isRelativeLink(link) {
479		out.WriteString(options.parameters.AbsolutePrefix)
480		if link[0] != '/' {
481			out.WriteByte('/')
482		}
483	}
484}
485
486func (options *Html) Image(out *bytes.Buffer, link []byte, title []byte, alt []byte) {
487	if options.flags&HTML_SKIP_IMAGES != 0 {
488		return
489	}
490
491	out.WriteString("<img src=\"")
492	options.maybeWriteAbsolutePrefix(out, link)
493	attrEscape(out, link)
494	out.WriteString("\" alt=\"")
495	if len(alt) > 0 {
496		attrEscape(out, alt)
497	}
498	if len(title) > 0 {
499		out.WriteString("\" title=\"")
500		attrEscape(out, title)
501	}
502
503	out.WriteByte('"')
504	out.WriteString(options.closeTag)
505	return
506}
507
508func (options *Html) LineBreak(out *bytes.Buffer) {
509	out.WriteString("<br")
510	out.WriteString(options.closeTag)
511}
512
513func (options *Html) Link(out *bytes.Buffer, link []byte, title []byte, content []byte) {
514	if options.flags&HTML_SKIP_LINKS != 0 {
515		// write the link text out but don't link it, just mark it with typewriter font
516		out.WriteString("<tt>")
517		attrEscape(out, content)
518		out.WriteString("</tt>")
519		return
520	}
521
522	if options.flags&HTML_SAFELINK != 0 && !isSafeLink(link) {
523		// write the link text out but don't link it, just mark it with typewriter font
524		out.WriteString("<tt>")
525		attrEscape(out, content)
526		out.WriteString("</tt>")
527		return
528	}
529
530	out.WriteString("<a href=\"")
531	options.maybeWriteAbsolutePrefix(out, link)
532	attrEscape(out, link)
533	if len(title) > 0 {
534		out.WriteString("\" title=\"")
535		attrEscape(out, title)
536	}
537	if options.flags&HTML_NOFOLLOW_LINKS != 0 && !isRelativeLink(link) {
538		out.WriteString("\" rel=\"nofollow")
539	}
540	// blank target only add to external link
541	if options.flags&HTML_HREF_TARGET_BLANK != 0 && !isRelativeLink(link) {
542		out.WriteString("\" target=\"_blank")
543	}
544
545	out.WriteString("\">")
546	out.Write(content)
547	out.WriteString("</a>")
548	return
549}
550
551func (options *Html) RawHtmlTag(out *bytes.Buffer, text []byte) {
552	if options.flags&HTML_SKIP_HTML != 0 {
553		return
554	}
555	if options.flags&HTML_SKIP_STYLE != 0 && isHtmlTag(text, "style") {
556		return
557	}
558	if options.flags&HTML_SKIP_LINKS != 0 && isHtmlTag(text, "a") {
559		return
560	}
561	if options.flags&HTML_SKIP_IMAGES != 0 && isHtmlTag(text, "img") {
562		return
563	}
564	out.Write(text)
565}
566
567func (options *Html) TripleEmphasis(out *bytes.Buffer, text []byte) {
568	out.WriteString("<strong><em>")
569	out.Write(text)
570	out.WriteString("</em></strong>")
571}
572
573func (options *Html) StrikeThrough(out *bytes.Buffer, text []byte) {
574	out.WriteString("<del>")
575	out.Write(text)
576	out.WriteString("</del>")
577}
578
579func (options *Html) FootnoteRef(out *bytes.Buffer, ref []byte, id int) {
580	slug := slugify(ref)
581	out.WriteString(`<sup class="footnote-ref" id="`)
582	out.WriteString(`fnref:`)
583	out.WriteString(options.parameters.FootnoteAnchorPrefix)
584	out.Write(slug)
585	out.WriteString(`"><a rel="footnote" href="#`)
586	out.WriteString(`fn:`)
587	out.WriteString(options.parameters.FootnoteAnchorPrefix)
588	out.Write(slug)
589	out.WriteString(`">`)
590	out.WriteString(strconv.Itoa(id))
591	out.WriteString(`</a></sup>`)
592}
593
594func (options *Html) Entity(out *bytes.Buffer, entity []byte) {
595	out.Write(entity)
596}
597
598func (options *Html) NormalText(out *bytes.Buffer, text []byte) {
599	if options.flags&HTML_USE_SMARTYPANTS != 0 {
600		options.Smartypants(out, text)
601	} else {
602		attrEscape(out, text)
603	}
604}
605
606func (options *Html) Smartypants(out *bytes.Buffer, text []byte) {
607	smrt := smartypantsData{false, false}
608
609	// first do normal entity escaping
610	var escaped bytes.Buffer
611	attrEscape(&escaped, text)
612	text = escaped.Bytes()
613
614	mark := 0
615	for i := 0; i < len(text); i++ {
616		if action := options.smartypants[text[i]]; action != nil {
617			if i > mark {
618				out.Write(text[mark:i])
619			}
620
621			previousChar := byte(0)
622			if i > 0 {
623				previousChar = text[i-1]
624			}
625			i += action(out, &smrt, previousChar, text[i:])
626			mark = i + 1
627		}
628	}
629
630	if mark < len(text) {
631		out.Write(text[mark:])
632	}
633}
634
635func (options *Html) DocumentHeader(out *bytes.Buffer) {
636	if options.flags&HTML_COMPLETE_PAGE == 0 {
637		return
638	}
639
640	ending := ""
641	if options.flags&HTML_USE_XHTML != 0 {
642		out.WriteString("<!DOCTYPE html PUBLIC \"-//W3C//DTD XHTML 1.0 Transitional//EN\" ")
643		out.WriteString("\"http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd\">\n")
644		out.WriteString("<html xmlns=\"http://www.w3.org/1999/xhtml\">\n")
645		ending = " /"
646	} else {
647		out.WriteString("<!DOCTYPE html>\n")
648		out.WriteString("<html>\n")
649	}
650	out.WriteString("<head>\n")
651	out.WriteString("  <title>")
652	options.NormalText(out, []byte(options.title))
653	out.WriteString("</title>\n")
654	out.WriteString("  <meta name=\"GENERATOR\" content=\"Blackfriday Markdown Processor v")
655	out.WriteString(VERSION)
656	out.WriteString("\"")
657	out.WriteString(ending)
658	out.WriteString(">\n")
659	out.WriteString("  <meta charset=\"utf-8\"")
660	out.WriteString(ending)
661	out.WriteString(">\n")
662	if options.css != "" {
663		out.WriteString("  <link rel=\"stylesheet\" type=\"text/css\" href=\"")
664		attrEscape(out, []byte(options.css))
665		out.WriteString("\"")
666		out.WriteString(ending)
667		out.WriteString(">\n")
668	}
669	out.WriteString("</head>\n")
670	out.WriteString("<body>\n")
671
672	options.tocMarker = out.Len()
673}
674
675func (options *Html) DocumentFooter(out *bytes.Buffer) {
676	// finalize and insert the table of contents
677	if options.flags&HTML_TOC != 0 {
678		options.TocFinalize()
679
680		// now we have to insert the table of contents into the document
681		var temp bytes.Buffer
682
683		// start by making a copy of everything after the document header
684		temp.Write(out.Bytes()[options.tocMarker:])
685
686		// now clear the copied material from the main output buffer
687		out.Truncate(options.tocMarker)
688
689		// corner case spacing issue
690		if options.flags&HTML_COMPLETE_PAGE != 0 {
691			out.WriteByte('\n')
692		}
693
694		// insert the table of contents
695		out.WriteString("<nav>\n")
696		out.Write(options.toc.Bytes())
697		out.WriteString("</nav>\n")
698
699		// corner case spacing issue
700		if options.flags&HTML_COMPLETE_PAGE == 0 && options.flags&HTML_OMIT_CONTENTS == 0 {
701			out.WriteByte('\n')
702		}
703
704		// write out everything that came after it
705		if options.flags&HTML_OMIT_CONTENTS == 0 {
706			out.Write(temp.Bytes())
707		}
708	}
709
710	if options.flags&HTML_COMPLETE_PAGE != 0 {
711		out.WriteString("\n</body>\n")
712		out.WriteString("</html>\n")
713	}
714
715}
716
717func (options *Html) TocHeaderWithAnchor(text []byte, level int, anchor string) {
718	for level > options.currentLevel {
719		switch {
720		case bytes.HasSuffix(options.toc.Bytes(), []byte("</li>\n")):
721			// this sublist can nest underneath a header
722			size := options.toc.Len()
723			options.toc.Truncate(size - len("</li>\n"))
724
725		case options.currentLevel > 0:
726			options.toc.WriteString("<li>")
727		}
728		if options.toc.Len() > 0 {
729			options.toc.WriteByte('\n')
730		}
731		options.toc.WriteString("<ul>\n")
732		options.currentLevel++
733	}
734
735	for level < options.currentLevel {
736		options.toc.WriteString("</ul>")
737		if options.currentLevel > 1 {
738			options.toc.WriteString("</li>\n")
739		}
740		options.currentLevel--
741	}
742
743	options.toc.WriteString("<li><a href=\"#")
744	if anchor != "" {
745		options.toc.WriteString(anchor)
746	} else {
747		options.toc.WriteString("toc_")
748		options.toc.WriteString(strconv.Itoa(options.headerCount))
749	}
750	options.toc.WriteString("\">")
751	options.headerCount++
752
753	options.toc.Write(text)
754
755	options.toc.WriteString("</a></li>\n")
756}
757
758func (options *Html) TocHeader(text []byte, level int) {
759	options.TocHeaderWithAnchor(text, level, "")
760}
761
762func (options *Html) TocFinalize() {
763	for options.currentLevel > 1 {
764		options.toc.WriteString("</ul></li>\n")
765		options.currentLevel--
766	}
767
768	if options.currentLevel > 0 {
769		options.toc.WriteString("</ul>\n")
770	}
771}
772
773func isHtmlTag(tag []byte, tagname string) bool {
774	found, _ := findHtmlTagPos(tag, tagname)
775	return found
776}
777
778// Look for a character, but ignore it when it's in any kind of quotes, it
779// might be JavaScript
780func skipUntilCharIgnoreQuotes(html []byte, start int, char byte) int {
781	inSingleQuote := false
782	inDoubleQuote := false
783	inGraveQuote := false
784	i := start
785	for i < len(html) {
786		switch {
787		case html[i] == char && !inSingleQuote && !inDoubleQuote && !inGraveQuote:
788			return i
789		case html[i] == '\'':
790			inSingleQuote = !inSingleQuote
791		case html[i] == '"':
792			inDoubleQuote = !inDoubleQuote
793		case html[i] == '`':
794			inGraveQuote = !inGraveQuote
795		}
796		i++
797	}
798	return start
799}
800
801func findHtmlTagPos(tag []byte, tagname string) (bool, int) {
802	i := 0
803	if i < len(tag) && tag[0] != '<' {
804		return false, -1
805	}
806	i++
807	i = skipSpace(tag, i)
808
809	if i < len(tag) && tag[i] == '/' {
810		i++
811	}
812
813	i = skipSpace(tag, i)
814	j := 0
815	for ; i < len(tag); i, j = i+1, j+1 {
816		if j >= len(tagname) {
817			break
818		}
819
820		if strings.ToLower(string(tag[i]))[0] != tagname[j] {
821			return false, -1
822		}
823	}
824
825	if i == len(tag) {
826		return false, -1
827	}
828
829	rightAngle := skipUntilCharIgnoreQuotes(tag, i, '>')
830	if rightAngle > i {
831		return true, rightAngle
832	}
833
834	return false, -1
835}
836
837func skipUntilChar(text []byte, start int, char byte) int {
838	i := start
839	for i < len(text) && text[i] != char {
840		i++
841	}
842	return i
843}
844
845func skipSpace(tag []byte, i int) int {
846	for i < len(tag) && isspace(tag[i]) {
847		i++
848	}
849	return i
850}
851
852func doubleSpace(out *bytes.Buffer) {
853	if out.Len() > 0 {
854		out.WriteByte('\n')
855	}
856}
857
858func isRelativeLink(link []byte) (yes bool) {
859	yes = false
860
861	// a tag begin with '#'
862	if link[0] == '#' {
863		yes = true
864	}
865
866	// link begin with '/' but not '//', the second maybe a protocol relative link
867	if len(link) >= 2 && link[0] == '/' && link[1] != '/' {
868		yes = true
869	}
870
871	// only the root '/'
872	if len(link) == 1 && link[0] == '/' {
873		yes = true
874	}
875	return
876}
877
878func (options *Html) ensureUniqueHeaderID(id string) string {
879	for count, found := options.headerIDs[id]; found; count, found = options.headerIDs[id] {
880		tmp := fmt.Sprintf("%s-%d", id, count+1)
881
882		if _, tmpFound := options.headerIDs[tmp]; !tmpFound {
883			options.headerIDs[id] = count + 1
884			id = tmp
885		} else {
886			id = id + "-1"
887		}
888	}
889
890	if _, found := options.headerIDs[id]; !found {
891		options.headerIDs[id] = 0
892	}
893
894	return id
895}