12345678910111213141516171819202122232425262728293031323334353637383940414243444546474849505152535455565758596061626364656667686970717273747576777879808182838485868788899091929394959697989910010110210310410510610710810911011111211311411511611711811912012112212312412512612712812913013113213313413513613713813914014114214314414514614714814915015115215315415515615715815916016116216316416516616716816917017117217317417517617717817918018118218318418518618718818919019119219319419519619719819920020120220320420520620720820921021121221321421521621721821922022122222322422522622722822923023123223323423523623723823924024124224324424524624724824925025125225325425525625725825926026126226326426526626726826927027127227327427527627727827928028128228328428528628728828929029129229329429529629729829930030130230330430530630730830931031131231331431531631731831932032132232332432532632732832933033133233333433533633733833934034134234334434534634734834935035135235335435535635735835936036136236336436536636736836937037137237337437537637737837938038138238338438538638738838939039139239339439539639739839940040140240340440540640740840941041141241341441541641741841942042142242342442542642742842943043143243343443543643743843944044144244344444544644744844945045145245345445545645745845946046146246346446546646746846947047147247347447547647747847948048148248348448548648748848949049149249349449549649749849950050150250350450550650750850951051151251351451551651751851952052152252352452552652752852953053153253353453553653753853954054154254354454554654754854955055155255355455555655755855956056156256356456556656756856957057157257357457557657757857958058158258358458558658758858959059159259359459559659759859960060160260360460560660760860961061161261361461561661761861962062162262362462562662762862963063163263363463563663763863964064164264364464564664764864965065165265365465565665765865966066166266366466566666766866967067167267367467567667767867968068168268368468568668768868969069169269369469569669769869970070170270370470570670770870971071171271371471571671771871972072172272372472572672772872973073173273373473573673773873974074174274374474574674774874975075175275375475575675775875976076176276376476576676776876977077177277377477577677777877978078178278378478578678778878979079179279379479579679779879980080180280380480580680780880981081181281381481581681781881982082182282382482582682782882983083183283383483583683783883984084184284384484584684784884985085185285385485585685785885986086186286386486586686786886987087187287387487587687787887988088188288388488588688788888989089189289389489589689789889990090190290390490590690790890991091191291391491591691791891992092192292392492592692792892993093193293393493593693793893994094194294394494594694794894995095195295395495595695795895996096196296396496596696796896997097197297397497597697797897998098198298398498598698798898999099199299399499599699799899910001001100210031004100510061007100810091010101110121013101410151016101710181019102010211022102310241025102610271028102910301031103210331034103510361037103810391040104110421043104410451046104710481049105010511052105310541055105610571058105910601061106210631064106510661067106810691070107110721073107410751076107710781079108010811082108310841085108610871088108910901091109210931094109510961097109810991100110111021103110411051106110711081109111011111112111311141115111611171118111911201121112211231124112511261127112811291130113111321133113411351136113711381139114011411142114311441145114611471148 |
- //
- // Blackfriday Markdown Processor
- // Available at http://github.com/russross/blackfriday
- //
- // Copyright © 2011 Russ Ross <russ@russross.com>.
- // Distributed under the Simplified BSD License.
- // See README.md for details.
- //
- //
- // Functions to parse inline elements.
- //
- package blackfriday
- import (
- "bytes"
- "regexp"
- "strconv"
- )
- var (
- urlRe = `((https?|ftp):\/\/|\/)[-A-Za-z0-9+&@#\/%?=~_|!:,.;\(\)]+`
- anchorRe = regexp.MustCompile(`^(<a\shref="` + urlRe + `"(\stitle="[^"<>]+")?\s?>` + urlRe + `<\/a>)`)
- )
- // Functions to parse text within a block
- // Each function returns the number of chars taken care of
- // data is the complete block being rendered
- // offset is the number of valid chars before the current cursor
- func (p *parser) inline(out *bytes.Buffer, data []byte) {
- // this is called recursively: enforce a maximum depth
- if p.nesting >= p.maxNesting {
- return
- }
- p.nesting++
- i, end := 0, 0
- for i < len(data) {
- // copy inactive chars into the output
- for end < len(data) && p.inlineCallback[data[end]] == nil {
- end++
- }
- p.r.NormalText(out, data[i:end])
- if end >= len(data) {
- break
- }
- i = end
- // call the trigger
- handler := p.inlineCallback[data[end]]
- if consumed := handler(p, out, data, i); consumed == 0 {
- // no action from the callback; buffer the byte for later
- end = i + 1
- } else {
- // skip past whatever the callback used
- i += consumed
- end = i
- }
- }
- p.nesting--
- }
- // single and double emphasis parsing
- func emphasis(p *parser, out *bytes.Buffer, data []byte, offset int) int {
- data = data[offset:]
- c := data[0]
- ret := 0
- if len(data) > 2 && data[1] != c {
- // whitespace cannot follow an opening emphasis;
- // strikethrough only takes two characters '~~'
- if c == '~' || isspace(data[1]) {
- return 0
- }
- if ret = helperEmphasis(p, out, data[1:], c); ret == 0 {
- return 0
- }
- return ret + 1
- }
- if len(data) > 3 && data[1] == c && data[2] != c {
- if isspace(data[2]) {
- return 0
- }
- if ret = helperDoubleEmphasis(p, out, data[2:], c); ret == 0 {
- return 0
- }
- return ret + 2
- }
- if len(data) > 4 && data[1] == c && data[2] == c && data[3] != c {
- if c == '~' || isspace(data[3]) {
- return 0
- }
- if ret = helperTripleEmphasis(p, out, data, 3, c); ret == 0 {
- return 0
- }
- return ret + 3
- }
- return 0
- }
- func codeSpan(p *parser, out *bytes.Buffer, data []byte, offset int) int {
- data = data[offset:]
- nb := 0
- // count the number of backticks in the delimiter
- for nb < len(data) && data[nb] == '`' {
- nb++
- }
- // find the next delimiter
- i, end := 0, 0
- for end = nb; end < len(data) && i < nb; end++ {
- if data[end] == '`' {
- i++
- } else {
- i = 0
- }
- }
- // no matching delimiter?
- if i < nb && end >= len(data) {
- return 0
- }
- // trim outside whitespace
- fBegin := nb
- for fBegin < end && data[fBegin] == ' ' {
- fBegin++
- }
- fEnd := end - nb
- for fEnd > fBegin && data[fEnd-1] == ' ' {
- fEnd--
- }
- // render the code span
- if fBegin != fEnd {
- p.r.CodeSpan(out, data[fBegin:fEnd])
- }
- return end
- }
- // newline preceded by two spaces becomes <br>
- // newline without two spaces works when EXTENSION_HARD_LINE_BREAK is enabled
- func lineBreak(p *parser, out *bytes.Buffer, data []byte, offset int) int {
- // remove trailing spaces from out
- outBytes := out.Bytes()
- end := len(outBytes)
- eol := end
- for eol > 0 && outBytes[eol-1] == ' ' {
- eol--
- }
- out.Truncate(eol)
- precededByTwoSpaces := offset >= 2 && data[offset-2] == ' ' && data[offset-1] == ' '
- precededByBackslash := offset >= 1 && data[offset-1] == '\\' // see http://spec.commonmark.org/0.18/#example-527
- precededByBackslash = precededByBackslash && p.flags&EXTENSION_BACKSLASH_LINE_BREAK != 0
- // should there be a hard line break here?
- if p.flags&EXTENSION_HARD_LINE_BREAK == 0 && !precededByTwoSpaces && !precededByBackslash {
- return 0
- }
- if precededByBackslash && eol > 0 {
- out.Truncate(eol - 1)
- }
- p.r.LineBreak(out)
- return 1
- }
- type linkType int
- const (
- linkNormal linkType = iota
- linkImg
- linkDeferredFootnote
- linkInlineFootnote
- )
- func isReferenceStyleLink(data []byte, pos int, t linkType) bool {
- if t == linkDeferredFootnote {
- return false
- }
- return pos < len(data)-1 && data[pos] == '[' && data[pos+1] != '^'
- }
- // '[': parse a link or an image or a footnote
- func link(p *parser, out *bytes.Buffer, data []byte, offset int) int {
- // no links allowed inside regular links, footnote, and deferred footnotes
- if p.insideLink && (offset > 0 && data[offset-1] == '[' || len(data)-1 > offset && data[offset+1] == '^') {
- return 0
- }
- var t linkType
- switch {
- // special case: ![^text] == deferred footnote (that follows something with
- // an exclamation point)
- case p.flags&EXTENSION_FOOTNOTES != 0 && len(data)-1 > offset && data[offset+1] == '^':
- t = linkDeferredFootnote
- // ![alt] == image
- case offset > 0 && data[offset-1] == '!':
- t = linkImg
- // ^[text] == inline footnote
- // [^refId] == deferred footnote
- case p.flags&EXTENSION_FOOTNOTES != 0:
- if offset > 0 && data[offset-1] == '^' {
- t = linkInlineFootnote
- } else if len(data)-1 > offset && data[offset+1] == '^' {
- t = linkDeferredFootnote
- }
- // [text] == regular link
- default:
- t = linkNormal
- }
- data = data[offset:]
- var (
- i = 1
- noteId int
- title, link, altContent []byte
- textHasNl = false
- )
- if t == linkDeferredFootnote {
- i++
- }
- brace := 0
- // look for the matching closing bracket
- for level := 1; level > 0 && i < len(data); i++ {
- switch {
- case data[i] == '\n':
- textHasNl = true
- case data[i-1] == '\\':
- continue
- case data[i] == '[':
- level++
- case data[i] == ']':
- level--
- if level <= 0 {
- i-- // compensate for extra i++ in for loop
- }
- }
- }
- if i >= len(data) {
- return 0
- }
- txtE := i
- i++
- // skip any amount of whitespace or newline
- // (this is much more lax than original markdown syntax)
- for i < len(data) && isspace(data[i]) {
- i++
- }
- switch {
- // inline style link
- case i < len(data) && data[i] == '(':
- // skip initial whitespace
- i++
- for i < len(data) && isspace(data[i]) {
- i++
- }
- linkB := i
- // look for link end: ' " ), check for new opening braces and take this
- // into account, this may lead for overshooting and probably will require
- // some fine-tuning.
- findlinkend:
- for i < len(data) {
- switch {
- case data[i] == '\\':
- i += 2
- case data[i] == '(':
- brace++
- i++
- case data[i] == ')':
- if brace <= 0 {
- break findlinkend
- }
- brace--
- i++
- case data[i] == '\'' || data[i] == '"':
- break findlinkend
- default:
- i++
- }
- }
- if i >= len(data) {
- return 0
- }
- linkE := i
- // look for title end if present
- titleB, titleE := 0, 0
- if data[i] == '\'' || data[i] == '"' {
- i++
- titleB = i
- findtitleend:
- for i < len(data) {
- switch {
- case data[i] == '\\':
- i += 2
- case data[i] == ')':
- break findtitleend
- default:
- i++
- }
- }
- if i >= len(data) {
- return 0
- }
- // skip whitespace after title
- titleE = i - 1
- for titleE > titleB && isspace(data[titleE]) {
- titleE--
- }
- // check for closing quote presence
- if data[titleE] != '\'' && data[titleE] != '"' {
- titleB, titleE = 0, 0
- linkE = i
- }
- }
- // remove whitespace at the end of the link
- for linkE > linkB && isspace(data[linkE-1]) {
- linkE--
- }
- // remove optional angle brackets around the link
- if data[linkB] == '<' {
- linkB++
- }
- if data[linkE-1] == '>' {
- linkE--
- }
- // build escaped link and title
- if linkE > linkB {
- link = data[linkB:linkE]
- }
- if titleE > titleB {
- title = data[titleB:titleE]
- }
- i++
- // reference style link
- case isReferenceStyleLink(data, i, t):
- var id []byte
- altContentConsidered := false
- // look for the id
- i++
- linkB := i
- for i < len(data) && data[i] != ']' {
- i++
- }
- if i >= len(data) {
- return 0
- }
- linkE := i
- // find the reference
- if linkB == linkE {
- if textHasNl {
- var b bytes.Buffer
- for j := 1; j < txtE; j++ {
- switch {
- case data[j] != '\n':
- b.WriteByte(data[j])
- case data[j-1] != ' ':
- b.WriteByte(' ')
- }
- }
- id = b.Bytes()
- } else {
- id = data[1:txtE]
- altContentConsidered = true
- }
- } else {
- id = data[linkB:linkE]
- }
- // find the reference with matching id
- lr, ok := p.getRef(string(id))
- if !ok {
- return 0
- }
- // keep link and title from reference
- link = lr.link
- title = lr.title
- if altContentConsidered {
- altContent = lr.text
- }
- i++
- // shortcut reference style link or reference or inline footnote
- default:
- var id []byte
- // craft the id
- if textHasNl {
- var b bytes.Buffer
- for j := 1; j < txtE; j++ {
- switch {
- case data[j] != '\n':
- b.WriteByte(data[j])
- case data[j-1] != ' ':
- b.WriteByte(' ')
- }
- }
- id = b.Bytes()
- } else {
- if t == linkDeferredFootnote {
- id = data[2:txtE] // get rid of the ^
- } else {
- id = data[1:txtE]
- }
- }
- if t == linkInlineFootnote {
- // create a new reference
- noteId = len(p.notes) + 1
- var fragment []byte
- if len(id) > 0 {
- if len(id) < 16 {
- fragment = make([]byte, len(id))
- } else {
- fragment = make([]byte, 16)
- }
- copy(fragment, slugify(id))
- } else {
- fragment = append([]byte("footnote-"), []byte(strconv.Itoa(noteId))...)
- }
- ref := &reference{
- noteId: noteId,
- hasBlock: false,
- link: fragment,
- title: id,
- }
- p.notes = append(p.notes, ref)
- link = ref.link
- title = ref.title
- } else {
- // find the reference with matching id
- lr, ok := p.getRef(string(id))
- if !ok {
- return 0
- }
- if t == linkDeferredFootnote {
- lr.noteId = len(p.notes) + 1
- p.notes = append(p.notes, lr)
- }
- // keep link and title from reference
- link = lr.link
- // if inline footnote, title == footnote contents
- title = lr.title
- noteId = lr.noteId
- }
- // rewind the whitespace
- i = txtE + 1
- }
- // build content: img alt is escaped, link content is parsed
- var content bytes.Buffer
- if txtE > 1 {
- if t == linkImg {
- content.Write(data[1:txtE])
- } else {
- // links cannot contain other links, so turn off link parsing temporarily
- insideLink := p.insideLink
- p.insideLink = true
- p.inline(&content, data[1:txtE])
- p.insideLink = insideLink
- }
- }
- var uLink []byte
- if t == linkNormal || t == linkImg {
- if len(link) > 0 {
- var uLinkBuf bytes.Buffer
- unescapeText(&uLinkBuf, link)
- uLink = uLinkBuf.Bytes()
- }
- // links need something to click on and somewhere to go
- if len(uLink) == 0 || (t == linkNormal && content.Len() == 0) {
- return 0
- }
- }
- // call the relevant rendering function
- switch t {
- case linkNormal:
- if len(altContent) > 0 {
- p.r.Link(out, uLink, title, altContent)
- } else {
- p.r.Link(out, uLink, title, content.Bytes())
- }
- case linkImg:
- outSize := out.Len()
- outBytes := out.Bytes()
- if outSize > 0 && outBytes[outSize-1] == '!' {
- out.Truncate(outSize - 1)
- }
- p.r.Image(out, uLink, title, content.Bytes())
- case linkInlineFootnote:
- outSize := out.Len()
- outBytes := out.Bytes()
- if outSize > 0 && outBytes[outSize-1] == '^' {
- out.Truncate(outSize - 1)
- }
- p.r.FootnoteRef(out, link, noteId)
- case linkDeferredFootnote:
- p.r.FootnoteRef(out, link, noteId)
- default:
- return 0
- }
- return i
- }
- func (p *parser) inlineHTMLComment(out *bytes.Buffer, data []byte) int {
- if len(data) < 5 {
- return 0
- }
- if data[0] != '<' || data[1] != '!' || data[2] != '-' || data[3] != '-' {
- return 0
- }
- i := 5
- // scan for an end-of-comment marker, across lines if necessary
- for i < len(data) && !(data[i-2] == '-' && data[i-1] == '-' && data[i] == '>') {
- i++
- }
- // no end-of-comment marker
- if i >= len(data) {
- return 0
- }
- return i + 1
- }
- // '<' when tags or autolinks are allowed
- func leftAngle(p *parser, out *bytes.Buffer, data []byte, offset int) int {
- data = data[offset:]
- altype := LINK_TYPE_NOT_AUTOLINK
- end := tagLength(data, &altype)
- if size := p.inlineHTMLComment(out, data); size > 0 {
- end = size
- }
- if end > 2 {
- if altype != LINK_TYPE_NOT_AUTOLINK {
- var uLink bytes.Buffer
- unescapeText(&uLink, data[1:end+1-2])
- if uLink.Len() > 0 {
- p.r.AutoLink(out, uLink.Bytes(), altype)
- }
- } else {
- p.r.RawHtmlTag(out, data[:end])
- }
- }
- return end
- }
- // '\\' backslash escape
- var escapeChars = []byte("\\`*_{}[]()#+-.!:|&<>~")
- func escape(p *parser, out *bytes.Buffer, data []byte, offset int) int {
- data = data[offset:]
- if len(data) > 1 {
- if bytes.IndexByte(escapeChars, data[1]) < 0 {
- return 0
- }
- p.r.NormalText(out, data[1:2])
- }
- return 2
- }
- func unescapeText(ob *bytes.Buffer, src []byte) {
- i := 0
- for i < len(src) {
- org := i
- for i < len(src) && src[i] != '\\' {
- i++
- }
- if i > org {
- ob.Write(src[org:i])
- }
- if i+1 >= len(src) {
- break
- }
- ob.WriteByte(src[i+1])
- i += 2
- }
- }
- // '&' escaped when it doesn't belong to an entity
- // valid entities are assumed to be anything matching &#?[A-Za-z0-9]+;
- func entity(p *parser, out *bytes.Buffer, data []byte, offset int) int {
- data = data[offset:]
- end := 1
- if end < len(data) && data[end] == '#' {
- end++
- }
- for end < len(data) && isalnum(data[end]) {
- end++
- }
- if end < len(data) && data[end] == ';' {
- end++ // real entity
- } else {
- return 0 // lone '&'
- }
- p.r.Entity(out, data[:end])
- return end
- }
- func linkEndsWithEntity(data []byte, linkEnd int) bool {
- entityRanges := htmlEntity.FindAllIndex(data[:linkEnd], -1)
- return entityRanges != nil && entityRanges[len(entityRanges)-1][1] == linkEnd
- }
- func autoLink(p *parser, out *bytes.Buffer, data []byte, offset int) int {
- // quick check to rule out most false hits on ':'
- if p.insideLink || len(data) < offset+3 || data[offset+1] != '/' || data[offset+2] != '/' {
- return 0
- }
- // Now a more expensive check to see if we're not inside an anchor element
- anchorStart := offset
- offsetFromAnchor := 0
- for anchorStart > 0 && data[anchorStart] != '<' {
- anchorStart--
- offsetFromAnchor++
- }
- anchorStr := anchorRe.Find(data[anchorStart:])
- if anchorStr != nil {
- out.Write(anchorStr[offsetFromAnchor:])
- return len(anchorStr) - offsetFromAnchor
- }
- // scan backward for a word boundary
- rewind := 0
- for offset-rewind > 0 && rewind <= 7 && isletter(data[offset-rewind-1]) {
- rewind++
- }
- if rewind > 6 { // longest supported protocol is "mailto" which has 6 letters
- return 0
- }
- origData := data
- data = data[offset-rewind:]
- if !isSafeLink(data) {
- return 0
- }
- linkEnd := 0
- for linkEnd < len(data) && !isEndOfLink(data[linkEnd]) {
- linkEnd++
- }
- // Skip punctuation at the end of the link
- if (data[linkEnd-1] == '.' || data[linkEnd-1] == ',') && data[linkEnd-2] != '\\' {
- linkEnd--
- }
- // But don't skip semicolon if it's a part of escaped entity:
- if data[linkEnd-1] == ';' && data[linkEnd-2] != '\\' && !linkEndsWithEntity(data, linkEnd) {
- linkEnd--
- }
- // See if the link finishes with a punctuation sign that can be closed.
- var copen byte
- switch data[linkEnd-1] {
- case '"':
- copen = '"'
- case '\'':
- copen = '\''
- case ')':
- copen = '('
- case ']':
- copen = '['
- case '}':
- copen = '{'
- default:
- copen = 0
- }
- if copen != 0 {
- bufEnd := offset - rewind + linkEnd - 2
- openDelim := 1
- /* Try to close the final punctuation sign in this same line;
- * if we managed to close it outside of the URL, that means that it's
- * not part of the URL. If it closes inside the URL, that means it
- * is part of the URL.
- *
- * Examples:
- *
- * foo http://www.pokemon.com/Pikachu_(Electric) bar
- * => http://www.pokemon.com/Pikachu_(Electric)
- *
- * foo (http://www.pokemon.com/Pikachu_(Electric)) bar
- * => http://www.pokemon.com/Pikachu_(Electric)
- *
- * foo http://www.pokemon.com/Pikachu_(Electric)) bar
- * => http://www.pokemon.com/Pikachu_(Electric))
- *
- * (foo http://www.pokemon.com/Pikachu_(Electric)) bar
- * => foo http://www.pokemon.com/Pikachu_(Electric)
- */
- for bufEnd >= 0 && origData[bufEnd] != '\n' && openDelim != 0 {
- if origData[bufEnd] == data[linkEnd-1] {
- openDelim++
- }
- if origData[bufEnd] == copen {
- openDelim--
- }
- bufEnd--
- }
- if openDelim == 0 {
- linkEnd--
- }
- }
- // we were triggered on the ':', so we need to rewind the output a bit
- if out.Len() >= rewind {
- out.Truncate(len(out.Bytes()) - rewind)
- }
- var uLink bytes.Buffer
- unescapeText(&uLink, data[:linkEnd])
- if uLink.Len() > 0 {
- p.r.AutoLink(out, uLink.Bytes(), LINK_TYPE_NORMAL)
- }
- return linkEnd - rewind
- }
- func isEndOfLink(char byte) bool {
- return isspace(char) || char == '<'
- }
- var validUris = [][]byte{[]byte("http://"), []byte("https://"), []byte("ftp://"), []byte("mailto://")}
- var validPaths = [][]byte{[]byte("/"), []byte("./"), []byte("../")}
- func isSafeLink(link []byte) bool {
- for _, path := range validPaths {
- if len(link) >= len(path) && bytes.Equal(link[:len(path)], path) {
- if len(link) == len(path) {
- return true
- } else if isalnum(link[len(path)]) {
- return true
- }
- }
- }
- for _, prefix := range validUris {
- // TODO: handle unicode here
- // case-insensitive prefix test
- if len(link) > len(prefix) && bytes.Equal(bytes.ToLower(link[:len(prefix)]), prefix) && isalnum(link[len(prefix)]) {
- return true
- }
- }
- return false
- }
- // return the length of the given tag, or 0 is it's not valid
- func tagLength(data []byte, autolink *int) int {
- var i, j int
- // a valid tag can't be shorter than 3 chars
- if len(data) < 3 {
- return 0
- }
- // begins with a '<' optionally followed by '/', followed by letter or number
- if data[0] != '<' {
- return 0
- }
- if data[1] == '/' {
- i = 2
- } else {
- i = 1
- }
- if !isalnum(data[i]) {
- return 0
- }
- // scheme test
- *autolink = LINK_TYPE_NOT_AUTOLINK
- // try to find the beginning of an URI
- for i < len(data) && (isalnum(data[i]) || data[i] == '.' || data[i] == '+' || data[i] == '-') {
- i++
- }
- if i > 1 && i < len(data) && data[i] == '@' {
- if j = isMailtoAutoLink(data[i:]); j != 0 {
- *autolink = LINK_TYPE_EMAIL
- return i + j
- }
- }
- if i > 2 && i < len(data) && data[i] == ':' {
- *autolink = LINK_TYPE_NORMAL
- i++
- }
- // complete autolink test: no whitespace or ' or "
- switch {
- case i >= len(data):
- *autolink = LINK_TYPE_NOT_AUTOLINK
- case *autolink != 0:
- j = i
- for i < len(data) {
- if data[i] == '\\' {
- i += 2
- } else if data[i] == '>' || data[i] == '\'' || data[i] == '"' || isspace(data[i]) {
- break
- } else {
- i++
- }
- }
- if i >= len(data) {
- return 0
- }
- if i > j && data[i] == '>' {
- return i + 1
- }
- // one of the forbidden chars has been found
- *autolink = LINK_TYPE_NOT_AUTOLINK
- }
- // look for something looking like a tag end
- for i < len(data) && data[i] != '>' {
- i++
- }
- if i >= len(data) {
- return 0
- }
- return i + 1
- }
- // look for the address part of a mail autolink and '>'
- // this is less strict than the original markdown e-mail address matching
- func isMailtoAutoLink(data []byte) int {
- nb := 0
- // address is assumed to be: [-@._a-zA-Z0-9]+ with exactly one '@'
- for i := 0; i < len(data); i++ {
- if isalnum(data[i]) {
- continue
- }
- switch data[i] {
- case '@':
- nb++
- case '-', '.', '_':
- // Do nothing.
- case '>':
- if nb == 1 {
- return i + 1
- } else {
- return 0
- }
- default:
- return 0
- }
- }
- return 0
- }
- // look for the next emph char, skipping other constructs
- func helperFindEmphChar(data []byte, c byte) int {
- i := 0
- for i < len(data) {
- for i < len(data) && data[i] != c && data[i] != '`' && data[i] != '[' {
- i++
- }
- if i >= len(data) {
- return 0
- }
- // do not count escaped chars
- if i != 0 && data[i-1] == '\\' {
- i++
- continue
- }
- if data[i] == c {
- return i
- }
- if data[i] == '`' {
- // skip a code span
- tmpI := 0
- i++
- for i < len(data) && data[i] != '`' {
- if tmpI == 0 && data[i] == c {
- tmpI = i
- }
- i++
- }
- if i >= len(data) {
- return tmpI
- }
- i++
- } else if data[i] == '[' {
- // skip a link
- tmpI := 0
- i++
- for i < len(data) && data[i] != ']' {
- if tmpI == 0 && data[i] == c {
- tmpI = i
- }
- i++
- }
- i++
- for i < len(data) && (data[i] == ' ' || data[i] == '\n') {
- i++
- }
- if i >= len(data) {
- return tmpI
- }
- if data[i] != '[' && data[i] != '(' { // not a link
- if tmpI > 0 {
- return tmpI
- } else {
- continue
- }
- }
- cc := data[i]
- i++
- for i < len(data) && data[i] != cc {
- if tmpI == 0 && data[i] == c {
- return i
- }
- i++
- }
- if i >= len(data) {
- return tmpI
- }
- i++
- }
- }
- return 0
- }
- func helperEmphasis(p *parser, out *bytes.Buffer, data []byte, c byte) int {
- i := 0
- // skip one symbol if coming from emph3
- if len(data) > 1 && data[0] == c && data[1] == c {
- i = 1
- }
- for i < len(data) {
- length := helperFindEmphChar(data[i:], c)
- if length == 0 {
- return 0
- }
- i += length
- if i >= len(data) {
- return 0
- }
- if i+1 < len(data) && data[i+1] == c {
- i++
- continue
- }
- if data[i] == c && !isspace(data[i-1]) {
- if p.flags&EXTENSION_NO_INTRA_EMPHASIS != 0 {
- if !(i+1 == len(data) || isspace(data[i+1]) || ispunct(data[i+1])) {
- continue
- }
- }
- var work bytes.Buffer
- p.inline(&work, data[:i])
- p.r.Emphasis(out, work.Bytes())
- return i + 1
- }
- }
- return 0
- }
- func helperDoubleEmphasis(p *parser, out *bytes.Buffer, data []byte, c byte) int {
- i := 0
- for i < len(data) {
- length := helperFindEmphChar(data[i:], c)
- if length == 0 {
- return 0
- }
- i += length
- if i+1 < len(data) && data[i] == c && data[i+1] == c && i > 0 && !isspace(data[i-1]) {
- var work bytes.Buffer
- p.inline(&work, data[:i])
- if work.Len() > 0 {
- // pick the right renderer
- if c == '~' {
- p.r.StrikeThrough(out, work.Bytes())
- } else {
- p.r.DoubleEmphasis(out, work.Bytes())
- }
- }
- return i + 2
- }
- i++
- }
- return 0
- }
- func helperTripleEmphasis(p *parser, out *bytes.Buffer, data []byte, offset int, c byte) int {
- i := 0
- origData := data
- data = data[offset:]
- for i < len(data) {
- length := helperFindEmphChar(data[i:], c)
- if length == 0 {
- return 0
- }
- i += length
- // skip whitespace preceded symbols
- if data[i] != c || isspace(data[i-1]) {
- continue
- }
- switch {
- case i+2 < len(data) && data[i+1] == c && data[i+2] == c:
- // triple symbol found
- var work bytes.Buffer
- p.inline(&work, data[:i])
- if work.Len() > 0 {
- p.r.TripleEmphasis(out, work.Bytes())
- }
- return i + 3
- case (i+1 < len(data) && data[i+1] == c):
- // double symbol found, hand over to emph1
- length = helperEmphasis(p, out, origData[offset-2:], c)
- if length == 0 {
- return 0
- } else {
- return length - 2
- }
- default:
- // single symbol found, hand over to emph2
- length = helperDoubleEmphasis(p, out, origData[offset-1:], c)
- if length == 0 {
- return 0
- } else {
- return length - 1
- }
- }
- }
- return 0
- }
|