/****************************************************************************** * * Copyright (C) 1997-2015 by Dimitri van Heesch. * * Permission to use, copy, modify, and distribute this software and its * documentation under the terms of the GNU General Public License is hereby * granted. No representations are made about the suitability of this software * for any purpose. It is provided "as is" without express or implied warranty. * See the GNU General Public License for more details. * * Documents produced by Doxygen are derivative works derived from the * input used in their production; they are not affected by this license. * */ /* Note: part of the code below is inspired by libupskirt written by * Natacha Porté. Original copyright message follows: * * Copyright (c) 2008, Natacha Porté * * Permission to use, copy, modify, and distribute this software for any * purpose with or without fee is hereby granted, provided that the above * copyright notice and this permission notice appear in all copies. * * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE. */ #include #include #include #include #include #include //#define USE_ORIGINAL_TABLES #include "markdown.h" #include "growbuf.h" #include "debug.h" #include "util.h" #include "doxygen.h" #include "commentscan.h" #include "entry.h" #include "bufstr.h" #include "commentcnv.h" #include "config.h" #include "section.h" #include "message.h" #include "portable.h" //----------- // is character at position i in data part of an identifier? #define isIdChar(i) \ ((data[i]>='a' && data[i]<='z') || \ (data[i]>='A' && data[i]<='Z') || \ (data[i]>='0' && data[i]<='9') || \ (((unsigned char)data[i])>=0x80)) // unicode characters #define extraChar(i) \ (data[i]=='-' || data[i]=='+' || data[i]=='!' || \ data[i]=='?' || data[i]=='$' || data[i]=='@' || \ data[i]=='&' || data[i]=='*' || data[i]=='%') // is character at position i in data allowed before an emphasis section #define isOpenEmphChar(i) \ (data[i]=='\n' || data[i]==' ' || data[i]=='\'' || data[i]=='<' || \ data[i]=='{' || data[i]=='(' || data[i]=='[' || data[i]==',' || \ data[i]==':' || data[i]==';') // is character at position i in data an escape that prevents ending an emphasis section // so for example *bla (*.txt) is cool* #define ignoreCloseEmphChar(i) \ (data[i]=='(' || data[i]=='{' || data[i]=='[' || data[i]=='<' || \ data[i]=='\\' || \ data[i]=='@') //---------- struct LinkRef { LinkRef(const QCString &l,const QCString &t) : link(l), title(t) {} QCString link; QCString title; }; struct TableCell { TableCell() : colSpan(false) {} QCString cellText; bool colSpan; }; typedef int (*action_t)(GrowBuf &out,const char *data,int offset,int size); enum Alignment { AlignNone, AlignLeft, AlignCenter, AlignRight }; //---------- static QDict g_linkRefs(257); static action_t g_actions[256]; static Entry *g_current; static QCString g_fileName; static int g_lineNr; static int g_indentLevel=0; // 0 is outside markdown, -1=page level static const char g_utf8_nbsp[3] = {'\xc2', '\xa0', '\0'}; // UTF-8 nbsp static const char *g_doxy_nsbp = "&_doxy_nbsp;"; // doxygen escape command for UTF-8 nbsp //---------- const int codeBlockIndent = 4; static void processInline(GrowBuf &out,const char *data,int size); // escape characters that have a special meaning later on. static QCString escapeSpecialChars(const QCString &s) { if (s.isEmpty()) return ""; bool insideQuote=FALSE; GrowBuf growBuf; const char *p=s; char c,pc='\0'; while ((c=*p++)) { switch (c) { case '"': if (pc!='\\') { insideQuote=!insideQuote; } growBuf.addChar(c); break; case '<': if (!insideQuote) { growBuf.addChar('\\'); } growBuf.addChar('<'); break; case '>': if (!insideQuote) { growBuf.addChar('\\'); } growBuf.addChar('>'); break; case '\\': if (!insideQuote) { growBuf.addChar('\\'); } growBuf.addChar('\\'); break; case '@': if (!insideQuote) { growBuf.addChar('\\'); } growBuf.addChar('@'); break; case '#': if (!insideQuote) { growBuf.addChar('\\'); } growBuf.addChar('#'); break; default: growBuf.addChar(c); break; } pc=c; } growBuf.addChar(0); return growBuf.get(); } static void convertStringFragment(QCString &result,const char *data,int size) { if (size<0) size=0; result.resize(size+1); memcpy(result.rawData(),data,size); result.at(size)='\0'; } /** helper function to convert presence of left and/or right alignment markers * to a alignment value */ static Alignment markersToAlignment(bool leftMarker,bool rightMarker) { //printf("markerToAlignment(%d,%d)\n",leftMarker,rightMarker); if (leftMarker && rightMarker) { return AlignCenter; } else if (leftMarker) { return AlignLeft; } else if (rightMarker) { return AlignRight; } else { return AlignNone; } } // Check if data contains a block command. If so returned the command // that ends the block. If not an empty string is returned. // Note When offset>0 character position -1 will be inspected. // // Checks for and skip the following block commands: // {@code .. { .. } .. } // \dot .. \enddot // \code .. \endcode // \msc .. \endmsc // \f$..\f$ // \f[..\f] // \f{..\f} // \verbatim..\endverbatim // \latexonly..\endlatexonly // \htmlonly..\endhtmlonly // \xmlonly..\endxmlonly // \rtfonly..\endrtfonly // \manonly..\endmanonly static QCString isBlockCommand(const char *data,int offset,int size) { bool openBracket = offset>0 && data[-1]=='{'; bool isEscaped = offset>0 && (data[-1]=='\\' || data[-1]=='@'); if (isEscaped) return QCString(); int end=1; while (end='a' && data[end]<='z')) end++; if (end==1) return QCString(); QCString blockName; convertStringFragment(blockName,data+1,end-1); if (blockName=="code" && openBracket) { return "}"; } else if (blockName=="dot" || blockName=="code" || blockName=="msc" || blockName=="verbatim" || blockName=="latexonly" || blockName=="htmlonly" || blockName=="xmlonly" || blockName=="rtfonly" || blockName=="manonly" || blockName=="docbookonly" ) { return "end"+blockName; } else if (blockName=="startuml") { return "enduml"; } else if (blockName=="f" && end0 && ignoreCloseEmphChar(i-1)) { i++; continue; } else { // get length of emphasis token int len = 0; while (i+len0) { if (len!=c_size || (i=size || data[i]=='\n') return 0; // empty line -> paragraph } else // should not get here! { i++; } } return 0; } /** process single emphasis */ static int processEmphasis1(GrowBuf &out, const char *data, int size, char c) { int i = 0, len; /* skipping one symbol if coming from emph3 */ if (size>1 && data[0]==c && data[1]==c) { i=1; } while (i=size) return 0; if (i+1"); processInline(out,data,i); out.addStr(""); return i+1; } } return 0; } /** process double emphasis */ static int processEmphasis2(GrowBuf &out, const char *data, int size, char c) { int i = 0, len; while (i"); else out.addStr(""); processInline(out,data,i); if (c == '~') out.addStr(""); else out.addStr(""); return i + 2; } i++; } return 0; } /** Parsing triple emphasis. * Finds the first closing tag, and delegates to the other emph */ static int processEmphasis3(GrowBuf &out, const char *data, int size, char c) { int i = 0, len; while (i"); processInline(out,data,i); out.addStr(""); return i+3; } else if (i+1=2 && off>=2 && qstrncmp(data-2,"')) return 0; // end HTML comment if (count==2 && (off<8 || qstrncmp(data-8,"operator",8)!=0)) // -- => ndash { out.addStr("–"); return 2; } else if (count==3) // --- => ndash { out.addStr("—"); return 3; } // not an ndash or mdash return 0; } /** Process quoted section "...", can contain one embedded newline */ static int processQuoted(GrowBuf &out,const char *data,int,int size) { int i=1; int nl=0; while (i.. are treated specially, in * the sense that all code inside is written unprocessed */ static int processHtmlTagWrite(GrowBuf &out,const char *data,int offset,int size,bool doWrite) { if (offset>0 && data[-1]=='\\') return 0; // escaped < // find the end of the html tag int i=1; int l=0; // compute length of the tag name while (i tag { bool insideStr=FALSE; while (i') { // found tag, copy from start to end of tag if (doWrite) out.addStr(data,i+6); //printf("found
..
[%d..%d]\n",0,i+6); return i+6; } } else if (insideStr && c=='"') { if (data[i-1]!='\\') insideStr=FALSE; } else if (c=='"') { insideStr=TRUE; } i++; } } else // some other html tag { if (l>0 && i') // { //printf("Found htmlTag={%s}\n",QCString(data).left(i+2).data()); if (doWrite) out.addStr(data,i+2); return i+2; } else if (data[i]=='>') // { //printf("Found htmlTag={%s}\n",QCString(data).left(i+1).data()); if (doWrite) out.addStr(data,i+1); return i+1; } else if (data[i]==' ') // ') // found end of tag { //printf("Found htmlTag={%s}\n",QCString(data).left(i+1).data()); if (doWrite) out.addStr(data,i+1); return i+1; } i++; } } } } //printf("Not a valid html tag\n"); return 0; } static int processHtmlTag(GrowBuf &out,const char *data,int offset,int size) { return processHtmlTagWrite(out,data,offset,size,true); } static int processEmphasis(GrowBuf &out,const char *data,int offset,int size) { if ((offset>0 && !isOpenEmphChar(-1)) || // invalid char before * or _ (size>1 && data[0]!=data[1] && !(isIdChar(1) || extraChar(1) || data[1]=='[')) || // invalid char after * or _ (size>2 && data[0]==data[1] && !(isIdChar(2) || extraChar(2) || data[2]=='['))) // invalid char after ** or __ { return 0; } char c = data[0]; int ret; if (size>2 && c!='~' && data[1]!=c) // _bla or *bla { // whitespace cannot follow an opening emphasis if (data[1]==' ' || data[1]=='\n' || (ret = processEmphasis1(out, data+1, size-1, c)) == 0) { return 0; } return ret+1; } if (size>3 && data[1]==c && data[2]!=c) // __bla or **bla { if (data[2]==' ' || data[2]=='\n' || (ret = processEmphasis2(out, data+2, size-2, c)) == 0) { return 0; } return ret+2; } if (size>4 && c!='~' && data[1]==c && data[2]==c && data[3]!=c) // ___bla or ***bla { if (data[3]==' ' || data[3]=='\n' || (ret = processEmphasis3(out, data+3, size-3, c)) == 0) { return 0; } return ret+3; } return 0; } static void writeMarkdownImage(GrowBuf &out, const char *fmt, bool explicitTitle, QCString title, QCString content, QCString link, FileDef *fd) { out.addStr("@image "); out.addStr(fmt); out.addStr(" "); out.addStr(link.mid(fd ? 0 : 5)); if (!explicitTitle && !content.isEmpty()) { out.addStr(" \""); out.addStr(content); out.addStr("\""); } else if ((content.isEmpty() || explicitTitle) && !title.isEmpty()) { out.addStr(" \""); out.addStr(title); out.addStr("\""); } out.addStr("\n"); } static int processLink(GrowBuf &out,const char *data,int,int size) { QCString content; QCString link; QCString title; int contentStart,contentEnd,linkStart,titleStart,titleEnd; bool isImageLink = FALSE; bool isToc = FALSE; int i=1; if (data[0]=='!') { isImageLink = TRUE; if (size<2 || data[1]!='[') { return 0; } i++; } contentStart=i; int level=1; int nl=0; // find the matching ] while (i1) return 0; // only allow one newline in the content } i++; } if (i>=size) return 0; // premature end of comment -> no link contentEnd=i; convertStringFragment(content,data+contentStart,contentEnd-contentStart); //printf("processLink: content={%s}\n",content.data()); if (!isImageLink && content.isEmpty()) return 0; // no link text i++; // skip over ] // skip whitespace while (i0) { if (data[i]=='\n') // unexpected EOL { nl++; if (nl>1) return 0; } else if (data[i]=='(') { braceCount++; } else if (data[i]==')') { braceCount--; } if (braceCount>0) { i++; } } if (i>=size || data[i]=='\n') return 0; convertStringFragment(link,data+linkStart,i-linkStart); link = link.stripWhiteSpace(); //printf("processLink: link={%s}\n",link.data()); if (link.isEmpty()) return 0; if (link.at(link.length()-1)=='>') link=link.left(link.length()-1); // optional title if (data[i]=='\'' || data[i]=='"') { char c = data[i]; i++; titleStart=i; nl=0; while (i1) return 0; nl++; } i++; } if (i>=size) { return 0; } titleEnd = i-1; // search back for closing marker while (titleEnd>titleStart && data[titleEnd]==' ') titleEnd--; if (data[titleEnd]==c) // found it { convertStringFragment(title,data+titleStart,titleEnd-titleStart); //printf("processLink: title={%s}\n",title.data()); } else { return 0; } } i++; } else if (i1) return 0; } i++; } if (i>=size) return 0; // extract link convertStringFragment(link,data+linkStart,i-linkStart); //printf("processLink: link={%s}\n",link.data()); link = link.stripWhiteSpace(); if (link.isEmpty()) // shortcut link { link=content; } // lookup reference LinkRef *lr = g_linkRefs.find(link.lower()); if (lr) // found it { link = lr->link; title = lr->title; //printf("processLink: ref: link={%s} title={%s}\n",link.data(),title.data()); } else // reference not found! { //printf("processLink: ref {%s} do not exist\n",link.lower().data()); return 0; } i++; } else if (ilink; title = lr->title; explicitTitle=TRUE; i=contentEnd; } else if (content=="TOC") { isToc=TRUE; i=contentEnd; } else { return 0; } i++; } else { return 0; } if (isToc) // special case for [TOC] { int level = Config_getInt(TOC_INCLUDE_HEADINGS); if (level > 0 && level <=5) { char levStr[10]; sprintf(levStr,"%d",level); out.addStr("@tableofcontents{html:"); out.addStr(levStr); out.addStr("}"); } } else if (isImageLink) { bool ambig; FileDef *fd=0; if (link.find("@ref ")!=-1 || link.find("\\ref ")!=-1 || (fd=findFileDef(Doxygen::imageNameDict,link,ambig))) // assume doxygen symbol link or local image link { writeMarkdownImage(out, "html", explicitTitle, title, content, link, fd); writeMarkdownImage(out, "latex", explicitTitle, title, content, link, fd); writeMarkdownImage(out, "rtf", explicitTitle, title, content, link, fd); writeMarkdownImage(out, "docbook", explicitTitle, title, content, link, fd); } else { out.addStr("\"");"); } } else { SrcLangExt lang = getLanguageFromFileName(link); int lp=-1; if ((lp=link.find("@ref "))!=-1 || (lp=link.find("\\ref "))!=-1 || (lang==SrcLangExt_Markdown && !isURL(link))) // assume doxygen symbol link { if (lp==-1) // link to markdown page { out.addStr("@ref "); } out.addStr(link); out.addStr(" \""); if (explicitTitle && !title.isEmpty()) { out.addStr(title); } else { out.addStr(content); } out.addStr("\""); } else if (link.find('/')!=-1 || link.find('.')!=-1 || link.find('#')!=-1) { // file/url link out.addStr(""); content = content.simplifyWhiteSpace(); processInline(out,content,content.length()); out.addStr(""); } else // avoid link to e.g. F[x](y) { //printf("no link for '%s'\n",link.data()); return 0; } } return i; } /** '`' parsing a code span (assuming codespan != 0) */ static int processCodeSpan(GrowBuf &out, const char *data, int /*offset*/, int size) { int end, nb = 0, i, f_begin, f_end; /* counting the number of backticks in the delimiter */ while (nb= size) { return 0; // no matching delimiter } if (nl==2) // too many newlines inside the span { return 0; } // trimming outside whitespaces f_begin = nb; while (f_begin < end && data[f_begin]==' ') { f_begin++; } f_end = end - nb; while (f_end > nb && data[f_end-1]==' ') { f_end--; } //printf("found code span '%s'\n",QCString(data+f_begin).left(f_end-f_begin).data()); /* real code span */ if (f_begin < f_end) { QCString codeFragment; convertStringFragment(codeFragment,data+f_begin,f_end-f_begin); out.addStr(""); //out.addStr(convertToHtml(codeFragment,TRUE)); out.addStr(escapeSpecialChars(codeFragment)); out.addStr(""); } return end; } static void addStrEscapeUtf8Nbsp(GrowBuf &out,const char *s,int len) { if (Portable::strnstr(s,g_doxy_nsbp,len)==0) // no escape needed -> fast { out.addStr(s,len); } else // escape needed -> slow { out.addStr(substitute(QCString(s).left(len),g_doxy_nsbp,g_utf8_nbsp)); } } static int processSpecialCommand(GrowBuf &out, const char *data, int offset, int size) { int i=1; QCString endBlockName = isBlockCommand(data,offset,size); if (!endBlockName.isEmpty()) { int l = endBlockName.length(); while (i1 && data[0]=='\\') { char c=data[1]; if (c=='[' || c==']' || c=='*' || c=='!' || c=='(' || c==')' || c=='`' || c=='_') { out.addChar(data[1]); return 2; } else if (c=='-' && size>3 && data[2]=='-' && data[3]=='-') // \--- { out.addStr(&data[1],3); return 4; } else if (c=='-' && size>2 && data[2]=='-') // \-- { out.addStr(&data[1],2); return 3; } } return 0; } static void processInline(GrowBuf &out,const char *data,int size) { int i=0, end=0; action_t action = 0; while (i=size) break; i=end; end = action(out,data+i,i,size-i); if (end<=0) { end=i+1-end; } else { i+=end; end=i; } } } /** returns whether the line is a setext-style hdr underline */ static int isHeaderline(const char *data, int size, bool allowAdjustLevel) { int i=0, c=0; while (i1 && (i>=size || data[i]=='\n')) ? 1 : 0; if (allowAdjustLevel && level==1 && g_indentLevel==-1) { // In case a page starts with a header line we use it as title, promoting it to @page. // We set g_indentLevel to -1 to promoting the other sections if they have a deeper // nesting level than the page header, i.e. @section..@subsection becomes @page..@section. // In case a section at the same level is found (@section..@section) however we need // to undo this (and the result will be @page..@section). g_indentLevel=0; } return g_indentLevel+level; } // test of level 2 header if (data[i]=='-') { while (i1 && (i>=size || data[i]=='\n')) ? g_indentLevel+2 : 0; } return 0; } /** returns TRUE if this line starts a block quote */ static bool isBlockQuote(const char *data,int size,int indent) { int i = 0; while (i's and skip spaces int level=0; while (i' || data[i]==' ')) { if (data[i]=='>') level++; i++; } // last characters should be a space or newline, // so a line starting with >= does not match return level>0 && i code block { return FALSE; } //return i' && i=size || data[i]!='[') return 0; i++; int refIdStart=i; while (i=size || data[i]!=']') return 0; convertStringFragment(refid,data+refIdStart,i-refIdStart); if (refid.isEmpty()) return 0; //printf(" isLinkRef: found refid='%s'\n",refid.data()); i++; if (i>=size || data[i]!=':') return 0; i++; // format: whitespace* \n? whitespace* ( | url) while (i=size) return 0; if (i') i++; if (linkStart==linkEnd) return 0; // empty link convertStringFragment(link,data+linkStart,linkEnd-linkStart); //printf(" isLinkRef: found link='%s'\n",link.data()); if (link=="@ref" || link=="\\ref") { int argStart=i; while (i=size) { //printf("end of isLinkRef while looking for title! i=%d\n",i); return i; // end of buffer while looking for the optional title } char c = data[i]; if (c=='\'' || c=='"' || c=='(') // optional title present? { //printf(" start of title found! char='%c'\n",c); i++; if (c=='(') c=')'; // replace c by end character int titleStart=i; // search for end of the line while (ititleStart && data[end]!=c) end--; if (end>titleStart) { convertStringFragment(title,data+titleStart,end-titleStart); } //printf(" title found: '%s'\n",title.data()); } while (i=size) return i; // end of buffer while ref id was found else if (eol) return eol; // end of line while ref id was found return 0; // invalid link ref } static int isHRuler(const char *data,int size) { int i=0; if (size>0 && data[size-1]=='\n') size--; // ignore newline character while (i=size) return 0; // empty line char c=data[i]; if (c!='*' && c!='-' && c!='_') { return 0; // not a hrule character } int n=0; while (i=3; // at least 3 characters needed for a hruler } static QCString extractTitleId(QCString &title, int level) { //static QRegExp r1("^[a-z_A-Z][a-z_A-Z0-9\\-]*:"); static QRegExp r2("\\{#[a-z_A-Z][a-z_A-Z0-9\\-]*\\}"); int l=0; int i = r2.match(title,0,&l); if (i!=-1 && title.mid(i+l).stripWhiteSpace().isEmpty()) // found {#id} style id { QCString id = title.mid(i+2,l-3); title = title.left(i); //printf("found id='%s' title='%s'\n",id.data(),title.data()); return id; } if ((level > 0) && (level <= Config_getInt(TOC_INCLUDE_HEADINGS))) { static int autoId = 0; QCString id; id.sprintf("autotoc_md%d",autoId++); //printf("auto-generated id='%s' title='%s'\n",id.data(),title.data()); return id; } //printf("no id found in title '%s'\n",title.data()); return ""; } static int isAtxHeader(const char *data,int size, QCString &header,QCString &id,bool allowAdjustLevel) { int i = 0, end; int level = 0, blanks=0; // find start of header text and determine heading level while (i=size || data[i]!='#') { return 0; } while (ii && (data[end-1]=='#' || data[end-1]==' ')) end--; // store result convertStringFragment(header,data+i,end-i); id = extractTitleId(header, level); if (!id.isEmpty()) // strip #'s between title and id { i=header.length()-1; while (i>=0 && (header.at(i)=='#' || header.at(i)==' ')) i--; header=header.left(i+1); } if (allowAdjustLevel && level==1 && g_indentLevel==-1) { // in case we find a `# Section` on a markdown page that started with the same level // header, we no longer need to artificially decrease the paragraph level. // So both // ------------------- // # heading 1 <-- here we set g_indentLevel to -1 // # heading 2 <-- here we set g_indentLevel back to 0 such that this will be a @section // ------------------- // and // ------------------- // # heading 1 <-- here we set g_indentLevel to -1 // ## heading 2 <-- here we keep g_indentLevel at -1 such that @subsection will be @section // ------------------- // will convert to // ------------------- // @page md_page Heading 1 // @section autotoc_md1 Heading 2 // ------------------- g_indentLevel=0; } return level+g_indentLevel; } static int isEmptyLine(const char *data,int size) { int i=0; while (i')) // compute the indent from the start of the input, excluding list markers // such as -, -#, *, +, 1., and
  • static int computeIndentExcludingListMarkers(const char *data,int size) { int i=0; int indent=0; bool isDigit=FALSE; bool isLi=FALSE; bool listMarkerSkipped=FALSE; while (i0 && data[i-1]=='-') || // -# item (isDigit=(data[i]>='1' && data[i]<='9')) || // ordered list marker? (isLi=(i tag ) ) ) ) { if (isDigit) // skip over ordered list marker '10. ' { int j=i+1; while (j='0' && data[j]<='9') || data[j]=='.')) { if (data[j]=='.') // should be end of the list marker { if (j indent+=3; listMarkerSkipped=TRUE; } else if (data[i]=='-' && i%d\n",QCString(data).left(size).data(),indent); return indent; } static bool isFencedCodeBlock(const char *data,int size,int refIndent, QCString &lang,int &start,int &end,int &offset) { // rules: at least 3 ~~~, end of the block same amount of ~~~'s, otherwise // return FALSE int i=0; int indent=0; int startTildes=0; while (i=refIndent+4) return FALSE; // part of code block char tildaChar='~'; if (iisCodeBlock: line is not indented enough %d<4\n",indent0); return FALSE; } if (indent0>=size || data[indent0]=='\n') // empty line does not start a code block { //printf("only spaces at the end of a comment block\n"); return FALSE; } i=offset; int nl=0; int nl_pos[3]; // search back 3 lines and remember the start of lines -1 and -2 while (i>0 && nl<3) { if (data[i-offset-1]=='\n') nl_pos[nl++]=i-offset; i--; } // if there are only 2 preceding lines, then line -2 starts at -offset if (i==0 && nl==2) nl_pos[nl++]=-offset; //printf(" nl=%d\n",nl); if (nl==3) // we have at least 2 preceding lines { //printf(" positions: nl_pos=[%d,%d,%d] line[-2]='%s' line[-1]='%s'\n", // nl_pos[0],nl_pos[1],nl_pos[2], // QCString(data+nl_pos[1]).left(nl_pos[0]-nl_pos[1]-1).data(), // QCString(data+nl_pos[2]).left(nl_pos[1]-nl_pos[2]-1).data()); // check that line -1 is empty if (!isEmptyLine(data+nl_pos[1],nl_pos[0]-nl_pos[1]-1)) { return FALSE; } // determine the indent of line -2 indent=computeIndentExcludingListMarkers(data+nl_pos[2],nl_pos[1]-nl_pos[2]); //printf(">isCodeBlock local_indent %d>=%d+4=%d\n", // indent0,indent2,indent0>=indent2+4); // if the difference is >4 spaces -> code block return indent0>=indent+codeBlockIndent; } else // not enough lines to determine the relative indent, use global indent { // check that line -1 is empty if (nl==1 && !isEmptyLine(data-offset,offset-1)) { return FALSE; } //printf(">isCodeBlock global indent %d>=%d+4=%d nl=%d\n", // indent0,indent,indent0>=indent+4,nl); return indent0>=indent+codeBlockIndent; } } /** Finds the location of the table's contains in the string \a data. * Only one line will be inspected. * @param[in] data pointer to the string buffer. * @param[in] size the size of the buffer. * @param[out] start offset of the first character of the table content * @param[out] end offset of the last character of the table content * @param[out] columns number of table columns found * @returns The offset until the next line in the buffer. */ int findTableColumns(const char *data,int size,int &start,int &end,int &columns) { int i=0,n=0; int eol; // find start character of the table line while (i0 && data[i]==' ') i--; if (i>0 && data[i-1]!='\\' && data[i]=='|') i--,n++; // trailing or escaped | does not count end = i; // count columns between start and end columns=0; if (end>start) { i=start; while (i<=end) // look for more column markers { if (data[i]=='|' && (i==0 || data[i-1]!='\\')) columns++; if (columns==1) columns++; // first | make a non-table into a two column table i++; } } if (n==2 && columns==0) // table row has | ... | { columns++; } //printf("findTableColumns(start=%d,end=%d,columns=%d) eol=%d\n", // start,end,columns,eol); return eol; } /** Returns TRUE iff data points to the start of a table block */ static bool isTableBlock(const char *data,int size) { int cc0,start,end; // the first line should have at least two columns separated by '|' int i = findTableColumns(data,size,start,end,cc0); if (i>=size || cc0<1) { //printf("isTableBlock: no |'s in the header\n"); return FALSE; } int cc1; int ret = findTableColumns(data+i,size-i,start,end,cc1); int j=i+start; // separator line should consist of |, - and : and spaces only while (j<=end+i) { if (data[j]!=':' && data[j]!='-' && data[j]!='|' && data[j]!=' ') { //printf("isTableBlock: invalid character '%c'\n",data[j]); return FALSE; // invalid characters in table separator } j++; } if (cc1!=cc0) // number of columns should be same as previous line { return FALSE; } i+=ret; // goto next line int cc2; findTableColumns(data+i,size-i,start,end,cc2); //printf("isTableBlock: %d\n",cc1==cc2); return cc1==cc2; } static int writeTableBlock(GrowBuf &out,const char *data,int size) { int i=0,j,k; int columns,start,end,cc; i = findTableColumns(data,size,start,end,columns); int headerStart = start; int headerEnd = end; #ifdef USE_ORIGINAL_TABLES out.addStr(""); // write table header, in range [start..end] out.addStr(""); #endif // read cell alignments int ret = findTableColumns(data+i,size-i,start,end,cc); k=0; Alignment *columnAlignment = new Alignment[columns]; bool leftMarker=FALSE,rightMarker=FALSE; bool startFound=FALSE; j=start+i; while (j<=end+i) { if (!startFound) { if (data[j]==':') { leftMarker=TRUE; startFound=TRUE; } if (data[j]=='-') startFound=TRUE; //printf(" data[%d]=%c startFound=%d\n",j,data[j],startFound); } if (data[j]=='-') rightMarker=FALSE; else if (data[j]==':') rightMarker=TRUE; if (j<=end+i && (data[j]=='|' && (j==0 || data[j-1]!='\\'))) { if (k"); while (m<=headerEnd && (data[m]!='|' || (m>0 && data[m-1]=='\\'))) { out.addChar(data[m++]); } m++; } out.addStr("\n\n"); // write table cells while (i"); j=start+i; int columnStart=j; k=0; while (j<=end+i) { if (j==columnStart) { out.addStr(""); } if (j<=end+i && (data[j]=='|' && (j==0 || data[j-1]!='\\'))) { columnStart=j+1; k++; } else { out.addChar(data[j]); } j++; } out.addChar('\n'); // proceed to next line i+=ret; } out.addStr("
    "); #else // Store the table cell information by row then column. This // allows us to handle row spanning. QVector > tableContents; tableContents.setAutoDelete(TRUE); int m=headerStart; QVector *headerContents = new QVector(columns); headerContents->setAutoDelete(TRUE); for (k=0;kinsert(k, new TableCell); while (m<=headerEnd && (data[m]!='|' || (m>0 && data[m-1]=='\\'))) { headerContents->at(k)->cellText += data[m++]; } m++; // do the column span test before stripping white space // || is spanning columns, | | is not headerContents->at(k)->colSpan = headerContents->at(k)->cellText.isEmpty(); headerContents->at(k)->cellText = headerContents->at(k)->cellText.stripWhiteSpace(); } // qvector doesn't have an append like std::vector, so we gotta do // extra work tableContents.resize(1); tableContents.insert(0, headerContents); // write table cells int rowNum = 1; while (i *rowContents = new QVector(columns); rowContents->setAutoDelete(TRUE); rowContents->insert(k, new TableCell); while (j<=end+i) { if (j<=end+i && (data[j]=='|' && (j==0 || data[j-1]!='\\'))) { // do the column span test before stripping white space // || is spanning columns, | | is not rowContents->at(k)->colSpan = rowContents->at(k)->cellText.isEmpty(); rowContents->at(k)->cellText = rowContents->at(k)->cellText.stripWhiteSpace(); k++; rowContents->insert(k, new TableCell); } // if (j<=end+i && (data[j]=='|' && (j==0 || data[j-1]!='\\'))) else { rowContents->at(k)->cellText += data[j]; } // else { if (j<=end+i && (data[j]=='|' && (j==0 || data[j-1]!='\\'))) } j++; } // while (j<=end+i) // do the column span test before stripping white space // || is spanning columns, | | is not rowContents->at(k)->colSpan = rowContents->at(k)->cellText.isEmpty(); rowContents->at(k)->cellText = rowContents->at(k)->cellText.stripWhiteSpace(); // qvector doesn't have an append like std::vector, so we gotta do // extra work tableContents.resize(tableContents.size()+1); tableContents.insert(rowNum++, rowContents); // proceed to next line i+=ret; } out.addStr(""); QCString cellTag("th"), cellClass("class=\"markdownTableHead"); for (unsigned row = 0; row < tableContents.size(); row++) { if (row) { if (row % 2) { out.addStr(""); } else { out.addStr(""); } } else { out.addStr(" "); } for (int c = 0; c < columns; c++) { // save the cell text for use after column span computation QCString cellText(tableContents[row]->at(c)->cellText); // Row span handling. Spanning rows will contain a caret ('^'). // If the current cell contains just a caret, this is part of an // earlier row's span and the cell should not be added to the // output. if (tableContents[row]->at(c)->cellText == "^") continue; unsigned rowSpan = 1, spanRow = row+1; while ((spanRow < tableContents.size()) && (tableContents[spanRow]->at(c)->cellText == "^")) { spanRow++; rowSpan++; } out.addStr(" <" + cellTag + " " + cellClass); // use appropriate alignment style switch (columnAlignment[c]) { case AlignLeft: out.addStr("Left\""); break; case AlignRight: out.addStr("Right\""); break; case AlignCenter: out.addStr("Center\""); break; case AlignNone: out.addStr("None\""); break; } if (rowSpan > 1) { QCString spanStr; spanStr.setNum(rowSpan); out.addStr(" rowspan=\"" + spanStr + "\""); } // Column span handling, assumes that column spans will have // empty strings, which would indicate the sequence "||", used // to signify spanning columns. unsigned colSpan = 1; while ((c < columns-1) && tableContents[row]->at(c+1)->colSpan) { c++; colSpan++; } if (colSpan > 1) { QCString spanStr; spanStr.setNum(colSpan); out.addStr(" colspan=\"" + spanStr + "\""); } // need at least one space on either side of the cell text in // order for doxygen to do other formatting out.addStr("> " + cellText + ""); } cellTag = "td"; cellClass = "class=\"markdownTableBody"; out.addStr(" \n"); } out.addStr("
    \n"); #endif delete[] columnAlignment; return i; } static int hasLineBreak(const char *data,int size) { int i=0; int j=0; // search for end of line and also check if it is not a completely blank while (i=size) return 0; // empty line if (i<2) return 0; // not long enough return (j>0 && data[i-1]==' ' && data[i-2]==' '); // non blank line with at two spaces at the end } void writeOneLineHeaderOrRuler(GrowBuf &out,const char *data,int size) { int level; QCString header; QCString id; if (isHRuler(data,size)) { out.addStr("\n
    \n"); } else if ((level=isAtxHeader(data,size,header,id,TRUE))) { QCString hTag; if (level<5 && !id.isEmpty()) { SectionType type = SectionType::Anchor; switch(level) { case 1: out.addStr("@section "); type=SectionType::Section; break; case 2: out.addStr("@subsection "); type=SectionType::Subsection; break; case 3: out.addStr("@subsubsection "); type=SectionType::Subsubsection; break; default: out.addStr("@paragraph "); type=SectionType::Paragraph; break; } out.addStr(id); out.addStr(" "); out.addStr(header); out.addStr("\n"); } else { if (!id.isEmpty()) { out.addStr("\\anchor "+id+"\n"); } hTag.sprintf("h%d",level); out.addStr("<"+hTag+">"); out.addStr(header); out.addStr("\n"); } } else // nothing interesting -> just output the line { out.addStr(data,size); if (hasLineBreak(data,size)) { out.addStr("
    \n"); } } } static int writeBlockQuote(GrowBuf &out,const char *data,int size) { int l; int i=0; int curLevel=0; int end=0; while (i')) { if (data[j]=='>') { level++; indent=j+1; } else if (j>0 && data[j-1]=='>') indent=j+1; j++; } if (j>0 && data[j-1]=='>' && !(j==size || data[j]=='\n')) // disqualify last > if not followed by space { indent--; j--; } if (level>curLevel) // quote level increased => add start markers { for (l=curLevel;l\n"); } } else if (level add end markers { for (l=level;l\n"); } } curLevel=level; if (level==0) break; // end of quote block // copy line without quotation marks out.addStr(data+indent,end-indent); // proceed with next line i=end; } // end of comment within blockquote => add end markers for (l=0;l\n"); } return i; } static int writeCodeBlock(GrowBuf &out,const char *data,int size,int refIndent) { int i=0,end; //printf("writeCodeBlock: data={%s}\n",QCString(data).left(size).data()); out.addStr("@verbatim\n"); int emptyLines=0; while (i=refIndent+codeBlockIndent) // enough indent to continue the code block { while (emptyLines>0) // write skipped empty lines { // add empty line out.addStr("\n"); emptyLines--; } // add code line minus the indent out.addStr(data+i+refIndent+codeBlockIndent,end-i-refIndent-codeBlockIndent); i=end; } else // end of code block { break; } } out.addStr("@endverbatim\n"); while (emptyLines>0) // write skipped empty lines { // add empty line out.addStr("\n"); emptyLines--; } //printf("i=%d\n",i); return i; } // start searching for the end of the line start at offset \a i // keeping track of possible blocks that need to be skipped. static void findEndOfLine(GrowBuf &out,const char *data,int size, int &pi,int&i,int &end) { // find end of the line int nb=0; end=i+1; while (end<=size && data[end-1]!='\n') { // while looking for the end of the line we might encounter a block // that needs to be passed unprocessed. if ((data[end-1]=='\\' || data[end-1]=='@') && // command (end<=1 || (data[end-2]!='\\' && data[end-2]!='@')) // not escaped ) { QCString endBlockName = isBlockCommand(data+end-1,end-1,size-(end-1)); end++; if (!endBlockName.isEmpty()) { int l = endBlockName.length(); for (;end') //
     tag
          {
            // skip part until including 
    end = end + processHtmlTagWrite(out,data+end-1,end-1,size-end+1,false) + 2; break; } else { end++; } } else if (nb==0 && data[end-1]=='`') { while (end<=size && data[end-1]=='`') end++,nb++; } else if (nb>0 && data[end-1]=='`') { int enb=0; while (end<=size && data[end-1]=='`') end++,enb++; if (enb==nb) nb=0; } else { end++; } } //printf("findEndOfLine pi=%d i=%d end=%d {%s}\n",pi,i,end,QCString(data+i).left(end-i).data()); } static void writeFencedCodeBlock(GrowBuf &out,const char *data,const char *lng, int blockStart,int blockEnd) { QCString lang = lng; if (!lang.isEmpty() && lang.at(0)=='.') lang=lang.mid(1); out.addStr("@code"); if (!lang.isEmpty()) { out.addStr("{"+lang+"}"); } addStrEscapeUtf8Nbsp(out,data+blockStart,blockEnd-blockStart); out.addStr("\n"); out.addStr("@endcode\n"); } static QCString processQuotations(const QCString &s,int refIndent) { GrowBuf out; const char *data = s.data(); int size = s.length(); int i=0,end=0,pi=-1; int blockStart,blockEnd,blockOffset; QCString lang; while (icodeBlockIndent && isCodeBlock(data,0,end,blockIndent)) { i=writeCodeBlock(out,data,size,blockIndent); end=i+1; pi=-1; } #endif // process each line while (i0) { //printf("Found header at %d-%d\n",i,end); while (pi":"

    "); out.addStr(header); out.addStr(level==1?"\n

    \n":"\n\n"); } } else { out.addStr("\n
    \n"); } pi=-1; i=end; end=i+1; continue; } else if ((ref=isLinkRef(data+pi,size-pi,id,link,title))) { //printf("found link ref: id='%s' link='%s' title='%s'\n", // id.data(),link.data(),title.data()); g_linkRefs.insert(id.lower(),new LinkRef(link,title)); i=ref+pi; pi=-1; end=i+1; } else if (isFencedCodeBlock(data+pi,size-pi,indent,lang,blockStart,blockEnd,blockOffset)) { //printf("Found FencedCodeBlock lang='%s' start=%d end=%d code={%s}\n", // lang.data(),blockStart,blockEnd,QCString(data+pi+blockStart).left(blockEnd-blockStart).data()); writeFencedCodeBlock(out,data+pi,lang,blockStart,blockEnd); i=pi+blockOffset; pi=-1; end=i+1; continue; } else if (isCodeBlock(data+i,i,end-i,blockIndent)) { // skip previous line (it is empty anyway) i+=writeCodeBlock(out,data+i,size-i,blockIndent); pi=-1; end=i+1; continue; } else if (isTableBlock(data+pi,size-pi)) { i=pi+writeTableBlock(out,data+pi,size-pi); pi=-1; end=i+1; continue; } else { writeOneLineHeaderOrRuler(out,data+pi,i-pi); } } pi=i; i=end; } //printf("last line %d size=%d\n",i,size); if (pi!=-1 && pi=size) return ""; int end1=i+1; while (end10) { docs=docs.mid(end1); } else { id = extractTitleId(title, 0); } //printf("extractPageTitle(title='%s' docs='%s' id='%s')\n",title.data(),docs.data(),id.data()); return title; } static QCString detab(const QCString &s,int &refIndent) { static int tabSize = Config_getInt(TAB_SIZE); int size = s.length(); GrowBuf out(size); const char *data = s.data(); int i=0; int col=0; const int maxIndent=1000000; // value representing infinity int minIndent=maxIndent; while (i update minIndent if (c<0 && i= 2 bytes if (((uchar)c&0xE0)==0xE0 && i",4)==0) p+=4; // skip over
    } if (p>result.data()) { // strip part of the input result = result.mid(p-result.data()); } return result; } else { return comment; } } //--------------------------------------------------------------------------- struct MarkdownOutlineParser::Private { CommentScanner commentScanner; }; MarkdownOutlineParser::MarkdownOutlineParser() : p(std::make_unique()) { } MarkdownOutlineParser::~MarkdownOutlineParser() { } void MarkdownOutlineParser::parseInput(const char *fileName, const char *fileBuf, const std::shared_ptr &root, bool /*sameTranslationUnit*/, QStrList & /*filesInSameTranslationUnit*/) { std::shared_ptr current = std::make_shared(); current->lang = SrcLangExt_Markdown; current->fileName = fileName; current->docFile = fileName; current->docLine = 1; QCString docs = fileBuf; QCString id; QCString title=extractPageTitle(docs,id).stripWhiteSpace(); if (id.startsWith("autotoc_md")) id = ""; g_indentLevel=title.isEmpty() ? 0 : -1; QCString titleFn = QFileInfo(fileName).baseName().utf8(); QCString fn = QFileInfo(fileName).fileName().utf8(); static QCString mdfileAsMainPage = Config_getString(USE_MDFILE_AS_MAINPAGE); if (id.isEmpty()) id = markdownFileNameToId(fileName); if (!isExplicitPage(docs)) { if (!mdfileAsMainPage.isEmpty() && (fn==mdfileAsMainPage || // name reference QFileInfo(fileName).absFilePath()== QFileInfo(mdfileAsMainPage).absFilePath()) // file reference with path ) { docs.prepend("@anchor " + id + "\n"); docs.prepend("@mainpage "+title+"\n"); } else if (id=="mainpage" || id=="index") { if (title.isEmpty()) title = titleFn; docs.prepend("@anchor " + id + "\n"); docs.prepend("@mainpage "+title+"\n"); } else { if (title.isEmpty()) title = titleFn; docs.prepend("@page "+id+" "+title+"\n"); } } int lineNr=1; // even without markdown support enabled, we still // parse markdown files as such bool markdownEnabled = Doxygen::markdownSupport; Doxygen::markdownSupport = TRUE; Protection prot=Public; bool needsEntry = FALSE; int position=0; QCString processedDocs = processMarkdownForCommentBlock(docs,fileName,lineNr); while (p->commentScanner.parseCommentBlock( this, current.get(), processedDocs, fileName, lineNr, FALSE, // isBrief FALSE, // javadoc autobrief FALSE, // inBodyDocs prot, // protection position, needsEntry)) { if (needsEntry) { QCString docFile = current->docFile; root->moveToSubEntryAndRefresh(current); current->lang = SrcLangExt_Markdown; current->docFile = docFile; current->docLine = lineNr; } } if (needsEntry) { root->moveToSubEntryAndKeep(current); } // restore setting Doxygen::markdownSupport = markdownEnabled; g_indentLevel=0; } void MarkdownOutlineParser::parsePrototype(const char *text) { OutlineParserInterface &intf = Doxygen::parserManager->getOutlineParser("*.cpp"); if (&intf!=this) { intf.parsePrototype(text); } } //------------------------------------------------------------------------