X-Git-Url: https://git.lyx.org/gitweb/?a=blobdiff_plain;f=src%2Foutput_xhtml.cpp;h=f1bcfb879b3cfba5b579fdde04d05b65bc9c41a8;hb=d9847302359b7f6d5ca02b97b3b323d7f6bc3df7;hp=d76a35379a160901a4bcd17ab4847dec186dd38b;hpb=5c32dc85d54b9cee8ceae77b7e4882447617358f;p=lyx.git diff --git a/src/output_xhtml.cpp b/src/output_xhtml.cpp index d76a35379a..f1bcfb879b 100644 --- a/src/output_xhtml.cpp +++ b/src/output_xhtml.cpp @@ -3,8 +3,8 @@ * This file is part of LyX, the document processor. * Licence details can be found in the file COPYING. * - * \author Richard Heck - * + * \author Richard Kimberly Heck + * * This code is based upon output_docbook.cpp * * Full author contact details are available in file CREDITS. @@ -15,26 +15,21 @@ #include "output_xhtml.h" #include "Buffer.h" -#include "buffer_funcs.h" #include "BufferParams.h" #include "Counters.h" #include "Font.h" #include "Layout.h" -#include "OutputParams.h" #include "Paragraph.h" #include "ParagraphList.h" #include "ParagraphParameters.h" -#include "sgml.h" +#include "xml.h" #include "Text.h" #include "TextClass.h" -#include "support/convert.h" -#include "support/debug.h" #include "support/lassert.h" -#include "support/lstrings.h" -#include "support/textutils.h" -#include +#include +#include // Uncomment to activate debugging code. // #define XHTML_DEBUG @@ -44,724 +39,135 @@ using namespace lyx::support; namespace lyx { -namespace html { - -docstring escapeChar(char_type c, XHTMLStream::EscapeSettings e) -{ - docstring str; - switch (e) { - case XHTMLStream::ESCAPE_NONE: - str += c; - break; - case XHTMLStream::ESCAPE_ALL: - if (c == '<') { - str += "<"; - break; - } else if (c == '>') { - str += ">"; - break; - } - // fall through - case XHTMLStream::ESCAPE_AND: - if (c == '&') - str += "&"; - else - str +=c ; - break; - } - return str; -} - - -// escape what needs escaping -docstring htmlize(docstring const & str, XHTMLStream::EscapeSettings e) -{ - odocstringstream d; - docstring::const_iterator it = str.begin(); - docstring::const_iterator en = str.end(); - for (; it != en; ++it) - d << escapeChar(*it, e); - return d.str(); -} - - -string escapeChar(char c, XHTMLStream::EscapeSettings e) -{ - string str; - switch (e) { - case XHTMLStream::ESCAPE_NONE: - str += c; - break; - case XHTMLStream::ESCAPE_ALL: - if (c == '<') { - str += "<"; - break; - } else if (c == '>') { - str += ">"; - break; - } - // fall through - case XHTMLStream::ESCAPE_AND: - if (c == '&') - str += "&"; - else - str +=c ; - break; - } - return str; -} - - -// escape what needs escaping -string htmlize(string const & str, XHTMLStream::EscapeSettings e) -{ - ostringstream d; - string::const_iterator it = str.begin(); - string::const_iterator en = str.end(); - for (; it != en; ++it) - d << escapeChar(*it, e); - return d.str(); -} - - -string cleanAttr(string const & str) -{ - string newname; - string::const_iterator it = str.begin(); - string::const_iterator en = str.end(); - for (; it != en; ++it) - newname += isAlnumASCII(*it) ? *it : '_'; - return newname; -} - - -docstring cleanAttr(docstring const & str) -{ - docstring newname; - docstring::const_iterator it = str.begin(); - docstring::const_iterator en = str.end(); - for (; it != en; ++it) { - char_type const c = *it; - newname += isAlnumASCII(c) ? c : char_type('_'); - } - return newname; -} - - -docstring StartTag::writeTag() const -{ - string output = "<" + tag_; - if (!attr_.empty()) - output += " " + html::htmlize(attr_, XHTMLStream::ESCAPE_NONE); - output += ">"; - return from_utf8(output); -} - - -docstring StartTag::writeEndTag() const -{ - string output = ""; - return from_utf8(output); -} - - -bool StartTag::operator==(FontTag const & rhs) const -{ - return rhs == *this; -} - - -docstring EndTag::writeEndTag() const -{ - string output = ""; - return from_utf8(output); -} - - -docstring ParTag::writeTag() const -{ - docstring output = StartTag::writeTag(); - - if (parid_.empty()) - return output; - - string const pattr = "id='" + parid_ + "'"; - output += html::CompTag("a", pattr).writeTag(); - return output; -} - -docstring CompTag::writeTag() const +docstring fontToHtmlTag(xml::FontTypes type) +{ + switch(type) { + case xml::FontTypes::FT_EMPH: + return from_utf8("em"); + case xml::FontTypes::FT_BOLD: + return from_utf8("b"); + case xml::FontTypes::FT_NOUN: + return from_utf8("dfn"); + case xml::FontTypes::FT_UBAR: + case xml::FontTypes::FT_WAVE: + case xml::FontTypes::FT_DBAR: + return from_utf8("u"); + case xml::FontTypes::FT_SOUT: + case xml::FontTypes::FT_XOUT: + return from_utf8("del"); + case xml::FontTypes::FT_ITALIC: + return from_utf8("i"); + case xml::FontTypes::FT_UPRIGHT: + case xml::FontTypes::FT_SLANTED: + case xml::FontTypes::FT_SMALLCAPS: + case xml::FontTypes::FT_ROMAN: + case xml::FontTypes::FT_SANS: + case xml::FontTypes::FT_TYPE: + case xml::FontTypes::FT_SIZE_TINY: + case xml::FontTypes::FT_SIZE_SCRIPT: + case xml::FontTypes::FT_SIZE_FOOTNOTE: + case xml::FontTypes::FT_SIZE_SMALL: + case xml::FontTypes::FT_SIZE_NORMAL: + case xml::FontTypes::FT_SIZE_LARGE: + case xml::FontTypes::FT_SIZE_LARGER: + case xml::FontTypes::FT_SIZE_LARGEST: + case xml::FontTypes::FT_SIZE_HUGE: + case xml::FontTypes::FT_SIZE_HUGER: + case xml::FontTypes::FT_SIZE_INCREASE: + case xml::FontTypes::FT_SIZE_DECREASE: + return from_utf8("span"); + } + // kill warning + return docstring(); +} + + +docstring fontToHtmlAttribute(xml::FontTypes type) { - string output = "<" + tag_; - if (!attr_.empty()) - output += " " + html::htmlize(attr_, XHTMLStream::ESCAPE_NONE); - output += " />"; - return from_utf8(output); -} - - - -namespace { - -string fontToTag(html::FontTypes type) - { - switch(type) { - case FT_EMPH: - return "em"; - case FT_BOLD: - return "b"; - case FT_NOUN: - return "dfn"; - case FT_UBAR: - case FT_WAVE: - case FT_DBAR: - return "u"; - case FT_SOUT: - return "del"; - case FT_ITALIC: - return "i"; - case FT_UPRIGHT: - case FT_SLANTED: - case FT_SMALLCAPS: - case FT_ROMAN: - case FT_SANS: - case FT_TYPE: - case FT_SIZE_TINY: - case FT_SIZE_SCRIPT: - case FT_SIZE_FOOTNOTE: - case FT_SIZE_SMALL: - case FT_SIZE_NORMAL: - case FT_SIZE_LARGE: - case FT_SIZE_LARGER: - case FT_SIZE_LARGEST: - case FT_SIZE_HUGE: - case FT_SIZE_HUGER: - case FT_SIZE_INCREASE: - case FT_SIZE_DECREASE: - return "span"; - } - // kill warning - return ""; -} - -StartTag fontToStartTag(html::FontTypes type) - { - string tag = fontToTag(type); switch(type) { - case FT_EMPH: - return html::StartTag(tag); - case FT_BOLD: - return html::StartTag(tag); - case FT_NOUN: - return html::StartTag(tag, "class='lyxnoun'"); - case FT_UBAR: - return html::StartTag(tag); - case FT_DBAR: - return html::StartTag(tag, "class='dline'"); - case FT_SOUT: - return html::StartTag(tag, "class='strikeout'"); - case FT_WAVE: - return html::StartTag(tag, "class='wline'"); - case FT_ITALIC: - return html::StartTag(tag); - case FT_UPRIGHT: - return html::StartTag(tag, "style='font-style:normal;'"); - case FT_SLANTED: - return html::StartTag(tag, "style='font-style:oblique;'"); - case FT_SMALLCAPS: - return html::StartTag(tag, "style='font-variant:small-caps;'"); - case FT_ROMAN: - return html::StartTag(tag, "style='font-family:serif;'"); - case FT_SANS: - return html::StartTag(tag, "style='font-family:sans-serif;'"); - case FT_TYPE: - return html::StartTag(tag, "style='font-family:monospace;'"); - case FT_SIZE_TINY: - case FT_SIZE_SCRIPT: - case FT_SIZE_FOOTNOTE: - return html::StartTag(tag, "style='font-size:x-small;'"); - case FT_SIZE_SMALL: - return html::StartTag(tag, "style='font-size:small;'"); - case FT_SIZE_NORMAL: - return html::StartTag(tag, "style='font-size:normal;'"); - case FT_SIZE_LARGE: - return html::StartTag(tag, "style='font-size:large;'"); - case FT_SIZE_LARGER: - case FT_SIZE_LARGEST: - return html::StartTag(tag, "style='font-size:x-large;'"); - case FT_SIZE_HUGE: - case FT_SIZE_HUGER: - return html::StartTag(tag, "style='font-size:xx-large;'"); - case FT_SIZE_INCREASE: - return html::StartTag(tag, "style='font-size:larger;'"); - case FT_SIZE_DECREASE: - return html::StartTag(tag, "style='font-size:smaller;'"); + case xml::FontTypes::FT_EMPH: + case xml::FontTypes::FT_BOLD: + return from_ascii(""); + case xml::FontTypes::FT_NOUN: + return from_ascii("class='lyxnoun'"); + case xml::FontTypes::FT_UBAR: + return from_ascii(""); + case xml::FontTypes::FT_DBAR: + return from_ascii("class='dline'"); + case xml::FontTypes::FT_XOUT: + case xml::FontTypes::FT_SOUT: + return from_ascii("class='strikeout'"); + case xml::FontTypes::FT_WAVE: + return from_ascii("class='wline'"); + case xml::FontTypes::FT_ITALIC: + return from_ascii(""); + case xml::FontTypes::FT_UPRIGHT: + return from_ascii("style='font-style:normal;'"); + case xml::FontTypes::FT_SLANTED: + return from_ascii("style='font-style:oblique;'"); + case xml::FontTypes::FT_SMALLCAPS: + return from_ascii("style='font-variant:small-caps;'"); + case xml::FontTypes::FT_ROMAN: + return from_ascii("style='font-family:serif;'"); + case xml::FontTypes::FT_SANS: + return from_ascii("style='font-family:sans-serif;'"); + case xml::FontTypes::FT_TYPE: + return from_ascii("style='font-family:monospace;'"); + case xml::FontTypes::FT_SIZE_TINY: + case xml::FontTypes::FT_SIZE_SCRIPT: + case xml::FontTypes::FT_SIZE_FOOTNOTE: + return from_ascii("style='font-size:x-small;'"); + case xml::FontTypes::FT_SIZE_SMALL: + return from_ascii("style='font-size:small;'"); + case xml::FontTypes::FT_SIZE_NORMAL: + return from_ascii("style='font-size:normal;'"); + case xml::FontTypes::FT_SIZE_LARGE: + return from_ascii("style='font-size:large;'"); + case xml::FontTypes::FT_SIZE_LARGER: + case xml::FontTypes::FT_SIZE_LARGEST: + return from_ascii("style='font-size:x-large;'"); + case xml::FontTypes::FT_SIZE_HUGE: + case xml::FontTypes::FT_SIZE_HUGER: + return from_ascii("style='font-size:xx-large;'"); + case xml::FontTypes::FT_SIZE_INCREASE: + return from_ascii("style='font-size:larger;'"); + case xml::FontTypes::FT_SIZE_DECREASE: + return from_ascii("style='font-size:smaller;'"); } // kill warning - return StartTag(""); -} - -} // end anonymous namespace - - -FontTag::FontTag(FontTypes type) - : StartTag(fontToStartTag(type)), font_type_(type) -{} - - -bool FontTag::operator==(StartTag const & tag) const -{ - FontTag const * const ftag = tag.asFontTag(); - if (!ftag) - return false; - return (font_type_ == ftag->font_type_); + return from_ascii(""); } -EndFontTag::EndFontTag(FontTypes type) - : EndTag(fontToTag(type)), font_type_(type) -{} - -} // namespace html - - - -//////////////////////////////////////////////////////////////// -/// -/// XHTMLStream -/// -//////////////////////////////////////////////////////////////// - -XHTMLStream::XHTMLStream(odocstream & os) - : os_(os), escape_(ESCAPE_ALL) -{} - - -#ifdef XHTML_DEBUG -void XHTMLStream::dumpTagStack(string const & msg) const +xml::FontTag xhtmlStartFontTag(xml::FontTypes type) { - writeError(msg + ": Tag Stack"); - TagStack::const_reverse_iterator it = tag_stack_.rbegin(); - TagStack::const_reverse_iterator en = tag_stack_.rend(); - for (; it != en; ++it) { - writeError(it->tag_); - } - writeError("Pending Tags"); - it = pending_tags_.rbegin(); - en = pending_tags_.rend(); - for (; it != en; ++it) { - writeError(it->tag_); - } - writeError("End Tag Stack"); + return xml::FontTag(fontToHtmlTag(type), fontToHtmlAttribute(type), type); } -#endif -void XHTMLStream::writeError(std::string const & s) const +xml::EndFontTag xhtmlEndFontTag(xml::FontTypes type) { - LYXERR0(s); - os_ << from_utf8("\n"); + return xml::EndFontTag(fontToHtmlTag(type), type); } - -namespace { - // an illegal tag for internal use - static html::StartTag const parsep_tag("&LyX_parsep_tag&"); -} - - -bool XHTMLStream::closeFontTags() -{ - if (isTagPending(parsep_tag)) - // we haven't had any content - return true; - - // this may be a useless check, since we ought at least to have - // the parsep_tag. but it can't hurt too much to be careful. - if (tag_stack_.empty()) - return true; - - // first, we close any open font tags we can close - TagPtr curtag = tag_stack_.back(); - while (curtag->asFontTag()) { - os_ << curtag->writeEndTag(); - tag_stack_.pop_back(); - // this shouldn't happen, since then the font tags - // weren't in any other tag. - LBUFERR(!tag_stack_.empty()); - curtag = tag_stack_.back(); - } - - if (*curtag == parsep_tag) - return true; - - // so we've hit a non-font tag. - writeError("Tags still open in closeFontTags(). Probably not a problem,\n" - "but you might want to check these tags:"); - TagDeque::const_reverse_iterator it = tag_stack_.rbegin(); - TagDeque::const_reverse_iterator const en = tag_stack_.rend(); - for (; it != en; ++it) { - if (**it == parsep_tag) - break; - writeError((*it)->tag_); - } - return false; -} - - -void XHTMLStream::startParagraph(bool keep_empty) -{ - pending_tags_.push_back(makeTagPtr(html::StartTag(parsep_tag))); - if (keep_empty) - clearTagDeque(); -} - - -void XHTMLStream::endParagraph() -{ - if (isTagPending(parsep_tag)) { - // this case is normal. it just means we didn't have content, - // so the parsep_tag never got moved onto the tag stack. - while (!pending_tags_.empty()) { - // clear all pending tags up to and including the parsep tag. - // note that we work from the back, because we want to get rid - // of everything that hasn't been used. - TagPtr const cur_tag = pending_tags_.back(); - pending_tags_.pop_back(); - if (*cur_tag == parsep_tag) - break; - } - return; - } - - if (!isTagOpen(parsep_tag)) { - writeError("No paragraph separation tag found in endParagraph()."); - return; - } - - // this case is also normal, if the parsep tag is the last one - // on the stack. otherwise, it's an error. - while (!tag_stack_.empty()) { - TagPtr const cur_tag = tag_stack_.back(); - tag_stack_.pop_back(); - if (*cur_tag == parsep_tag) - break; - writeError("Tag `" + cur_tag->tag_ + "' still open at end of paragraph. Closing."); - os_ << cur_tag->writeEndTag(); - } -} - - -void XHTMLStream::clearTagDeque() -{ - while (!pending_tags_.empty()) { - TagPtr const tag = pending_tags_.front(); - if (*tag != parsep_tag) - // tabs? - os_ << tag->writeTag(); - tag_stack_.push_back(tag); - pending_tags_.pop_front(); - } -} - - -XHTMLStream & XHTMLStream::operator<<(docstring const & d) -{ - clearTagDeque(); - os_ << html::htmlize(d, escape_); - escape_ = ESCAPE_ALL; - return *this; -} - - -XHTMLStream & XHTMLStream::operator<<(const char * s) -{ - clearTagDeque(); - docstring const d = from_ascii(s); - os_ << html::htmlize(d, escape_); - escape_ = ESCAPE_ALL; - return *this; -} - - -XHTMLStream & XHTMLStream::operator<<(char_type c) -{ - clearTagDeque(); - os_ << html::escapeChar(c, escape_); - escape_ = ESCAPE_ALL; - return *this; -} - - -XHTMLStream & XHTMLStream::operator<<(char c) -{ - clearTagDeque(); - string const d = html::escapeChar(c, escape_); - escape_ = ESCAPE_ALL; - return *this; -} - - -XHTMLStream & XHTMLStream::operator<<(int i) -{ - clearTagDeque(); - os_ << i; - escape_ = ESCAPE_ALL; - return *this; -} - - -XHTMLStream & XHTMLStream::operator<<(EscapeSettings e) -{ - escape_ = e; - return *this; -} - - -XHTMLStream & XHTMLStream::operator<<(html::StartTag const & tag) -{ - if (tag.tag_.empty()) - return *this; - pending_tags_.push_back(makeTagPtr(tag)); - if (tag.keepempty_) - clearTagDeque(); - return *this; -} - - -XHTMLStream & XHTMLStream::operator<<(html::ParTag const & tag) -{ - if (tag.tag_.empty()) - return *this; - pending_tags_.push_back(makeTagPtr(tag)); - return *this; -} - - -XHTMLStream & XHTMLStream::operator<<(html::CompTag const & tag) -{ - if (tag.tag_.empty()) - return *this; - clearTagDeque(); - os_ << tag.writeTag(); - *this << html::CR(); - return *this; -} - - -XHTMLStream & XHTMLStream::operator<<(html::FontTag const & tag) -{ - if (tag.tag_.empty()) - return *this; - pending_tags_.push_back(makeTagPtr(tag)); - return *this; -} - - -XHTMLStream & XHTMLStream::operator<<(html::CR const &) -{ - // tabs? - os_ << from_ascii("\n"); - return *this; -} - - -bool XHTMLStream::isTagOpen(html::StartTag const & stag) const -{ - TagDeque::const_iterator sit = tag_stack_.begin(); - TagDeque::const_iterator const sen = tag_stack_.end(); - for (; sit != sen; ++sit) - if (**sit == stag) - return true; - return false; -} - - -bool XHTMLStream::isTagOpen(html::EndTag const & etag) const -{ - TagDeque::const_iterator sit = tag_stack_.begin(); - TagDeque::const_iterator const sen = tag_stack_.end(); - for (; sit != sen; ++sit) - if (etag == **sit) - return true; - return false; -} - - -bool XHTMLStream::isTagPending(html::StartTag const & stag) const -{ - TagDeque::const_iterator sit = pending_tags_.begin(); - TagDeque::const_iterator const sen = pending_tags_.end(); - for (; sit != sen; ++sit) - if (**sit == stag) - return true; - return false; -} - - -// this is complicated, because we want to make sure that -// everything is properly nested. the code ought to make -// sure of that, but we won't assert (yet) if we run into -// a problem. we'll just output error messages and try our -// best to make things work. -XHTMLStream & XHTMLStream::operator<<(html::EndTag const & etag) -{ - if (etag.tag_.empty()) - return *this; - - // if this tag is pending, we can simply discard it. - if (!pending_tags_.empty()) { - - if (etag == *pending_tags_.back()) { - // we have , so we discard it and remove it - // from the pending_tags_. - pending_tags_.pop_back(); - return *this; - } - - // there is a pending tag that isn't the one we are trying - // to close. - - // is this tag itself pending? - // non-const iterators because we may call erase(). - TagDeque::iterator dit = pending_tags_.begin(); - TagDeque::iterator const den = pending_tags_.end(); - for (; dit != den; ++dit) { - if (etag == **dit) { - // it was pending, so we just erase it - writeError("Tried to close pending tag `" + etag.tag_ - + "' when other tags were pending. Last pending tag is `" - + to_utf8(pending_tags_.back()->writeTag()) - + "'. Tag discarded."); - pending_tags_.erase(dit); - return *this; - } - } - // so etag isn't itself pending. is it even open? - if (!isTagOpen(etag)) { - writeError("Tried to close `" + etag.tag_ - + "' when tag was not open. Tag discarded."); - return *this; - } - // ok, so etag is open. - // our strategy will be as below: we will do what we need to - // do to close this tag. - string estr = "Closing tag `" + etag.tag_ - + "' when other tags are pending. Discarded pending tags:\n"; - for (dit = pending_tags_.begin(); dit != den; ++dit) - estr += to_utf8(html::htmlize((*dit)->writeTag(), XHTMLStream::ESCAPE_ALL)) + "\n"; - writeError(estr); - // clear the pending tags... - pending_tags_.clear(); - // ...and then just fall through. - } - - // make sure there are tags to be closed - if (tag_stack_.empty()) { - writeError("Tried to close `" + etag.tag_ - + "' when no tags were open!"); - return *this; - } - - // is the tag we are closing the last one we opened? - if (etag == *tag_stack_.back()) { - // output it... - os_ << etag.writeEndTag(); - // ...and forget about it - tag_stack_.pop_back(); - return *this; - } - - // we are trying to close a tag other than the one last opened. - // let's first see if this particular tag is still open somehow. - if (!isTagOpen(etag)) { - writeError("Tried to close `" + etag.tag_ - + "' when tag was not open. Tag discarded."); - return *this; - } - - // so the tag was opened, but other tags have been opened since - // and not yet closed. - // if it's a font tag, though... - if (etag.asFontTag()) { - // it won't be a problem if the other tags open since this one - // are also font tags. - TagDeque::const_reverse_iterator rit = tag_stack_.rbegin(); - TagDeque::const_reverse_iterator ren = tag_stack_.rend(); - for (; rit != ren; ++rit) { - if (etag == **rit) - break; - if (!(*rit)->asFontTag()) { - // we'll just leave it and, presumably, have to close it later. - writeError("Unable to close font tag `" + etag.tag_ - + "' due to open non-font tag `" + (*rit)->tag_ + "'."); - return *this; - } - } - - // so we have e.g.: - // this is bold - // and are being asked to closed em. we want: - // this is bold - // first, we close the intervening tags... - TagPtr curtag = tag_stack_.back(); - // ...remembering them in a stack. - TagDeque fontstack; - while (etag != *curtag) { - os_ << curtag->writeEndTag(); - fontstack.push_back(curtag); - tag_stack_.pop_back(); - curtag = tag_stack_.back(); - } - os_ << etag.writeEndTag(); - tag_stack_.pop_back(); - - // ...and restore the other tags. - rit = fontstack.rbegin(); - ren = fontstack.rend(); - for (; rit != ren; ++rit) - pending_tags_.push_back(*rit); - return *this; - } - - // it wasn't a font tag. - // so other tags were opened before this one and not properly closed. - // so we'll close them, too. that may cause other issues later, but it - // at least guarantees proper nesting. - writeError("Closing tag `" + etag.tag_ - + "' when other tags are open, namely:"); - TagPtr curtag = tag_stack_.back(); - while (etag != *curtag) { - writeError(curtag->tag_); - if (*curtag != parsep_tag) - os_ << curtag->writeEndTag(); - tag_stack_.pop_back(); - curtag = tag_stack_.back(); - } - // curtag is now the one we actually want. - os_ << curtag->writeEndTag(); - tag_stack_.pop_back(); - - return *this; -} - -// End code for XHTMLStream - namespace { // convenience functions -inline void openParTag(XHTMLStream & xs, Layout const & lay, - std::string parlabel) +inline void openParTag(XMLStream & xs, Layout const & lay, + const std::string & parlabel) { - xs << html::ParTag(lay.htmltag(), lay.htmlattr(), parlabel); + string attrs = lay.htmlattr(); + if (!parlabel.empty()) + attrs += " id='" + parlabel + "'"; + xs << xml::ParTag(lay.htmltag(), attrs); } -void openParTag(XHTMLStream & xs, Layout const & lay, +void openParTag(XMLStream & xs, Layout const & lay, ParagraphParameters const & params, - std::string parlabel) + const std::string & parlabel) { // FIXME Are there other things we should handle here? string const align = alignmentToCSS(params.align()); @@ -770,35 +176,37 @@ void openParTag(XHTMLStream & xs, Layout const & lay, return; } string attrs = lay.htmlattr() + " style='text-align: " + align + ";'"; - xs << html::ParTag(lay.htmltag(), attrs, parlabel); + if (!parlabel.empty()) + attrs += " id='" + parlabel + "'"; + xs << xml::ParTag(lay.htmltag(), attrs); } -inline void closeTag(XHTMLStream & xs, Layout const & lay) +inline void closeTag(XMLStream & xs, Layout const & lay) { - xs << html::EndTag(lay.htmltag()); + xs << xml::EndTag(lay.htmltag()); } -inline void openLabelTag(XHTMLStream & xs, Layout const & lay) +inline void openLabelTag(XMLStream & xs, Layout const & lay) { - xs << html::StartTag(lay.htmllabeltag(), lay.htmllabelattr()); + xs << xml::StartTag(lay.htmllabeltag(), lay.htmllabelattr()); } -inline void closeLabelTag(XHTMLStream & xs, Layout const & lay) +inline void closeLabelTag(XMLStream & xs, Layout const & lay) { - xs << html::EndTag(lay.htmllabeltag()); + xs << xml::EndTag(lay.htmllabeltag()); } -inline void openItemTag(XHTMLStream & xs, Layout const & lay) +inline void openItemTag(XMLStream & xs, Layout const & lay) { - xs << html::StartTag(lay.htmlitemtag(), lay.htmlitemattr(), true); + xs << xml::StartTag(lay.htmlitemtag(), lay.htmlitemattr(), true); } -void openItemTag(XHTMLStream & xs, Layout const & lay, +void openItemTag(XMLStream & xs, Layout const & lay, ParagraphParameters const & params) { // FIXME Are there other things we should handle here? @@ -808,13 +216,13 @@ void openItemTag(XHTMLStream & xs, Layout const & lay, return; } string attrs = lay.htmlattr() + " style='text-align: " + align + ";'"; - xs << html::StartTag(lay.htmlitemtag(), attrs); + xs << xml::StartTag(lay.htmlitemtag(), attrs); } -inline void closeItemTag(XHTMLStream & xs, Layout const & lay) +inline void closeItemTag(XMLStream & xs, Layout const & lay) { - xs << html::EndTag(lay.htmlitemtag()); + xs << xml::EndTag(lay.htmlitemtag()); } // end of convenience functions @@ -831,7 +239,7 @@ ParagraphList::const_iterator findLastParagraph( ParagraphList::const_iterator findEndOfEnvironment( - ParagraphList::const_iterator const pstart, + ParagraphList::const_iterator const & pstart, ParagraphList::const_iterator const & pend) { ParagraphList::const_iterator p = pstart; @@ -855,7 +263,7 @@ ParagraphList::const_iterator findEndOfEnvironment( // FIXME I am not sure about the first check. // Surely we *could* have different layouts that count as - // LATEX_PARAGRAPH, right? + // LATEX_PARAGRAPH, right? if (style.latextype == LATEX_PARAGRAPH || style != bstyle) return p; } @@ -864,7 +272,7 @@ ParagraphList::const_iterator findEndOfEnvironment( ParagraphList::const_iterator makeParagraphs(Buffer const & buf, - XHTMLStream & xs, + XMLStream & xs, OutputParams const & runparams, Text const & text, ParagraphList::const_iterator const & pbegin, @@ -877,42 +285,72 @@ ParagraphList::const_iterator makeParagraphs(Buffer const & buf, if (!lay.counter.empty()) buf.masterBuffer()->params(). documentClass().counters().step(lay.counter, OutputUpdate); + // FIXME We should see if there's a label to be output and // do something with it. if (par != pbegin) - xs << html::CR(); - - // If we are already in a paragraph, and this is the first one, then we - // do not want to open the paragraph tag. - // we also do not want to open it if the current layout does not permit - // multiple paragraphs. - bool const opened = runparams.html_make_pars && - (par != pbegin || !runparams.html_in_par); - bool const make_parid = !runparams.for_toc && runparams.html_make_pars; - - if (opened) - openParTag(xs, lay, par->params(), - make_parid ? par->magicLabel() : ""); + xs << xml::CR(); + + // We want to open the paragraph tag if: + // (i) the current layout permits multiple paragraphs + // (ii) we are either not already inside a paragraph (HTMLIsBlock) OR + // we are, but this is not the first paragraph + // + // But there is also a special case, and we first see whether we are in it. + // We do not want to open the paragraph tag if this paragraph contains + // only one item, and that item is "inline", i.e., not HTMLIsBlock (such + // as a branch). On the other hand, if that single item has a font change + // applied to it, then we still do need to open the paragraph. + // + // Obviously, this is very fragile. The main reason we need to do this is + // because of branches, e.g., a branch that contains an entire new section. + // We do not really want to wrap that whole thing in a
...
. + bool special_case = false; + Inset const * specinset = par->size() == 1 ? par->getInset(0) : nullptr; + if (specinset && !specinset->getLayout().htmlisblock()) { + Layout const & style = par->layout(); + FontInfo const first_font = style.labeltype == LABEL_MANUAL ? + style.labelfont : style.font; + FontInfo const our_font = + par->getFont(buf.masterBuffer()->params(), 0, + text.outerFont(distance(begin, par))).fontInfo(); + if (first_font == our_font) + special_case = true; + } - docstring const deferred = - par->simpleLyXHTMLOnePar(buf, xs, runparams, text.outerFont(distance(begin, par))); + bool const open_par = runparams.html_make_pars + && (!runparams.html_in_par || par != pbegin) + && !special_case; // We want to issue the closing tag if either: // (i) We opened it, and either html_in_par is false, // or we're not in the last paragraph, anyway. - // (ii) We didn't open it and html_in_par is true, + // (ii) We didn't open it and html_in_par is true, // but we are in the first par, and there is a next par. ParagraphList::const_iterator nextpar = par; ++nextpar; - bool const needclose = - (opened && (!runparams.html_in_par || nextpar != pend)) - || (!opened && runparams.html_in_par && par == pbegin && nextpar != pend); - if (needclose) { + bool const close_par = + (open_par && (!runparams.html_in_par || nextpar != pend)) + || (!open_par && runparams.html_in_par && par == pbegin && nextpar != pend); + + if (open_par) { + // We do not issue the paragraph id if we are doing + // this for the TOC (or some similar purpose) + openParTag(xs, lay, par->params(), + runparams.for_toc ? "" : par->magicLabel()); + } + + docstring const deferred = par->simpleLyXHTMLOnePar(buf, xs, + runparams, text.outerFont(distance(begin, par)), + open_par, close_par); + + if (close_par) { closeTag(xs, lay); - xs << html::CR(); + xs << xml::CR(); } + if (!deferred.empty()) { - xs << XHTMLStream::ESCAPE_NONE << deferred << html::CR(); + xs << XMLStream::ESCAPE_NONE << deferred << xml::CR(); } } return pend; @@ -920,22 +358,22 @@ ParagraphList::const_iterator makeParagraphs(Buffer const & buf, ParagraphList::const_iterator makeBibliography(Buffer const & buf, - XHTMLStream & xs, + XMLStream & xs, OutputParams const & runparams, Text const & text, ParagraphList::const_iterator const & pbegin, - ParagraphList::const_iterator const & pend) + ParagraphList::const_iterator const & pend) { // FIXME XHTML // Use TextClass::htmlTOCLayout() to figure out how we should look. - xs << html::StartTag("h2", "class='bibliography'") + xs << xml::StartTag("h2", "class='bibliography'") << pbegin->layout().labelstring(false) - << html::EndTag("h2") - << html::CR() - << html::StartTag("div", "class='bibliography'") - << html::CR(); + << xml::EndTag("h2") + << xml::CR() + << xml::StartTag("div", "class='bibliography'") + << xml::CR(); makeParagraphs(buf, xs, runparams, text, pbegin, pend); - xs << html::EndTag("div"); + xs << xml::EndTag("div"); return pend; } @@ -946,13 +384,13 @@ bool isNormalEnv(Layout const & lay) || lay.latextype == LATEX_BIB_ENVIRONMENT; } - + ParagraphList::const_iterator makeEnvironment(Buffer const & buf, - XHTMLStream & xs, + XMLStream & xs, OutputParams const & runparams, Text const & text, ParagraphList::const_iterator const & pbegin, - ParagraphList::const_iterator const & pend) + ParagraphList::const_iterator const & pend) { ParagraphList::const_iterator const begin = text.paragraphs().begin(); ParagraphList::const_iterator par = pbegin; @@ -961,10 +399,10 @@ ParagraphList::const_iterator makeEnvironment(Buffer const & buf, // open tag for this environment openParTag(xs, bstyle, pbegin->magicLabel()); - xs << html::CR(); + xs << xml::CR(); // we will on occasion need to remember a layout from before. - Layout const * lastlay = 0; + Layout const * lastlay = nullptr; while (par != pend) { Layout const & style = par->layout(); @@ -976,62 +414,62 @@ ParagraphList::const_iterator makeEnvironment(Buffer const & buf, // "ii", etc, as with enum. Counters & cnts = buf.masterBuffer()->params().documentClass().counters(); docstring const & cntr = style.counter; - if (!style.counter.empty() - && (par == pbegin || !isNormalEnv(style)) + if (!style.counter.empty() + && (par == pbegin || !isNormalEnv(style)) && cnts.hasCounter(cntr) ) cnts.step(cntr, OutputUpdate); ParagraphList::const_iterator send; - // this will be positive, if we want to skip the initial word - // (if it's been taken for the label). - pos_type sep = 0; switch (style.latextype) { case LATEX_ENVIRONMENT: case LATEX_LIST_ENVIRONMENT: case LATEX_ITEM_ENVIRONMENT: { - // There are two possiblities in this case. - // One is that we are still in the environment in which we + // There are two possibilities in this case. + // One is that we are still in the environment in which we // started---which we will be if the depth is the same. if (par->params().depth() == origdepth) { LATTEST(bstyle == style); - if (lastlay != 0) { + if (lastlay != nullptr) { closeItemTag(xs, *lastlay); - lastlay = 0; + lastlay = nullptr; } - + + // this will be positive, if we want to skip the + // initial word (if it's been taken for the label). + pos_type sep = 0; bool const labelfirst = style.htmllabelfirst(); if (!labelfirst) openItemTag(xs, style, par->params()); - + // label output - if (style.labeltype != LABEL_NO_LABEL && + if (style.labeltype != LABEL_NO_LABEL && style.htmllabeltag() != "NONE") { if (isNormalEnv(style)) { - // in this case, we print the label only for the first + // in this case, we print the label only for the first // paragraph (as in a theorem). if (par == pbegin) { - docstring const lbl = + docstring const lbl = pbegin->params().labelString(); if (!lbl.empty()) { openLabelTag(xs, style); xs << lbl; closeLabelTag(xs, style); } - xs << html::CR(); + xs << xml::CR(); } - } else { // some kind of list + } else { // some kind of list if (style.labeltype == LABEL_MANUAL) { openLabelTag(xs, style); sep = par->firstWordLyXHTML(xs, runparams); closeLabelTag(xs, style); - xs << html::CR(); + xs << xml::CR(); } else { openLabelTag(xs, style); xs << par->params().labelString(); closeLabelTag(xs, style); - xs << html::CR(); + xs << xml::CR(); } } } // end label output @@ -1039,13 +477,14 @@ ParagraphList::const_iterator makeEnvironment(Buffer const & buf, if (labelfirst) openItemTag(xs, style, par->params()); - par->simpleLyXHTMLOnePar(buf, xs, runparams, - text.outerFont(distance(begin, par)), sep); + docstring deferred = par->simpleLyXHTMLOnePar(buf, xs, runparams, + text.outerFont(distance(begin, par)), true, true, sep); + xs << XMLStream::ESCAPE_NONE << deferred; ++par; // We may not want to close the tag yet, in particular: // If we're not at the end... - if (par != pend + if (par != pend // and are doing items... && !isNormalEnv(style) // and if the depth has changed... @@ -1055,7 +494,7 @@ ParagraphList::const_iterator makeEnvironment(Buffer const & buf, lastlay = &style; } else closeItemTag(xs, style); - xs << html::CR(); + xs << xml::CR(); } // The other possibility is that the depth has increased, in which // case we need to recurse. @@ -1082,16 +521,16 @@ ParagraphList::const_iterator makeEnvironment(Buffer const & buf, } } - if (lastlay != 0) + if (lastlay != nullptr) closeItemTag(xs, *lastlay); closeTag(xs, bstyle); - xs << html::CR(); + xs << xml::CR(); return pend; } void makeCommand(Buffer const & buf, - XHTMLStream & xs, + XMLStream & xs, OutputParams const & runparams, Text const & text, ParagraphList::const_iterator const & pbegin) @@ -1121,7 +560,7 @@ void makeCommand(Buffer const & buf, pbegin->simpleLyXHTMLOnePar(buf, xs, runparams, text.outerFont(distance(begin, pbegin))); closeTag(xs, style); - xs << html::CR(); + xs << xml::CR(); } } // end anonymous namespace @@ -1129,7 +568,7 @@ void makeCommand(Buffer const & buf, void xhtmlParagraphs(Text const & text, Buffer const & buf, - XHTMLStream & xs, + XMLStream & xs, OutputParams const & runparams) { ParagraphList const & paragraphs = text.paragraphs(); @@ -1140,14 +579,16 @@ void xhtmlParagraphs(Text const & text, pit_type bpit = runparams.par_begin; pit_type const epit = runparams.par_end; LASSERT(bpit < epit, - { xs << XHTMLStream::ESCAPE_NONE << "\n"; return; }); + { xs << XMLStream::ESCAPE_NONE << "\n"; return; }); OutputParams ourparams = runparams; ParagraphList::const_iterator const pend = (epit == (int) paragraphs.size()) ? - paragraphs.end() : paragraphs.constIterator(epit); + paragraphs.end() : paragraphs.iterator_at(epit); + std::stack headerLevels; + while (bpit < epit) { - ParagraphList::const_iterator par = paragraphs.constIterator(bpit); + ParagraphList::const_iterator par = paragraphs.iterator_at(bpit); if (par->params().startOfAppendix()) { // We want to reset the counter corresponding to toplevel sectioning Layout const & lay = @@ -1163,11 +604,40 @@ void xhtmlParagraphs(Text const & text, ParagraphList::const_iterator const lastpar = par; ParagraphList::const_iterator send; + // Think about adding
and/or
s. + // Document title is not in Sectioning, but rather in FrontMatter, so that it does not need to be taken + // into account. + if (style.category() == from_utf8("Sectioning")) { + int level = style.toclevel; + + // Need to close a previous section if it has the same level or a higher one (close
if opening a + //

after a

,

,

,

or
). More examples: + // - current: h2; back: h1; do not close any
+ // - current: h1; back: h2; close two
(first the

, then the

, so a new

can come) + while (!headerLevels.empty() && level <= headerLevels.top()) { + // Output the tag only if it corresponds to a legit section. + int stackLevel = headerLevels.top(); + if (stackLevel != Layout::NOT_IN_TOC) { + xs << xml::EndTag("section"); + xs << xml::CR(); + } + headerLevels.pop(); + } + + // Open the new section: first push it onto the stack, then output it in XHTML. + headerLevels.push(level); + // Some sectioning-like elements should not be output (such as FrontMatter). + if (level != Layout::NOT_IN_TOC ) { + xs << xml::StartTag("section"); + xs << xml::CR(); + } + } + switch (style.latextype) { case LATEX_COMMAND: { // The files with which we are working never have more than // one paragraph in a command structure. - // FIXME + // FIXME // if (ourparams.html_in_par) // fix it so we don't get sections inside standard, e.g. // note that we may then need to make runparams not const, so we @@ -1199,6 +669,14 @@ void xhtmlParagraphs(Text const & text, } bpit += distance(lastpar, par); } + + // If need be, close
s, but only at the end of the document (otherwise, dealt with at the beginning + // of the loop). + while (!headerLevels.empty() && headerLevels.top() != Layout::NOT_IN_TOC) { + headerLevels.pop(); + xs << xml::EndTag("section"); + xs << xml::CR(); + } }