]> git.lyx.org Git - lyx.git/blobdiff - src/output_xhtml.cpp
Update my email and status.
[lyx.git] / src / output_xhtml.cpp
index 1c822a9bf529d4164462ff0e8d752ef4b74ea812..50ba0f4b0b96774166636d48e4ed5696f6cb73e6 100644 (file)
@@ -32,6 +32,7 @@
 #include "support/debug.h"
 #include "support/lassert.h"
 #include "support/lstrings.h"
+#include "support/textutils.h"
 
 #include <vector>
 
@@ -42,72 +43,80 @@ namespace lyx {
 
 namespace html {
 
-docstring escapeChar(char_type c)
+docstring escapeChar(char_type c, XHTMLStream::EscapeSettings e)
 {
        docstring str;
-       switch (c) {
-       case ' ':
-               str += " ";
-               break;
-       case '&':
-               str += "&amp;";
-               break;
-       case '<':
-               str += "&lt;";
-               break;
-       case '>':
-               str += "&gt;";
-               break;
-       default:
+       switch (e) {
+       case XHTMLStream::ESCAPE_NONE:
                str += c;
                break;
+       case XHTMLStream::ESCAPE_ALL:
+               if (c == '<') {
+                       str += "&lt;";
+                       break;
+               } else if (c == '>') {
+                       str += "&gt;";
+                       break;
+               }
+       // fall through
+       case XHTMLStream::ESCAPE_AND:
+               if (c == '&')
+                       str += "&amp;";
+               else
+                       str     +=c ;
+               break;
        }
        return str;
 }
 
 
 // escape what needs escaping
-docstring htmlize(docstring const & str) {
+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);
+               d << escapeChar(*it, e);
        return d.str();
 }
 
 
-string escapeChar(char c)
+string escapeChar(char c, XHTMLStream::EscapeSettings e)
 {
        string str;
-       switch (c) {
-       case ' ':
-               str += " ";
-               break;
-       case '&':
-               str += "&amp;";
-               break;
-       case '<':
-               str += "&lt;";
-               break;
-       case '>':
-               str += "&gt;";
-               break;
-       default:
+       switch (e) {
+       case XHTMLStream::ESCAPE_NONE:
                str += c;
                break;
+       case XHTMLStream::ESCAPE_ALL:
+               if (c == '<') {
+                       str += "&lt;";
+                       break;
+               } else if (c == '>') {
+                       str += "&gt;";
+                       break;
+               }
+       // fall through
+       case XHTMLStream::ESCAPE_AND:
+               if (c == '&')
+                       str += "&amp;";
+               else
+                       str     +=c ;
+               break;
        }
        return str;
 }
 
 
 // escape what needs escaping
-string htmlize(string const & str) {
+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);
+               d << escapeChar(*it, e);
        return d.str();
 }
 
@@ -118,7 +127,7 @@ string cleanAttr(string const & str)
        string::const_iterator it = str.begin();
        string::const_iterator en = str.end();
        for (; it != en; ++it)
-               newname += isalnum(*it) ? *it : '_';
+               newname += isAlnumASCII(*it) ? *it : '_';
        return newname; 
 }
 
@@ -128,18 +137,18 @@ docstring cleanAttr(docstring const & str)
        docstring newname;
        docstring::const_iterator it = str.begin();
        docstring::const_iterator en = str.end();
-       for (; it != en; ++it)
-               if (isalnum(*it))
-                       newname += *it;
-               else
-                       newname += '_';
+       for (; it != en; ++it) {
+               char_type const c = *it;
+               newname += isAlnumASCII(c) ? c : char_type('_');
+       }
        return newname; 
 }
 
 
 bool isFontTag(string const & s)
 {
-       return s == "em" || s == "strong"; // others?
+       // others?
+       return s == "em" || s == "strong" || s == "i" || s == "b";
 }
 
 
@@ -147,7 +156,7 @@ docstring StartTag::asTag() const
 {
        string output = "<" + tag_;
        if (!attr_.empty())
-               output += " " + html::htmlize(attr_);
+               output += " " + html::htmlize(attr_, XHTMLStream::ESCAPE_NONE);
        output += ">";
        return from_utf8(output);
 }
@@ -171,7 +180,7 @@ docstring CompTag::asTag() const
 {
        string output = "<" + tag_;
        if (!attr_.empty())
-               output += " " + html::htmlize(attr_);
+               output += " " + html::htmlize(attr_, XHTMLStream::ESCAPE_NONE);
        output += " />";
        return from_utf8(output);
 }
@@ -187,28 +196,54 @@ docstring CompTag::asTag() const
 ////////////////////////////////////////////////////////////////
 
 XHTMLStream::XHTMLStream(odocstream & os) 
-               : os_(os), nextraw_(false)
+               : os_(os), escape_(ESCAPE_ALL)
 {}
 
 
-void XHTMLStream::cr() 
+#if 0
+void XHTMLStream::dumpTagStack(string const & msg) const
 {
-       // tabs?
-       os_ << from_ascii("\n");
+       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");
 }
+#endif
 
 
-void XHTMLStream::writeError(std::string const & s)
+void XHTMLStream::writeError(std::string const & s) const
 {
        LYXERR0(s);
-       os_ << from_utf8("<!-- Output Error: " + s + " -->");
+       os_ << from_utf8("<!-- Output Error: " + s + " -->\n");
+}
+
+
+namespace {
+       // an illegal tag for internal use
+       static string 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
        html::StartTag curtag = tag_stack_.back();
        while (html::isFontTag(curtag.tag_)) {
@@ -221,19 +256,67 @@ bool XHTMLStream::closeFontTags()
                        return true;
                curtag = tag_stack_.back();
        }
-       // so we've hit a non-font tag. let's see if any of the
-       // remaining tags are font tags.
-       TagStack::const_iterator it = tag_stack_.begin();
-       TagStack::const_iterator en = tag_stack_.end();
-       bool noFontTags = true;
+       
+       if (curtag.tag_ == 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:");
+       TagStack::const_reverse_iterator it = tag_stack_.rbegin();
+       TagStack::const_reverse_iterator const en = tag_stack_.rend();
        for (; it != en; ++it) {
-               if (html::isFontTag(it->tag_)) {
-                       writeError("Font tag `" + it->tag_ + "' still open in closeFontTags().\n"
-                               "This is likely not a problem, but you might want to check.");
-                       noFontTags = false;
+               string const tagname = it->tag_;
+               if (tagname == parsep_tag)
+                       break;
+               writeError(it->tag_);
+       }
+       return false;
+}
+
+
+void XHTMLStream::startParagraph(bool keep_empty)
+{
+       pending_tags_.push_back(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.
+                       html::StartTag const cur_tag = pending_tags_.back();
+                       string const & tag = cur_tag.tag_;
+                       pending_tags_.pop_back();
+                       if (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()) {
+               html::StartTag const cur_tag = tag_stack_.back();
+               string const & tag = cur_tag.tag_;
+               tag_stack_.pop_back();
+               if (tag == parsep_tag)
+                       break;
+               writeError("Tag `" + tag + "' still open at end of paragraph. Closing.");
+               os_ << cur_tag.asEndTag();
        }
-       return noFontTags;
 }
 
 
@@ -241,8 +324,9 @@ void XHTMLStream::clearTagDeque()
 {
        while (!pending_tags_.empty()) {
                html::StartTag const & tag = pending_tags_.front();
-               // tabs?
-               os_ << tag.asTag();
+               if (tag.tag_ != parsep_tag)
+                       // tabs?
+                       os_ << tag.asTag();
                tag_stack_.push_back(tag);
                pending_tags_.pop_front();
        }
@@ -252,11 +336,8 @@ void XHTMLStream::clearTagDeque()
 XHTMLStream & XHTMLStream::operator<<(docstring const & d)
 {
        clearTagDeque();
-       if (nextraw_) {
-               os_ << d;
-               nextraw_ = false;
-       } else
-               os_ << html::htmlize(d);
+       os_ << html::htmlize(d, escape_);
+       escape_ = ESCAPE_ALL;
        return *this;
 }
 
@@ -265,11 +346,8 @@ XHTMLStream & XHTMLStream::operator<<(const char * s)
 {
        clearTagDeque();
        docstring const d = from_ascii(s);
-       if (nextraw_) {
-               os_ << d;
-               nextraw_ = false;
-       } else
-               os_ << html::htmlize(d);
+       os_ << html::htmlize(d, escape_);
+       escape_ = ESCAPE_ALL;
        return *this;
 }
 
@@ -277,11 +355,8 @@ XHTMLStream & XHTMLStream::operator<<(const char * s)
 XHTMLStream & XHTMLStream::operator<<(char_type c)
 {
        clearTagDeque();
-       if (nextraw_) {
-               os_ << c;
-               nextraw_ = false;
-       } else
-               os_ << html::escapeChar(c);
+       os_ << html::escapeChar(c, escape_);
+       escape_ = ESCAPE_ALL;
        return *this;
 }
 
@@ -289,11 +364,8 @@ XHTMLStream & XHTMLStream::operator<<(char_type c)
 XHTMLStream & XHTMLStream::operator<<(char c)
 {
        clearTagDeque();
-       if (nextraw_) {
-               os_ << c;
-               nextraw_ = false;
-       } else
-               os_ << html::escapeChar(c);
+       string const d = html::escapeChar(c, escape_);
+       escape_ = ESCAPE_ALL;
        return *this;
 }
 
@@ -302,14 +374,14 @@ XHTMLStream & XHTMLStream::operator<<(int i)
 {
        clearTagDeque();
        os_ << i;
-       nextraw_ = false;
+       escape_ = ESCAPE_ALL;
        return *this;
 }
 
 
-XHTMLStream & XHTMLStream::operator<<(NextRaw const &) 
+XHTMLStream & XHTMLStream::operator<<(EscapeSettings e)
 { 
-       nextraw_ = true; 
+       escape_ = e;
        return *this;
 }
 
@@ -332,12 +404,20 @@ XHTMLStream & XHTMLStream::operator<<(html::CompTag const & tag)
        clearTagDeque();
        // tabs?
        os_ << tag.asTag();
-       cr();
+       *this << html::CR();
+       return *this;
+}
+
+
+XHTMLStream & XHTMLStream::operator<<(html::CR const &)
+{
+       // tabs?
+       os_ << from_ascii("\n");
        return *this;
 }
 
 
-bool XHTMLStream::isTagOpen(string const & stag)
+bool XHTMLStream::isTagOpen(string const & stag) const
 {
        TagStack::const_iterator sit = tag_stack_.begin();
        TagStack::const_iterator const sen = tag_stack_.end();
@@ -348,6 +428,17 @@ bool XHTMLStream::isTagOpen(string const & stag)
 }
 
 
+bool XHTMLStream::isTagPending(string const & stag) const
+{
+       TagStack::const_iterator sit = pending_tags_.begin();
+       TagStack::const_iterator const sen = pending_tags_.end();
+       for (; sit != sen; ++sit)
+               if (sit->tag_ == 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
@@ -378,8 +469,8 @@ XHTMLStream & XHTMLStream::operator<<(html::EndTag const & etag)
                // 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();
+               TagStack::iterator dit = pending_tags_.begin();
+               TagStack::iterator const den = pending_tags_.end();
                for (; dit != den; ++dit) {
                        if (dit->tag_ == etag.tag_) {
                                // it was pending, so we just erase it
@@ -480,7 +571,8 @@ XHTMLStream & XHTMLStream::operator<<(html::EndTag const & etag)
        html::StartTag curtag = tag_stack_.back();
        while (curtag.tag_ != etag.tag_) {
                writeError(curtag.tag_);
-               os_ << curtag.asEndTag();
+               if (curtag.tag_ != parsep_tag)
+                       os_ << curtag.asEndTag();
                tag_stack_.pop_back();
                curtag = tag_stack_.back();
        }
@@ -503,7 +595,8 @@ inline void openTag(XHTMLStream & xs, Layout const & lay)
 }
 
 
-void openTag(XHTMLStream & xs, Layout const & lay, ParagraphParameters const & params)
+void openTag(XHTMLStream & xs, Layout const & lay, 
+             ParagraphParameters const & params)
 {
        // FIXME Are there other things we should handle here?
        string const align = alignmentToCSS(params.align());
@@ -540,6 +633,20 @@ inline void openItemTag(XHTMLStream & xs, Layout const & lay)
 }
 
 
+void openItemTag(XHTMLStream & xs, Layout const & lay, 
+             ParagraphParameters const & params)
+{
+       // FIXME Are there other things we should handle here?
+       string const align = alignmentToCSS(params.align());
+       if (align.empty()) {
+               openItemTag(xs, lay);
+               return;
+       }
+       string attrs = lay.htmlattr() + " style='text-align: " + align + ";'";
+       xs << html::StartTag(lay.htmlitemtag(), attrs);
+}
+
+
 inline void closeItemTag(XHTMLStream & xs, Layout const & lay)
 {
        xs << html::EndTag(lay.htmlitemtag());
@@ -603,7 +710,7 @@ ParagraphList::const_iterator makeParagraphs(Buffer const & buf,
                // FIXME We should see if there's a label to be output and
                // do something with it.
                if (par != pbegin)
-                       xs.cr();
+                       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.
@@ -622,17 +729,16 @@ ParagraphList::const_iterator makeParagraphs(Buffer const & buf,
                //   (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++;
+               ++nextpar;
                bool const needclose = 
                        (opened && (!runparams.html_in_par || nextpar != pend))
                        || (!opened && runparams.html_in_par && par == pbegin && nextpar != pend);
                if (needclose) {
                        closeTag(xs, lay);
-                       xs.cr();
+                       xs << html::CR();
                }
                if (!deferred.empty()) {
-                       xs << XHTMLStream::NextRaw() << deferred;
-                       xs.cr();
+                       xs << XHTMLStream::ESCAPE_NONE << deferred << html::CR();
                }
        }
        return pend;
@@ -648,12 +754,12 @@ ParagraphList::const_iterator makeBibliography(Buffer const & buf,
 {
        // FIXME XHTML
        // Use TextClass::htmlTOCLayout() to figure out how we should look.
-       xs << html::StartTag("h2", "class='bibliography'");
-       xs << pbegin->layout().labelstring(false);
-       xs << html::EndTag("h2");
-       xs.cr();
-       xs << html::StartTag("div", "class='bibliography'");
-       xs.cr();
+       xs << html::StartTag("h2", "class='bibliography'")
+          << pbegin->layout().labelstring(false)
+          << html::EndTag("h2")
+          << html::CR()
+          << html::StartTag("div", "class='bibliography'")
+          << html::CR();
        makeParagraphs(buf, xs, runparams, text, pbegin, pend);
        xs << html::EndTag("div");
        return pend;
@@ -681,7 +787,7 @@ ParagraphList::const_iterator makeEnvironmentHtml(Buffer const & buf,
 
        // open tag for this environment
        openTag(xs, bstyle);
-       xs.cr();
+       xs << html::CR();
 
        // we will on occasion need to remember a layout from before.
        Layout const * lastlay = 0;
@@ -722,7 +828,7 @@ ParagraphList::const_iterator makeEnvironmentHtml(Buffer const & buf,
                                
                                bool const labelfirst = style.htmllabelfirst();
                                if (!labelfirst)
-                                       openItemTag(xs, style);
+                                       openItemTag(xs, style, par->params());
                                
                                // label output
                                if (style.labeltype != LABEL_NO_LABEL && 
@@ -738,26 +844,26 @@ ParagraphList::const_iterator makeEnvironmentHtml(Buffer const & buf,
                                                                xs << lbl;
                                                                closeLabelTag(xs, style);
                                                        }
-                                                       xs.cr();
+                                                       xs << html::CR();
                                                }
                                        }       else { // some kind of list
                                                if (style.labeltype == LABEL_MANUAL) {
                                                        openLabelTag(xs, style);
                                                        sep = par->firstWordLyXHTML(xs, runparams);
                                                        closeLabelTag(xs, style);
-                                                       xs.cr();
+                                                       xs << html::CR();
                                                }
                                                else {
                                                        openLabelTag(xs, style);
                                                        xs << par->params().labelString();
                                                        closeLabelTag(xs, style);
-                                                       xs.cr();
+                                                       xs << html::CR();
                                                }
                                        }
                                } // end label output
 
                                if (labelfirst)
-                                       openItemTag(xs, style);
+                                       openItemTag(xs, style, par->params());
 
                                par->simpleLyXHTMLOnePar(buf, xs, runparams, 
                                        text.outerFont(distance(begin, par)), sep);
@@ -775,7 +881,7 @@ ParagraphList::const_iterator makeEnvironmentHtml(Buffer const & buf,
                                        lastlay = &style;
                                } else
                                        closeItemTag(xs, style);
-                               xs.cr();
+                               xs << html::CR();
                        }
                        // The other possibility is that the depth has increased, in which
                        // case we need to recurse.
@@ -805,22 +911,22 @@ ParagraphList::const_iterator makeEnvironmentHtml(Buffer const & buf,
        if (lastlay != 0)
                closeItemTag(xs, *lastlay);
        closeTag(xs, bstyle);
-       xs.cr();
+       xs << html::CR();
        return pend;
 }
 
 
 void makeCommand(Buffer const & buf,
-                                         XHTMLStream & xs,
-                                         OutputParams const & runparams,
-                                         Text const & text,
-                                         ParagraphList::const_iterator const & pbegin)
+                XHTMLStream & xs,
+                OutputParams const & runparams,
+                Text const & text,
+                ParagraphList::const_iterator const & pbegin)
 {
        Layout const & style = pbegin->layout();
        if (!style.counter.empty())
                buf.params().documentClass().counters().step(style.counter, OutputUpdate);
 
-       openTag(xs, style);
+       openTag(xs, style, pbegin->params());
 
        // Label around sectioning number:
        // FIXME Probably need to account for LABEL_MANUAL
@@ -836,7 +942,7 @@ void makeCommand(Buffer const & buf,
        pbegin->simpleLyXHTMLOnePar(buf, xs, runparams,
                        text.outerFont(distance(begin, pbegin)));
        closeTag(xs, style);
-       xs.cr();
+       xs << html::CR();
 }
 
 } // end anonymous namespace
@@ -848,20 +954,33 @@ void xhtmlParagraphs(Text const & text,
                       OutputParams const & runparams)
 {
        ParagraphList const & paragraphs = text.paragraphs();
-       ParagraphList::const_iterator par = paragraphs.begin();
-       ParagraphList::const_iterator pend = paragraphs.end();
+       if (runparams.par_begin == runparams.par_end) {
+               runparams.par_begin = 0;
+               runparams.par_end = paragraphs.size();
+       }
+       pit_type bpit = runparams.par_begin;
+       pit_type const epit = runparams.par_end;
+       LASSERT(bpit < epit, /* */);
 
        OutputParams ourparams = runparams;
-       while (par != pend) {
+       ParagraphList::const_iterator const pend =
+               (epit == (int) paragraphs.size()) ?
+                       paragraphs.end() : paragraphs.constIterator(epit);
+       while (bpit < epit) {
+               ParagraphList::const_iterator par = paragraphs.constIterator(bpit);
                if (par->params().startOfAppendix()) {
-                       // FIXME: only the counter corresponding to toplevel
-                       // sectioning should be reset
-                       Counters & cnts = buf.masterBuffer()->params().documentClass().counters();
-                       cnts.reset();
-                       cnts.appendix(true);
+                       // We want to reset the counter corresponding to toplevel sectioning
+                       Layout const & lay =
+                               buf.masterBuffer()->params().documentClass().getTOCLayout();
+                       docstring const cnt = lay.counter;
+                       if (!cnt.empty()) {
+                               Counters & cnts =
+                                       buf.masterBuffer()->params().documentClass().counters();
+                               cnts.reset(cnt);
+                       }
                }
                Layout const & style = par->layout();
-               ParagraphList::const_iterator lastpar = par;
+               ParagraphList::const_iterator const lastpar = par;
                ParagraphList::const_iterator send;
 
                switch (style.latextype) {
@@ -898,15 +1017,13 @@ void xhtmlParagraphs(Text const & text,
                        par = makeParagraphs(buf, xs, ourparams, text, par, send);
                        break;
                }
-               // FIXME??
-               // makeEnvironment may process more than one paragraphs and bypass pend
-               if (distance(lastpar, par) >= distance(lastpar, pend))
-                       break;
+               bpit += distance(lastpar, par);
        }
 }
 
 
-string alignmentToCSS(LyXAlignment align) {
+string alignmentToCSS(LyXAlignment align)
+{
        switch (align) {
        case LYX_ALIGN_BLOCK:
                // we are NOT going to use text-align: justify!!