]> git.lyx.org Git - lyx.git/blobdiff - src/output_xhtml.cpp
fix parsing of \def macros (was broken with the spreadsheet template support)
[lyx.git] / src / output_xhtml.cpp
index c4792ae739d38092cdd11c19be4e22b6bfdc7e9c..dee3060e59131c45f1f46a1f53208ae15874ddab 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 +43,78 @@ 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();
 }
 
@@ -128,18 +135,18 @@ 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 +154,7 @@ 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 +178,7 @@ 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);
 }
@@ -187,17 +194,10 @@ docstring CompTag::asTag() const
 ////////////////////////////////////////////////////////////////
 
 XHTMLStream::XHTMLStream(odocstream & os) 
-               : os_(os), nextraw_(false)
+               : os_(os), escape_(ESCAPE_ALL)
 {}
 
 
-void XHTMLStream::cr() 
-{
-       // tabs?
-       os_ << from_ascii("\n");
-}
-
-
 void XHTMLStream::writeError(std::string const & s)
 {
        LYXERR0(s);
@@ -205,6 +205,12 @@ void XHTMLStream::writeError(std::string const & s)
 }
 
 
+namespace {
+       // an illegal tag for internal use
+       static string const parsep_tag = "&LyX_parsep_tag&";
+}
+
+
 bool XHTMLStream::closeFontTags()
 {
        if (tag_stack_.empty())
@@ -221,19 +227,77 @@ 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 (!isTagOpen(parsep_tag)) {
+               // is it pending?
+               TagStack::const_iterator dit = pending_tags_.begin();
+               TagStack::const_iterator const den = pending_tags_.end();
+               bool found = false;
+               for (; dit != den; ++dit) {
+                       if (dit->tag_ == parsep_tag) {
+                               found = true;
+                               break;
+                       }
                }
+
+               if (!found) {
+                       writeError("No paragraph separation tag found in endParagraph().");
+                       return;
+               }
+               
+               // this case is normal.
+               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 hasnt' been used.
+                       html::StartTag const cur_tag = pending_tags_.back();
+                       string const & tag = cur_tag.tag_;
+                       tag_stack_.pop_back();
+                       if (tag == parsep_tag)
+                               break;
+               }
+               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 +305,9 @@ 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 +317,8 @@ 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 +327,8 @@ 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 +336,8 @@ 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 +345,8 @@ 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 +355,14 @@ 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;
 }
 
@@ -332,7 +385,15 @@ 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;
 }
 
@@ -378,8 +439,8 @@ 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 +541,8 @@ 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();
        }
@@ -618,7 +680,7 @@ 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.
@@ -643,11 +705,10 @@ ParagraphList::const_iterator makeParagraphs(Buffer const & buf,
                        || (!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;
@@ -663,12 +724,12 @@ 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;
@@ -696,7 +757,7 @@ 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;
@@ -753,20 +814,20 @@ 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
@@ -790,7 +851,7 @@ 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.
@@ -820,7 +881,7 @@ ParagraphList::const_iterator makeEnvironmentHtml(Buffer const & buf,
        if (lastlay != 0)
                closeItemTag(xs, *lastlay);
        closeTag(xs, bstyle);
-       xs.cr();
+       xs << html::CR();
        return pend;
 }
 
@@ -851,7 +912,7 @@ 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
@@ -863,11 +924,20 @@ 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
@@ -876,7 +946,7 @@ void xhtmlParagraphs(Text const & text,
                        cnts.appendix(true);
                }
                Layout const & style = par->layout();
-               ParagraphList::const_iterator lastpar = par;
+               ParagraphList::const_iterator const lastpar = par;
                ParagraphList::const_iterator send;
 
                switch (style.latextype) {
@@ -913,10 +983,7 @@ 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);
        }
 }