2 * \file output_xhtml.cpp
3 * This file is part of LyX, the document processor.
4 * Licence details can be found in the file COPYING.
8 * This code is based upon output_docbook.cpp
10 * Full author contact details are available in file CREDITS.
15 #include "output_xhtml.h"
18 #include "buffer_funcs.h"
19 #include "BufferParams.h"
22 #include "OutputParams.h"
23 #include "Paragraph.h"
24 #include "ParagraphList.h"
25 #include "ParagraphParameters.h"
28 #include "TextClass.h"
30 #include "support/lassert.h"
31 #include "support/debug.h"
32 #include "support/lstrings.h"
37 using namespace lyx::support;
43 docstring escapeChar(char_type c)
67 // escape what needs escaping
68 docstring htmlize(docstring const & str) {
70 docstring::const_iterator it = str.begin();
71 docstring::const_iterator en = str.end();
72 for (; it != en; ++it)
78 bool isFontTag(string const & s)
80 return s == "em" || s == "strong"; // others?
85 docstring StartTag::asTag() const
87 string output = "<" + tag_;
89 output += " " + attr_;
91 return from_utf8(output);
95 docstring StartTag::asEndTag() const
97 string output = "</" + tag_ + ">";
98 return from_utf8(output);
102 docstring EndTag::asEndTag() const
104 string output = "</" + tag_ + ">";
105 return from_utf8(output);
109 docstring CompTag::asTag() const
111 string output = "<" + tag_;
113 output += " " + attr_;
115 return from_utf8(output);
119 ////////////////////////////////////////////////////////////////
123 ////////////////////////////////////////////////////////////////
125 XHTMLStream::XHTMLStream(odocstream & os)
130 void XHTMLStream::cr()
133 os_ << from_ascii("\n");
137 bool XHTMLStream::closeFontTags()
139 // first, we close any open font tags we can close
140 StartTag curtag = tag_stack_.back();
141 while (html::isFontTag(curtag.tag_)) {
142 os_ << curtag.asEndTag();
143 tag_stack_.pop_back();
144 if (tag_stack_.empty())
145 // this probably shouldn't happen, since then the
146 // font tags weren't in any other tag. but that
147 // problem will likely be caught elsewhere.
149 curtag = tag_stack_.back();
151 // so we've hit a non-font tag. let's see if any of the
152 // remaining tags are font tags.
153 TagStack::const_iterator it = tag_stack_.begin();
154 TagStack::const_iterator en = tag_stack_.end();
155 bool noFontTags = true;
156 for (; it != en; ++it) {
157 if (html::isFontTag(it->tag_)) {
158 LYXERR0("Font tag `" << it->tag_ << "' still open in closeFontTags().");
166 void XHTMLStream::clearTagDeque()
168 while (!pending_tags_.empty()) {
169 StartTag const & tag = pending_tags_.front();
172 tag_stack_.push_back(tag);
173 pending_tags_.pop_front();
178 XHTMLStream & XHTMLStream::operator<<(docstring const & d)
181 os_ << html::htmlize(d);
186 XHTMLStream & XHTMLStream::operator<<(const char * s)
189 os_ << html::htmlize(from_ascii(s));
194 XHTMLStream & XHTMLStream::operator<<(char_type c)
197 os_ << html::escapeChar(c);
202 XHTMLStream & XHTMLStream::operator<<(StartTag const & tag)
204 if (tag.tag_.empty())
206 pending_tags_.push_back(tag);
213 XHTMLStream & XHTMLStream::operator<<(CompTag const & tag)
215 if (tag.tag_.empty())
224 bool XHTMLStream::isTagOpen(string const & stag)
226 TagStack::const_iterator sit = tag_stack_.begin();
227 TagStack::const_iterator const sen = tag_stack_.end();
228 for (; sit != sen; ++sit)
229 // we could check for the
230 if (sit->tag_ == stag)
236 // this is complicated, because we want to make sure that
237 // everything is properly nested. the code ought to make
238 // sure of that, but we won't assert (yet) if we run into
239 // a problem. we'll just output error messages and try our
240 // best to make things work.
241 XHTMLStream & XHTMLStream::operator<<(EndTag const & etag)
243 if (etag.tag_.empty())
245 // first make sure we're not closing an empty tag
246 if (!pending_tags_.empty()) {
247 StartTag const & stag = pending_tags_.back();
248 if (etag.tag_ == stag.tag_) {
249 // we have <tag></tag>, so we discard it and remove it
250 // from the pending_tags_.
251 pending_tags_.pop_back();
254 // there is a pending tag that isn't the one we are trying
256 // is this tag itself pending?
257 // non-const iterators because we may call erase().
258 TagDeque::iterator dit = pending_tags_.begin();
259 TagDeque::iterator const den = pending_tags_.end();
260 for (; dit != den; ++dit) {
261 if (dit->tag_ == etag.tag_) {
262 // it was pending, so we just erase it
263 LYXERR0("Tried to close pending tag `" << etag.tag_
264 << "' when other tags were pending. Tag discarded.");
265 pending_tags_.erase(dit);
269 // so etag isn't itself pending. is it even open?
270 if (!isTagOpen(etag.tag_)) {
271 LYXERR0("Tried to close `" << etag.tag_
272 << "' when tag was not open. Tag discarded.");
275 // ok, so etag is open.
276 // our strategy will be as below: we will do what we need to
277 // do to close this tag.
278 LYXERR0("Closing tag `" << etag.tag_
279 << "' when other tags are pending. Discarded pending tags:");
280 for (dit = pending_tags_.begin(); dit != den; ++dit)
282 // clear the pending tags...
283 pending_tags_.clear();
284 // ...and then just fall through.
287 // is the tag we are closing the last one we opened?
288 if (etag.tag_ == tag_stack_.back().tag_) {
290 os_ << etag.asEndTag();
291 // ...and forget about it
292 tag_stack_.pop_back();
296 // we are trying to close a tag other than the one last opened.
297 // let's first see if this particular tag is still open somehow.
298 if (!isTagOpen(etag.tag_)) {
299 LYXERR0("Tried to close `" << etag.tag_
300 << "' when tag was not open. Tag discarded.");
304 // so the tag was opened, but other tags have been opened since
305 // and not yet closed.
306 // if it's a font tag, though...
307 if (html::isFontTag(etag.tag_)) {
308 // it won't be a problem if the other tags open since this one
309 // are also font tags.
310 TagStack::const_reverse_iterator rit = tag_stack_.rbegin();
311 TagStack::const_reverse_iterator ren = tag_stack_.rend();
312 for (; rit != ren; ++rit) {
313 if (rit->tag_ == etag.tag_)
315 if (!html::isFontTag(rit->tag_)) {
316 // we'll just leave it and, presumably, have to close it later.
317 LYXERR0("Unable to close font tag `" << etag.tag_
318 << "' due to open non-font tag `" << rit->tag_ << "'.");
324 // <em>this is <strong>bold
325 // and are being asked to closed em. we want:
326 // <em>this is <strong>bold</strong></em><strong>
327 // first, we close the intervening tags...
328 StartTag curtag = tag_stack_.back();
329 // ...remembering them in a stack.
331 while (curtag.tag_ != etag.tag_) {
332 os_ << curtag.asEndTag();
333 fontstack.push_back(curtag);
334 tag_stack_.pop_back();
335 curtag = tag_stack_.back();
337 // now close our tag...
338 os_ << etag.asEndTag();
339 tag_stack_.pop_back();
341 // ...and restore the other tags.
342 rit = fontstack.rbegin();
343 ren = fontstack.rend();
344 for (; rit != ren; ++rit)
345 pending_tags_.push_back(*rit);
349 // it wasn't a font tag.
350 // so other tags were opened before this one and not properly closed.
351 // so we'll close them, too. that may cause other issues later, but it
352 // at least guarantees proper nesting.
353 LYXERR0("Closing tag `" << etag.tag_
354 << "' when other tags are open, namely:");
355 StartTag curtag = tag_stack_.back();
356 while (curtag.tag_ != etag.tag_) {
357 LYXERR0(curtag.tag_);
358 os_ << curtag.asEndTag();
359 tag_stack_.pop_back();
360 curtag = tag_stack_.back();
362 // curtag is now the one we actually want.
363 os_ << curtag.asEndTag();
364 tag_stack_.pop_back();
369 // End code for XHTMLStream
373 // convenience functions
375 inline void openTag(XHTMLStream & xs, Layout const & lay)
377 xs << StartTag(lay.htmltag(), lay.htmlattr());
381 inline void closeTag(XHTMLStream & xs, Layout const & lay)
383 xs << EndTag(lay.htmltag());
387 inline void openLabelTag(XHTMLStream & xs, Layout const & lay)
389 xs << StartTag(lay.htmllabeltag(), lay.htmllabelattr());
393 inline void closeLabelTag(XHTMLStream & xs, Layout const & lay)
395 xs << EndTag(lay.htmllabeltag());
399 inline void openItemTag(XHTMLStream & xs, Layout const & lay)
401 xs << StartTag(lay.htmlitemtag(), lay.htmlitemattr(), true);
405 inline void closeItemTag(XHTMLStream & xs, Layout const & lay)
407 xs << EndTag(lay.htmlitemtag());
410 // end of convenience functions
412 ParagraphList::const_iterator searchParagraphHtml(
413 ParagraphList::const_iterator p,
414 ParagraphList::const_iterator const & pend)
416 for (++p; p != pend && p->layout().latextype == LATEX_PARAGRAPH; ++p)
423 ParagraphList::const_iterator searchEnvironmentHtml(
424 ParagraphList::const_iterator const pstart,
425 ParagraphList::const_iterator const & pend)
427 ParagraphList::const_iterator p = pstart;
428 Layout const & bstyle = p->layout();
429 size_t const depth = p->params().depth();
430 for (++p; p != pend; ++p) {
431 Layout const & style = p->layout();
432 // It shouldn't happen that e.g. a section command occurs inside
433 // a quotation environment, at a higher depth, but as of 6/2009,
434 // it can happen. We pretend that it's just at lowest depth.
435 if (style.latextype == LATEX_COMMAND)
437 // If depth is down, we're done
438 if (p->params().depth() < depth)
440 // If depth is up, we're not done
441 if (p->params().depth() > depth)
443 // Now we know we are at the same depth
444 if (style.latextype == LATEX_PARAGRAPH
445 || style.latexname() != bstyle.latexname())
452 ParagraphList::const_iterator makeParagraphs(Buffer const & buf,
454 OutputParams const & runparams,
456 ParagraphList::const_iterator const & pbegin,
457 ParagraphList::const_iterator const & pend)
459 ParagraphList::const_iterator const begin = text.paragraphs().begin();
460 ParagraphList::const_iterator par = pbegin;
461 for (; par != pend; ++par) {
462 Layout const & lay = par->layout();
463 if (!lay.counter.empty())
464 buf.params().documentClass().counters().step(lay.counter);
465 // FIXME We should see if there's a label to be output and
466 // do something with it.
470 // FIXME Should we really allow anything other than 'p' here?
472 // If we are already in a paragraph, and this is the first one, then we
473 // do not want to open the paragraph tag.
475 (par == pbegin && runparams.html_in_par) ? false : true;
478 docstring const deferred =
479 par->simpleLyXHTMLOnePar(buf, xs, runparams, text.outerFont(distance(begin, par)));
481 // We want to issue the closing tag if either:
482 // (i) We opened it, and either html_in_par is false,
483 // or we're not in the last paragraph, anyway.
484 // (ii) We didn't open it and html_in_par is true,
485 // but we are in the first par, and there is a next par.
486 ParagraphList::const_iterator nextpar = par;
488 bool const needclose =
489 (opened && (!runparams.html_in_par || nextpar != pend))
490 || (!opened && runparams.html_in_par && par == pbegin && nextpar != pend);
495 if (!deferred.empty()) {
504 ParagraphList::const_iterator makeBibliography(Buffer const & buf,
506 OutputParams const & runparams,
508 ParagraphList::const_iterator const & pbegin,
509 ParagraphList::const_iterator const & pend)
511 xs << StartTag("h2", "class='bibliography'");
512 xs << pbegin->layout().labelstring(false);
515 xs << StartTag("div", "class='bibliography'");
517 makeParagraphs(buf, xs, runparams, text, pbegin, pend);
523 bool isNormalEnv(Layout const & lay)
525 return lay.latextype == LATEX_ENVIRONMENT;
529 ParagraphList::const_iterator makeEnvironmentHtml(Buffer const & buf,
531 OutputParams const & runparams,
533 ParagraphList::const_iterator const & pbegin,
534 ParagraphList::const_iterator const & pend)
536 ParagraphList::const_iterator const begin = text.paragraphs().begin();
537 ParagraphList::const_iterator par = pbegin;
538 Layout const & bstyle = par->layout();
539 depth_type const origdepth = pbegin->params().depth();
541 // open tag for this environment
545 // we will on occasion need to remember a layout from before.
546 Layout const * lastlay = 0;
548 while (par != pend) {
549 Layout const & style = par->layout();
550 // the counter only gets stepped if we're in some kind of list,
551 // or if it's the first time through.
552 if (!style.counter.empty() && (par == pbegin || !isNormalEnv(style)))
553 buf.params().documentClass().counters().step(style.counter);
554 ParagraphList::const_iterator send;
555 // this will be positive, if we want to skip the initial word
556 // (if it's been taken for the label).
559 switch (style.latextype) {
560 case LATEX_ENVIRONMENT:
561 case LATEX_LIST_ENVIRONMENT:
562 case LATEX_ITEM_ENVIRONMENT: {
563 // There are two possiblities in this case.
564 // One is that we are still in the environment in which we
565 // started---which we will be if the depth is the same.
566 if (par->params().depth() == origdepth) {
567 LASSERT(bstyle == style, /* */);
569 closeItemTag(xs, *lastlay);
572 bool const labelfirst = style.htmllabelfirst();
573 if (isNormalEnv(style)) {
574 // in this case, we print the label only for the first
575 // paragraph (as in a theorem).
576 openItemTag(xs, style);
577 if (par == pbegin && style.htmllabeltag() != "NONE") {
578 docstring const lbl =
579 pbegin->expandLabel(style, buf.params(), false);
581 openLabelTag(xs, style);
583 closeLabelTag(xs, style);
587 } else { // some kind of list
589 openItemTag(xs, style);
590 if (style.labeltype == LABEL_MANUAL
591 && style.htmllabeltag() != "NONE") {
592 openLabelTag(xs, style);
593 // sep = par->firstWordLyXHTML(xs, runparams);
594 closeLabelTag(xs, style);
597 else if (style.labeltype != LABEL_NO_LABEL
598 && style.htmllabeltag() != "NONE") {
599 openLabelTag(xs, style);
600 xs << par->expandLabel(style, buf.params(), false);
601 closeLabelTag(xs, style);
605 openItemTag(xs, style);
607 xs << StartTag("span", "class='" + to_utf8(style.name()) + " inneritem'");
609 par->simpleLyXHTMLOnePar(buf, xs, runparams,
610 text.outerFont(distance(begin, par)), sep);
611 if (!isNormalEnv(style) && !labelfirst)
612 xs << EndTag("span");
614 // We may not want to close the tag yet, in particular,
615 // if we're not at the end...
617 // and are doing items...
618 && style.latextype == LATEX_ITEM_ENVIRONMENT
619 // and if the depth has changed...
620 && par->params().depth() != origdepth) {
621 // then we'll save this layout for later, and close it when
622 // we get another item.
625 closeItemTag(xs, style);
628 // The other possibility is that the depth has increased, in which
629 // case we need to recurse.
631 send = searchEnvironmentHtml(par, pend);
632 par = makeEnvironmentHtml(buf, xs, runparams, text, par, send);
636 case LATEX_PARAGRAPH:
637 send = searchParagraphHtml(par, pend);
638 par = makeParagraphs(buf, xs, runparams, text, par, send);
641 case LATEX_BIB_ENVIRONMENT:
644 par = makeParagraphs(buf, xs, runparams, text, par, send);
654 closeItemTag(xs, *lastlay);
655 closeTag(xs, bstyle);
661 void makeCommand(Buffer const & buf,
663 OutputParams const & runparams,
665 ParagraphList::const_iterator const & pbegin)
667 Layout const & style = pbegin->layout();
668 if (!style.counter.empty())
669 buf.params().documentClass().counters().step(style.counter);
673 // Label around sectioning number:
674 // FIXME Probably need to account for LABEL_MANUAL
675 if (style.labeltype != LABEL_NO_LABEL) {
676 openLabelTag(xs, style);
677 xs << pbegin->expandLabel(style, buf.params(), false);
678 closeLabelTag(xs, style);
679 // Otherwise the label might run together with the text
680 xs << from_ascii(" ");
683 ParagraphList::const_iterator const begin = text.paragraphs().begin();
684 pbegin->simpleLyXHTMLOnePar(buf, xs, runparams,
685 text.outerFont(distance(begin, pbegin)));
690 } // end anonymous namespace
693 void xhtmlParagraphs(Text const & text,
696 OutputParams const & runparams)
698 ParagraphList const & paragraphs = text.paragraphs();
699 ParagraphList::const_iterator par = paragraphs.begin();
700 ParagraphList::const_iterator pend = paragraphs.end();
702 OutputParams ourparams = runparams;
703 while (par != pend) {
704 Layout const & style = par->layout();
705 ParagraphList::const_iterator lastpar = par;
706 ParagraphList::const_iterator send;
708 switch (style.latextype) {
709 case LATEX_COMMAND: {
710 // The files with which we are working never have more than
711 // one paragraph in a command structure.
713 // if (ourparams.html_in_par)
714 // fix it so we don't get sections inside standard, e.g.
715 // note that we may then need to make runparams not const, so we
716 // can communicate that back.
717 // FIXME Maybe this fix should be in the routines themselves, in case
718 // they are called from elsewhere.
719 makeCommand(buf, xs, ourparams, text, par);
723 case LATEX_ENVIRONMENT:
724 case LATEX_LIST_ENVIRONMENT:
725 case LATEX_ITEM_ENVIRONMENT: {
726 // FIXME Same fix here.
727 send = searchEnvironmentHtml(par, pend);
728 par = makeEnvironmentHtml(buf, xs, ourparams, text, par, send);
731 case LATEX_BIB_ENVIRONMENT: {
732 // FIXME Same fix here.
733 send = searchEnvironmentHtml(par, pend);
734 par = makeBibliography(buf, xs, ourparams, text, par, send);
737 case LATEX_PARAGRAPH:
738 send = searchParagraphHtml(par, pend);
739 par = makeParagraphs(buf, xs, ourparams, text, par, send);
743 // makeEnvironment may process more than one paragraphs and bypass pend
744 if (distance(lastpar, par) >= distance(lastpar, pend))