]> git.lyx.org Git - lyx.git/blobdiff - src/Paragraph.cpp
Rename a couple routines in preparation for more.
[lyx.git] / src / Paragraph.cpp
index e7f1721ca5a9ccc557d414a1bd5986cc3e83b626..92706f0c46009347c227d9ee58c55ac6b1e15dc8 100644 (file)
@@ -4,13 +4,13 @@
  * Licence details can be found in the file COPYING.
  *
  * \author Asger Alstrup
- * \author Lars Gullik Bjønnes
+ * \author Lars Gullik Bjønnes
  * \author Jean-Marc Lasgouttes
  * \author Angus Leeming
  * \author John Levon
- * \author André Pönitz
+ * \author André Pönitz
  * \author Dekel Tsur
- * \author Jürgen Vigna
+ * \author Jürgen Vigna
  *
  * Full author contact details are available in file CREDITS.
  */
 #include "LyXRC.h"
 #include "OutputParams.h"
 #include "output_latex.h"
+#include "output_xhtml.h"
 #include "paragraph_funcs.h"
 #include "ParagraphParameters.h"
+#include "SpellChecker.h"
 #include "sgml.h"
 #include "TextClass.h"
 #include "TexRow.h"
 #include "Text.h"
 #include "VSpace.h"
+#include "WordLangTuple.h"
 #include "WordList.h"
 
 #include "frontends/alert.h"
@@ -50,7 +53,6 @@
 #include "insets/InsetLabel.h"
 
 #include "support/lassert.h"
-#include "support/convert.h"
 #include "support/debug.h"
 #include "support/ExceptionMessage.h"
 #include "support/gettext.h"
@@ -286,6 +288,15 @@ void Paragraph::addChangesToToc(DocIterator const & cdit,
 }
 
 
+bool Paragraph::isFullyDeleted(pos_type start, pos_type end) const
+{
+       LASSERT(start >= 0 && start <= size(), /**/);
+       LASSERT(end > start && end <= size() + 1, /**/);
+
+       return d->changes_.isFullyDeleted(start, end);
+}
+
+
 bool Paragraph::isChanged(pos_type start, pos_type end) const
 {
        LASSERT(start >= 0 && start <= size(), /**/);
@@ -757,16 +768,16 @@ void Paragraph::Private::latexInset(
        }
 
        bool close = false;
-       odocstringstream ods;
+       odocstream::pos_type const len = os.tellp();
 
        if (inset->forceLTR() 
            && running_font.isRightToLeft()
                // ERT is an exception, it should be output with no decorations at all
                && inset->lyxCode() != ERT_CODE) {
                if (running_font.language()->lang() == "farsi")
-                       ods << "\\beginL{}";
+                       os << "\\beginL{}";
                else
-                       ods << "\\L{";
+                       os << "\\L{";
                close = true;
        }
 
@@ -785,7 +796,7 @@ void Paragraph::Private::latexInset(
        if (open_font && inset->noFontChange()) {
                bool closeLanguage = arabtex
                        || basefont.isRightToLeft() == running_font.isRightToLeft();
-               unsigned int count = running_font.latexWriteEndChanges(ods,
+               unsigned int count = running_font.latexWriteEndChanges(os,
                        bparams, runparams, basefont, basefont, closeLanguage);
                column += count;
                // if any font properties were closed, update the running_font, 
@@ -808,7 +819,7 @@ void Paragraph::Private::latexInset(
        int tmp;
 
        try {
-               tmp = inset->latex(ods, runparams);
+               tmp = inset->latex(os, runparams);
        } catch (EncodingException & e) {
                // add location information and throw again.
                e.par_id = id_;
@@ -818,13 +829,11 @@ void Paragraph::Private::latexInset(
 
        if (close) {
                if (running_font.language()->lang() == "farsi")
-                               ods << "\\endL{}";
+                               os << "\\endL{}";
                        else
-                               ods << '}';
+                               os << '}';
        }
 
-       os << ods.str();
-
        if (tmp) {
                for (int j = 0; j < tmp; ++j)
                        texrow.newline();
@@ -832,7 +841,7 @@ void Paragraph::Private::latexInset(
                texrow.start(owner_->id(), i + 1);
                column = 0;
        } else {
-               column += ods.str().size();
+               column += os.tellp() - len;
        }
 
        if (owner_->lookupChange(i).type == Change::DELETED)
@@ -866,9 +875,15 @@ void Paragraph::Private::latexSpecialChar(
                return;
        }
 
-       if (lyxrc.fontenc == "T1" && latexSpecialT1(c, os, i, column))
+       // If T1 font encoding is used, use the special
+       // characters it provides.
+       // NOTE: some languages reset the font encoding
+       // internally
+       if (!running_font.language()->internalFontEncoding()
+           && lyxrc.fontenc == "T1" && latexSpecialT1(c, os, i, column))
                return;
 
+       // \tt font needs special treatment
        if (running_font.fontInfo().family() == TYPEWRITER_FAMILY
                && latexSpecialTypewriter(c, os, i, column))
                return;
@@ -917,7 +932,9 @@ void Paragraph::Private::latexSpecialChar(
                column += 17;
                break;
 
-       case '*': case '[':
+       case '*':
+       case '[':
+       case ']':
                // avoid being mistaken for optional arguments
                os << '{';
                os.put(c);
@@ -1277,7 +1294,7 @@ void Paragraph::appendString(docstring const & s, Font const & font,
        d->text_.append(s);
 
        // FIXME: Optimize this!
-       for (pos_type i = oldsize; i != newsize; ++i) {
+       for (size_t i = oldsize; i != newsize; ++i) {
                // track change
                d->changes_.insert(change, i);
        }
@@ -1573,7 +1590,8 @@ docstring const & Paragraph::labelString() const
 // the next two functions are for the manual labels
 docstring const Paragraph::getLabelWidthString() const
 {
-       if (d->layout_->margintype == MARGIN_MANUAL)
+       if (d->layout_->margintype == MARGIN_MANUAL
+           || d->layout_->latextype == LATEX_BIB_ENVIRONMENT)
                return d->params_.labelWidthString();
        else
                return _("Senseless with this layout!");
@@ -1614,7 +1632,7 @@ docstring Paragraph::expandLabel(Layout const & layout,
 
        if (fmt.empty() && layout.labeltype == LABEL_COUNTER 
            && !layout.counter.empty())
-               fmt = "\\the" + layout.counter;
+               return tclass.counters().theCounter(layout.counter);
 
        // handle 'inherited level parts' in 'fmt',
        // i.e. the stuff between '@' in   '@Section@.\arabic{subsection}'
@@ -1907,9 +1925,10 @@ int Paragraph::Private::endTeXParParams(BufferParams const & bparams,
 
 // This one spits out the text of the paragraph
 bool Paragraph::latex(BufferParams const & bparams,
-                               Font const & outerfont,
-                               odocstream & os, TexRow & texrow,
-                               OutputParams const & runparams) const
+       Font const & outerfont,
+       odocstream & os, TexRow & texrow,
+       OutputParams const & runparams,
+       int start_pos, int end_pos) const
 {
        LYXERR(Debug::LATEX, "Paragraph::latex...     " << this);
 
@@ -2081,7 +2100,8 @@ bool Paragraph::latex(BufferParams const & bparams,
                                os << fontchange;
                }
 
-               if (c == ' ') {
+               // FIXME: think about end_pos implementation...
+               if (c == ' ' && i >= start_pos && (end_pos == -1 || i < end_pos)) {
                        // FIXME: integrate this case in latexSpecialChar
                        // Do not print the separation of the optional argument
                        // if style.pass_thru is false. This works because
@@ -2109,16 +2129,19 @@ bool Paragraph::latex(BufferParams const & bparams,
                // Two major modes:  LaTeX or plain
                // Handle here those cases common to both modes
                // and then split to handle the two modes separately.
-               if (c == META_INSET)
-                       d->latexInset(bparams, os,
-                                       texrow, rp, running_font,
-                                       basefont, outerfont, open_font,
-                                       runningChange, style, i, column);
-               else {
-                       try {
-                               d->latexSpecialChar(os, rp, running_font, runningChange,
-                                       style, i, column);
-                       } catch (EncodingException & e) {
+               if (c == META_INSET) {
+                       if (i >= start_pos && (end_pos == -1 || i < end_pos)) {
+                               d->latexInset(bparams, os,
+                                               texrow, rp, running_font,
+                                               basefont, outerfont, open_font,
+                                               runningChange, style, i, column);
+                       }
+               } else {
+                       if (i >= start_pos && (end_pos == -1 || i < end_pos)) {
+                               try {
+                                       d->latexSpecialChar(os, rp, running_font, runningChange,
+                                               style, i, column);
+                               } catch (EncodingException & e) {
                                if (runparams.dryrun) {
                                        os << "<" << _("LyX Warning: ")
                                           << _("uncodable character") << " '";
@@ -2132,6 +2155,7 @@ bool Paragraph::latex(BufferParams const & bparams,
                                }
                        }
                }
+               }
 
                // Set the encoding to that returned from latexSpecialChar (see
                // comment for encoding member in OutputParams.h)
@@ -2218,7 +2242,7 @@ string Paragraph::getID(Buffer const & buf, OutputParams const & runparams)
 }
 
 
-pos_type Paragraph::firstWord(odocstream & os, OutputParams const & runparams)
+pos_type Paragraph::firstWordDocBook(odocstream & os, OutputParams const & runparams)
        const
 {
        pos_type i;
@@ -2236,6 +2260,24 @@ pos_type Paragraph::firstWord(odocstream & os, OutputParams const & runparams)
 }
 
 
+pos_type Paragraph::firstWordLyXHTML(odocstream & os, OutputParams const & runparams)
+       const
+{
+       pos_type i;
+       for (i = 0; i < size(); ++i) {
+               if (Inset const * inset = getInset(i)) {
+                       inset->xhtml(os, runparams);
+               } else {
+                       char_type c = d->text_[i];
+                       if (c == ' ')
+                               break;
+                       os << html::escapeChar(c);
+               }
+       }
+       return i;
+}
+
+
 bool Paragraph::Private::onlyText(Buffer const & buf, Font const & outerfont, pos_type initial) const
 {
        Font font_old;
@@ -2307,6 +2349,91 @@ void Paragraph::simpleDocBookOnePar(Buffer const & buf,
 }
 
 
+void Paragraph::simpleLyXHTMLOnePar(Buffer const & buf,
+                                   odocstream & os,
+                                   OutputParams const & runparams,
+                                   Font const & outerfont,
+                                   pos_type initial) const
+{
+       // FIXME We really need to manage the tag nesting here.
+       // Probably in the same sort of way as in output_xhtml.
+       bool emph_flag = false;
+       bool bold_flag = false;
+       std::string closing_tag;
+
+       Layout const & style = *d->layout_;
+       FontInfo font_old =
+               style.labeltype == LABEL_MANUAL ? style.labelfont : style.font;
+
+       //if (style.pass_thru && !d->onlyText(buf, outerfont, initial))
+       //      os << "]]>";
+
+       // parsing main loop
+       for (pos_type i = initial; i < size(); ++i) {
+               Font font = getFont(buf.params(), i, outerfont);
+
+               // emphasis
+               if (font_old.emph() != font.fontInfo().emph()) {
+                       if (font.fontInfo().emph() == FONT_ON) {
+                               os << "<em>";
+                               emph_flag = true;
+                       } else if (emph_flag && i != initial) {
+                               os << "</em>";
+                               emph_flag = false;
+                       }
+               }
+               // bold
+               if (font_old.series() != font.fontInfo().series()) {
+                       if (font.fontInfo().series() == BOLD_SERIES) {
+                               os << "<strong>";
+                               bold_flag = true;
+                       } else if (bold_flag && i != initial) {
+                               os << "</strong>";
+                               bold_flag = false;
+                       }
+               }
+               // FIXME Other such tags? 
+
+               if (Inset const * inset = getInset(i)) {
+                       inset->xhtml(os, runparams);
+               } else {
+                       char_type c = d->text_[i];
+
+                       if (style.pass_thru)
+                               os.put(c);
+                       else if (c == '-') {
+                               int j = i + 1;
+                               if (j < size() && d->text_[j] == '-') {
+                                       j += 1;
+                                       if (j < size() && d->text_[j] == '-') {
+                                               os << from_ascii("&mdash;");
+                                               i += 2;
+                                       } else {
+                                               os << from_ascii("&ndash;");
+                                               i += 1;
+                                       }
+                               }
+                               else
+                                       os << c;
+                       } else
+                               os << html::escapeChar(c);
+               }
+               font_old = font.fontInfo();
+       }
+
+       // FIXME This could be out of order. See above.
+       if (emph_flag)
+               os << "</em>";
+       if (bold_flag)
+               os << "</strong>";
+
+//     if (style.free_spacing)
+//             os << '\n';
+//     if (style.pass_thru && !d->onlyText(buf, outerfont, initial))
+//             os << "<![CDATA[";
+}
+
+
 bool Paragraph::isHfill(pos_type pos) const
 {
        Inset const * inset = getInset(pos);
@@ -2338,7 +2465,12 @@ bool Paragraph::isLetter(pos_type pos) const
        if (Inset const * inset = getInset(pos))
                return inset->isLetter();
        char_type const c = d->text_[pos];
-       return isLetterChar(c) || isDigit(c);
+    // We want to pass the ' and escape chars to the spellchecker
+       static docstring const quote = from_utf8(lyxrc.spellchecker_esc_chars + '\'');
+       return (isLetterChar(c) || isDigit(c) || contains(quote, c))
+               && (!d->inset_owner_ || d->inset_owner_->allowSpellCheck())
+               && pos != size()
+               && !isDeleted(pos);
 }
 
 
@@ -2429,7 +2561,30 @@ docstring Paragraph::asString(pos_type beg, pos_type end, int options) const
                    || (c == '\n' && options & AS_STR_NEWLINES))
                        os.put(c);
                else if (c == META_INSET && options & AS_STR_INSETS)
-                       getInset(i)->textString(os);
+                       getInset(i)->tocString(os);
+       }
+
+       return os.str();
+}
+
+
+docstring Paragraph::stringify(pos_type beg, pos_type end, int options, OutputParams & runparams) const
+{
+       odocstringstream os;
+
+       if (beg == 0 
+               && options & AS_STR_LABEL
+               && !d->params_.labelString().empty())
+               os << d->params_.labelString() << ' ';
+
+       for (pos_type i = beg; i < end; ++i) {
+               char_type const c = d->text_[i];
+               if (isPrintable(c) || c == '\t'
+                   || (c == '\n' && options & AS_STR_NEWLINES))
+                       os.put(c);
+               else if (c == META_INSET && options & AS_STR_INSETS) {
+                       getInset(i)->plaintext(os, runparams);
+               }
        }
 
        return os.str();
@@ -2484,7 +2639,7 @@ void Paragraph::setPlainOrDefaultLayout(DocumentClass const & tclass)
 Inset const & Paragraph::inInset() const
 {
        LASSERT(d->inset_owner_, throw ExceptionMessage(BufferException,
-               _("Memory problem"), _("Paragraph not properly initiliazed")));
+               _("Memory problem"), _("Paragraph not properly initialized")));
        return *d->inset_owner_;
 }
 
@@ -2831,35 +2986,76 @@ void Paragraph::deregisterWords()
 }
 
 
-void Paragraph::collectWords(CursorSlice const & sl)
+void Paragraph::locateWord(pos_type & from, pos_type & to,
+       word_location const loc) const
 {
-       // find new words
-       bool inword = false;
+       switch (loc) {
+       case WHOLE_WORD_STRICT:
+               if (from == 0 || from == size()
+                   || !isLetter(from)
+                   || !isLetter(from - 1)) {
+                       to = from;
+                       return;
+               }
+               // no break here, we go to the next
+
+       case WHOLE_WORD:
+               // If we are already at the beginning of a word, do nothing
+               if (!from || !isLetter(from - 1))
+                       break;
+               // no break here, we go to the next
+
+       case PREVIOUS_WORD:
+               // always move the cursor to the beginning of previous word
+               while (from && isLetter(from - 1))
+                       --from;
+               break;
+       case NEXT_WORD:
+               LYXERR0("Paragraph::locateWord: NEXT_WORD not implemented yet");
+               break;
+       case PARTIAL_WORD:
+               // no need to move the 'from' cursor
+               break;
+       }
+       to = from;
+       while (to < size() && isLetter(to))
+               ++to;
+}
+
+
+void Paragraph::collectWords()
+{
+       SpellChecker * speller = theSpellChecker();
 
        //lyxerr << "Words: ";
        pos_type n = size();
-       for (pos_type pos = 0; pos != n; ++pos) {
-               if (isDeleted(pos))
+       for (pos_type pos = 0; pos < n; ++pos) {
+               if (!isLetter(pos))
                        continue;
-
-               if (!isLetter(pos)) {
-                       inword = false;
+               pos_type from = pos;
+               locateWord(from, pos, WHOLE_WORD);
+               if (!lyxrc.spellcheck_continuously && pos - from < 6)
                        continue;
-               }
 
-               if (inword)
-                       continue;
+               docstring word = asString(from, pos, false);
+               if (pos - from >= 6)
+                       d->words_.insert(word);
 
-               inword = true;
-               CursorSlice from = sl;
-               CursorSlice to = sl;
-               from.pos() = pos;
-               to.pos() = pos;
-               from.text()->getWord(from, to, WHOLE_WORD);
-               if (to.pos() - from.pos() < 6)
+               if (!lyxrc.spellcheck_continuously || !speller)
                        continue;
-               docstring word = asString(from.pos(), to.pos(), false);
-               d->words_.insert(word);
+               
+               string lang_code = lyxrc.spellchecker_use_alt_lang
+                     ? lyxrc.spellchecker_alt_lang
+                     : getFontSettings(d->inset_owner_->buffer().params(), from).language()->code();
+               WordLangTuple wl(word, lang_code);
+               SpellChecker::Result res = speller->check(wl);
+               // ... just ignore any error that the spellchecker reports.
+               if (!speller->error().empty())
+                       continue;
+               bool const misspelled = res != SpellChecker::OK
+                       && res != SpellChecker::IGNORED_WORD;
+               d->fontlist_.setMisspelled(from, pos, misspelled);
+
                //lyxerr << word << " ";
        }
        //lyxerr << std::endl;
@@ -2875,12 +3071,39 @@ void Paragraph::registerWords()
 }
 
 
-void Paragraph::updateWords(CursorSlice const & sl)
+void Paragraph::updateWords()
 {
-       LASSERT(&sl.paragraph() == this, /**/);
        deregisterWords();
-       collectWords(sl);
+       collectWords();
        registerWords();
 }
 
+
+bool Paragraph::isMisspelled(pos_type pos) const
+{
+       SpellChecker * speller = theSpellChecker();
+       pos_type from = pos;
+       pos_type to = pos;
+       locateWord(from, to, WHOLE_WORD);
+       docstring word = asString(from, to, false);
+       if (!speller)
+               return false;
+               
+       string lang_code = lyxrc.spellchecker_use_alt_lang
+             ? lyxrc.spellchecker_alt_lang
+             : getFontSettings(d->inset_owner_->buffer().params(), from).language()->code();
+       WordLangTuple wl(word, lang_code);
+       SpellChecker::Result res = speller->check(wl);
+       // ... just ignore any error that the spellchecker reports.
+       if (!speller->error().empty())
+               return false;
+
+       bool const misspelled = res != SpellChecker::OK
+               && res != SpellChecker::IGNORED_WORD;
+       if (lyxrc.spellcheck_continuously)
+               d->fontlist_.setMisspelled(from, pos, misspelled);
+       return misspelled;
+}
+
+
 } // namespace lyx