X-Git-Url: https://git.lyx.org/gitweb/?a=blobdiff_plain;f=src%2FParagraph.cpp;h=5bad896dfdf4e13f0a09bc9d1597e52290332e16;hb=26821705565c9d7bd8b534b86580c62d896c5200;hp=c9a234d01fbed9aefa93e7292d3fe91130edfffd;hpb=6dc1884d05ee0d51dd896c7b924fdec758ea3d8e;p=lyx.git diff --git a/src/Paragraph.cpp b/src/Paragraph.cpp index c9a234d01f..5bad896dfd 100644 --- a/src/Paragraph.cpp +++ b/src/Paragraph.cpp @@ -314,13 +314,15 @@ public: /// Output the surrogate pair formed by \p c and \p next to \p os. /// \return the number of characters written. - int latexSurrogatePair(otexstream & os, char_type c, char_type next, + int latexSurrogatePair(BufferParams const &, otexstream & os, + char_type c, char_type next, OutputParams const &); /// Output a space in appropriate formatting (or a surrogate pair /// if the next character is a combining character). /// \return whether a surrogate pair was output. - bool simpleTeXBlanks(OutputParams const &, + bool simpleTeXBlanks(BufferParams const &, + OutputParams const &, otexstream &, pos_type i, unsigned int & column, @@ -330,8 +332,11 @@ public: /// Output consecutive unicode chars, belonging to the same script as /// specified by the latex macro \p ltx, to \p os starting from \p i. /// \return the number of characters written. - int writeScriptChars(otexstream & os, docstring const & ltx, - Change const &, Encoding const &, pos_type & i); + int writeScriptChars(BufferParams const &, OutputParams const &, + otexstream & os, + docstring const & ltx, + Change const &, Encoding const &, + std::string const, pos_type & i); /// This could go to ParagraphParameters if we want to. int startTeXParParams(BufferParams const &, otexstream &, @@ -865,8 +870,9 @@ int Paragraph::eraseChars(pos_type start, pos_type end, bool trackChanges) } -int Paragraph::Private::latexSurrogatePair(otexstream & os, char_type c, - char_type next, OutputParams const & runparams) +int Paragraph::Private::latexSurrogatePair(BufferParams const & bparams, + otexstream & os, char_type c, char_type next, + OutputParams const & runparams) { // Writing next here may circumvent a possible font change between // c and next. Since next is only output if it forms a surrogate pair @@ -883,23 +889,63 @@ int Paragraph::Private::latexSurrogatePair(otexstream & os, char_type c, latex1 = from_ascii(tipashortcut); } } - docstring const latex2 = encoding.latexChar(c).first; + docstring latex2 = encoding.latexChar(c).first; + if (docstring(1, next) == latex1) { - // the encoding supports the combination + // The encoding supports the combination: + // output as is (combining char after base char). os << latex2 << latex1; return latex1.length() + latex2.length(); - } else if (runparams.local_font && - runparams.local_font->language()->lang() == "polutonikogreek") { - // polutonikogreek only works without the brackets - os << latex1 << latex2; - return latex1.length() + latex2.length(); - } else - os << latex1 << '{' << latex2 << '}'; - return latex1.length() + latex2.length() + 2; + } + + // Handle combining characters in "script" context (i.e., \textgreek and \textcyrillic) + docstring::size_type const brace1 = latex2.find_first_of(from_ascii("{")); + docstring::size_type const brace2 = latex2.find_last_of(from_ascii("}")); + string script = to_ascii(latex2.substr(1, brace1 - 1)); + // "Script chars" need to embraced in \textcyrillic and \textgreek notwithstanding + // whether they are encodable or not (it only depends on the font encoding) + if (!runparams.isFullUnicode()) + // This will get us a script value to deal with below + Encodings::isKnownScriptChar(c, script); + int pos = 0; + int length = brace2; + string fontenc; + if (runparams.local_font) + fontenc = runparams.local_font->language()->fontenc(bparams); + else + fontenc = bparams.language->fontenc(bparams); + docstring scriptmacro; + docstring cb; + if (script == "textgreek" || script == "textcyrillic") { + // We separate the script macro (\text[greek|cyr]) from the rest, + // since we need to include the combining char in it (#6463). + // This is "the rest": + pos = brace1 + 1; + length -= pos; + latex2 = latex2.substr(pos, length); + // We only need the script macro with non-native font encodings + if (Encodings::needsScriptWrapper(script, fontenc)) { + scriptmacro = from_ascii("\\" + script + "{"); + cb = from_ascii("}"); + } + } + + docstring lb; + docstring rb; + // polutonikogreek does not play nice with brackets + if (!runparams.local_font + || runparams.local_font->language()->lang() != "polutonikogreek") { + lb = from_ascii("{"); + rb = from_ascii("}"); + } + + os << scriptmacro << latex1 << lb << latex2 << rb << cb; + return latex1.length() + latex2.length() + lb.length() + rb.length() + cb.length(); } -bool Paragraph::Private::simpleTeXBlanks(OutputParams const & runparams, +bool Paragraph::Private::simpleTeXBlanks(BufferParams const & bparams, + OutputParams const & runparams, otexstream & os, pos_type i, unsigned int & column, @@ -913,7 +959,7 @@ bool Paragraph::Private::simpleTeXBlanks(OutputParams const & runparams, char_type next = text_[i + 1]; if (Encodings::isCombiningChar(next)) { // This space has an accent, so we must always output it. - column += latexSurrogatePair(os, ' ', next, runparams) - 1; + column += latexSurrogatePair(bparams, os, ' ', next, runparams) - 1; return true; } } @@ -944,10 +990,13 @@ bool Paragraph::Private::simpleTeXBlanks(OutputParams const & runparams, } -int Paragraph::Private::writeScriptChars(otexstream & os, +int Paragraph::Private::writeScriptChars(BufferParams const & bparams, + OutputParams const & runparams, + otexstream & os, docstring const & ltx, Change const & runningChange, Encoding const & encoding, + string const fontenc, pos_type & i) { // FIXME: modifying i here is not very nice... @@ -955,10 +1004,6 @@ int Paragraph::Private::writeScriptChars(otexstream & os, // We only arrive here when character text_[i] could not be translated // into the current latex encoding (or its latex translation has been forced,) // and it belongs to a known script. - // TODO: We need \textcyr and \textgreek wrappers also for characters - // that can be encoded in the "LaTeX encoding" but not in the - // current *font encoding*. - // (See #9681 for details and test) // Parameter ltx contains the latex translation of text_[i] as specified // in the unicodesymbols file and is something like "\textXXX{}". // The latex macro name "textXXX" specifies the script to which text_[i] @@ -972,12 +1017,8 @@ int Paragraph::Private::writeScriptChars(otexstream & os, int pos = 0; int length = brace2; bool closing_brace = true; - if (script == "textgreek" && encoding.latexName() == "iso-8859-7") { - // Correct encoding is being used, so we can avoid \textgreek. - // TODO: wrong test: we need to check the *font encoding* - // (i.e. the active language and its FontEncoding tag) - // instead of the LaTeX *input encoding*! - // See #9637 for details and test-cases. + if (!Encodings::needsScriptWrapper(script, fontenc)) { + // Correct font encoding is being used, so we can avoid \text[greek|cyr]. pos = brace1 + 1; length -= pos; closing_brace = false; @@ -1006,6 +1047,18 @@ int Paragraph::Private::writeScriptChars(otexstream & os, // Stop here if there is a font attribute or encoding change. if (found && cit != end && prev_font != cit->font()) break; + + // Check whether we have a combining pair + char_type next_next = '\0'; + if (i + 2 < size) { + next_next = text_[i + 2]; + if (Encodings::isCombiningChar(next_next)) { + length += latexSurrogatePair(bparams, os, next, next_next, runparams) - 1; + i += 2; + continue; + } + } + docstring const latex = encoding.latexChar(next).first; docstring::size_type const b1 = latex.find_first_of(from_ascii("{")); @@ -1090,13 +1143,15 @@ void Paragraph::Private::latexInset(BufferParams const & bparams, odocstream::pos_type const len = os.os().tellp(); if (inset->forceLTR() - && !runparams.use_polyglossia && 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") - os << "\\beginL" << termcmd; + if (runparams.use_polyglossia) { + os << "\\LRE{"; + } else if (running_font.language()->lang() == "farsi" + || running_font.language()->lang() == "arabic_arabi") + os << "\\textLR{" << termcmd; else os << "\\L{"; close = true; @@ -1152,12 +1207,8 @@ void Paragraph::Private::latexInset(BufferParams const & bparams, throw(e); } - if (close) { - if (running_font.language()->lang() == "farsi") - os << "\\endL" << termcmd; - else - os << '}'; - } + if (close) + os << '}'; if (os.texrow().rows() > previous_row_count) { os.texrow().start(owner_->id(), i + 1); @@ -1181,10 +1232,7 @@ void Paragraph::Private::latexSpecialChar(otexstream & os, pos_type end_pos, unsigned int & column) { - // With polyglossia, brackets and stuff need not be reversed - // in RTL scripts (see bug #8251) - char_type const c = (runparams.use_polyglossia) ? - owner_->getUChar(bparams, i) : text_[i]; + char_type const c = owner_->getUChar(bparams, runparams, i); if (style.pass_thru || runparams.pass_thru || contains(style.pass_thru_chars, c) @@ -1238,10 +1286,8 @@ void Paragraph::Private::latexSpecialChar(otexstream & os, if (i + 1 < static_cast(text_.size()) && (end_pos == -1 || i + 1 < end_pos) && text_[i+1] == '-') { - // Prevent "--" becoming an endash and "---" becoming - // an emdash. - // Within \ttfamily, "--" is merged to "-" (no endash) - // so we avoid this rather irritating ligature as well + // Prevent "--" becoming an en dash and "---" an em dash. + // (Within \ttfamily, "---" is merged to en dash + hyphen.) os << "{}"; column += 2; } @@ -1289,7 +1335,13 @@ void Paragraph::Private::latexSpecialChar(otexstream & os, case 0x2013: case 0x2014: - if (bparams.use_dash_ligatures && !bparams.useNonTeXFonts) { + // XeTeX's dash behaviour is determined via a global setting + if (bparams.use_dash_ligatures + && owner_->getFontSettings(bparams, i).fontInfo().family() != TYPEWRITER_FAMILY + && !runparams.inIPA + // TODO #10961: && not in inset Flex Code + // TODO #10961: && not in layout LyXCode + && (!bparams.useNonTeXFonts || runparams.flavor != OutputParams::XETEX)) { if (c == 0x2013) { // en-dash os << "--"; @@ -1311,7 +1363,7 @@ void Paragraph::Private::latexSpecialChar(otexstream & os, if (i + 1 < int(text_.size())) { next = text_[i + 1]; if (Encodings::isCombiningChar(next)) { - column += latexSurrogatePair(os, c, next, runparams) - 1; + column += latexSurrogatePair(bparams, os, c, next, runparams) - 1; ++i; break; } @@ -1337,11 +1389,18 @@ void Paragraph::Private::latexSpecialChar(otexstream & os, tipas = true; } } - if (Encodings::isKnownScriptChar(c, script) - && prefixIs(latex.first, from_ascii("\\" + script))) - column += writeScriptChars(os, latex.first, - running_change, encoding, i) - 1; - else if (latex.second + string fontenc; + fontenc = running_font.language()->fontenc(bparams); + // "Script chars" need to embraced in \textcyrillic and \textgreek notwithstanding + // whether they are encodable or not (it only depends on the font encoding) + if (!runparams.isFullUnicode() && Encodings::isKnownScriptChar(c, script)) { + docstring const wrapper = from_ascii("\\" + script + "{"); + docstring ltx = latex.first; + if (!prefixIs(ltx, wrapper)) + ltx = wrapper + latex.first + from_ascii("}"); + column += writeScriptChars(bparams, runparams, os, ltx, running_change, + encoding, fontenc, i) - 1; + } else if (latex.second && ((!prefixIs(nextlatex, '\\') && !prefixIs(nextlatex, '{') && !prefixIs(nextlatex, '}')) @@ -1426,43 +1485,18 @@ bool Paragraph::Private::latexSpecialT3(char_type const c, otexstream & os, void Paragraph::Private::validate(LaTeXFeatures & features) const { if (layout_->inpreamble && inset_owner_) { - bool const is_command = layout_->latextype == LATEX_COMMAND; - Buffer const & buf = inset_owner_->buffer(); - BufferParams const & bp = features.runparams().is_child - ? buf.masterParams() : buf.params(); - Font f; - // Using a string stream here circumvents the encoding + // FIXME: Using a string stream here circumvents the encoding // switching machinery of odocstream. Therefore the // output is wrong if this paragraph contains content // that needs to switch encoding. + Buffer const & buf = inset_owner_->buffer(); otexstringstream os; os << layout_->preamble(); - if (is_command) { - os << '\\' << from_ascii(layout_->latexname()); - // we have to provide all the optional arguments here, even though - // the last one is the only one we care about. - // Separate handling of optional argument inset. - if (!layout_->latexargs().empty()) { - OutputParams rp = features.runparams(); - rp.local_font = &owner_->getFirstFontSettings(bp); - latexArgInsets(*owner_, os, rp, layout_->latexargs()); - } - os << from_ascii(layout_->latexparam()); - } size_t const length = os.length(); - // this will output "{" at the beginning, but not at the end - owner_->latex(bp, f, os, features.runparams(), 0, -1, true); - if (os.length() > length) { - if (is_command) { - os << '}'; - if (!layout_->postcommandargs().empty()) { - OutputParams rp = features.runparams(); - rp.local_font = &owner_->getFirstFontSettings(bp); - latexArgInsets(*owner_, os, rp, layout_->postcommandargs(), "post:"); - } - } + TeXOnePar(buf, buf.text(), buf.getParFromID(owner_->id()).pit(), os, + features.runparams(), string(), 0, -1, true); + if (os.length() > length) features.addPreambleSnippet(os.release(), true); - } } if (features.runparams().flavor == OutputParams::HTML @@ -1490,6 +1524,16 @@ void Paragraph::Private::validate(LaTeXFeatures & features) const for (; icit != iend; ++icit) { if (icit->inset) { features.inDeletedInset(owner_->isDeleted(icit->pos)); + if (icit->inset->lyxCode() == FOOT_CODE) { + // FIXME: an item inset would make things much easier. + if ((layout_->latextype == LATEX_LIST_ENVIRONMENT + || (layout_->latextype == LATEX_ITEM_ENVIRONMENT + && layout_->margintype == MARGIN_FIRST_DYNAMIC)) + && (icit->pos < begin_of_body_ + || (icit->pos == begin_of_body_ + && (icit->pos == 0 || text_[icit->pos - 1] != ' ')))) + features.saveNoteEnv("description"); + } icit->inset->validate(features); features.inDeletedInset(false); if (layout_->needprotect && @@ -1499,16 +1543,60 @@ void Paragraph::Private::validate(LaTeXFeatures & features) const } // then the contents - BufferParams const bp = features.buffer().masterParams(); + BufferParams const bp = features.runparams().is_child + ? features.buffer().masterParams() : features.buffer().params(); for (pos_type i = 0; i < int(text_.size()) ; ++i) { char_type c = text_[i]; + CharInfo const & ci = Encodings::unicodeCharInfo(c); if (c == 0x0022) { if (features.runparams().isFullUnicode() && bp.useNonTeXFonts) features.require("textquotedblp"); else if (bp.main_font_encoding() != "T1" || ((&owner_->getFontSettings(bp, i))->language()->internalFontEncoding())) features.require("textquotedbl"); - } + } else if (ci.textfeature() && contains(ci.textpreamble(), '=')) { + // features that depend on the font or input encoding + string feats = ci.textpreamble(); + string fontenc = (&owner_->getFontSettings(bp, i))->language()->fontenc(bp); + if (fontenc.empty()) + fontenc = features.runparams().main_fontenc; + while (!feats.empty()) { + string feat; + feats = split(feats, feat, ','); + if (contains(feat, "!=")) { + // a feature that is required except for the spcified + // font or input encodings + string realfeature; + string const contexts = ltrim(split(feat, realfeature, '!'), "="); + // multiple encodings are separated by semicolon + vector context = getVectorFromString(contexts, ";"); + // require feature if the context matches neither current font + // nor input encoding + if (std::find(context.begin(), context.end(), fontenc) == context.end() + && std::find(context.begin(), context.end(), + features.runparams().encoding->name()) == context.end()) + features.require(realfeature); + } else if (contains(feat, '=')) { + // a feature that is required only for the spcified + // font or input encodings + string realfeature; + string const contexts = split(feat, realfeature, '='); + // multiple encodings are separated by semicolon + vector context = getVectorFromString(contexts, ";"); + // require feature if the context matches either current font + // or input encoding + if (std::find(context.begin(), context.end(), fontenc) != context.end() + || std::find(context.begin(), context.end(), + features.runparams().encoding->name()) != context.end()) + features.require(realfeature); + } + } + } else if (!bp.use_dash_ligatures + && (c == 0x2013 || c == 0x2014) + && bp.useNonTeXFonts + && features.runparams().flavor == OutputParams::XETEX) + // XeTeX's dash behaviour is determined via a global setting + features.require("xetexdashbreakstate"); BufferEncodings::validate(c, features); } } @@ -1698,6 +1786,10 @@ void Paragraph::write(ostream & os, BufferParams const & bparams, void Paragraph::validate(LaTeXFeatures & features) const { d->validate(features); + bool fragile = features.runparams().moving_arg; + fragile |= layout().needprotect; + if (needsCProtection(fragile)) + features.require("cprotect"); } @@ -1900,32 +1992,57 @@ Font const Paragraph::getLayoutFont } -char_type Paragraph::getUChar(BufferParams const & bparams, pos_type pos) const +char_type Paragraph::getUChar(BufferParams const & bparams, + OutputParams const & rp, + pos_type pos) const { char_type c = d->text_[pos]; - if (!getFontSettings(bparams, pos).isRightToLeft()) + + // Return unchanged character in LTR languages + // or if we use poylglossia/bidi. + if (rp.use_polyglossia || !getFontSettings(bparams, pos).isRightToLeft()) return c; - // FIXME: The arabic special casing is due to the difference of arabic - // round brackets input introduced in r18599. Check if this should be - // unified with Hebrew or at least if all bracket types should be - // handled the same (file format change in either case). + // Without polyglossia/bidi, we need to account for some special cases. + // FIXME This needs to be audited! + // Check if: + // * The input is as expected for all delimiters + // => checked for Hebrew! + // * The output matches the display in the LyX workarea + // => checked for Hebrew! + // * The special cases below are really necessary + // => checked for Hebrew! + // * In arabic_arabi, brackets are transformed to Arabic + // Ornate Parentheses. Is this is really wanted? + string const & lang = getFontSettings(bparams, pos).language()->lang(); - bool const arabic = lang == "arabic_arabtex" || lang == "arabic_arabi" - || lang == "farsi"; char_type uc = c; + + // 1. In the following languages, parentheses need to be reversed. + bool const reverseparens = lang == "hebrew"; + + // 2. In the following languages, brackets don't need to be reversed. + bool const reversebrackets = lang != "arabic_arabtex" + && lang != "arabic_arabi" + && lang != "farsi"; + + // Now swap delimiters if needed. switch (c) { case '(': - uc = arabic ? c : ')'; + if (reverseparens) + uc = ')'; break; case ')': - uc = arabic ? c : '('; + if (reverseparens) + uc = '('; break; case '[': - uc = ']'; + if (reversebrackets) + uc = ']'; break; case ']': - uc = '['; + if (reversebrackets) + uc = '['; break; case '{': uc = '}'; @@ -2261,13 +2378,13 @@ int Paragraph::Private::startTeXParParams(BufferParams const & bparams, case LYX_ALIGN_DECIMAL: break; case LYX_ALIGN_LEFT: { - if (owner_->getParLanguage(bparams)->babel() != "hebrew") + if (!owner_->getParLanguage(bparams)->rightToLeft()) corrected_env(os, begin_tag, "flushleft", code, lastpar, column); else corrected_env(os, begin_tag, "flushright", code, lastpar, column); break; } case LYX_ALIGN_RIGHT: { - if (owner_->getParLanguage(bparams)->babel() != "hebrew") + if (!owner_->getParLanguage(bparams)->rightToLeft()) corrected_env(os, begin_tag, "flushright", code, lastpar, column); else corrected_env(os, begin_tag, "flushleft", code, lastpar, column); @@ -2319,13 +2436,13 @@ bool Paragraph::Private::endTeXParParams(BufferParams const & bparams, case LYX_ALIGN_DECIMAL: break; case LYX_ALIGN_LEFT: { - if (owner_->getParLanguage(bparams)->babel() != "hebrew") + if (!owner_->getParLanguage(bparams)->rightToLeft()) output = corrected_env(os, end_tag, "flushleft", code, lastpar, col); else output = corrected_env(os, end_tag, "flushright", code, lastpar, col); break; } case LYX_ALIGN_RIGHT: { - if (owner_->getParLanguage(bparams)->babel() != "hebrew") + if (!owner_->getParLanguage(bparams)->rightToLeft()) output = corrected_env(os, end_tag, "flushright", code, lastpar, col); else output = corrected_env(os, end_tag, "flushleft", code, lastpar, col); @@ -2398,7 +2515,9 @@ void Paragraph::latex(BufferParams const & bparams, // if the paragraph is empty, the loop will not be entered at all if (empty()) { - if (style.isCommand()) { + // For InTitle commands, we have already opened a group + // in output_latex::TeXOnePar. + if (style.isCommand() && !style.intitle) { os << '{'; ++column; } @@ -2436,7 +2555,9 @@ void Paragraph::latex(BufferParams const & bparams, os << "}] "; column +=3; } - if (style.isCommand()) { + // For InTitle commands, we have already opened a group + // in output_latex::TeXOnePar. + if (style.isCommand() && !style.intitle) { os << '{'; ++column; } @@ -2454,6 +2575,8 @@ void Paragraph::latex(BufferParams const & bparams, runparams.wasDisplayMath = runparams.inDisplayMath; runparams.inDisplayMath = false; bool deleted_display_math = false; + Change const & change = runparams.inDeletedInset + ? runparams.changeOfDeletedInset : lookupChange(i); // Check whether a display math inset follows if (d->text_[i] == META_INSET @@ -2468,11 +2591,28 @@ void Paragraph::latex(BufferParams const & bparams, // cannot set it here because it is a counter. deleted_display_math = isDeleted(i); } + if (bparams.output_changes && deleted_display_math + && runningChange == change + && change.type == Change::DELETED + && !os.afterParbreak()) { + // A display math in the same paragraph follows. + // We have to close and then reopen \lyxdeleted, + // otherwise the math will be shifted up. + OutputParams rp = runparams; + if (open_font) { + bool needPar = false; + column += running_font.latexWriteEndChanges( + os, bparams, rp, basefont, + basefont, needPar); + open_font = false; + } + basefont = getLayoutFont(bparams, outerfont); + running_font = basefont; + column += Changes::latexMarkChange(os, bparams, + Change(Change::INSERTED), change, rp); + } } - Change const & change = runparams.inDeletedInset - ? runparams.changeOfDeletedInset : lookupChange(i); - if (bparams.output_changes && runningChange != change) { if (open_font) { bool needPar = false; @@ -2523,6 +2663,7 @@ void Paragraph::latex(BufferParams const & bparams, bool const using_begin_end = runparams.use_polyglossia || !lang_end_command.empty(); if (!running_lang.empty() && + (!using_begin_end || running_lang == openLanguageName()) && current_font.language()->encoding()->package() == Encoding::CJK) { string end_tag = subst(lang_end_command, "$$lang", @@ -2549,7 +2690,7 @@ void Paragraph::latex(BufferParams const & bparams, char_type const c = d->text_[i]; // A display math inset inside an ulem command will be output - // as a box of width \columnwidth, so we have to either disable + // as a box of width \linewidth, so we have to either disable // indentation if the inset starts a paragraph, or start a new // line to accommodate such box. This has to be done before // writing any font changing commands. @@ -2604,7 +2745,7 @@ void Paragraph::latex(BufferParams const & bparams, // latexSpecialChar ignores spaces if // style.pass_thru is false. if (i != body_pos - 1) { - if (d->simpleTeXBlanks(runparams, os, + if (d->simpleTeXBlanks(bparams, runparams, os, i, column, current_font, style)) { // A surrogate pair was output. We // must not call latexSpecialChar @@ -2641,7 +2782,7 @@ void Paragraph::latex(BufferParams const & bparams, ParagraphList const & pars = textinset->text().paragraphs(); pit_type const pit = pars.size() - 1; - Font const last_font = + Font const lastfont = pit < 0 || pars[pit].empty() ? pars[pit].getLayoutFont( bparams, @@ -2649,7 +2790,7 @@ void Paragraph::latex(BufferParams const & bparams, : pars[pit].getFont(bparams, pars[pit].size() - 1, outerfont); - if (last_font.fontInfo().size() != + if (lastfont.fontInfo().size() != basefont.fontInfo().size()) { ++parInline; incremented = true; @@ -3266,7 +3407,8 @@ docstring Paragraph::simpleLyXHTMLOnePar(Buffer const & buf, retval += inset->xhtml(xs, np); } } else { - char_type c = getUChar(buf.masterBuffer()->params(), i); + char_type c = getUChar(buf.masterBuffer()->params(), + runparams, i); xs << c; } font_old = font.fontInfo(); @@ -3356,6 +3498,41 @@ bool Paragraph::isHardHyphenOrApostrophe(pos_type pos) const } +bool Paragraph::needsCProtection(bool const fragile) const +{ + // first check the layout of the paragraph, but only in insets + InsetText const * textinset = inInset().asInsetText(); + bool const maintext = textinset + ? textinset->text().isMainText() + : false; + + if (!maintext && layout().needcprotect) { + // Environments need cprotection regardless the content + if (layout().latextype == LATEX_ENVIRONMENT) + return true; + + // Commands need cprotection if they contain specific chars + int const nchars_escape = 9; + static char_type const chars_escape[nchars_escape] = { + '&', '_', '$', '%', '#', '^', '{', '}', '\\'}; + + docstring const pars = asString(); + for (int k = 0; k < nchars_escape; k++) { + if (contains(pars, chars_escape[k])) + return true; + } + } + + // now check whether we have insets that need cprotection + pos_type size = d->text_.size(); + for (pos_type i = 0; i < size; ++i) + if (isInset(i) && getInset(i)->needsCProtection(maintext, fragile)) + return true; + + return false; +} + + FontSpan const & Paragraph::getSpellRange(pos_type pos) const { return d->speller_state_.getRange(pos); @@ -3415,28 +3592,22 @@ void Paragraph::changeLanguage(BufferParams const & bparams, bool Paragraph::isMultiLingual(BufferParams const & bparams) const { Language const * doc_language = bparams.language; - FontList::const_iterator cit = d->fontlist_.begin(); - FontList::const_iterator end = d->fontlist_.end(); - - for (; cit != end; ++cit) - if (cit->font().language() != ignore_language && - cit->font().language() != latex_language && - cit->font().language() != doc_language) + for (auto const & f : d->fontlist_) + if (f.font().language() != ignore_language && + f.font().language() != latex_language && + f.font().language() != doc_language) return true; return false; } -void Paragraph::getLanguages(std::set & languages) const +void Paragraph::getLanguages(std::set & langs) const { - FontList::const_iterator cit = d->fontlist_.begin(); - FontList::const_iterator end = d->fontlist_.end(); - - for (; cit != end; ++cit) { - Language const * lang = cit->font().language(); + for (auto const & f : d->fontlist_) { + Language const * lang = f.font().language(); if (lang != ignore_language && lang != latex_language) - languages.insert(lang); + langs.insert(lang); } } @@ -3483,6 +3654,8 @@ void Paragraph::forOutliner(docstring & os, size_t const maxlen, size_t tmplen = shorten ? maxlen + 1 : maxlen; if (label && !labelString().empty()) os += labelString() + ' '; + if (!layout().isTocCaption()) + return; for (pos_type i = 0; i < size() && os.length() < tmplen; ++i) { if (isDeleted(i)) continue; @@ -3673,7 +3846,7 @@ InsetList const & Paragraph::insetList() const } -void Paragraph::setBuffer(Buffer & b) +void Paragraph::setInsetBuffers(Buffer & b) { d->insetlist_.setBuffer(b); } @@ -4079,6 +4252,9 @@ SpellChecker::Result Paragraph::spellCheck(pos_type & from, pos_type & to, docstring word = asString(from, to, AS_STR_INSETS | AS_STR_SKIPDELETE); Language * lang = d->getSpellLanguage(from); + if (getFontSettings(d->inset_owner_->buffer().params(), from).fontInfo().nospellcheck() == FONT_ON) + return result; + wl = WordLangTuple(word, lang); if (word.empty()) @@ -4128,6 +4304,15 @@ SpellChecker::Result Paragraph::spellCheck(pos_type & from, pos_type & to, } +void Paragraph::anonymize() +{ + // This is a very crude anonymization for now + for (char_type & c : d->text_) + if (isLetterChar(c) || isNumber(c)) + c = 'a'; +} + + void Paragraph::Private::markMisspelledWords( pos_type const & first, pos_type const & last, SpellChecker::Result result,