X-Git-Url: https://git.lyx.org/gitweb/?a=blobdiff_plain;f=src%2FParagraph.cpp;h=5bad896dfdf4e13f0a09bc9d1597e52290332e16;hb=26821705565c9d7bd8b534b86580c62d896c5200;hp=daac791831525dd8df6c2d3a8fcbe39a5fa47c81;hpb=31e25c8ec695f864bec3679c3e11495e3011a0e2;p=lyx.git diff --git a/src/Paragraph.cpp b/src/Paragraph.cpp index daac791831..5bad896dfd 100644 --- a/src/Paragraph.cpp +++ b/src/Paragraph.cpp @@ -52,6 +52,9 @@ #include "insets/InsetBibitem.h" #include "insets/InsetLabel.h" #include "insets/InsetSpecialChar.h" +#include "insets/InsetText.h" + +#include "mathed/InsetMathHull.h" #include "support/debug.h" #include "support/docstring_list.h" @@ -61,12 +64,22 @@ #include "support/lstrings.h" #include "support/textutils.h" +#include #include #include using namespace std; using namespace lyx::support; +// OSX clang, gcc < 4.8.0, and msvc < 2015 do not support C++11 thread_local +#if defined(__APPLE__) || (defined(__GNUC__) && __GNUC__ == 4 && __GNUC_MINOR__ < 8) +#define THREAD_LOCAL_STATIC static __thread +#elif defined(_MSC_VER) && (_MSC_VER < 1900) +#define THREAD_LOCAL_STATIC static __declspec(thread) +#else +#define THREAD_LOCAL_STATIC thread_local static +#endif + namespace lyx { namespace { @@ -74,7 +87,7 @@ namespace { /// Inset identifier (above 0x10ffff, for ucs-4) char_type const META_INSET = 0x200001; -} +} // namespace ///////////////////////////////////////////////////////////////////// @@ -283,10 +296,11 @@ private: class Paragraph::Private { - // Enforce our own "copy" constructor by declaring the standard one and - // the assignment operator private without implementing them. - Private(Private const &); - Private & operator=(Private const &); + // Enforce our own "copy" constructor + Private(Private const &) = delete; + Private & operator=(Private const &) = delete; + // Unique ID generator + static int make_id(); public: /// Private(Paragraph * owner, Layout const & layout); @@ -300,13 +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, @@ -316,8 +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 &, @@ -359,6 +378,12 @@ public: pos_type i, unsigned int & column); /// + bool latexSpecialTU( + char_type const c, + otexstream & os, + pos_type i, + unsigned int & column); + /// bool latexSpecialT3( char_type const c, otexstream & os, @@ -377,7 +402,7 @@ public: typedef SkipPositions::const_iterator SkipPositionsIterator; void appendSkipPosition(SkipPositions & skips, pos_type const pos) const; - + Language * getSpellLanguage(pos_type const from) const; Language * locateSpellRange(pos_type & from, pos_type & to, @@ -392,7 +417,7 @@ public: } bool ignoreWord(docstring const & word) const ; - + void setMisspelled(pos_type from, pos_type to, SpellChecker::Result state) { pos_type textsize = owner_->size(); @@ -506,35 +531,37 @@ Paragraph::Private::Private(Paragraph * owner, Layout const & layout) } -// Initialization of the counter for the paragraph id's, -// -// FIXME: There should be a more intelligent way to generate and use the -// paragraph ids per buffer instead a global static counter for all InsetText -// in the running program. -// However, this per-session id is used in LFUN_PARAGRAPH_GOTO to -// switch to a different buffer, as used in the outliner for instance. -static int paragraph_id = -1; +//static +int Paragraph::Private::make_id() +{ + // The id is unique per session across buffers because it is used in + // LFUN_PARAGRAPH_GOTO to switch to a different buffer, for instance in the + // outliner. + // (thread-safe) + static atomic_uint next_id(0); + return next_id++; +} + Paragraph::Private::Private(Private const & p, Paragraph * owner) : owner_(owner), inset_owner_(p.inset_owner_), fontlist_(p.fontlist_), + id_(make_id()), params_(p.params_), changes_(p.changes_), insetlist_(p.insetlist_), begin_of_body_(p.begin_of_body_), text_(p.text_), words_(p.words_), layout_(p.layout_) { - id_ = ++paragraph_id; requestSpellCheck(p.text_.size()); } Paragraph::Private::Private(Private const & p, Paragraph * owner, pos_type beg, pos_type end) - : owner_(owner), inset_owner_(p.inset_owner_), + : owner_(owner), inset_owner_(p.inset_owner_), id_(make_id()), params_(p.params_), changes_(p.changes_), insetlist_(p.insetlist_, beg, end), begin_of_body_(p.begin_of_body_), words_(p.words_), layout_(p.layout_) { - id_ = ++paragraph_id; if (beg >= pos_type(p.text_.size())) return; text_ = p.text_.substr(beg, end - beg); @@ -556,10 +583,10 @@ Paragraph::Private::Private(Private const & p, Paragraph * owner, } -void Paragraph::addChangesToToc(DocIterator const & cdit, - Buffer const & buf, bool output_active) const +void Paragraph::addChangesToToc(DocIterator const & cdit, Buffer const & buf, + bool output_active, TocBackend & backend) const { - d->changes_.addToToc(cdit, buf, output_active); + d->changes_.addToToc(cdit, buf, output_active, backend); } @@ -843,8 +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 @@ -861,23 +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, @@ -891,7 +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; } } @@ -922,10 +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... @@ -933,10 +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] @@ -950,12 +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; @@ -984,6 +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("{")); @@ -1032,9 +1107,10 @@ void Paragraph::Private::latexInset(BufferParams const & bparams, os << '\n'; } else { if (open_font) { + bool needPar = false; column += running_font.latexWriteEndChanges( os, bparams, runparams, - basefont, basefont); + basefont, basefont, needPar); open_font = false; } @@ -1067,13 +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{}"; + 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; @@ -1092,10 +1170,12 @@ void Paragraph::Private::latexInset(BufferParams const & bparams, bool arabtex = basefont.language()->lang() == "arabic_arabtex" || running_font.language()->lang() == "arabic_arabtex"; if (open_font && !inset->inheritFont()) { + bool needPar = false; bool closeLanguage = arabtex || basefont.isRightToLeft() == running_font.isRightToLeft(); unsigned int count = running_font.latexWriteEndChanges(os, - bparams, runparams, basefont, basefont, closeLanguage); + bparams, runparams, basefont, basefont, + needPar, closeLanguage); column += count; // if any font properties were closed, update the running_font, // making sure, however, to leave the language as it was @@ -1114,7 +1194,7 @@ void Paragraph::Private::latexInset(BufferParams const & bparams, } } - int prev_rows = os.texrow().rows(); + size_t const previous_row_count = os.texrow().rows(); try { runparams.lastid = id_; @@ -1127,14 +1207,10 @@ void Paragraph::Private::latexInset(BufferParams const & bparams, throw(e); } - if (close) { - if (running_font.language()->lang() == "farsi") - os << "\\endL{}"; - else - os << '}'; - } + if (close) + os << '}'; - if (os.texrow().rows() > prev_rows) { + if (os.texrow().rows() > previous_row_count) { os.texrow().start(owner_->id(), i + 1); column = 0; } else { @@ -1156,10 +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) @@ -1182,25 +1255,30 @@ void Paragraph::Private::latexSpecialChar(otexstream & os, // non-standard font encoding. If we are using such a language, // we do not output special T1 chars. if (!runparams.inIPA && !running_font.language()->internalFontEncoding() - && bparams.font_encoding() == "T1" && latexSpecialT1(c, os, i, column)) + && !runparams.isFullUnicode() && bparams.main_font_encoding() == "T1" + && latexSpecialT1(c, os, i, column)) return; + // NOTE: XeTeX and LuaTeX use EU1/2 (pre 2017) or TU (as of 2017) encoding + else if (!runparams.inIPA && !running_font.language()->internalFontEncoding() + && runparams.isFullUnicode() && latexSpecialTU(c, os, i, column)) + return; // Otherwise, we use what LaTeX provides us. switch (c) { case '\\': - os << "\\textbackslash{}"; + os << "\\textbackslash" << termcmd; column += 15; break; case '<': - os << "\\textless{}"; + os << "\\textless" << termcmd; column += 10; break; case '>': - os << "\\textgreater{}"; + os << "\\textgreater" << termcmd; column += 13; break; case '|': - os << "\\textbar{}"; + os << "\\textbar" << termcmd; column += 9; break; case '-': @@ -1208,17 +1286,15 @@ 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; } break; case '\"': - os << "\\char`\\\"{}"; - column += 9; + os << "\\textquotedbl" << termcmd; + column += 14; break; case '$': case '&': @@ -1230,12 +1306,12 @@ void Paragraph::Private::latexSpecialChar(otexstream & os, break; case '~': - os << "\\textasciitilde{}"; + os << "\\textasciitilde" << termcmd; column += 16; break; case '^': - os << "\\textasciicircum{}"; + os << "\\textasciicircum" << termcmd; column += 17; break; @@ -1257,6 +1333,27 @@ void Paragraph::Private::latexSpecialChar(otexstream & os, // written. (Asger) break; + case 0x2013: + case 0x2014: + // 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 << "--"; + column +=2; + } else { + // em-dash + os << "---"; + column +=3; + } + break; + } + // fall through default: if (c == '\0') return; @@ -1266,7 +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; } @@ -1292,11 +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, '}')) @@ -1333,7 +1437,7 @@ bool Paragraph::Private::latexSpecialT1(char_type const c, otexstream & os, // but we should avoid ligatures if (i + 1 >= int(text_.size()) || text_[i + 1] != c) return true; - os << "\\textcompwordmark{}"; + os << "\\textcompwordmark" << termcmd; column += 19; return true; case '|': @@ -1341,7 +1445,7 @@ bool Paragraph::Private::latexSpecialT1(char_type const c, otexstream & os, return true; case '\"': // soul.sty breaks with \char`\" - os << "\\textquotedbl{}"; + os << "\\textquotedbl" << termcmd; column += 14; return true; default: @@ -1350,6 +1454,14 @@ bool Paragraph::Private::latexSpecialT1(char_type const c, otexstream & os, } +bool Paragraph::Private::latexSpecialTU(char_type const c, otexstream & os, + pos_type i, unsigned int & column) +{ + // TU encoding is currently on par with T1. + return latexSpecialT1(c, os, i, column); +} + + bool Paragraph::Private::latexSpecialT3(char_type const c, otexstream & os, pos_type /*i*/, unsigned int & column) { @@ -1361,7 +1473,7 @@ bool Paragraph::Private::latexSpecialT3(char_type const c, otexstream & os, os.put(c); return true; case '|': - os << "\\textvertline{}"; + os << "\\textvertline" << termcmd; column += 14; return true; default: @@ -1373,44 +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. - odocstringstream ods; - otexstream os(ods, false); - 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()); - } - docstring::size_type const length = ods.str().length(); - // this will output "{" at the beginning, but not at the end - owner_->latex(bp, f, os, features.runparams(), 0, -1, true); - if (ods.str().length() > length) { - if (is_command) { - ods << '}'; - if (!layout_->postcommandargs().empty()) { - OutputParams rp = features.runparams(); - rp.local_font = &owner_->getFirstFontSettings(bp); - latexArgInsets(*owner_, os, rp, layout_->postcommandargs(), "post:"); - } - } - string const snippet = to_utf8(ods.str()); - features.addPreambleSnippet(snippet, true); - } + Buffer const & buf = inset_owner_->buffer(); + otexstringstream os; + os << layout_->preamble(); + size_t const length = os.length(); + 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 @@ -1437,7 +1523,19 @@ void Paragraph::Private::validate(LaTeXFeatures & features) const InsetList::const_iterator iend = insetlist_.end(); 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 && icit->inset->lyxCode() == FOOT_CODE) features.require("NeedLyXFootnoteCode"); @@ -1445,8 +1543,61 @@ void Paragraph::Private::validate(LaTeXFeatures & features) const } // then the contents + BufferParams const bp = features.runparams().is_child + ? features.buffer().masterParams() : features.buffer().params(); for (pos_type i = 0; i < int(text_.size()) ; ++i) { - BufferEncodings::validate(text_[i], features); + 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); } } @@ -1515,7 +1666,7 @@ void flushString(ostream & os, docstring & s) s.erase(); } -} +} // namespace void Paragraph::write(ostream & os, BufferParams const & bparams, @@ -1635,6 +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"); } @@ -1837,32 +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 = '}'; @@ -1942,7 +2122,7 @@ depth_type Paragraph::getMaxDepthAfter() const } -char Paragraph::getAlign() const +LyXAlignment Paragraph::getAlign() const { if (d->params_.align() == LYX_ALIGN_LAYOUT) return d->layout_->align; @@ -2146,7 +2326,7 @@ bool corrected_env(otexstream & os, string const & suffix, string const & env, return true; } -} // namespace anon +} // namespace int Paragraph::Private::startTeXParParams(BufferParams const & bparams, @@ -2198,13 +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); @@ -2256,13 +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); @@ -2335,7 +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; } @@ -2347,14 +2529,19 @@ void Paragraph::latex(BufferParams const & bparams, column += d->startTeXParParams(bparams, os, runparams); } + // Whether a \par can be issued for insets typeset inline with text. + // Yes if greater than 0. This has to be static. + THREAD_LOCAL_STATIC int parInline = 0; + for (pos_type i = 0; i < size(); ++i) { // First char in paragraph or after label? if (i == body_pos) { if (body_pos > 0) { if (open_font) { + bool needPar = false; column += running_font.latexWriteEndChanges( os, bparams, runparams, - basefont, basefont); + basefont, basefont, needPar); open_font = false; } basefont = getLayoutFont(bparams, outerfont); @@ -2368,7 +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; } @@ -2383,18 +2572,57 @@ void Paragraph::latex(BufferParams const & bparams, runparams); } + 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 + && i >= start_pos && (end_pos == -1 || i < end_pos)) { + InsetMath const * im = getInset(i)->asInsetMath(); + if (im && im->asHullInset() + && im->asHullInset()->outerDisplay()) { + runparams.inDisplayMath = true; + // runparams.inDeletedInset will be set by + // latexInset later, but we need this info + // before it is called. On the other hand, we + // 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); + } + } + if (bparams.output_changes && runningChange != change) { if (open_font) { + bool needPar = false; column += running_font.latexWriteEndChanges( - os, bparams, runparams, basefont, basefont); + os, bparams, runparams, + basefont, basefont, needPar); open_font = false; } basefont = getLayoutFont(bparams, outerfont); running_font = basefont; - column += Changes::latexMarkChange(os, bparams, runningChange, change, runparams); runningChange = change; @@ -2409,18 +2637,20 @@ void Paragraph::latex(BufferParams const & bparams, ++column; // Fully instantiated font - Font const font = getFont(bparams, i, outerfont); + Font const current_font = getFont(bparams, i, outerfont); Font const last_font = running_font; // Do we need to close the previous font? if (open_font && - (font != running_font || - font.language() != running_font.language())) + (current_font != running_font || + current_font.language() != running_font.language())) { + bool needPar = false; column += running_font.latexWriteEndChanges( - os, bparams, runparams, basefont, - (i == body_pos-1) ? basefont : font); + os, bparams, runparams, basefont, + (i == body_pos-1) ? basefont : current_font, + needPar); running_font = basefont; open_font = false; } @@ -2430,40 +2660,67 @@ void Paragraph::latex(BufferParams const & bparams, // close babel's font environment before opening CJK. string const lang_end_command = runparams.use_polyglossia ? "\\end{$$lang}" : lyxrc.language_command_end; + bool const using_begin_end = runparams.use_polyglossia || + !lang_end_command.empty(); if (!running_lang.empty() && - font.language()->encoding()->package() == Encoding::CJK) { + (!using_begin_end || running_lang == openLanguageName()) && + current_font.language()->encoding()->package() == Encoding::CJK) { string end_tag = subst(lang_end_command, "$$lang", running_lang); os << from_ascii(end_tag); column += end_tag.length(); + if (using_begin_end) + popLanguageName(); } // Switch file encoding if necessary (and allowed) if (!runparams.pass_thru && !style.pass_thru && runparams.encoding->package() != Encoding::none && - font.language()->encoding()->package() != Encoding::none) { + current_font.language()->encoding()->package() != Encoding::none) { pair const enc_switch = switchEncoding(os.os(), bparams, runparams, - *(font.language()->encoding())); + *(current_font.language()->encoding())); if (enc_switch.first) { column += enc_switch.second; - runparams.encoding = font.language()->encoding(); + runparams.encoding = current_font.language()->encoding(); } } char_type const c = d->text_[i]; + // A display math inset inside an ulem command will be output + // 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. + if (runparams.inDisplayMath && !deleted_display_math + && runparams.inulemcmd) { + if (os.afterParbreak()) + os << "\\noindent"; + else + os << "\\\\\n"; + } + // Do we need to change font? - if ((font != running_font || - font.language() != running_font.language()) && + if ((current_font != running_font || + current_font.language() != running_font.language()) && i != body_pos - 1) { odocstringstream ods; - column += font.latexWriteStartChanges(ods, bparams, + column += current_font.latexWriteStartChanges(ods, bparams, runparams, basefont, last_font); - running_font = font; + // Check again for display math in ulem commands as a + // font change may also occur just before a math inset. + if (runparams.inDisplayMath && !deleted_display_math + && runparams.inulemcmd) { + if (os.afterParbreak()) + os << "\\noindent"; + else + os << "\\\\\n"; + } + running_font = current_font; open_font = true; docstring fontchange = ods.str(); // check whether the fontchange ends with a \\textcolor @@ -2488,8 +2745,8 @@ 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, - i, column, font, style)) { + if (d->simpleTeXBlanks(bparams, runparams, os, + i, column, current_font, style)) { // A surrogate pair was output. We // must not call latexSpecialChar // in this iteration, since it would output @@ -2502,7 +2759,7 @@ void Paragraph::latex(BufferParams const & bparams, OutputParams rp = runparams; rp.free_spacing = style.free_spacing; - rp.local_font = &font; + rp.local_font = ¤t_font; rp.intitle = style.intitle; // Two major modes: LaTeX or plain @@ -2510,16 +2767,47 @@ void Paragraph::latex(BufferParams const & bparams, // and then split to handle the two modes separately. if (c == META_INSET) { if (i >= start_pos && (end_pos == -1 || i < end_pos)) { + // Greyedout notes and, in general, all insets + // with InsetLayout::isDisplay() == false, + // are typeset inline with the text. So, we + // can add a \par to the last paragraph of + // such insets only if nothing else follows. + bool incremented = false; + Inset const * inset = getInset(i); + InsetText const * textinset = inset + ? inset->asInsetText() + : 0; + if (i + 1 == size() && textinset + && !inset->getLayout().isDisplay()) { + ParagraphList const & pars = + textinset->text().paragraphs(); + pit_type const pit = pars.size() - 1; + Font const lastfont = + pit < 0 || pars[pit].empty() + ? pars[pit].getLayoutFont( + bparams, + outerfont) + : pars[pit].getFont(bparams, + pars[pit].size() - 1, + outerfont); + if (lastfont.fontInfo().size() != + basefont.fontInfo().size()) { + ++parInline; + incremented = true; + } + } d->latexInset(bparams, os, rp, running_font, basefont, outerfont, open_font, runningChange, style, i, column); + if (incremented) + --parInline; } - } else { - if (i >= start_pos && (end_pos == -1 || i < end_pos)) { - try { - d->latexSpecialChar(os, bparams, rp, running_font, runningChange, - style, i, end_pos, column); - } catch (EncodingException & e) { + } else if (i >= start_pos && (end_pos == -1 || i < end_pos)) { + try { + d->latexSpecialChar(os, bparams, rp, + running_font, runningChange, + style, i, end_pos, column); + } catch (EncodingException & e) { if (runparams.dryrun) { os << "<" << _("LyX Warning: ") << _("uncodable character") << " '"; @@ -2533,31 +2821,75 @@ void Paragraph::latex(BufferParams const & bparams, } } } - } // Set the encoding to that returned from latexSpecialChar (see // comment for encoding member in OutputParams.h) runparams.encoding = rp.encoding; + + // Also carry on the info on a closed ulem command for insets + // such as Note that do not produce any output, so that no + // command is ever executed but its opening was recorded. + runparams.inulemcmd = rp.inulemcmd; } // If we have an open font definition, we have to close it if (open_font) { + // Make sure that \\par is done with the font of the last + // character if this has another size as the default. + // This is necessary because LaTeX (and LyX on the screen) + // calculates the space between the baselines according + // to this font. (Matthias) + // + // We must not change the font for the last paragraph + // of non-multipar insets, tabular cells or commands, + // since this produces unwanted whitespace. + + Font const font = empty() + ? getLayoutFont(bparams, outerfont) + : getFont(bparams, size() - 1, outerfont); + + InsetText const * textinset = inInset().asInsetText(); + + bool const maintext = textinset + ? textinset->text().isMainText() + : false; + + size_t const numpars = textinset + ? textinset->text().paragraphs().size() + : 0; + + bool needPar = false; + + if (style.resfont.size() != font.fontInfo().size() + && (!runparams.isLastPar || maintext + || (numpars > 1 && d->ownerCode() != CELL_CODE + && (inInset().getLayout().isDisplay() + || parInline))) + && !style.isCommand()) { + needPar = true; + } #ifdef FIXED_LANGUAGE_END_DETECTION if (next_) { running_font.latexWriteEndChanges(os, bparams, runparams, basefont, - next_->getFont(bparams, 0, outerfont)); + next_->getFont(bparams, 0, outerfont), + needPar); } else { running_font.latexWriteEndChanges(os, bparams, - runparams, basefont, basefont); + runparams, basefont, basefont, needPar); } #else //FIXME: For now we ALWAYS have to close the foreign font settings if they are //FIXME: there as we start another \selectlanguage with the next paragraph if //FIXME: we are in need of this. This should be fixed sometime (Jug) running_font.latexWriteEndChanges(os, bparams, runparams, - basefont, basefont); + basefont, basefont, needPar); #endif + if (needPar) { + // The \par could not be inserted at the same nesting + // level of the font size change, so do it now. + os << "{\\" << font.latexSize() << "\\par}"; + } } column += Changes::latexMarkChange(os, bparams, runningChange, @@ -2746,13 +3078,13 @@ void doFontSwitch(vector & tagsToOpen, flag = false; } } -} +} // namespace docstring Paragraph::simpleLyXHTMLOnePar(Buffer const & buf, XHTMLStream & xs, OutputParams const & runparams, - Font const & outerfont, + Font const & outerfont, bool start_paragraph, bool close_paragraph, pos_type initial) const { @@ -2765,6 +3097,7 @@ docstring Paragraph::simpleLyXHTMLOnePar(Buffer const & buf, bool ubar_flag = false; bool dbar_flag = false; bool sout_flag = false; + bool xout_flag = false; bool wave_flag = false; // shape tags bool shap_flag = false; @@ -2776,7 +3109,7 @@ docstring Paragraph::simpleLyXHTMLOnePar(Buffer const & buf, Layout const & style = *d->layout_; if (start_paragraph) - xs.startParagraph(allowEmpty()); + xs.startDivision(allowEmpty()); FontInfo font_old = style.labeltype == LABEL_MANUAL ? style.labelfont : style.font; @@ -2784,13 +3117,13 @@ docstring Paragraph::simpleLyXHTMLOnePar(Buffer const & buf, FontShape curr_fs = INHERIT_SHAPE; FontFamily curr_fam = INHERIT_FAMILY; FontSize curr_size = FONT_SIZE_INHERIT; - - string const default_family = - buf.masterBuffer()->params().fonts_default_family; + + string const default_family = + buf.masterBuffer()->params().fonts_default_family; vector tagsToOpen; vector tagsToClose; - + // parsing main loop for (pos_type i = initial; i < size(); ++i) { // let's not show deleted material in the output @@ -2813,12 +3146,17 @@ docstring Paragraph::simpleLyXHTMLOnePar(Buffer const & buf, curstate = font.fontInfo().underbar(); if (font_old.underbar() != curstate) doFontSwitch(tagsToOpen, tagsToClose, ubar_flag, curstate, html::FT_UBAR); - + // strikeout curstate = font.fontInfo().strikeout(); if (font_old.strikeout() != curstate) doFontSwitch(tagsToOpen, tagsToClose, sout_flag, curstate, html::FT_SOUT); + // xout + curstate = font.fontInfo().xout(); + if (font_old.xout() != curstate) + doFontSwitch(tagsToOpen, tagsToClose, xout_flag, curstate, html::FT_XOUT); + // double underbar curstate = font.fontInfo().uuline(); if (font_old.uuline() != curstate) @@ -3069,7 +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(); @@ -3080,7 +3419,7 @@ docstring Paragraph::simpleLyXHTMLOnePar(Buffer const & buf, // wrapped in some font stuff. I think that will not work. xs.closeFontTags(); if (close_paragraph) - xs.endParagraph(); + xs.endDivision(); return retval; } @@ -3095,6 +3434,11 @@ bool Paragraph::isHfill(pos_type pos) const bool Paragraph::isNewline(pos_type pos) const { + // U+2028 LINE SEPARATOR + // U+2029 PARAGRAPH SEPARATOR + char_type const c = d->text_[pos]; + if (c == 0x2028 || c == 0x2029) + return true; Inset const * inset = getInset(pos); return inset && inset->lyxCode() == NEWLINE_CODE; } @@ -3154,6 +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); @@ -3213,28 +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); } } @@ -3276,11 +3649,13 @@ docstring Paragraph::asString(pos_type beg, pos_type end, int options, const Out void Paragraph::forOutliner(docstring & os, size_t const maxlen, - bool const shorten) const + bool const shorten, bool const label) const { size_t tmplen = shorten ? maxlen + 1 : maxlen; - if (!d->params_.labelString().empty()) - os += d->params_.labelString() + ' '; + 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; @@ -3434,7 +3809,7 @@ int Paragraph::fixBiblio(Buffer const & buffer) InsetCommandParams(BIBITEM_CODE)); Font font(inherit_font, buffer.params().language); - insertInset(0, inset, font, Change(track_changes ? Change::INSERTED + insertInset(0, inset, font, Change(track_changes ? Change::INSERTED : Change::UNCHANGED)); return 1; @@ -3471,12 +3846,18 @@ InsetList const & Paragraph::insetList() const } -void Paragraph::setBuffer(Buffer & b) +void Paragraph::setInsetBuffers(Buffer & b) { d->insetlist_.setBuffer(b); } +void Paragraph::resetBuffer() +{ + d->insetlist_.resetBuffer(); +} + + Inset * Paragraph::releaseInset(pos_type pos) { Inset * inset = d->insetlist_.release(pos); @@ -3642,11 +4023,11 @@ void Paragraph::deregisterWords() Private::LangWordsMap::const_iterator itl = d->words_.begin(); Private::LangWordsMap::const_iterator ite = d->words_.end(); for (; itl != ite; ++itl) { - WordList * wl = theWordList(itl->first); + WordList & wl = theWordList(itl->first); Private::Words::const_iterator it = (itl->second).begin(); Private::Words::const_iterator et = (itl->second).end(); for (; it != et; ++it) - wl->remove(*it); + wl.remove(*it); } d->words_.clear(); } @@ -3663,13 +4044,13 @@ void Paragraph::locateWord(pos_type & from, pos_type & to, to = from; return; } - // no break here, we go to the next + // fall through case WHOLE_WORD: // If we are already at the beginning of a word, do nothing if (!from || isWordSeparator(from - 1)) break; - // no break here, we go to the next + // fall through case PREVIOUS_WORD: // always move the cursor to the beginning of previous word @@ -3722,11 +4103,11 @@ void Paragraph::registerWords() Private::LangWordsMap::const_iterator itl = d->words_.begin(); Private::LangWordsMap::const_iterator ite = d->words_.end(); for (; itl != ite; ++itl) { - WordList * wl = theWordList(itl->first); + WordList & wl = theWordList(itl->first); Private::Words::const_iterator it = (itl->second).begin(); Private::Words::const_iterator et = (itl->second).end(); for (; it != et; ++it) - wl->insert(*it); + wl.insert(*it); } } @@ -3871,6 +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()) @@ -3920,6 +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,