X-Git-Url: https://git.lyx.org/gitweb/?a=blobdiff_plain;f=src%2FParagraph.cpp;h=94cf9313d53d6eebc9e901114221cadf0ab56db6;hb=3bc08a76c42cd350a3141f00f37082bc9fab8967;hp=e7b2c78564a446497cddfb1d140eeda11d0eff2a;hpb=e5e8bff75b283de174ed8d1c141a14e991bab2c7;p=lyx.git diff --git a/src/Paragraph.cpp b/src/Paragraph.cpp index e7b2c78564..94cf9313d5 100644 --- a/src/Paragraph.cpp +++ b/src/Paragraph.cpp @@ -40,6 +40,7 @@ #include "ParagraphParameters.h" #include "SpellChecker.h" #include "sgml.h" +#include "texstream.h" #include "TextClass.h" #include "TexRow.h" #include "Text.h" @@ -60,6 +61,7 @@ #include "support/lstrings.h" #include "support/textutils.h" +#include #include #include @@ -282,10 +284,11 @@ 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); @@ -363,19 +366,6 @@ public: otexstream & os, pos_type i, unsigned int & column); - /// - bool latexSpecialTypewriter( - char_type const c, - otexstream & os, - pos_type i, - unsigned int & column); - /// - bool latexSpecialPhrase( - otexstream & os, - pos_type & i, - pos_type end_pos, - unsigned int & column, - OutputParams const & runparams); /// void validate(LaTeXFeatures & features) const; @@ -384,9 +374,6 @@ public: bool onlyText(Buffer const & buf, Font const & outerfont, pos_type initial) const; - /// match a string against a particular point in the paragraph - bool isTextAt(string const & str, pos_type pos) const; - /// a vector of speller skip positions typedef vector SkipPositions; typedef SkipPositions::const_iterator SkipPositionsIterator; @@ -514,26 +501,6 @@ public: }; -namespace { - -struct special_phrase { - string phrase; - docstring macro; - bool builtin; -}; - -special_phrase const special_phrases[] = { - { "LyX", from_ascii("\\LyX{}"), false }, - { "TeX", from_ascii("\\TeX{}"), true }, - { "LaTeX2e", from_ascii("\\LaTeXe{}"), true }, - { "LaTeX", from_ascii("\\LaTeX{}"), true }, -}; - -size_t const phrases_nr = sizeof(special_phrases)/sizeof(special_phrase); - -} // namespace anon - - Paragraph::Private::Private(Paragraph * owner, Layout const & layout) : owner_(owner), inset_owner_(0), id_(-1), begin_of_body_(0), layout_(&layout) { @@ -541,33 +508,37 @@ 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. -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); @@ -596,6 +567,18 @@ void Paragraph::addChangesToToc(DocIterator const & cdit, } +void Paragraph::addChangesToBuffer(Buffer const & buf) const +{ + d->changes_.updateBuffer(buf); +} + + +bool Paragraph::isChangeUpdateRequired() const +{ + return d->changes_.isUpdateRequired(); +} + + bool Paragraph::isDeleted(pos_type start, pos_type end) const { LASSERT(start >= 0 && start <= size(), return false); @@ -620,7 +603,7 @@ bool Paragraph::isMergedOnEndOfParDeletion(bool trackChanges) const if (!trackChanges) return true; - Change const change = d->changes_.lookup(size()); + Change const & change = d->changes_.lookup(size()); return change.inserted() && change.currentAuthor(); } @@ -951,10 +934,13 @@ int Paragraph::Private::writeScriptChars(otexstream & os, { // FIXME: modifying i here is not very nice... - // We only arrive here when a proper language for character text_[i] has - // not been specified (i.e., it could not be translated in the current - // latex encoding) or its latex translation has been forced, and it - // belongs to a known script. + // 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] @@ -970,6 +956,10 @@ int Paragraph::Private::writeScriptChars(otexstream & os, 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. pos = brace1 + 1; length -= pos; closing_brace = false; @@ -1016,26 +1006,6 @@ int Paragraph::Private::writeScriptChars(otexstream & os, } -bool Paragraph::Private::isTextAt(string const & str, pos_type pos) const -{ - pos_type const len = str.length(); - - // is the paragraph large enough? - if (pos + len > int(text_.size())) - return false; - - // does the wanted text start at point? - for (string::size_type i = 0; i < str.length(); ++i) { - // Caution: direct comparison of characters works only - // because str is pure ASCII. - if (str[i] != text_[pos + i]) - return false; - } - - return fontlist_.hasChangeInRange(pos, len); -} - - void Paragraph::Private::latexInset(BufferParams const & bparams, otexstream & os, OutputParams & runparams, @@ -1195,7 +1165,9 @@ void Paragraph::Private::latexSpecialChar(otexstream & os, char_type const c = (runparams.use_polyglossia) ? owner_->getUChar(bparams, i) : text_[i]; - if (style.pass_thru || runparams.pass_thru) { + if (style.pass_thru || runparams.pass_thru + || contains(style.pass_thru_chars, c) + || contains(runparams.pass_thru_chars, c)) { if (c != '\0') { Encoding const * const enc = runparams.encoding; if (enc && !enc->encodable(c)) @@ -1210,16 +1182,11 @@ void Paragraph::Private::latexSpecialChar(otexstream & os, return; // If T1 font encoding is used, use the special // characters it provides. - // NOTE: some languages reset the font encoding - // internally + // NOTE: Some languages reset the font encoding internally to a + // 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() - && lyxrc.fontenc == "T1" && latexSpecialT1(c, os, i, column)) - return; - - // \tt font needs special treatment - if (!runparams.inIPA - && running_font.fontInfo().family() == TYPEWRITER_FAMILY - && latexSpecialTypewriter(c, os, i, column)) + && bparams.font_encoding() == "T1" && latexSpecialT1(c, os, i, column)) return; // Otherwise, we use what LaTeX provides us. @@ -1242,6 +1209,16 @@ void Paragraph::Private::latexSpecialChar(otexstream & os, break; case '-': 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 + os << "{}"; + column += 2; + } break; case '\"': os << "\\char`\\\"{}"; @@ -1285,10 +1262,6 @@ void Paragraph::Private::latexSpecialChar(otexstream & os, break; default: - // LyX, LaTeX etc. - if (latexSpecialPhrase(os, i, end_pos, column, runparams)) - return; - if (c == '\0') return; @@ -1401,55 +1374,6 @@ bool Paragraph::Private::latexSpecialT3(char_type const c, otexstream & os, } -bool Paragraph::Private::latexSpecialTypewriter(char_type const c, otexstream & os, - pos_type i, unsigned int & column) -{ - switch (c) { - case '-': - // within \ttfamily, "--" is merged to "-" (no endash) - // so we avoid this rather irritating ligature - if (i + 1 < int(text_.size()) && text_[i + 1] == '-') { - os << "-{}"; - column += 2; - } else - os << '-'; - return true; - - // everything else has to be checked separately - // (depending on the encoding) - default: - return false; - } -} - - -/// \param end_pos -/// If [start_pos, end_pos) does not include entirely the special phrase, then -/// do not apply the macro transformation. -bool Paragraph::Private::latexSpecialPhrase(otexstream & os, pos_type & i, pos_type end_pos, - unsigned int & column, OutputParams const & runparams) -{ - // FIXME: if we have "LaTeX" with a font - // change in the middle (before the 'T', then - // the "TeX" part is still special cased. - // Really we should only operate this on - // "words" for some definition of word - - for (size_t pnr = 0; pnr < phrases_nr; ++pnr) { - if (!isTextAt(special_phrases[pnr].phrase, i) - || (end_pos != -1 && i + int(special_phrases[pnr].phrase.size()) > end_pos)) - continue; - if (runparams.moving_arg) - os << "\\protect"; - os << special_phrases[pnr].macro; - i += special_phrases[pnr].phrase.length() - 1; - column += special_phrases[pnr].macro.length() - 1; - return true; - } - return false; -} - - void Paragraph::Private::validate(LaTeXFeatures & features) const { if (layout_->inpreamble && inset_owner_) { @@ -1458,13 +1382,12 @@ void Paragraph::Private::validate(LaTeXFeatures & features) const BufferParams const & bp = features.runparams().is_child ? buf.masterParams() : buf.params(); Font f; - TexRow texrow; // 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, texrow); + otexstream os(ods); if (is_command) { os << '\\' << from_ascii(layout_->latexname()); // we have to provide all the optional arguments here, even though @@ -1490,7 +1413,7 @@ void Paragraph::Private::validate(LaTeXFeatures & features) const } } string const snippet = to_utf8(ods.str()); - features.addPreambleSnippet(snippet); + features.addPreambleSnippet(snippet, true); } } @@ -1521,19 +1444,12 @@ void Paragraph::Private::validate(LaTeXFeatures & features) const icit->inset->validate(features); if (layout_->needprotect && icit->inset->lyxCode() == FOOT_CODE) - features.require("footmisc"); + features.require("NeedLyXFootnoteCode"); } } // then the contents for (pos_type i = 0; i < int(text_.size()) ; ++i) { - for (size_t pnr = 0; pnr < phrases_nr; ++pnr) { - if (!special_phrases[pnr].builtin - && isTextAt(special_phrases[pnr].phrase, i)) { - features.require(special_phrases[pnr].phrase); - break; - } - } BufferEncodings::validate(text_[i], features); } } @@ -1640,7 +1556,7 @@ void Paragraph::write(ostream & os, BufferParams const & bparams, int column = 0; for (pos_type i = 0; i <= size(); ++i) { - Change const change = lookupChange(i); + Change const & change = lookupChange(i); if (change != running_change) flushString(os, write_buffer); Changes::lyxMarkChange(os, bparams, column, running_change, change); @@ -1833,7 +1749,10 @@ Font const & Paragraph::getFontSettings(BufferParams const & bparams, FontSpan Paragraph::fontSpan(pos_type pos) const { - LBUFERR(pos < size()); + LBUFERR(pos <= size()); + + if (pos == size()) + return FontSpan(pos, pos); pos_type start = 0; FontList::const_iterator cit = d->fontlist_.begin(); @@ -1922,14 +1841,6 @@ Font const Paragraph::getLayoutFont } -/// Returns the height of the highest font in range -FontSize Paragraph::highestFontInRange - (pos_type startpos, pos_type endpos, FontSize def_size) const -{ - return d->fontlist_.highestInRange(startpos, endpos, def_size); -} - - char_type Paragraph::getUChar(BufferParams const & bparams, pos_type pos) const { char_type c = d->text_[pos]; @@ -2101,7 +2012,7 @@ docstring Paragraph::expandParagraphLabel(Layout const & layout, docstring parent(fmt, i + 1, j - i - 1); docstring label = from_ascii("??"); if (tclass.hasLayout(parent)) - docstring label = expandParagraphLabel(tclass[parent], bparams, + label = expandParagraphLabel(tclass[parent], bparams, process_appendix); fmt = docstring(fmt, 0, i) + label + docstring(fmt, j + 1, docstring::npos); @@ -2145,14 +2056,12 @@ void Paragraph::setBeginOfBody() pos_type end = size(); if (i < end && !(isNewline(i) || isEnvSeparator(i))) { ++i; - char_type previous_char = 0; - char_type temp = 0; if (i < end) { - previous_char = d->text_[i]; + char_type previous_char = d->text_[i]; if (!(isNewline(i) || isEnvSeparator(i))) { ++i; while (i < end && previous_char != ' ') { - temp = d->text_[i]; + char_type temp = d->text_[i]; if (isNewline(i) || isEnvSeparator(i)) break; ++i; @@ -2249,8 +2158,12 @@ int Paragraph::Private::startTeXParParams(BufferParams const & bparams, { int column = 0; - if (params_.noindent() && !layout_->pass_thru - && (layout_->toggle_indent != ITOGGLE_NEVER)) { + bool canindent = + (bparams.paragraph_separation == BufferParams::ParagraphIndentSeparation) ? + (layout_->toggle_indent != ITOGGLE_NEVER) : + (layout_->toggle_indent == ITOGGLE_ALWAYS); + + if (canindent && params_.noindent() && !layout_->pass_thru) { os << "\\noindent "; column += 10; } @@ -2295,7 +2208,6 @@ int Paragraph::Private::startTeXParParams(BufferParams const & bparams, corrected_env(os, begin_tag, "flushright", code, lastpar, column); break; } case LYX_ALIGN_RIGHT: { - string output; if (owner_->getParLanguage(bparams)->babel() != "hebrew") corrected_env(os, begin_tag, "flushright", code, lastpar, column); else @@ -2501,18 +2413,18 @@ 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())) { column += running_font.latexWriteEndChanges( os, bparams, runparams, basefont, - (i == body_pos-1) ? basefont : font); + (i == body_pos-1) ? basefont : current_font); running_font = basefont; open_font = false; } @@ -2523,7 +2435,7 @@ void Paragraph::latex(BufferParams const & bparams, string const lang_end_command = runparams.use_polyglossia ? "\\end{$$lang}" : lyxrc.language_command_end; if (!running_lang.empty() && - font.language()->encoding()->package() == Encoding::CJK) { + current_font.language()->encoding()->package() == Encoding::CJK) { string end_tag = subst(lang_end_command, "$$lang", running_lang); @@ -2534,28 +2446,28 @@ void Paragraph::latex(BufferParams const & bparams, // 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]; // 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; + running_font = current_font; open_font = true; docstring fontchange = ods.str(); // check whether the fontchange ends with a \\textcolor @@ -2581,7 +2493,7 @@ void Paragraph::latex(BufferParams const & bparams, // style.pass_thru is false. if (i != body_pos - 1) { if (d->simpleTeXBlanks(runparams, os, - i, column, font, style)) { + i, column, current_font, style)) { // A surrogate pair was output. We // must not call latexSpecialChar // in this iteration, since it would output @@ -2594,7 +2506,7 @@ 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 @@ -2668,8 +2580,7 @@ void Paragraph::latex(BufferParams const & bparams, if (allowcust && d->endTeXParParams(bparams, os, runparams) && runparams.encoding != prev_encoding) { runparams.encoding = prev_encoding; - if (!runparams.isFullUnicode()) - os << setEncoding(prev_encoding->iconvName()); + os << setEncoding(prev_encoding->iconvName()); } LYXERR(Debug::LATEX, "Paragraph::latex... done " << this); @@ -2846,6 +2757,7 @@ docstring Paragraph::simpleLyXHTMLOnePar(Buffer const & buf, XHTMLStream & xs, OutputParams const & runparams, Font const & outerfont, + bool start_paragraph, bool close_paragraph, pos_type initial) const { docstring retval; @@ -2867,7 +2779,8 @@ docstring Paragraph::simpleLyXHTMLOnePar(Buffer const & buf, Layout const & style = *d->layout_; - xs.startParagraph(allowEmpty()); + if (start_paragraph) + xs.startDivision(allowEmpty()); FontInfo font_old = style.labeltype == LABEL_MANUAL ? style.labelfont : style.font; @@ -3153,42 +3066,26 @@ docstring Paragraph::simpleLyXHTMLOnePar(Buffer const & buf, if (!runparams.for_toc || inset->isInToc()) { OutputParams np = runparams; np.local_font = &font; - if (!inset->getLayout().htmlisblock()) + // If the paragraph has size 1, then we are in the "special + // case" where we do not output the containing paragraph info + if (!inset->getLayout().htmlisblock() && size() != 1) np.html_in_par = true; retval += inset->xhtml(xs, np); } } else { char_type c = getUChar(buf.masterBuffer()->params(), i); - - if (style.pass_thru || runparams.pass_thru) - xs << c; - else if (c == '-') { - docstring str; - int j = i + 1; - if (j < size() && d->text_[j] == '-') { - j += 1; - if (j < size() && d->text_[j] == '-') { - str += from_ascii("—"); - i += 2; - } else { - str += from_ascii("–"); - i += 1; - } - } - else - str += c; - // We don't want to escape the entities. Note that - // it is safe to do this, since str can otherwise - // only be "-". E.g., it can't be "<". - xs << XHTMLStream::ESCAPE_NONE << str; - } else - xs << c; + xs << c; } font_old = font.fontInfo(); } + // FIXME XHTML + // I'm worried about what happens if a branch, say, is itself + // wrapped in some font stuff. I think that will not work. xs.closeFontTags(); - xs.endParagraph(); + if (close_paragraph) + xs.endDivision(); + return retval; } @@ -3257,16 +3154,13 @@ bool Paragraph::isHardHyphenOrApostrophe(pos_type pos) const if ((nextpos == psize || isSpace(nextpos)) && (pos == 0 || isSpace(prevpos))) return false; - return c == '\'' - || ((nextpos == psize || d->text_[nextpos] != '-') - && (pos == 0 || d->text_[prevpos] != '-')); + return true; } -bool Paragraph::isSameSpellRange(pos_type pos1, pos_type pos2) const +FontSpan const & Paragraph::getSpellRange(pos_type pos) const { - return pos1 == pos2 - || d->speller_state_.getRange(pos1) == d->speller_state_.getRange(pos2); + return d->speller_state_.getRange(pos); } @@ -3385,21 +3279,23 @@ docstring Paragraph::asString(pos_type beg, pos_type end, int options, const Out } -void Paragraph::forOutliner(docstring & os, size_t maxlen) const +void Paragraph::forOutliner(docstring & os, size_t const maxlen, + bool const shorten) const { + size_t tmplen = shorten ? maxlen + 1 : maxlen; if (!d->params_.labelString().empty()) os += d->params_.labelString() + ' '; - for (pos_type i = 0; i < size() && os.length() < maxlen; ++i) { + for (pos_type i = 0; i < size() && os.length() < tmplen; ++i) { if (isDeleted(i)) continue; char_type const c = d->text_[i]; if (isPrintable(c)) os += c; - else if (c == '\t' || c == '\n') - os += ' '; else if (c == META_INSET) - getInset(i)->forOutliner(os, maxlen); + getInset(i)->forOutliner(os, tmplen, false); } + if (shorten) + Text::shortenForOutliner(os, maxlen); }