/// 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(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 &,
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();
+ else
+ fontenc = runparams.main_fontenc;
+ 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();
}
}
-int Paragraph::Private::writeScriptChars(otexstream & os,
+int Paragraph::Private::writeScriptChars(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...
// 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{<spec>}".
// The latex macro name "textXXX" specifies the script to which text_[i]
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;
// 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(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("{"));
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;
+ if (running_font.language()->lang() == bparams.language->lang())
+ fontenc = runparams.main_fontenc;
+ else
+ fontenc = running_font.language()->fontenc();
+ // "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(runparams, os, ltx, running_change,
+ encoding, fontenc, i) - 1;
+ } else if (latex.second
&& ((!prefixIs(nextlatex, '\\')
&& !prefixIs(nextlatex, '{')
&& !prefixIs(nextlatex, '}'))
// then the contents
BufferParams const bp = features.runparams().is_child
? features.buffer().masterParams() : features.buffer().params();
+ string bscript = "textbaltic";
for (pos_type i = 0; i < int(text_.size()) ; ++i) {
char_type c = text_[i];
if (c == 0x0022) {
else if (bp.main_font_encoding() != "T1"
|| ((&owner_->getFontSettings(bp, i))->language()->internalFontEncoding()))
features.require("textquotedbl");
- }
- if (!bp.use_dash_ligatures
- && (c == 0x2013 || c == 0x2014)
- && bp.useNonTeXFonts
- && features.runparams().flavor == OutputParams::XETEX)
+ } else if (Encodings::isKnownScriptChar(c, bscript)){
+ string fontenc = (&owner_->getFontSettings(bp, i))->language()->fontenc();
+ if (fontenc.empty())
+ fontenc = features.runparams().main_fontenc;
+ if (Encodings::needsScriptWrapper("textbaltic", fontenc))
+ features.require("textbalticdefs");
+ } 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);
void Paragraph::validate(LaTeXFeatures & features) const
{
d->validate(features);
+ if (needsCProtection())
+ features.require("cprotect");
}
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,
: pars[pit].getFont(bparams,
pars[pit].size() - 1,
outerfont);
- if (last_font.fontInfo().size() !=
+ if (lastfont.fontInfo().size() !=
basefont.fontInfo().size()) {
++parInline;
incremented = true;
}
+bool Paragraph::needsCProtection() const
+{
+ // first check the layout of the paragraph
+ if (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())
+ return true;
+
+ return false;
+}
+
+
FontSpan const & Paragraph::getSpellRange(pos_type pos) const
{
return d->speller_state_.getRange(pos);
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<Language const *> & languages) const
+void Paragraph::getLanguages(std::set<Language const *> & 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);
}
}
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;