/// 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,
+ int writeScriptChars(OutputParams const &, otexstream & os,
+ docstring const & ltx,
Change const &, Encoding const &,
std::string const, pos_type & i);
}
docstring latex2 = encoding.latexChar(c).first;
+ if (docstring(1, next) == latex1) {
+ // The encoding supports the combination:
+ // output as is (combining char after base char).
+ os << latex2 << latex1;
+ return latex1.length() + latex2.length();
+ }
+
+ // 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;
fontenc = runparams.local_font->language()->fontenc();
else
fontenc = runparams.main_fontenc;
- if ((script == "textgreek" && fontenc == "LGR")
- ||(script == "textcyr" && (fontenc == "T2A" || fontenc == "T2B"
- || fontenc == "T2C" || fontenc == "X2"))) {
- // Correct font encoding is being used, so we can avoid \text[greek|cyr].
+ 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("}");
+ }
}
- if (docstring(1, next) == latex1) {
- // the encoding supports the combination
- 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;
+ 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,
int pos = 0;
int length = brace2;
bool closing_brace = true;
- if ((script == "textgreek" && fontenc == "LGR")
- ||(script == "textcyr" && (fontenc == "T2A" || fontenc == "T2B"
- || fontenc == "T2C" || fontenc == "X2"))) {
+ if (!Encodings::needsScriptWrapper(script, fontenc)) {
// Correct font encoding is being used, so we can avoid \text[greek|cyr].
pos = brace1 + 1;
length -= pos;
// 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("{"));
}
}
string fontenc;
- if (running_font.language()->lang() == runparams.document_language)
+ if (running_font.language()->lang() == bparams.language->lang())
fontenc = runparams.main_fontenc;
else
fontenc = running_font.language()->fontenc();
- if (Encodings::isKnownScriptChar(c, script)
- && prefixIs(latex.first, from_ascii("\\" + script)))
- column += writeScriptChars(os, latex.first,
- running_change, encoding,
- fontenc, i) - 1;
- else if (latex.second
+ // "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);