Update latex export of combining characters.

* Remove exception for polytonikogreek (drop braces)
  no longer required since the fixes to #6463 and #9637.
* Update comments.
This commit is contained in:
Günter Milde 2019-03-02 18:58:53 +01:00
parent 99bacf006e
commit 7dfc6c7f8b
2 changed files with 14 additions and 25 deletions

View File

@ -580,7 +580,7 @@ string const Encodings::TIPAShortcut(char_type c)
return string();
}
// Return true, if `c` is a supported Greek or Cyrillic letter.
bool Encodings::isKnownScriptChar(char_type const c, string & preamble)
{
CharInfoMap::const_iterator const it = unicodesymbols.find(c);
@ -602,6 +602,8 @@ bool Encodings::isKnownScriptChar(char_type const c, string & preamble)
bool Encodings::needsScriptWrapper(string const & script, string const & fontenc)
{
// Note: the wrapper is not required with Unicode font encoding "TU".
// However, this function is not called with non-TeX (Unicode) fonts.
if (script == "textgreek")
return (fontenc != "LGR");
if (script == "textcyrillic") {

View File

@ -869,7 +869,7 @@ int Paragraph::eraseChars(pos_type start, pos_type end, bool trackChanges)
return end - i;
}
// Handle combining characters
int Paragraph::Private::latexSurrogatePair(BufferParams const & bparams,
otexstream & os, char_type c, char_type next,
OutputParams const & runparams)
@ -895,16 +895,15 @@ int Paragraph::Private::latexSurrogatePair(BufferParams const & bparams,
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),
// except if we are using fontspec.
// Greek and Cyrillic letters need to be wrapped in \textcyrillic and \textgreek if they
// are not encodable in the current font encoding (regardless of the input encoding).
bool scriptchar = false;
if (!bparams.useNonTeXFonts)
// This will get us a script value to deal with below
if (!bparams.useNonTeXFonts) // With non-TeX fonts the font encoding is Unicode.
scriptchar = Encodings::isKnownScriptChar(c, script);
if (!scriptchar && docstring(1, next) == latex1) {
// The encoding supports the combination:
// Font and input encoding support the combination:
// output as is (combining char after base char).
os << latex2 << latex1;
return latex1.length() + latex2.length();
@ -920,32 +919,20 @@ int Paragraph::Private::latexSurrogatePair(BufferParams const & 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":
// Strip the \text(greek|cyrillic) script macro ...
pos = brace1 + 1;
length -= pos;
latex2 = latex2.substr(pos, length);
// We only need the script macro with non-native font encodings
// and with XeTeX/LuaTeX (with TeX fonts)
// and place it before the accent macro if required (#6463)
if (Encodings::needsScriptWrapper(script, fontenc)
|| runparams.isFullUnicode()) {
|| runparams.isFullUnicode()) {
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();
os << scriptmacro << latex1 << "{" << latex2 << "}" << cb;
return latex1.length() + 1 + latex2.length() + 1 + cb.length();
}