size_t const phrases_nr = sizeof(special_phrases)/sizeof(special_phrase);
-
-bool isEncoding(BufferParams const & bparams, LyXFont const & font,
- string const & encoding)
-{
- // We do ignore bparams.inputenc == "default" here because characters
- // in this encoding could be treated by TeX as something different,
- // e.g. if they are inside a CJK environment. See also
- // http://bugzilla.lyx.org/show_bug.cgi?id=3043.
- return (bparams.inputenc == encoding
- || (bparams.inputenc == "auto"
- && font.language()->encoding()->latexName() == encoding));
-}
-
} // namespace anon
}
-void Paragraph::Pimpl::acceptChanges(pos_type start, pos_type end)
+void Paragraph::Pimpl::acceptChanges(BufferParams const & bparams, pos_type start, pos_type end)
{
BOOST_ASSERT(start >= 0 && start <= size());
BOOST_ASSERT(end > start && end <= size() + 1);
case Change::UNCHANGED:
// accept changes in nested inset
if (pos < size() && owner_->isInset(pos)) {
- owner_->getInset(pos)->acceptChanges();
+ owner_->getInset(pos)->acceptChanges(bparams);
}
break;
changes_.set(Change(Change::UNCHANGED), pos);
// also accept changes in nested inset
if (pos < size() && owner_->isInset(pos)) {
- owner_->getInset(pos)->acceptChanges();
+ owner_->getInset(pos)->acceptChanges(bparams);
}
break;
}
-void Paragraph::Pimpl::rejectChanges(pos_type start, pos_type end)
+void Paragraph::Pimpl::rejectChanges(BufferParams const & bparams, pos_type start, pos_type end)
{
BOOST_ASSERT(start >= 0 && start <= size());
BOOST_ASSERT(end > start && end <= size() + 1);
case Change::UNCHANGED:
// reject changes in nested inset
if (pos < size() && owner_->isInset(pos)) {
- owner_->getInset(pos)->rejectChanges();
+ owner_->getInset(pos)->rejectChanges(bparams);
}
break;
}
-void Paragraph::Pimpl::simpleTeXBlanks(odocstream & os, TexRow & texrow,
- pos_type const i,
+int Paragraph::Pimpl::latexSurrogatePair(odocstream & os, value_type c,
+ value_type next, Encoding const & encoding)
+{
+ // Writing next here may circumvent a possible font change between
+ // c and next. Since next is only output if it forms a surrogate pair
+ // with c we can ignore this:
+ // A font change inside a surrogate pair does not make sense and is
+ // hopefully impossible to input.
+ // FIXME: change tracking
+ // Is this correct WRT change tracking?
+ docstring const latex1 = encoding.latexChar(next);
+ docstring const latex2 = encoding.latexChar(c);
+ os << latex1 << '{' << latex2 << '}';
+ return latex1.length() + latex2.length() + 2;
+}
+
+
+bool Paragraph::Pimpl::simpleTeXBlanks(Encoding const & encoding,
+ odocstream & os, TexRow & texrow,
+ pos_type & i,
unsigned int & column,
LyXFont const & font,
LyXLayout const & style)
{
if (style.pass_thru)
- return;
+ return false;
- if (column > lyxrc.ascii_linelen
+ if (i < size() - 1) {
+ char_type next = getChar(i + 1);
+ if (Encodings::isCombiningChar(next)) {
+ // This space has an accent, so we must always output it.
+ column += latexSurrogatePair(os, ' ', next, encoding) - 1;
+ ++i;
+ return true;
+ }
+ }
+
+ if (lyxrc.plaintext_linelen > 0
+ && column > lyxrc.plaintext_linelen
&& i
&& getChar(i - 1) != ' '
&& (i < size() - 1)
} else {
os << ' ';
}
+ 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] != owner_->text_[pos + i])
return false;
}
odocstream & os,
TexRow & texrow,
OutputParams const & runparams,
- LyXFont & font,
LyXFont & running_font,
LyXFont & basefont,
LyXFont const & outerfont,
if (style.pass_thru) {
if (c != Paragraph::META_INSET) {
if (c != '\0')
+ // FIXME UNICODE: This can fail if c cannot
+ // be encoded in the current encoding.
os.put(c);
} else
owner_->getInset(i)->plaintext(buf, os, runparams);
} else {
if (open_font) {
column += running_font.latexWriteEndChanges(
- os, basefont, basefont, bparams);
+ os, basefont, basefont);
open_font = false;
}
- basefont = owner_->getLayoutFont(bparams, outerfont);
- running_font = basefont;
- if (font.family() == LyXFont::TYPEWRITER_FAMILY)
+ if (running_font.family() == LyXFont::TYPEWRITER_FAMILY)
os << '~';
+ basefont = owner_->getLayoutFont(bparams, outerfont);
+ running_font = basefont;
+
if (runparams.moving_arg)
os << "\\protect ";
// some insets cannot be inside a font change command
if (open_font && inset->noFontChange()) {
column += running_font.latexWriteEndChanges(
- os, basefont, basefont, bparams);
+ os, basefont, basefont);
open_font = false;
basefont = owner_->getLayoutFont(bparams, outerfont);
running_font = basefont;
// would be wrongly converted on systems where char is signed, so we give
// the code points.
// This also makes us independant from the encoding of this source file.
- case 0xb1: // ± PLUS-MINUS SIGN
- case 0xb2: // ² SUPERSCRIPT TWO
- case 0xb3: // ³ SUPERSCRIPT THREE
- case 0xd7: // × MULTIPLICATION SIGN
- case 0xf7: // ÷ DIVISION SIGN
- case 0xb9: // ¹ SUPERSCRIPT ONE
- case 0xac: // ¬ NOT SIGN
- case 0xb5: // µ MICRO SIGN
- if (isEncoding(bparams, font, "latin1")
- || isEncoding(bparams, font, "latin9")) {
- os << "\\ensuremath{";
- os.put(c);
- os << '}';
- column += 13;
- } else {
- os.put(c);
- }
- break;
-
case '|': case '<': case '>':
// In T1 encoding, these characters exist
if (lyxrc.fontenc == "T1") {
&& i <= size() - 2
&& getChar(i + 1) == c) {
//os << "\\textcompwordmark{}";
+ //column += 19;
// Jean-Marc, have a look at
// this. I think this works
// equally well:
os << "\\,{}";
// Lgb
- column += 19;
+ column += 3;
}
break;
}
// Typewriter font also has them
- if (font.family() == LyXFont::TYPEWRITER_FAMILY) {
+ if (running_font.family() == LyXFont::TYPEWRITER_FAMILY) {
os.put(c);
break;
}
break;
case '-': // "--" in Typewriter mode -> "-{}-"
- if (i <= size() - 2
- && getChar(i + 1) == '-'
- && font.family() == LyXFont::TYPEWRITER_FAMILY) {
+ if (i <= size() - 2 &&
+ getChar(i + 1) == '-' &&
+ running_font.family() == LyXFont::TYPEWRITER_FAMILY) {
os << "-{}";
column += 2;
} else {
column += 9;
break;
- case 0xa3: // £ POUND SIGN
- if (bparams.inputenc == "default") {
- os << "\\pounds{}";
- column += 8;
- } else {
- os.put(c);
- }
- break;
-
- case 0x20ac: // EURO SIGN
- if (isEncoding(bparams, font, "latin9")
- || isEncoding(bparams, font, "cp1251")
- || isEncoding(bparams, font, "utf8")
- || isEncoding(bparams, font, "latin10")
- || isEncoding(bparams, font, "cp858")) {
- os.put(c);
- } else {
- os << "\\texteuro{}";
- column += 10;
- }
- break;
-
- // These characters are covered by latin1, but not
- // by latin9 (a.o.). We have to support them because
- // we switched the default of latin1-languages to latin9
- case 0xa4: // CURRENCY SYMBOL
- case 0xa6: // BROKEN BAR
- case 0xa8: // DIAERESIS
- case 0xb4: // ACUTE ACCENT
- case 0xb8: // CEDILLA
- case 0xbd: // 1/2 FRACTION
- case 0xbc: // 1/4 FRACTION
- case 0xbe: // 3/4 FRACTION
- if (isEncoding(bparams, font, "latin1")
- || isEncoding(bparams, font, "latin5")
- || isEncoding(bparams, font, "utf8")) {
- os.put(c);
- break;
- } else {
- switch (c) {
- case 0xa4:
- os << "\\textcurrency{}";
- column += 15;
- break;
- case 0xa6:
- os << "\\textbrokenbar{}";
- column += 16;
- break;
- case 0xa8:
- os << "\\textasciidieresis{}";
- column += 20;
- break;
- case 0xb4:
- os << "\\textasciiacute{}";
- column += 17;
- break;
- case 0xb8: // from latin1.def:
- os << "\\c\\ ";
- column += 3;
- break;
- case 0xbd:
- os << "\\textonehalf{}";
- column += 14;
- break;
- case 0xbc:
- os << "\\textonequarter{}";
- column += 17;
- break;
- case 0xbe:
- os << "\\textthreequarters{}";
- column += 20;
- break;
- }
- break;
- }
-
case '$': case '&':
case '%': case '#': case '{':
case '}': case '_':
default:
// I assume this is hack treating typewriter as verbatim
- if (font.family() == LyXFont::TYPEWRITER_FAMILY) {
+ // FIXME UNICODE: This can fail if c cannot be encoded
+ // in the current encoding.
+ if (running_font.family() == LyXFont::TYPEWRITER_FAMILY) {
if (c != '\0') {
os.put(c);
}
}
if (pnr == phrases_nr && c != '\0') {
- os.put(c);
+ Encoding const & encoding = *(runparams.encoding);
+ if (i < size() - 1) {
+ char_type next = getChar(i + 1);
+ if (Encodings::isCombiningChar(next)) {
+ column += latexSurrogatePair(os, c, next, encoding) - 1;
+ ++i;
+ break;
+ }
+ }
+ docstring const latex = encoding.latexChar(c);
+ if (latex.length() > 1 &&
+ latex[latex.length() - 1] != '}') {
+ // Prevent eating of a following
+ // space or command corruption by
+ // following characters
+ column += latex.length() + 1;
+ os << latex << "{}";
+ } else {
+ column += latex.length() - 1;
+ os << latex;
+ }
}
break;
}
FontList::const_iterator fend = fontlist.end();
for (; fcit != fend; ++fcit) {
if (fcit->font().noun() == LyXFont::ON) {
- lyxerr[Debug::LATEX] << "font.noun: "
+ LYXERR(Debug::LATEX) << "font.noun: "
<< fcit->font().noun()
<< endl;
features.require("noun");
- lyxerr[Debug::LATEX] << "Noun enabled. Font: "
+ LYXERR(Debug::LATEX) << "Noun enabled. Font: "
<< to_utf8(fcit->font().stateText(0))
<< endl;
}
break;
default:
features.require("color");
- lyxerr[Debug::LATEX] << "Color enabled. Font: "
+ LYXERR(Debug::LATEX) << "Color enabled. Font: "
<< to_utf8(fcit->font().stateText(0))
<< endl;
}
language != latex_language)
{
features.useLanguage(language);
- lyxerr[Debug::LATEX] << "Found language "
+ LYXERR(Debug::LATEX) << "Found language "
<< language->babel() << endl;
}
}
break;
}
}
- // these glyphs require the textcomp package
- if (getChar(i) == 0x20ac || getChar(i) == 0xa4
- || getChar(i) == 0xa6 || getChar(i) == 0xa8
- || getChar(i) == 0xb4 || getChar(i) == 0xbd
- || getChar(i) == 0xbc || getChar(i) == 0xbe)
- features.require("textcomp");
+ Encodings::validate(getChar(i), features);
}
}