lyx_mirror/src/mathed/InsetMathChar.cpp
Guillaume Munch 03a4b8c932 Implement rule 17
* Convert the kerning into a proper right margin for mathnormal, mathscr and
  mathcal (fonts with \fontdimen2 == 0 as per rule 17 from TeXBook).

* Simulate the fact that characters in mathnormal fonts have a 0-width left
  bearing.

* Implement subscript positioning in the case of rule 17 using negative italic
  correction (kerning_).
2017-01-31 22:55:00 +01:00

309 lines
7.6 KiB
C++

/**
* \file InsetMathChar.cpp
* This file is part of LyX, the document processor.
* Licence details can be found in the file COPYING.
*
* \author Alejandro Aguilar Sierra
* \author André Pönitz
*
* Full author contact details are available in file CREDITS.
*/
#include <config.h>
#include "InsetMathChar.h"
#include "MathParser.h"
#include "MathSupport.h"
#include "MathStream.h"
#include "MetricsInfo.h"
#include "Dimension.h"
#include "BufferEncodings.h"
#include "LaTeXFeatures.h"
#include "TextPainter.h"
#include "frontends/FontMetrics.h"
#include "support/debug.h"
#include "support/lstrings.h"
#include "support/textutils.h"
#include <algorithm>
using namespace std;
namespace lyx {
extern bool has_math_fonts;
namespace {
latexkeys const * makeSubstitute(char_type c)
{
std::string name;
switch (c) {
// Latex replaces ', *, -, and : with specific symbols. With unicode-math,
// these symbols are replaced respectively by ^U+2032, U+2217, U+2212 and
// U+2236 (the latter substitution can be turned off with a package
// option). Unicode-math also replaces ` with \backprime.
// prime needs to be placed in superscript unless an opentype font is used.
//case '\'':
//name = "prime";
//break;
case '*':
name = "ast";
break;
case '-':
name = "lyxminus";// unicode-math: "minus"
break;
case ':':
name = "ordinarycolon";// unicode-math: "mathratio"
break;
// The remaining replacements are not real character substitutions (from a
// unicode point of view) but are done here: 1. for cosmetic reasons, in the
// context of being stuck with CM fonts at the moment, to ensure consistency
// with related symbols: -, \leq, \geq, etc. 2. to get the proper spacing
// as defined in lib/symbols.
case '+':
name = "lyxplus";//unicode-math: "mathplus"
break;
case '>':
name = "lyxgt";//unicode-math: "greater"
break;
case '<':
name = "lyxlt";//unicode-math: "less"
break;
case '=':
name = "lyxeqrel";//unicode-math: "equal"
break;
//case ','://unicode-math: "mathcomma"
//case ';'://unicode-math: "mathsemicolon"
default:
return nullptr;
}
return in_word_set(from_ascii(name));
}
} //anonymous namespace
static bool slanted(char_type c)
{
return isAlphaASCII(c) || Encodings::isMathAlpha(c);
}
InsetMathChar::InsetMathChar(char_type c)
: char_(c), kerning_(0), subst_(makeSubstitute(c))
{}
Inset * InsetMathChar::clone() const
{
return new InsetMathChar(*this);
}
void InsetMathChar::metrics(MetricsInfo & mi, Dimension & dim) const
{
string const & f = mi.base.fontname;
if (isMathFont(f) && subst_) {
// If the char has a substitute, draw the replacement symbol
// instead, but only in math mode.
mathedSymbolDim(mi.base, dim, subst_);
kerning_ = mathed_char_kerning(mi.base.font, *subst_->draw.rbegin());
return;
} else if (!slanted(char_) && f == "mathnormal") {
Changer dummy = mi.base.font.changeShape(UP_SHAPE);
dim = theFontMetrics(mi.base.font).dimension(char_);
kerning_ = 0;
} else {
frontend::FontMetrics const & fm = theFontMetrics(mi.base.font);
dim = fm.dimension(char_);
kerning_ = mathed_char_kerning(mi.base.font, char_);
// cmmi has almost no left bearing: simulate this
if (f == "mathnormal") {
dim.wid += max(-fm.lbearing(char_), 0);
}
// Rule 17 from Appendix G
// These are the fonts with fontdimen(2)==0.
// To properly handle the case fontdimen(2)!=0 (that is for all other
// math fonts), where the italic correction must also be converted into
// kerning but only at the end of strings of characters with the same
// font, one would need a more elaborate implementation in MathRow. For
// now the case fontdimen(2)==0 is the most important.
if (f == "mathnormal" || f == "mathscr" || f == "mathcal") {
dim.wid += kerning_;
// We use a negative value to tell InsetMathScript to move the
// subscript leftwards instead of the superscript rightwards
kerning_ = -kerning_;
}
}
}
void InsetMathChar::draw(PainterInfo & pi, int x, int y) const
{
//lyxerr << "drawing '" << char_ << "' font: " << pi.base.fontname << std::endl;
if (isMathFont(pi.base.fontname)) {
if (subst_) {
// If the char has a substitute, draw the replacement symbol
// instead, but only in math mode.
mathedSymbolDraw(pi, x, y, subst_);
return;
} else if (!slanted(char_) && pi.base.fontname == "mathnormal") {
Changer dummy = pi.base.font.changeShape(UP_SHAPE);
pi.draw(x, y, char_);
return;
}
}
// cmmi has almost no left bearing: simulate this
if (pi.base.fontname == "mathnormal") {
x += max(-theFontMetrics(pi.base.font).lbearing(char_), 0);
}
pi.draw(x, y, char_);
}
void InsetMathChar::metricsT(TextMetricsInfo const &, Dimension & dim) const
{
dim.wid = 1;
dim.asc = 1;
dim.des = 0;
}
void InsetMathChar::drawT(TextPainter & pain, int x, int y) const
{
//lyxerr << "drawing text '" << char_ << "' code: " << code_ << endl;
pain.draw(x, y, char_);
}
void InsetMathChar::write(WriteStream & os) const
{
os.os().put(char_);
}
void InsetMathChar::validate(LaTeXFeatures & features) const
{
if (!isASCII(char_))
BufferEncodings::validate(char_, features, true);
}
void InsetMathChar::normalize(NormalStream & os) const
{
os << "[char ";
os.os().put(char_);
os << " mathalpha]";
}
void InsetMathChar::octave(OctaveStream & os) const
{
os.os().put(char_);
}
// We have a bit of a problem here. MathML wants to know whether the
// character represents an "identifier" or an "operator", and we have
// no general way of telling. So we shall guess: If it's alpha or
// mathalpha, then we'll treat it as an identifier, otherwise as an
// operator.
// Worst case: We get bad spacing, or bad italics.
void InsetMathChar::mathmlize(MathStream & ms) const
{
std::string entity;
switch (char_) {
case '<': entity = "&lt;"; break;
case '>': entity = "&gt;"; break;
case '&': entity = "&amp;"; break;
case ' ': {
ms << from_ascii("&nbsp;");
return;
}
default: break;
}
if (ms.inText()) {
if (entity.empty())
ms.os().put(char_);
else
ms << from_ascii(entity);
return;
}
if (!entity.empty()) {
ms << "<mo>" << from_ascii(entity) << "</mo>";
return;
}
char const * type =
(isAlphaASCII(char_) || Encodings::isMathAlpha(char_))
? "mi" : "mo";
// we don't use MTag and ETag because we do not want the spacing
ms << "<" << type << ">" << char_type(char_) << "</" << type << ">";
}
void InsetMathChar::htmlize(HtmlStream & ms) const
{
std::string entity;
// Not taking subst_ into account here because the MathML output of
// <>=+-* looks correct as it is. FIXME: ' is not output as ^\prime
switch (char_) {
case '<': entity = "&lt;"; break;
case '>': entity = "&gt;"; break;
case '&': entity = "&amp;"; break;
case ' ': entity = "&nbsp;"; break;
default: break;
}
bool have_entity = !entity.empty();
if (ms.inText()) {
if (have_entity)
ms << from_ascii(entity);
else
ms.os().put(char_);
return;
}
if (have_entity) {
// an operator, so give some space
ms << ' ' << from_ascii(entity) << ' ';
return;
}
if (isAlphaASCII(char_) || Encodings::isMathAlpha(char_))
// we don't use MTag and ETag because we do not want the spacing
ms << MTag("i") << char_type(char_) << ETag("i");
else
// an operator, so give some space
ms << " " << char_type(char_) << " ";
}
MathClass InsetMathChar::mathClass() const
{
// this information comes from fontmath.ltx in LaTeX source.
char const ch = static_cast<char>(char_);
if (subst_)
return string_to_class(subst_->extra);
else if (support::contains(",;", ch))
return MC_PUNCT;
else if (support::contains("([", ch))
return MC_OPEN;
else if (support::contains(")]!?", ch))
return MC_CLOSE;
else return MC_ORD;
}
} // namespace lyx