PDF4QT/Pdf4QtLib/sources/pdffont.cpp
2021-08-10 19:22:56 +02:00

2404 lines
92 KiB
C++

// Copyright (C) 2019-2021 Jakub Melka
//
// This file is part of PDF4QT.
//
// PDF4QT is free software: you can redistribute it and/or modify
// it under the terms of the GNU Lesser General Public License as published by
// the Free Software Foundation, either version 3 of the License, or
// with the written consent of the copyright owner, any later version.
//
// PDF4QT is distributed in the hope that it will be useful,
// but WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
// GNU Lesser General Public License for more details.
//
// You should have received a copy of the GNU Lesser General Public License
// along with PDF4QT. If not, see <https://www.gnu.org/licenses/>.
#include "pdffont.h"
#include "pdfdocument.h"
#include "pdfparser.h"
#include "pdfnametounicode.h"
#include "pdfexception.h"
#include <ft2build.h>
#include <freetype/freetype.h>
#include <freetype/ftglyph.h>
#include <freetype/fterrors.h>
#include <freetype/ftoutln.h>
#include <freetype/t1tables.h>
#include <QMutex>
#include <QReadWriteLock>
#include <QPainterPath>
#include <QDataStream>
#include <QTreeWidgetItem>
#ifdef Q_OS_WIN
#include "Windows.h"
#pragma comment(lib, "Gdi32")
#pragma comment(lib, "User32")
#endif
namespace pdf
{
/// Storage class for system fonts
class PDFSystemFontInfoStorage
{
public:
/// Returns instance of storage
static const PDFSystemFontInfoStorage* getInstance();
/// Loads font from descriptor
/// \param descriptor Descriptor describing the font
QByteArray loadFont(const FontDescriptor* descriptor, StandardFontType standardFontType, PDFRenderErrorReporter* reporter) const;
private:
explicit PDFSystemFontInfoStorage();
#ifdef Q_OS_WIN
/// Callback for enumerating fonts
static int CALLBACK enumerateFontProc(const LOGFONT* font, const TEXTMETRIC* textMetrics, DWORD fontType, LPARAM lParam);
/// Retrieves font data for desired font
static QByteArray getFontData(const LOGFONT* font, HDC hdc);
/// Create a postscript name for comparation purposes
static QString getFontPostscriptName(QString fontName);
struct FontInfo
{
QString faceName;
QString faceNameAdjusted;
LOGFONT logFont;
TEXTMETRIC textMetric;
};
struct CallbackInfo
{
PDFSystemFontInfoStorage* storage = nullptr;
HDC hdc = nullptr;
};
std::vector<FontInfo> m_fontInfos;
#endif
};
const PDFSystemFontInfoStorage* PDFSystemFontInfoStorage::getInstance()
{
static PDFSystemFontInfoStorage instance;
return &instance;
}
QByteArray PDFSystemFontInfoStorage::loadFont(const FontDescriptor* descriptor, StandardFontType standardFontType, PDFRenderErrorReporter* reporter) const
{
QByteArray result;
#ifdef Q_OS_WIN
HDC hdc = GetDC(NULL);
const BYTE lfItalic = (descriptor->italicAngle != 0.0 ? TRUE : FALSE);
// Exact match font face name
QString fontName;
switch (standardFontType)
{
case StandardFontType::TimesRoman:
case StandardFontType::TimesRomanBold:
case StandardFontType::TimesRomanItalics:
case StandardFontType::TimesRomanBoldItalics:
{
fontName = "TimesNewRoman";
break;
}
case StandardFontType::Helvetica:
case StandardFontType::HelveticaBold:
case StandardFontType::HelveticaOblique:
case StandardFontType::HelveticaBoldOblique:
{
fontName = "Arial";
break;
}
case StandardFontType::Courier:
case StandardFontType::CourierBold:
case StandardFontType::CourierOblique:
case StandardFontType::CourierBoldOblique:
{
fontName = "CourierNew";
break;
}
case StandardFontType::Symbol:
case StandardFontType::ZapfDingbats:
{
fontName = "Symbol";
break;
}
default:
{
fontName = getFontPostscriptName(descriptor->fontName);
break;
}
}
if (!fontName.isEmpty())
{
for (const FontInfo& fontInfo : m_fontInfos)
{
if (fontInfo.faceNameAdjusted == fontName &&
fontInfo.logFont.lfWeight == descriptor->fontWeight &&
fontInfo.logFont.lfItalic == lfItalic)
{
result = getFontData(&fontInfo.logFont, hdc);
if (!result.isEmpty())
{
break;
}
}
}
// Match for font family
if (result.isEmpty())
{
for (const FontInfo& fontInfo : m_fontInfos)
{
if (fontInfo.faceNameAdjusted == fontName)
{
LOGFONT logFont = fontInfo.logFont;
logFont.lfWeight = descriptor->fontWeight;
logFont.lfItalic = lfItalic;
result = getFontData(&logFont, hdc);
if (!result.isEmpty())
{
break;
}
}
}
}
}
// Exact match for font, if font can't be exact matched, then match font family
// and try to set weight
QString fontFamily = QString::fromLatin1(descriptor->fontFamily);
if (!fontFamily.isEmpty() && result.isEmpty())
{
for (const FontInfo& fontInfo : m_fontInfos)
{
if (fontInfo.faceName.contains(fontFamily) &&
fontInfo.logFont.lfWeight == descriptor->fontWeight &&
fontInfo.logFont.lfItalic == lfItalic)
{
result = getFontData(&fontInfo.logFont, hdc);
if (!result.isEmpty())
{
reporter->reportRenderError(RenderErrorType::Warning, PDFTranslationContext::tr("Inexact font substitution: font %1 replaced by %2 using font family %3.").arg(fontName, fontInfo.faceNameAdjusted, fontFamily));
break;
}
}
}
// Match for font family
if (result.isEmpty())
{
for (const FontInfo& fontInfo : m_fontInfos)
{
if (fontInfo.faceName.contains(fontFamily))
{
LOGFONT logFont = fontInfo.logFont;
logFont.lfWeight = descriptor->fontWeight;
logFont.lfItalic = lfItalic;
result = getFontData(&logFont, hdc);
if (!result.isEmpty())
{
reporter->reportRenderError(RenderErrorType::Warning, PDFTranslationContext::tr("Inexact font substitution: font %1 replaced by %2 using font family %3.").arg(fontName, fontInfo.faceNameAdjusted, fontFamily));
break;
}
}
}
}
}
// Try to inexact match for font name - find similar font
if (!fontName.isEmpty() && result.isEmpty())
{
for (const FontInfo& fontInfo : m_fontInfos)
{
if (fontInfo.faceNameAdjusted.contains(fontName))
{
LOGFONT logFont = fontInfo.logFont;
logFont.lfWeight = descriptor->fontWeight;
logFont.lfItalic = lfItalic;
result = getFontData(&logFont, hdc);
if (!result.isEmpty())
{
reporter->reportRenderError(RenderErrorType::Warning, PDFTranslationContext::tr("Inexact font substitution: font %1 replaced by %2.").arg(fontName, fontInfo.faceNameAdjusted));
break;
}
}
}
}
ReleaseDC(NULL, hdc);
#endif
if (result.isEmpty() && standardFontType == StandardFontType::Invalid)
{
reporter->reportRenderError(RenderErrorType::Warning, PDFTranslationContext::tr("Inexact font substitution: font %1 replaced by standard font Times New Roman.").arg(fontName));
result = loadFont(descriptor, StandardFontType::TimesRoman, reporter);
}
return result;
}
PDFSystemFontInfoStorage::PDFSystemFontInfoStorage()
{
#ifdef Q_OS_WIN
LOGFONT logfont;
std::memset(&logfont, 0, sizeof(logfont));
logfont.lfCharSet = DEFAULT_CHARSET;
logfont.lfFaceName[0] = 0;
logfont.lfPitchAndFamily = 0;
HDC hdc = GetDC(NULL);
CallbackInfo callbackInfo{ this, hdc};
EnumFontFamiliesEx(hdc, &logfont, &PDFSystemFontInfoStorage::enumerateFontProc, reinterpret_cast<LPARAM>(&callbackInfo), 0);
ReleaseDC(NULL, hdc);
#endif
}
#ifdef Q_OS_WIN
int PDFSystemFontInfoStorage::enumerateFontProc(const LOGFONT* font, const TEXTMETRIC* textMetrics, DWORD fontType, LPARAM lParam)
{
if ((fontType & TRUETYPE_FONTTYPE) && (font->lfCharSet == ANSI_CHARSET))
{
CallbackInfo* callbackInfo = reinterpret_cast<CallbackInfo*>(lParam);
FontInfo fontInfo;
fontInfo.logFont = *font;
fontInfo.textMetric = *textMetrics;
fontInfo.faceName = QString::fromWCharArray(font->lfFaceName);
fontInfo.faceNameAdjusted = getFontPostscriptName(fontInfo.faceName);
callbackInfo->storage->m_fontInfos.push_back(qMove(fontInfo));
// For debug purposes only!
#if 0
QByteArray byteArray = getFontData(font, callbackInfo->hdc);
qDebug() << "Font: " << QString::fromWCharArray(font->lfFaceName) << ", italic = " << font->lfItalic << ", weight = " << font->lfWeight << ", data size = " << byteArray.size();
#endif
}
return TRUE;
}
QByteArray PDFSystemFontInfoStorage::getFontData(const LOGFONT* font, HDC hdc)
{
QByteArray byteArray;
if (HFONT fontHandle = ::CreateFontIndirect(font))
{
HGDIOBJ oldFont = ::SelectObject(hdc, fontHandle);
DWORD size = ::GetFontData(hdc, 0, 0, nullptr, 0);
if (size != GDI_ERROR)
{
byteArray.resize(static_cast<int>(size));
::GetFontData(hdc, 0, 0, byteArray.data(), byteArray.size());
}
::SelectObject(hdc, oldFont);
::DeleteObject(fontHandle);
}
return byteArray;
}
QString PDFSystemFontInfoStorage::getFontPostscriptName(QString fontName)
{
for (const char* string : { "PS", "MT", "Regular", "Bold", "Italic", "Oblique" })
{
fontName.remove(QLatin1String(string), Qt::CaseInsensitive);
}
return fontName.remove(QChar(' ')).remove(QChar('-')).remove(QChar(',')).trimmed();
}
#endif
PDFFont::PDFFont(FontDescriptor fontDescriptor) :
m_fontDescriptor(qMove(fontDescriptor))
{
}
class IRealizedFontImpl
{
public:
explicit IRealizedFontImpl() = default;
virtual ~IRealizedFontImpl() = default;
/// Fills the text sequence by interpreting byte array according font data and
/// produces glyphs for the font.
/// \param byteArray Array of bytes to be interpreted
/// \param textSequence Text sequence to be filled
virtual void fillTextSequence(const QByteArray& byteArray, TextSequence& textSequence, PDFRenderErrorReporter* reporter) = 0;
/// Returns true, if font has horizontal writing system
virtual bool isHorizontalWritingSystem() const = 0;
/// Dumps information about the font
virtual void dumpFontToTreeItem(QTreeWidgetItem* item) const { Q_UNUSED(item); }
/// Returns postscript name of the font
virtual QString getPostScriptName() const { return QString(); }
/// Returns character info
virtual CharacterInfos getCharacterInfos() const = 0;
};
/// Implementation of the PDFRealizedFont class using PIMPL pattern for Type 3 fonts
class PDFRealizedType3FontImpl : public IRealizedFontImpl
{
public:
explicit PDFRealizedType3FontImpl(PDFFontPointer parentFont, PDFReal pixelSize) : m_pixelSize(pixelSize), m_parentFont(parentFont) { }
virtual ~PDFRealizedType3FontImpl() override = default;
PDFReal getPixelSize() const { return m_pixelSize; }
virtual void fillTextSequence(const QByteArray& byteArray, TextSequence& textSequence, PDFRenderErrorReporter* reporter) override;
virtual bool isHorizontalWritingSystem() const override;
virtual CharacterInfos getCharacterInfos() const override;
private:
/// Pixel size of the font
PDFReal m_pixelSize = 0.0;
/// Parent font
PDFFontPointer m_parentFont;
};
/// Implementation of the PDFRealizedFont class using PIMPL pattern
class PDFRealizedFontImpl : public IRealizedFontImpl
{
public:
explicit PDFRealizedFontImpl();
virtual ~PDFRealizedFontImpl();
virtual void fillTextSequence(const QByteArray& byteArray, TextSequence& textSequence, PDFRenderErrorReporter* reporter) override;
virtual bool isHorizontalWritingSystem() const override { return !m_isVertical; }
virtual void dumpFontToTreeItem(QTreeWidgetItem* item) const override;
virtual QString getPostScriptName() const override { return m_postScriptName; }
virtual CharacterInfos getCharacterInfos() const override;
static constexpr const PDFReal PIXEL_SIZE_MULTIPLIER = 100.0;
private:
friend class PDFRealizedFont;
static constexpr const PDFReal FONT_WIDTH_MULTIPLIER = 1.0 / 1000.0;
static constexpr const PDFReal FORMAT_26_6_MULTIPLIER = 1 / 64.0;
static constexpr const PDFReal FONT_MULTIPLIER = FORMAT_26_6_MULTIPLIER / PIXEL_SIZE_MULTIPLIER;
struct Glyph
{
QPainterPath glyph;
PDFReal advance = 0.0;
};
static int outlineMoveTo(const FT_Vector* to, void* user);
static int outlineLineTo(const FT_Vector* to, void* user);
static int outlineConicTo(const FT_Vector* control, const FT_Vector* to, void* user);
static int outlineCubicTo(const FT_Vector* control1, const FT_Vector* control2, const FT_Vector* to, void* user);
/// Get glyph for glyph index
const Glyph& getGlyph(unsigned int glyphIndex);
/// Function checks, if error occured, and if yes, then exception is thrown
static void checkFreeTypeError(FT_Error error);
/// Read/write lock for accessing the glyph data
QReadWriteLock m_readWriteLock;
/// Glyph cache, must be protected by the mutex above
std::unordered_map<unsigned int, Glyph> m_glyphCache;
/// For embedded fonts, this byte array contains embedded font data
QByteArray m_embeddedFontData;
/// For system fonts, this byte array contains system font data
QByteArray m_systemFontData;
/// Instance of FreeType library assigned to this font
FT_Library m_library;
/// Face of the font
FT_Face m_face;
/// Pixel size of the font
PDFReal m_pixelSize;
/// Parent font
PDFFontPointer m_parentFont;
/// True, if font is embedded
bool m_isEmbedded;
/// True, if font has vertical writing system
bool m_isVertical;
/// Postscript name of the font
QString m_postScriptName;
};
PDFRealizedFontImpl::PDFRealizedFontImpl() :
m_library(nullptr),
m_face(nullptr),
m_pixelSize(0.0),
m_parentFont(nullptr),
m_isEmbedded(false),
m_isVertical(false)
{
}
PDFRealizedFontImpl::~PDFRealizedFontImpl()
{
if (m_face)
{
FT_Done_Face(m_face);
m_face = nullptr;
}
if (m_library)
{
FT_Done_FreeType(m_library);
m_library = nullptr;
}
}
void PDFRealizedFontImpl::fillTextSequence(const QByteArray& byteArray, TextSequence& textSequence, PDFRenderErrorReporter* reporter)
{
switch (m_parentFont->getFontType())
{
case FontType::Type1:
case FontType::TrueType:
case FontType::MMType1:
{
// We can use encoding
Q_ASSERT(dynamic_cast<PDFSimpleFont*>(m_parentFont.get()));
const PDFSimpleFont* font = static_cast<PDFSimpleFont*>(m_parentFont.get());
const encoding::EncodingTable* encoding = font->getEncoding();
const GlyphIndices* glyphIndices = font->getGlyphIndices();
textSequence.items.reserve(textSequence.items.size() + byteArray.size());
for (int i = 0, count = byteArray.size(); i < count; ++i)
{
GID glyphIndex = (*glyphIndices)[static_cast<uint8_t>(byteArray[i])];
if (!glyphIndex)
{
// Try to obtain glyph index from unicode
if (m_face->charmap && m_face->charmap->encoding == FT_ENCODING_UNICODE)
{
glyphIndex = FT_Get_Char_Index(m_face, (*encoding)[static_cast<uint8_t>(byteArray[i])].unicode());
}
}
const PDFReal glyphWidth = font->getGlyphAdvance(static_cast<uint8_t>(byteArray[i]));
if (glyphIndex)
{
const Glyph& glyph = getGlyph(glyphIndex);
textSequence.items.emplace_back(&glyph.glyph, (*encoding)[static_cast<uint8_t>(byteArray[i])], glyph.advance);
}
else
{
reporter->reportRenderError(RenderErrorType::Warning, PDFTranslationContext::tr("Glyph for simple font character code '%1' not found.").arg(static_cast<uint8_t>(byteArray[i])));
if (glyphWidth > 0)
{
const QPainterPath* nullpath = nullptr;
textSequence.items.emplace_back(nullpath, QChar(), glyphWidth * m_pixelSize * FONT_WIDTH_MULTIPLIER);
}
}
}
break;
}
case FontType::Type0:
{
Q_ASSERT(dynamic_cast<PDFType0Font*>(m_parentFont.get()));
const PDFType0Font* font = static_cast<PDFType0Font*>(m_parentFont.get());
const PDFFontCMap* cmap = font->getCMap();
const PDFFontCMap* toUnicode = font->getToUnicode();
const PDFCIDtoGIDMapper* CIDtoGIDmapper = font->getCIDtoGIDMapper();
std::vector<CID> cids = cmap->interpret(byteArray);
textSequence.items.reserve(textSequence.items.size() + cids.size());
for (CID cid : cids)
{
const GID glyphIndex = CIDtoGIDmapper->map(cid);
const PDFReal glyphWidth = font->getGlyphAdvance(cid);
if (glyphIndex)
{
QChar character = toUnicode->getToUnicode(cid);
const Glyph& glyph = getGlyph(glyphIndex);
textSequence.items.emplace_back(&glyph.glyph, character, glyph.advance);
}
else
{
if (cid > 0)
{
// Character with CID == 0 is treated as default whitespace, it hasn't glyph
reporter->reportRenderError(RenderErrorType::Warning, PDFTranslationContext::tr("Glyph for composite font character with cid '%1' not found.").arg(cid));
}
if (glyphWidth > 0)
{
// We do not multiply advance with font size and FONT_WIDTH_MULTIPLIER, because in the code,
// "advance" is treated as in font space.
const QPainterPath* nullpath = nullptr;
textSequence.items.emplace_back(nullpath, QChar(), -glyphWidth);
}
}
}
break;
}
default:
{
// Unhandled font type
Q_ASSERT(false);
break;
}
}
}
CharacterInfos PDFRealizedFontImpl::getCharacterInfos() const
{
CharacterInfos result;
switch (m_parentFont->getFontType())
{
case FontType::Type1:
case FontType::TrueType:
case FontType::MMType1:
{
// We can use encoding
Q_ASSERT(dynamic_cast<PDFSimpleFont*>(m_parentFont.get()));
const PDFSimpleFont* font = static_cast<PDFSimpleFont*>(m_parentFont.get());
const encoding::EncodingTable* encoding = font->getEncoding();
const GlyphIndices* glyphIndices = font->getGlyphIndices();
for (size_t i = 0; i < encoding->size(); ++i)
{
QChar character = (*encoding)[i];
GID glyphIndex = (*glyphIndices)[static_cast<uint8_t>(i)];
if (!glyphIndex)
{
// Try to obtain glyph index from unicode
if (m_face->charmap && m_face->charmap->encoding == FT_ENCODING_UNICODE)
{
glyphIndex = FT_Get_Char_Index(m_face, character.unicode());
}
}
if (glyphIndex)
{
CharacterInfo info;
info.gid = glyphIndex;
info.character = character;
result.emplace_back(qMove(info));
}
}
break;
}
case FontType::Type0:
{
Q_ASSERT(dynamic_cast<PDFType0Font*>(m_parentFont.get()));
const PDFType0Font* font = static_cast<PDFType0Font*>(m_parentFont.get());
const PDFFontCMap* toUnicode = font->getToUnicode();
const PDFCIDtoGIDMapper* CIDtoGIDmapper = font->getCIDtoGIDMapper();
FT_UInt index = 0;
FT_ULong character = FT_Get_First_Char(m_face, &index);
while (index != 0)
{
const GID gid = index;
const CID cid = CIDtoGIDmapper->unmap(gid);
CharacterInfo info;
info.gid = gid;
info.character = toUnicode->getToUnicode(cid);
result.emplace_back(qMove(info));
character = FT_Get_Next_Char(m_face, character, &index);
}
if (result.empty())
{
// We will try all reasonable high CIDs
for (CID cid = 0; cid < QChar::LastValidCodePoint; ++cid)
{
const GID gid = CIDtoGIDmapper->map(cid);
if (!gid)
{
continue;
}
if (!FT_Load_Glyph(m_face, gid, FT_LOAD_NO_BITMAP | FT_LOAD_NO_HINTING))
{
CharacterInfo info;
info.gid = gid;
info.character = toUnicode->getToUnicode(cid);
result.emplace_back(qMove(info));
}
}
}
break;
}
default:
{
// Unhandled font type
Q_ASSERT(false);
break;
}
}
return result;
}
void PDFRealizedFontImpl::dumpFontToTreeItem(QTreeWidgetItem* item) const
{
QTreeWidgetItem* root = new QTreeWidgetItem(item, { PDFTranslationContext::tr("Details") });
if (m_face->family_name)
{
new QTreeWidgetItem(root, { PDFTranslationContext::tr("Font"), QString::fromLatin1(m_face->family_name) });
}
if (m_face->style_name)
{
new QTreeWidgetItem(root, { PDFTranslationContext::tr("Style"), QString::fromLatin1(m_face->style_name) });
}
QString yesString = PDFTranslationContext::tr("Yes");
QString noString = PDFTranslationContext::tr("No");
new QTreeWidgetItem(root, { PDFTranslationContext::tr("Glyph count"), QString::number(m_face->num_glyphs) });
new QTreeWidgetItem(root, { PDFTranslationContext::tr("Is CID keyed"), (m_face->face_flags & FT_FACE_FLAG_CID_KEYED) ? yesString : noString });
new QTreeWidgetItem(root, { PDFTranslationContext::tr("Is bold"), (m_face->style_flags & FT_STYLE_FLAG_BOLD) ? yesString : noString });
new QTreeWidgetItem(root, { PDFTranslationContext::tr("Is italics"), (m_face->style_flags & FT_STYLE_FLAG_ITALIC) ? yesString : noString });
new QTreeWidgetItem(root, { PDFTranslationContext::tr("Has vertical writing system"), (m_face->face_flags & FT_FACE_FLAG_VERTICAL) ? yesString : noString });
new QTreeWidgetItem(root, { PDFTranslationContext::tr("Has SFNT storage scheme"), (m_face->face_flags & FT_FACE_FLAG_SFNT) ? yesString : noString });
new QTreeWidgetItem(root, { PDFTranslationContext::tr("Has glyph names"), (m_face->face_flags & FT_FACE_FLAG_GLYPH_NAMES) ? yesString : noString });
if (m_face->num_charmaps > 0)
{
QTreeWidgetItem* encodingRoot = new QTreeWidgetItem(item, { PDFTranslationContext::tr("Encoding") });
for (FT_Int i = 0; i < m_face->num_charmaps; ++i)
{
FT_CharMap charMap = m_face->charmaps[i];
const FT_Encoding encoding = charMap->encoding;
QString encodingName;
switch (encoding)
{
case FT_ENCODING_NONE:
encodingName = PDFTranslationContext::tr("None");
break;
case FT_ENCODING_UNICODE:
encodingName = PDFTranslationContext::tr("Unicode");
break;
case FT_ENCODING_MS_SYMBOL:
encodingName = PDFTranslationContext::tr("MS Symbol");
break;
case FT_ENCODING_SJIS:
encodingName = PDFTranslationContext::tr("Japanese Shift JIS");
break;
case FT_ENCODING_PRC:
encodingName = PDFTranslationContext::tr("PRC - Simplified Chinese");
break;
case FT_ENCODING_BIG5:
encodingName = PDFTranslationContext::tr("Traditional Chinese");
break;
case FT_ENCODING_WANSUNG:
encodingName = PDFTranslationContext::tr("Korean Extended Wansung");
break;
case FT_ENCODING_JOHAB:
encodingName = PDFTranslationContext::tr("Korean Standard");
break;
case FT_ENCODING_ADOBE_STANDARD:
encodingName = PDFTranslationContext::tr("Adobe Standard");
break;
case FT_ENCODING_ADOBE_EXPERT:
encodingName = PDFTranslationContext::tr("Adobe Expert");
break;
case FT_ENCODING_ADOBE_CUSTOM:
encodingName = PDFTranslationContext::tr("Adobe Custom");
break;
case FT_ENCODING_ADOBE_LATIN_1:
encodingName = PDFTranslationContext::tr("Adobe Latin 1");
break;
case FT_ENCODING_OLD_LATIN_2:
encodingName = PDFTranslationContext::tr("Old Latin 1");
break;
case FT_ENCODING_APPLE_ROMAN:
encodingName = PDFTranslationContext::tr("Apple Roman");
break;
default:
encodingName = PDFTranslationContext::tr("Unknown");
break;
}
QString encodingString = PDFTranslationContext::tr("Platform/Encoding = %1 %2").arg(charMap->platform_id).arg(charMap->encoding_id);
new QTreeWidgetItem(encodingRoot, { encodingName, encodingString });
}
}
}
int PDFRealizedFontImpl::outlineMoveTo(const FT_Vector* to, void* user)
{
Glyph* glyph = reinterpret_cast<Glyph*>(user);
glyph->glyph.moveTo(to->x * FONT_MULTIPLIER, to->y * FONT_MULTIPLIER);
return 0;
}
int PDFRealizedFontImpl::outlineLineTo(const FT_Vector* to, void* user)
{
Glyph* glyph = reinterpret_cast<Glyph*>(user);
glyph->glyph.lineTo(to->x * FONT_MULTIPLIER, to->y * FONT_MULTIPLIER);
return 0;
}
int PDFRealizedFontImpl::outlineConicTo(const FT_Vector* control, const FT_Vector* to, void* user)
{
Glyph* glyph = reinterpret_cast<Glyph*>(user);
glyph->glyph.quadTo(control->x * FONT_MULTIPLIER, control->y * FONT_MULTIPLIER, to->x * FONT_MULTIPLIER, to->y * FONT_MULTIPLIER);
return 0;
}
int PDFRealizedFontImpl::outlineCubicTo(const FT_Vector* control1, const FT_Vector* control2, const FT_Vector* to, void* user)
{
Glyph* glyph = reinterpret_cast<Glyph*>(user);
glyph->glyph.cubicTo(control1->x * FONT_MULTIPLIER, control1->y * FONT_MULTIPLIER, control2->x * FONT_MULTIPLIER, control2->y * FONT_MULTIPLIER, to->x * FONT_MULTIPLIER, to->y * FONT_MULTIPLIER);
return 0;
}
const PDFRealizedFontImpl::Glyph& PDFRealizedFontImpl::getGlyph(unsigned int glyphIndex)
{
if (glyphIndex)
{
{
QReadLocker readLock(&m_readWriteLock);
// First look into cache
auto it = m_glyphCache.find(glyphIndex);
if (it != m_glyphCache.cend())
{
return it->second;
}
}
QWriteLocker writeLock(&m_readWriteLock);
Glyph glyph;
FT_Outline_Funcs glyphOutlineInterface;
glyphOutlineInterface.delta = 0;
glyphOutlineInterface.shift = 0;
glyphOutlineInterface.move_to = PDFRealizedFontImpl::outlineMoveTo;
glyphOutlineInterface.line_to = PDFRealizedFontImpl::outlineLineTo;
glyphOutlineInterface.conic_to = PDFRealizedFontImpl::outlineConicTo;
glyphOutlineInterface.cubic_to = PDFRealizedFontImpl::outlineCubicTo;
checkFreeTypeError(FT_Load_Glyph(m_face, glyphIndex, FT_LOAD_NO_BITMAP | FT_LOAD_NO_HINTING));
checkFreeTypeError(FT_Outline_Decompose(&m_face->glyph->outline, &glyphOutlineInterface, &glyph));
glyph.glyph.closeSubpath();
glyph.advance = !m_isVertical ? m_face->glyph->advance.x : m_face->glyph->advance.y;
glyph.advance *= FONT_MULTIPLIER;
auto it = m_glyphCache.find(glyphIndex);
if (it == m_glyphCache.cend())
{
it = m_glyphCache.insert(std::make_pair(glyphIndex, qMove(glyph))).first;
}
return it->second;
}
static Glyph dummy;
return dummy;
}
void PDFRealizedFontImpl::checkFreeTypeError(FT_Error error)
{
if (error)
{
QString message;
if (const char* errorString = FT_Error_String(error))
{
message = QString::fromLatin1(errorString);
}
throw PDFException(PDFTranslationContext::tr("FreeType error code %1: %2").arg(error).arg(message));
}
}
PDFRealizedFont::~PDFRealizedFont()
{
delete m_impl;
}
void PDFRealizedFont::fillTextSequence(const QByteArray& byteArray, TextSequence& textSequence, PDFRenderErrorReporter* reporter)
{
m_impl->fillTextSequence(byteArray, textSequence, reporter);
}
bool PDFRealizedFont::isHorizontalWritingSystem() const
{
return m_impl->isHorizontalWritingSystem();
}
void PDFRealizedFont::dumpFontToTreeItem(QTreeWidgetItem* item) const
{
m_impl->dumpFontToTreeItem(item);
}
QString PDFRealizedFont::getPostScriptName() const
{
return m_impl->getPostScriptName();
}
CharacterInfos PDFRealizedFont::getCharacterInfos() const
{
return m_impl->getCharacterInfos();
}
PDFRealizedFontPointer PDFRealizedFont::createRealizedFont(PDFFontPointer font, PDFReal pixelSize, PDFRenderErrorReporter* reporter)
{
PDFRealizedFontPointer result;
if (font->getFontType() == FontType::Type3)
{
result.reset(new PDFRealizedFont(new PDFRealizedType3FontImpl(font, pixelSize)));
}
else
{
std::unique_ptr<PDFRealizedFontImpl> implPtr(new PDFRealizedFontImpl());
PDFRealizedFontImpl* impl = implPtr.get();
impl->m_parentFont = font;
impl->m_pixelSize = pixelSize;
const FontDescriptor* descriptor = font->getFontDescriptor();
if (descriptor->isEmbedded())
{
PDFRealizedFontImpl::checkFreeTypeError(FT_Init_FreeType(&impl->m_library));
const QByteArray* embeddedFontData = descriptor->getEmbeddedFontData();
Q_ASSERT(embeddedFontData);
impl->m_embeddedFontData = *embeddedFontData;
// At this time, embedded font data should not be empty!
Q_ASSERT(!impl->m_embeddedFontData.isEmpty());
PDFRealizedFontImpl::checkFreeTypeError(FT_New_Memory_Face(impl->m_library, reinterpret_cast<const FT_Byte*>(impl->m_embeddedFontData.constData()), impl->m_embeddedFontData.size(), 0, &impl->m_face));
FT_Select_Charmap(impl->m_face, FT_ENCODING_UNICODE); // We try to select unicode encoding, but if it fails, we don't do anything (use glyph indices instead)
PDFRealizedFontImpl::checkFreeTypeError(FT_Set_Pixel_Sizes(impl->m_face, 0, qRound(pixelSize * PDFRealizedFontImpl::PIXEL_SIZE_MULTIPLIER)));
impl->m_isVertical = impl->m_face->face_flags & FT_FACE_FLAG_VERTICAL;
impl->m_isEmbedded = true;
result.reset(new PDFRealizedFont(implPtr.release()));
}
else
{
StandardFontType standardFontType = StandardFontType::Invalid;
if (font->getFontType() == FontType::Type1 || font->getFontType() == FontType::MMType1)
{
Q_ASSERT(dynamic_cast<const PDFType1Font*>(font.get()));
const PDFType1Font* type1Font = static_cast<const PDFType1Font*>(font.get());
standardFontType = type1Font->getStandardFontType();
}
const PDFSystemFontInfoStorage* fontStorage = PDFSystemFontInfoStorage::getInstance();
impl->m_systemFontData = fontStorage->loadFont(descriptor, standardFontType, reporter);
if (impl->m_systemFontData.isEmpty())
{
throw PDFException(PDFTranslationContext::tr("Can't load system font '%1'.").arg(QString::fromLatin1(descriptor->fontName)));
}
PDFRealizedFontImpl::checkFreeTypeError(FT_Init_FreeType(&impl->m_library));
PDFRealizedFontImpl::checkFreeTypeError(FT_New_Memory_Face(impl->m_library, reinterpret_cast<const FT_Byte*>(impl->m_systemFontData.constData()), impl->m_systemFontData.size(), 0, &impl->m_face));
FT_Select_Charmap(impl->m_face, FT_ENCODING_UNICODE); // We try to select unicode encoding, but if it fails, we don't do anything (use glyph indices instead)
PDFRealizedFontImpl::checkFreeTypeError(FT_Set_Pixel_Sizes(impl->m_face, 0, qRound(pixelSize * PDFRealizedFontImpl::PIXEL_SIZE_MULTIPLIER)));
impl->m_isVertical = impl->m_face->face_flags & FT_FACE_FLAG_VERTICAL;
impl->m_isEmbedded = false;
if (const char* postScriptName = FT_Get_Postscript_Name(impl->m_face))
{
impl->m_postScriptName = QString::fromLatin1(postScriptName);
}
result.reset(new PDFRealizedFont(implPtr.release()));
}
}
return result;
}
FontDescriptor PDFFont::readFontDescriptor(const PDFObject& fontDescriptorObject, const PDFDocument* document)
{
FontDescriptor fontDescriptor;
PDFDocumentDataLoaderDecorator fontLoader(document);
if (fontDescriptorObject.isDictionary())
{
const PDFDictionary* fontDescriptorDictionary = fontDescriptorObject.getDictionary();
fontDescriptor.fontName = fontLoader.readNameFromDictionary(fontDescriptorDictionary, "FontName");
fontDescriptor.fontFamily = fontLoader.readStringFromDictionary(fontDescriptorDictionary, "FontFamily");
constexpr const std::array<std::pair<const char*, QFont::Stretch>, 9> stretches = {
std::pair<const char*, QFont::Stretch>{ "UltraCondensed", QFont::UltraCondensed },
std::pair<const char*, QFont::Stretch>{ "ExtraCondensed", QFont::ExtraCondensed },
std::pair<const char*, QFont::Stretch>{ "Condensed", QFont::Condensed },
std::pair<const char*, QFont::Stretch>{ "SemiCondensed", QFont::SemiCondensed },
std::pair<const char*, QFont::Stretch>{ "Normal", QFont::Unstretched },
std::pair<const char*, QFont::Stretch>{ "SemiExpanded", QFont::SemiExpanded },
std::pair<const char*, QFont::Stretch>{ "Expanded", QFont::Expanded },
std::pair<const char*, QFont::Stretch>{ "ExtraExpanded", QFont::ExtraExpanded },
std::pair<const char*, QFont::Stretch>{ "UltraExpanded", QFont::UltraExpanded }
};
fontDescriptor.fontStretch = fontLoader.readEnumByName(fontDescriptorDictionary->get("FontStretch"), stretches.cbegin(), stretches.cend(), QFont::Unstretched);
fontDescriptor.fontWeight = fontLoader.readNumberFromDictionary(fontDescriptorDictionary, "FontWeight", 500);
fontDescriptor.italicAngle = fontLoader.readNumberFromDictionary(fontDescriptorDictionary, "ItalicAngle", 0.0);
fontDescriptor.ascent = fontLoader.readNumberFromDictionary(fontDescriptorDictionary, "Ascent", 0.0);
fontDescriptor.descent = fontLoader.readNumberFromDictionary(fontDescriptorDictionary, "Descent", 0.0);
fontDescriptor.leading = fontLoader.readNumberFromDictionary(fontDescriptorDictionary, "Leading", 0.0);
fontDescriptor.capHeight = fontLoader.readNumberFromDictionary(fontDescriptorDictionary, "CapHeight", 0.0);
fontDescriptor.xHeight = fontLoader.readNumberFromDictionary(fontDescriptorDictionary, "XHeight", 0.0);
fontDescriptor.stemV = fontLoader.readNumberFromDictionary(fontDescriptorDictionary, "StemV", 0.0);
fontDescriptor.stemH = fontLoader.readNumberFromDictionary(fontDescriptorDictionary, "StemH", 0.0);
fontDescriptor.avgWidth = fontLoader.readNumberFromDictionary(fontDescriptorDictionary, "AvgWidth", 0.0);
fontDescriptor.maxWidth = fontLoader.readNumberFromDictionary(fontDescriptorDictionary, "MaxWidth", 0.0);
fontDescriptor.missingWidth = fontLoader.readNumberFromDictionary(fontDescriptorDictionary, "MissingWidth", 0.0);
fontDescriptor.flags = fontLoader.readIntegerFromDictionary(fontDescriptorDictionary, "Flags", 0);
fontDescriptor.boundingBox = fontLoader.readRectangle(fontDescriptorDictionary->get("FontBBox"), QRectF());
fontDescriptor.charset = fontLoader.readStringFromDictionary(fontDescriptorDictionary, "Charset");
auto loadStream = [fontDescriptorDictionary, document](QByteArray& byteArray, const char* name)
{
if (fontDescriptorDictionary->hasKey(name))
{
const PDFObject& streamObject = document->getObject(fontDescriptorDictionary->get(name));
if (streamObject.isStream())
{
byteArray = document->getDecodedStream(streamObject.getStream());
}
}
};
loadStream(fontDescriptor.fontFile, "FontFile");
loadStream(fontDescriptor.fontFile2, "FontFile2");
loadStream(fontDescriptor.fontFile3, "FontFile3");
}
return fontDescriptor;
}
PDFFontPointer PDFFont::createFont(const PDFObject& object, const PDFDocument* document)
{
const PDFObject& dereferencedFontDictionary = document->getObject(object);
if (!dereferencedFontDictionary.isDictionary())
{
throw PDFException(PDFTranslationContext::tr("Font object must be a dictionary."));
}
const PDFDictionary* fontDictionary = dereferencedFontDictionary.getDictionary();
PDFDocumentDataLoaderDecorator fontLoader(document);
// First, determine the font subtype
constexpr const std::array fontTypes = {
std::pair<const char*, FontType>{ "Type0", FontType::Type0 },
std::pair<const char*, FontType>{ "Type1", FontType::Type1 },
std::pair<const char*, FontType>{ "TrueType", FontType::TrueType },
std::pair<const char*, FontType>{ "Type3", FontType::Type3},
std::pair<const char*, FontType>{ "MMType1", FontType::MMType1 }
};
const FontType fontType = fontLoader.readEnumByName(fontDictionary->get("Subtype"), fontTypes.cbegin(), fontTypes.cend(), FontType::Invalid);
if (fontType == FontType::Invalid)
{
throw PDFException(PDFTranslationContext::tr("Invalid font type."));
}
QByteArray name = fontLoader.readNameFromDictionary(fontDictionary, "Name");
QByteArray baseFont = fontLoader.readNameFromDictionary(fontDictionary, "BaseFont");
const PDFInteger firstChar = fontLoader.readIntegerFromDictionary(fontDictionary, "FirstChar", 0);
const PDFInteger lastChar = fontLoader.readIntegerFromDictionary(fontDictionary, "LastChar", 255);
std::vector<PDFInteger> widths = fontLoader.readIntegerArrayFromDictionary(fontDictionary, "Widths");
// Read standard font
constexpr const std::array<std::pair<const char*, StandardFontType>, 14> standardFonts = {
std::pair<const char*, StandardFontType>{ "Times-Roman", StandardFontType::TimesRoman },
std::pair<const char*, StandardFontType>{ "Times-Bold", StandardFontType::TimesRomanBold },
std::pair<const char*, StandardFontType>{ "Times-Italic", StandardFontType::TimesRomanItalics },
std::pair<const char*, StandardFontType>{ "Times-BoldItalic", StandardFontType::TimesRomanBoldItalics },
std::pair<const char*, StandardFontType>{ "Helvetica", StandardFontType::Helvetica },
std::pair<const char*, StandardFontType>{ "Helvetica-Bold", StandardFontType::HelveticaBold },
std::pair<const char*, StandardFontType>{ "Helvetica-Oblique", StandardFontType::HelveticaOblique },
std::pair<const char*, StandardFontType>{ "Helvetica-BoldOblique", StandardFontType::HelveticaBoldOblique },
std::pair<const char*, StandardFontType>{ "Courier", StandardFontType::Courier },
std::pair<const char*, StandardFontType>{ "Courier-Bold", StandardFontType::CourierBold },
std::pair<const char*, StandardFontType>{ "Courier-Oblique", StandardFontType::CourierOblique },
std::pair<const char*, StandardFontType>{ "Courier-BoldOblique", StandardFontType::CourierBoldOblique },
std::pair<const char*, StandardFontType>{ "Symbol", StandardFontType::Symbol },
std::pair<const char*, StandardFontType>{ "ZapfDingbats", StandardFontType::ZapfDingbats }
};
const StandardFontType standardFont = fontLoader.readEnumByName(fontDictionary->get("BaseFont"), standardFonts.cbegin(), standardFonts.cend(), StandardFontType::Invalid);
// Read Font Descriptor
const PDFObject& fontDescriptorObject = document->getObject(fontDictionary->get("FontDescriptor"));
FontDescriptor fontDescriptor = readFontDescriptor(fontDescriptorObject, document);
// Read Font Encoding
// The font encoding for the simple font is determined by this algorithm:
// 1) Try to use Encoding dictionary to determine base encoding
// (it can be MacRomanEncoding, MacExpertEncoding, WinAnsiEncoding or StandardEncoding)
// 2) If it is not present, then try to obtain built-in encoding from the font file (usually, this is not possible)
// 3) Use default encoding for the font depending on the font type
// - one of the 14 base fonts - use builtin encoding for the font type
// - TrueType - use WinAnsiEncoding
// - all others - use StandardEncoding
// 4) Merge with Differences, if present
// 5) Fill missing characters from StandardEncoding
// After the encoding is obtained, try to extract glyph indices for embedded font.
PDFEncoding::Encoding encoding = PDFEncoding::Encoding::Invalid;
encoding::EncodingTable simpleFontEncodingTable = { };
GlyphIndices glyphIndexArray = { };
switch (fontType)
{
case FontType::Type1:
case FontType::MMType1:
case FontType::TrueType:
{
bool hasDifferences = false;
encoding::EncodingTable differences = { };
if (fontDictionary->hasKey("Encoding"))
{
constexpr const std::array<std::pair<const char*, PDFEncoding::Encoding>, 3> encodings = {
std::pair<const char*, PDFEncoding::Encoding>{ "MacRomanEncoding", PDFEncoding::Encoding::MacRoman },
std::pair<const char*, PDFEncoding::Encoding>{ "MacExpertEncoding", PDFEncoding::Encoding::MacExpert },
std::pair<const char*, PDFEncoding::Encoding>{ "WinAnsiEncoding", PDFEncoding::Encoding::WinAnsi }
};
const PDFObject& encodingObject = document->getObject(fontDictionary->get("Encoding"));
if (encodingObject.isName())
{
// Decode name of the encoding
encoding = fontLoader.readEnumByName(encodingObject, encodings.cbegin(), encodings.cend(), PDFEncoding::Encoding::Invalid);
}
else if (encodingObject.isDictionary())
{
// Dictionary with base encoding and differences (all optional)
const PDFDictionary* encodingDictionary = encodingObject.getDictionary();
if (encodingDictionary->hasKey("BaseEncoding"))
{
encoding = fontLoader.readEnumByName(encodingDictionary->get("BaseEncoding"), encodings.cbegin(), encodings.cend(), PDFEncoding::Encoding::Invalid);
}
else
{
// We get encoding for the standard font. If we have invalid standard font,
// then we get standard encoding. So we shouldn't test it.
encoding = getEncodingForStandardFont(standardFont);
}
if (encodingDictionary->hasKey("Differences"))
{
const PDFObject& differencesArray = document->getObject(encodingDictionary->get("Differences"));
if (differencesArray.isArray())
{
hasDifferences = true;
const PDFArray* array = differencesArray.getArray();
size_t currentOffset = 0;
for (size_t i = 0, count = array->getCount(); i < count; ++i)
{
const PDFObject& item = document->getObject(array->getItem(i));
if (item.isInt())
{
currentOffset = static_cast<size_t>(item.getInteger());
}
else if (item.isName())
{
if (currentOffset >= differences.size())
{
throw PDFException(PDFTranslationContext::tr("Invalid differences in encoding entry of the font."));
}
QChar character = PDFNameToUnicode::getUnicodeUsingResolvedName(item.getString());
differences[currentOffset] = character;
++currentOffset;
}
else
{
throw PDFException(PDFTranslationContext::tr("Invalid differences in encoding entry of the font."));
}
}
}
else
{
throw PDFException(PDFTranslationContext::tr("Invalid differences in encoding entry of the font."));
}
}
}
else
{
throw PDFException(PDFTranslationContext::tr("Invalid encoding entry of the font."));
}
}
if (encoding == PDFEncoding::Encoding::Invalid)
{
// We get encoding for the standard font. If we have invalid standard font,
// then we get standard encoding. So we shouldn't test it.
encoding = getEncodingForStandardFont(standardFont);
}
if (encoding == PDFEncoding::Encoding::Invalid)
{
throw PDFException(PDFTranslationContext::tr("Invalid encoding entry of the font."));
}
simpleFontEncodingTable = *PDFEncoding::getTableForEncoding(encoding);
auto finishFont = [&]
{
// Fill in differences
if (hasDifferences)
{
for (size_t i = 0; i < differences.size(); ++i)
{
if (!differences[i].isNull())
{
simpleFontEncodingTable[i] = differences[i];
}
}
// Set the encoding to custom
encoding = PDFEncoding::Encoding::Custom;
}
// Fill in missing characters from standard encoding
const encoding::EncodingTable& standardEncoding = *PDFEncoding::getTableForEncoding(PDFEncoding::Encoding::Standard);
for (size_t i = 0; i < standardEncoding.size(); ++i)
{
if ((simpleFontEncodingTable[i].isNull() || simpleFontEncodingTable[i] == QChar(QChar::SpecialCharacter::ReplacementCharacter)) &&
(!standardEncoding[i].isNull() && standardEncoding[i] != QChar(QChar::SpecialCharacter::ReplacementCharacter)))
{
simpleFontEncodingTable[i] = standardEncoding[i];
}
}
};
if (fontDescriptor.isEmbedded())
{
// Return encoding from the embedded font
const QByteArray* embeddedFontData = fontDescriptor.getEmbeddedFontData();
Q_ASSERT(embeddedFontData);
FT_Library library;
if (!FT_Init_FreeType(&library))
{
FT_Face face;
if (!FT_New_Memory_Face(library, reinterpret_cast<const FT_Byte*>(embeddedFontData->constData()), embeddedFontData->size(), 0, &face))
{
if (FT_Has_PS_Glyph_Names(face))
{
for (FT_Int i = 0; i < face->num_charmaps; ++i)
{
FT_CharMap charMap = face->charmaps[i];
switch (charMap->encoding)
{
case FT_ENCODING_ADOBE_STANDARD:
case FT_ENCODING_ADOBE_LATIN_1:
case FT_ENCODING_ADOBE_CUSTOM:
case FT_ENCODING_ADOBE_EXPERT:
{
// Try to load data from the encoding
if (!FT_Set_Charmap(face, charMap))
{
for (size_t i = 0; i < simpleFontEncodingTable.size(); ++i)
{
FT_UInt glyphIndex = FT_Get_Char_Index(face, static_cast<FT_ULong>(i));
if (glyphIndex == 0)
{
glyphIndex = FT_Get_Char_Index(face, static_cast<FT_ULong>(i + 0xF000));
}
if (glyphIndex == 0)
{
glyphIndex = FT_Get_Char_Index(face, static_cast<FT_ULong>(i + 0xF100));
}
if (glyphIndex > 0)
{
// Fill the glyph index array
glyphIndexArray[i] = glyphIndex;
// Set mapping to unicode
char buffer[128] = { };
if (!FT_Get_Glyph_Name(face, glyphIndex, buffer, static_cast<FT_ULong>(std::size(buffer))))
{
QByteArray byteArrayBuffer(buffer);
QChar character = PDFNameToUnicode::getUnicodeForName(byteArrayBuffer);
if (character.isNull())
{
character = PDFNameToUnicode::getUnicodeForNameZapfDingbats(byteArrayBuffer);
}
if (!character.isNull())
{
encoding = PDFEncoding::Encoding::Custom;
simpleFontEncodingTable[i] = character;
}
}
}
}
}
break;
}
default:
break;
}
}
}
else if (!FT_Select_Charmap(face, FT_ENCODING_APPLE_ROMAN))
{
// We have (1, 0) Mac Roman Encoding, which is slightly different, than Mac Roman Encoding defined
// in PDF (for 15 characters).
simpleFontEncodingTable = *PDFEncoding::getTableForEncoding(PDFEncoding::Encoding::MacOsRoman);
encoding = PDFEncoding::Encoding::Custom;
for (size_t i = 0; i < simpleFontEncodingTable.size(); ++i)
{
FT_UInt glyphIndex = FT_Get_Char_Index(face, static_cast<FT_ULong>(i));
if (glyphIndex > 0)
{
glyphIndexArray[i] = glyphIndex;
}
}
}
finishFont();
// Fill the glyph index array from unicode, if we have unicode mapping
if (!FT_Select_Charmap(face, FT_ENCODING_UNICODE))
{
for (size_t i = 0; i < simpleFontEncodingTable.size(); ++i)
{
QChar character = simpleFontEncodingTable[i];
if (!character.isNull() && character != QChar(QChar::SpecialCharacter::ReplacementCharacter))
{
const FT_UInt glyphIndex = FT_Get_Char_Index(face, character.unicode());
if (glyphIndex > 0)
{
glyphIndexArray[i] = glyphIndex;
}
}
}
}
FT_Done_Face(face);
}
FT_Done_FreeType(library);
}
}
else
{
// Finish font - fill differences
finishFont();
}
break;
}
case FontType::Type0:
{
// This is composite font (CID keyed font)
// Load CMAP
PDFFontCMap cmap;
const PDFObject& cmapObject = document->getObject(fontDictionary->get("Encoding"));
if (cmapObject.isName())
{
cmap = PDFFontCMap::createFromName(cmapObject.getString());
}
else if (cmapObject.isStream())
{
const PDFStream* stream = cmapObject.getStream();
QByteArray decodedStream = document->getDecodedStream(stream);
cmap = PDFFontCMap::createFromData(decodedStream);
}
if (!cmap.isValid())
{
throw PDFException(PDFTranslationContext::tr("Invalid CMAP in CID-keyed font."));
}
const PDFObject& descendantFonts = document->getObject(fontDictionary->get("DescendantFonts"));
if (!descendantFonts.isArray())
{
throw PDFException(PDFTranslationContext::tr("Invalid descendant font in CID-keyed font."));
}
const PDFArray* descendantFontsArray = descendantFonts.getArray();
if (descendantFontsArray->getCount() != 1)
{
throw PDFException(PDFTranslationContext::tr("Invalid number (%1) of descendant fonts in CID-keyed font - exactly one is required.").arg(descendantFontsArray->getCount()));
}
const PDFObject& descendantFont = document->getObject(descendantFontsArray->getItem(0));
if (!descendantFont.isDictionary())
{
throw PDFException(PDFTranslationContext::tr("Invalid descendant font in CID-keyed font."));
}
const PDFDictionary* descendantFontDictionary = descendantFont.getDictionary();
const PDFObject& fontDescriptorObjectForCompositeFont = document->getObject(descendantFontDictionary->get("FontDescriptor"));
fontDescriptor = readFontDescriptor(fontDescriptorObjectForCompositeFont, document);
QByteArray cidToGidMapping;
const PDFObject& cidToGidMappingObject = document->getObject(descendantFontDictionary->get("CIDtoGIDMap"));
if (cidToGidMappingObject.isStream())
{
const PDFStream* cidToGidMappingStream = cidToGidMappingObject.getStream();
cidToGidMapping = document->getDecodedStream(cidToGidMappingStream);
}
PDFCIDtoGIDMapper cidToGidMapper(qMove(cidToGidMapping));
baseFont = fontLoader.readNameFromDictionary(descendantFontDictionary, "BaseFont");
// Read default advance
PDFReal dw = fontLoader.readNumberFromDictionary(descendantFontDictionary, "DW", 1000.0);
std::array<PDFReal, 2> dw2 = { };
fontLoader.readNumberArrayFromDictionary(descendantFontDictionary, "DW2", dw2.begin(), dw2.end());
PDFReal defaultWidth = descendantFontDictionary->hasKey("DW") ? dw : dw2.back();
// Read horizontal advances
std::unordered_map<CID, PDFReal> advances;
if (descendantFontDictionary->hasKey("W"))
{
const PDFObject& wArrayObject = document->getObject(descendantFontDictionary->get("W"));
if (wArrayObject.isArray())
{
const PDFArray* wArray = wArrayObject.getArray();
const size_t size = wArray->getCount();
for (size_t i = 0; i < size;)
{
CID startCID = fontLoader.readInteger(wArray->getItem(i++), 0);
const PDFObject& arrayOrCID = document->getObject(wArray->getItem(i++));
if (arrayOrCID.isInt())
{
CID endCID = arrayOrCID.getInteger();
PDFReal width = fontLoader.readInteger(wArray->getItem(i++), 0);
for (CID currentCID = startCID; currentCID <= endCID; ++currentCID)
{
advances[currentCID] = width;
}
}
else if (arrayOrCID.isArray())
{
const PDFArray* widthArray = arrayOrCID.getArray();
const size_t widthArraySize = widthArray->getCount();
for (size_t widthArrayIndex = 0; widthArrayIndex < widthArraySize; ++widthArrayIndex)
{
PDFReal width = fontLoader.readNumber(widthArray->getItem(widthArrayIndex), 0);
advances[startCID + static_cast<CID>(widthArrayIndex)] = width;
}
}
}
}
}
PDFFontCMap toUnicodeCMap;
const PDFObject& toUnicode = document->getObject(fontDictionary->get("ToUnicode"));
if (toUnicode.isName())
{
toUnicodeCMap = PDFFontCMap::createFromName(toUnicode.getString());
}
else if (toUnicode.isStream())
{
const PDFStream* stream = toUnicode.getStream();
QByteArray decodedStream = document->getDecodedStream(stream);
toUnicodeCMap = PDFFontCMap::createFromData(decodedStream);
}
return PDFFontPointer(new PDFType0Font(qMove(fontDescriptor), qMove(cmap), qMove(toUnicodeCMap), qMove(cidToGidMapper), defaultWidth, qMove(advances)));
}
case FontType::Type3:
{
// Read the font matrix
std::vector<PDFReal> fontMatrixValues = fontLoader.readNumberArrayFromDictionary(fontDictionary, "FontMatrix");
if (fontMatrixValues.size() != 6)
{
throw PDFException(PDFTranslationContext::tr("Invalid Type 3 font matrix."));
}
QMatrix fontMatrix(fontMatrixValues[0], fontMatrixValues[1], fontMatrixValues[2], fontMatrixValues[3], fontMatrixValues[4], fontMatrixValues[5]);
PDFObject charProcs = document->getObject(fontDictionary->get("CharProcs"));
if (!charProcs.isDictionary())
{
throw PDFException(PDFTranslationContext::tr("Invalid Type 3 font character content streams."));
}
const PDFDictionary* charProcsDictionary = charProcs.getDictionary();
PDFInteger firstChar = fontLoader.readIntegerFromDictionary(fontDictionary, "FirstChar", -1);
PDFInteger lastChar = fontLoader.readIntegerFromDictionary(fontDictionary, "LastChar", -1);
if (firstChar < 0 || lastChar > 255 || firstChar > lastChar)
{
throw PDFException(PDFTranslationContext::tr("Invalid Type 3 font character range (from %1 to %2).").arg(firstChar).arg(lastChar));
}
const PDFObject& encoding = document->getObject(fontDictionary->get("Encoding"));
if (!encoding.isDictionary())
{
throw PDFException(PDFTranslationContext::tr("Invalid Type 3 font encoding."));
}
const PDFDictionary* encodingDictionary = encoding.getDictionary();
const PDFObject& differences = document->getObject(encodingDictionary->get("Differences"));
if (!differences.isArray())
{
throw PDFException(PDFTranslationContext::tr("Invalid Type 3 font encoding."));
}
std::map<int, QByteArray> characterContentStreams;
const PDFArray* differencesArray = differences.getArray();
size_t currentOffset = 0;
for (size_t i = 0, count = differencesArray->getCount(); i < count; ++i)
{
const PDFObject& item = document->getObject(differencesArray->getItem(i));
if (item.isInt())
{
currentOffset = static_cast<size_t>(item.getInteger());
}
else if (item.isName())
{
if (currentOffset > 255)
{
throw PDFException(PDFTranslationContext::tr("Invalid differences in encoding entry of type 3 font."));
}
QByteArray characterName = item.getString();
const PDFObject& characterContentStreamObject = document->getObject(charProcsDictionary->get(characterName));
if (characterContentStreamObject.isStream())
{
QByteArray contentStream = document->getDecodedStream(characterContentStreamObject.getStream());
characterContentStreams[static_cast<int>(currentOffset)] = qMove(contentStream);
}
++currentOffset;
}
else
{
throw PDFException(PDFTranslationContext::tr("Invalid differences in encoding entry of type 3 font."));
}
}
PDFFontCMap toUnicodeCMap;
const PDFObject& toUnicode = document->getObject(fontDictionary->get("ToUnicode"));
if (toUnicode.isName())
{
toUnicodeCMap = PDFFontCMap::createFromName(toUnicode.getString());
}
else if (toUnicode.isStream())
{
const PDFStream* stream = toUnicode.getStream();
QByteArray decodedStream = document->getDecodedStream(stream);
toUnicodeCMap = PDFFontCMap::createFromData(decodedStream);
}
std::vector<PDFReal> widths = fontLoader.readNumberArrayFromDictionary(fontDictionary, "Widths");
return PDFFontPointer(new PDFType3Font(qMove(fontDescriptor), firstChar, lastChar, fontMatrix, qMove(characterContentStreams), qMove(widths), document->getObject(fontDictionary->get("Resources")), qMove(toUnicodeCMap)));
}
default:
{
Q_ASSERT(false);
break;
}
}
switch (fontType)
{
case FontType::Type1:
case FontType::MMType1:
return PDFFontPointer(new PDFType1Font(fontType, qMove(fontDescriptor), qMove(name), qMove(baseFont), firstChar, lastChar, qMove(widths), encoding, simpleFontEncodingTable, standardFont, glyphIndexArray));
case FontType::TrueType:
return PDFFontPointer(new PDFTrueTypeFont(qMove(fontDescriptor), qMove(name), qMove(baseFont), firstChar, lastChar, qMove(widths), encoding, simpleFontEncodingTable, glyphIndexArray));
default:
{
Q_ASSERT(false);
break;
}
}
return PDFFontPointer();
}
PDFSimpleFont::PDFSimpleFont(FontDescriptor fontDescriptor,
QByteArray name,
QByteArray baseFont,
PDFInteger firstChar,
PDFInteger lastChar,
std::vector<PDFInteger> widths,
PDFEncoding::Encoding encodingType,
encoding::EncodingTable encoding,
GlyphIndices glyphIndices) :
PDFFont(qMove(fontDescriptor)),
m_name(qMove(name)),
m_baseFont(qMove(baseFont)),
m_firstChar(firstChar),
m_lastChar(lastChar),
m_widths(qMove(widths)),
m_encodingType(encodingType),
m_encoding(encoding),
m_glyphIndices(glyphIndices)
{
}
PDFInteger PDFSimpleFont::getGlyphAdvance(size_t index) const
{
const size_t min = m_firstChar;
const size_t max = m_lastChar;
if (index >= min && index <= max)
{
const size_t adjustedIndex = index - min;
if (adjustedIndex < m_widths.size())
{
return m_widths[adjustedIndex];
}
}
return 0;
}
void PDFSimpleFont::dumpFontToTreeItem(QTreeWidgetItem* item) const
{
BaseClass::dumpFontToTreeItem(item);
QString encodingTypeString;
switch (m_encodingType)
{
case PDFEncoding::Encoding::Standard:
encodingTypeString = PDFTranslationContext::tr("Standard");
break;
case PDFEncoding::Encoding::MacRoman:
encodingTypeString = PDFTranslationContext::tr("Mac Roman");
break;
case PDFEncoding::Encoding::WinAnsi:
encodingTypeString = PDFTranslationContext::tr("Win Ansi");
break;
case PDFEncoding::Encoding::PDFDoc:
encodingTypeString = PDFTranslationContext::tr("PDF Doc");
break;
case PDFEncoding::Encoding::MacExpert:
encodingTypeString = PDFTranslationContext::tr("Mac Expert");
break;
case PDFEncoding::Encoding::Symbol:
encodingTypeString = PDFTranslationContext::tr("Symbol");
break;
case PDFEncoding::Encoding::ZapfDingbats:
encodingTypeString = PDFTranslationContext::tr("Zapf Dingbats");
break;
case PDFEncoding::Encoding::MacOsRoman:
encodingTypeString = PDFTranslationContext::tr("Mac OS Roman");
break;
case PDFEncoding::Encoding::Custom:
encodingTypeString = PDFTranslationContext::tr("Custom");
break;
default:
{
Q_ASSERT(false);
break;
}
}
new QTreeWidgetItem(item, { PDFTranslationContext::tr("Encoding"), encodingTypeString });
}
PDFType1Font::PDFType1Font(FontType fontType,
FontDescriptor fontDescriptor,
QByteArray name,
QByteArray baseFont,
PDFInteger firstChar,
PDFInteger lastChar,
std::vector<PDFInteger> widths,
PDFEncoding::Encoding encodingType,
encoding::EncodingTable encoding,
StandardFontType standardFontType,
GlyphIndices glyphIndices) :
PDFSimpleFont(qMove(fontDescriptor), qMove(name), qMove(baseFont), firstChar, lastChar, qMove(widths), encodingType, encoding, glyphIndices),
m_fontType(fontType),
m_standardFontType(standardFontType)
{
}
FontType PDFType1Font::getFontType() const
{
return m_fontType;
}
void PDFType1Font::dumpFontToTreeItem(QTreeWidgetItem* item) const
{
BaseClass::dumpFontToTreeItem(item);
if (m_standardFontType != StandardFontType::Invalid)
{
QString standardFontTypeString;
switch (m_standardFontType)
{
case StandardFontType::TimesRoman:
case StandardFontType::TimesRomanBold:
case StandardFontType::TimesRomanItalics:
case StandardFontType::TimesRomanBoldItalics:
standardFontTypeString = PDFTranslationContext::tr("Times Roman");
break;
case StandardFontType::Helvetica:
case StandardFontType::HelveticaBold:
case StandardFontType::HelveticaOblique:
case StandardFontType::HelveticaBoldOblique:
standardFontTypeString = PDFTranslationContext::tr("Helvetica");
break;
case StandardFontType::Courier:
case StandardFontType::CourierBold:
case StandardFontType::CourierOblique:
case StandardFontType::CourierBoldOblique:
standardFontTypeString = PDFTranslationContext::tr("Courier");
break;
case StandardFontType::Symbol:
standardFontTypeString = PDFTranslationContext::tr("Symbol");
break;
case StandardFontType::ZapfDingbats:
standardFontTypeString = PDFTranslationContext::tr("Zapf Dingbats");
break;
default:
Q_ASSERT(false);
break;
}
new QTreeWidgetItem(item, { PDFTranslationContext::tr("Standard font"), standardFontTypeString });
}
}
FontType PDFTrueTypeFont::getFontType() const
{
return FontType::TrueType;
}
void PDFFontCache::setDocument(const PDFModifiedDocument& document)
{
QMutexLocker lock(&m_mutex);
if (m_document != document)
{
m_document = document;
// Jakub Melka: If document has not reset flag, then fonts of the
// document remains the same. So it is not needed to clear font cache.
if (document.hasReset())
{
m_fontCache.clear();
m_realizedFontCache.clear();
}
}
}
PDFFontPointer PDFFontCache::getFont(const PDFObject& fontObject) const
{
if (fontObject.isReference())
{
// Font is object reference. Look in the cache, if we have it, then return it.
QMutexLocker lock(&m_mutex);
PDFObjectReference reference = fontObject.getReference();
auto it = m_fontCache.find(reference);
if (it == m_fontCache.cend())
{
// We must create the font
PDFFontPointer font = PDFFont::createFont(fontObject, m_document);
if (m_fontCacheShrinkDisabledObjects.empty() && m_fontCache.size() >= m_fontCacheLimit)
{
// We have exceeded the cache limit. Clear the cache.
m_fontCache.clear();
}
it = m_fontCache.insert(std::make_pair(reference, qMove(font))).first;
}
return it->second;
}
else
{
// Object is not a reference. Create font directly and return it.
return PDFFont::createFont(fontObject, m_document);
}
}
PDFRealizedFontPointer PDFFontCache::getRealizedFont(const PDFFontPointer& font, PDFReal size, PDFRenderErrorReporter* reporter) const
{
Q_ASSERT(font);
QMutexLocker lock(&m_mutex);
auto it = m_realizedFontCache.find(std::make_pair(font, size));
if (it == m_realizedFontCache.cend())
{
// We must create the realized font
PDFRealizedFontPointer realizedFont = PDFRealizedFont::createRealizedFont(font, size, reporter);
if (m_fontCacheShrinkDisabledObjects.empty() && m_realizedFontCache.size() >= m_realizedFontCacheLimit)
{
m_realizedFontCache.clear();
}
it = m_realizedFontCache.insert(std::make_pair(std::make_pair(font, size), qMove(realizedFont))).first;
}
return it->second;
}
void PDFFontCache::setCacheShrinkEnabled(const void* source, bool enabled)
{
QMutexLocker lock(&m_mutex);
if (enabled)
{
m_fontCacheShrinkDisabledObjects.erase(source);
lock.unlock();
shrink();
}
else
{
m_fontCacheShrinkDisabledObjects.insert(source);
}
}
void PDFFontCache::setCacheLimits(int fontCacheLimit, int instancedFontCacheLimit)
{
if (m_fontCacheLimit != fontCacheLimit || m_realizedFontCacheLimit != instancedFontCacheLimit)
{
m_fontCacheLimit = fontCacheLimit;
m_realizedFontCacheLimit = instancedFontCacheLimit;
shrink();
}
}
void PDFFontCache::shrink()
{
QMutexLocker lock(&m_mutex);
if (m_fontCacheShrinkDisabledObjects.empty())
{
if (m_fontCache.size() >= m_fontCacheLimit)
{
m_fontCache.clear();
}
if (m_realizedFontCache.size() >= m_realizedFontCacheLimit)
{
m_realizedFontCache.clear();
}
}
}
const QByteArray* FontDescriptor::getEmbeddedFontData() const
{
if (!fontFile.isEmpty())
{
return &fontFile;
}
else if (!fontFile2.isEmpty())
{
return &fontFile2;
}
else if (!fontFile3.isEmpty())
{
return &fontFile3;
}
return nullptr;
}
PDFFontCMap PDFFontCMap::createFromName(const QByteArray& name)
{
QFile file(QString(":/cmaps/%1").arg(QString::fromLatin1(name)));
if (file.exists())
{
QByteArray data;
if (file.open(QFile::ReadOnly))
{
data = file.readAll();
file.close();
}
return createFromData(data);
}
throw PDFException(PDFTranslationContext::tr("Can't load CID font mapping named '%1'.").arg(QString::fromLatin1(name)));
return PDFFontCMap();
}
PDFFontCMap PDFFontCMap::createFromData(const QByteArray& data)
{
Entries entries;
entries.reserve(1024); // Arbitrary number, we have enough memory, better than perform reallocation each time
std::vector<PDFFontCMap> additionalMappings;
PDFLexicalAnalyzer parser(data.constBegin(), data.constEnd());
bool vertical = false;
PDFLexicalAnalyzer::Token previousToken;
while (!parser.isAtEnd())
{
PDFLexicalAnalyzer::Token token = parser.fetch();
if (token.type == PDFLexicalAnalyzer::TokenType::Name && token.data.toByteArray() == "WMode")
{
PDFLexicalAnalyzer::Token valueToken = parser.fetch();
vertical = valueToken.type == PDFLexicalAnalyzer::TokenType::Integer && valueToken.data.value<PDFInteger>() == 1;
continue;
}
auto fetchCode = [] (const PDFLexicalAnalyzer::Token& currentToken) -> std::pair<unsigned int, unsigned int>
{
if (currentToken.type == PDFLexicalAnalyzer::TokenType::String)
{
QByteArray byteArray = currentToken.data.toByteArray();
unsigned int codeValue = 0;
for (int i = 0; i < byteArray.size(); ++i)
{
codeValue = (codeValue << 8) + static_cast<unsigned char>(byteArray[i]);
}
return std::make_pair(codeValue, byteArray.size());
}
throw PDFException(PDFTranslationContext::tr("Can't fetch code from CMap definition."));
return std::pair<unsigned int, unsigned int>();
};
auto fetchCID = [] (const PDFLexicalAnalyzer::Token& currentToken) -> CID
{
if (currentToken.type == PDFLexicalAnalyzer::TokenType::Integer)
{
return currentToken.data.value<PDFInteger>();
}
throw PDFException(PDFTranslationContext::tr("Can't fetch CID from CMap definition."));
return 0;
};
auto fetchUnicode = [](const PDFLexicalAnalyzer::Token& currentToken) -> CID
{
if (currentToken.type == PDFLexicalAnalyzer::TokenType::String)
{
QByteArray byteArray = currentToken.data.toByteArray();
if (byteArray.size() == 2)
{
CID unicodeValue = 0;
for (int i = 0; i < byteArray.size(); ++i)
{
unicodeValue = (unicodeValue << 8) + static_cast<unsigned char>(byteArray[i]);
}
return unicodeValue;
}
}
return 0;
};
if (token.type == PDFLexicalAnalyzer::TokenType::Command)
{
QByteArray command = token.data.toByteArray();
if (command == "usecmap")
{
if (previousToken.type == PDFLexicalAnalyzer::TokenType::Name)
{
additionalMappings.emplace_back(createFromName(previousToken.data.toByteArray()));
}
else
{
throw PDFException(PDFTranslationContext::tr("Can't use cmap inside cmap file."));
}
}
else if (command == "beginbfrange")
{
PDFLexicalAnalyzer::Token token1 = parser.fetch();
if (token1.type == PDFLexicalAnalyzer::TokenType::Command &&
token1.data.toByteArray() == "endbfrange")
{
break;
}
PDFLexicalAnalyzer::Token token2 = parser.fetch();
PDFLexicalAnalyzer::Token token3 = parser.fetch();
std::pair<unsigned int, unsigned int> from = fetchCode(token1);
std::pair<unsigned int, unsigned int> to = fetchCode(token2);
CID cid = fetchUnicode(token3);
entries.emplace_back(from.first, to.first, qMax(from.second, to.second), cid);
}
else if (command == "begincidrange")
{
while (true)
{
PDFLexicalAnalyzer::Token token1 = parser.fetch();
if (token1.type == PDFLexicalAnalyzer::TokenType::Command &&
token1.data.toByteArray() == "endcidrange")
{
break;
}
PDFLexicalAnalyzer::Token token2 = parser.fetch();
PDFLexicalAnalyzer::Token token3 = parser.fetch();
std::pair<unsigned int, unsigned int> from = fetchCode(token1);
std::pair<unsigned int, unsigned int> to = fetchCode(token2);
CID cid = fetchCID(token3);
entries.emplace_back(from.first, to.first, qMax(from.second, to.second), cid);
}
}
else if (command == "begincidchar")
{
while (true)
{
PDFLexicalAnalyzer::Token token1 = parser.fetch();
if (token1.type == PDFLexicalAnalyzer::TokenType::Command &&
token1.data.toByteArray() == "endcidchar")
{
break;
}
PDFLexicalAnalyzer::Token token2 = parser.fetch();
std::pair<unsigned int, unsigned int> code = fetchCode(token1);
CID cid = fetchCID(token2);
entries.emplace_back(code.first, code.first, code.second, cid);
}
}
else if (command == "beginbfchar")
{
while (true)
{
PDFLexicalAnalyzer::Token token1 = parser.fetch();
if (token1.type == PDFLexicalAnalyzer::TokenType::Command &&
token1.data.toByteArray() == "endbfchar")
{
break;
}
PDFLexicalAnalyzer::Token token2 = parser.fetch();
std::pair<unsigned int, unsigned int> code = fetchCode(token1);
CID cid = fetchUnicode(token2);
entries.emplace_back(code.first, code.first, code.second, cid);
}
}
}
previousToken = token;
}
std::sort(entries.begin(), entries.end());
entries = optimize(entries);
if (!additionalMappings.empty())
{
for (const PDFFontCMap& map : additionalMappings)
{
entries.insert(entries.cend(), map.m_entries.cbegin(), map.m_entries.cend());
}
}
return PDFFontCMap(qMove(entries), vertical);
}
QByteArray PDFFontCMap::serialize() const
{
QByteArray result;
{
QDataStream stream(&result, QIODevice::WriteOnly);
stream << m_maxKeyLength;
stream << m_vertical;
stream << m_entries.size();
for (const Entry& entry : m_entries)
{
stream << entry.from;
stream << entry.to;
stream << entry.byteCount;
stream << entry.cid;
}
}
return qCompress(result, 9);
}
PDFFontCMap PDFFontCMap::deserialize(const QByteArray& byteArray)
{
PDFFontCMap result;
QByteArray decompressed = qUncompress(byteArray);
QDataStream stream(&decompressed, QIODevice::ReadOnly);
stream >> result.m_maxKeyLength;
stream >> result.m_vertical;
Entries::size_type size = 0;
stream >> size;
result.m_entries.reserve(size);
for (Entries::size_type i = 0; i < size; ++i)
{
Entry entry;
stream >> entry.from;
stream >> entry.to;
stream >> entry.byteCount;
stream >> entry.cid;
result.m_entries.push_back(entry);
}
return result;
}
std::vector<CID> PDFFontCMap::interpret(const QByteArray& byteArray) const
{
std::vector<CID> result;
result.reserve(byteArray.size() / m_maxKeyLength);
unsigned int value = 0;
int scannedBytes = 0;
for (int i = 0, size = byteArray.size(); i < size; ++i)
{
value = (value << 8) + static_cast<unsigned char>(byteArray[i]);
++scannedBytes;
// Find suitable mapping
auto it = std::find_if(m_entries.cbegin(), m_entries.cend(), [value, scannedBytes](const Entry& entry) { return entry.from <= value && entry.to >= value && entry.byteCount == scannedBytes; });
if (it != m_entries.cend())
{
const Entry& entry = *it;
const CID cid = value - entry.from + entry.cid;
result.push_back(cid);
value = 0;
scannedBytes = 0;
}
else if (scannedBytes == m_maxKeyLength)
{
// This means error occured - fill empty CID
result.push_back(0);
value = 0;
scannedBytes = 0;
}
}
return result;
}
QChar PDFFontCMap::getToUnicode(CID cid) const
{
if (isValid())
{
auto it = std::find_if(m_entries.cbegin(), m_entries.cend(), [cid](const Entry& entry) { return entry.from <= cid && entry.to >= cid; });
if (it != m_entries.cend())
{
const Entry& entry = *it;
const CID unicodeCID = cid - entry.from + entry.cid;
return QChar(unicodeCID);
}
}
return QChar();
}
PDFFontCMap::PDFFontCMap(Entries&& entries, bool vertical) :
m_entries(qMove(entries)),
m_maxKeyLength(0),
m_vertical(vertical)
{
m_maxKeyLength = std::accumulate(m_entries.cbegin(), m_entries.cend(), 0, [](unsigned int a, const Entry& b) { return qMax(a, b.byteCount); });
}
PDFFontCMap::Entries PDFFontCMap::optimize(const PDFFontCMap::Entries& entries)
{
Entries result;
result.reserve(entries.size());
if (!entries.empty())
{
Entry current = entries.front();
for (size_t i = 1, count = entries.size(); i < count; ++i)
{
Entry toMerge = entries[i];
if (current.canMerge(toMerge))
{
current = current.merge(toMerge);
}
else
{
result.emplace_back(current);
current = toMerge;
}
}
result.emplace_back(current);
}
result.shrink_to_fit();
return result;
}
PDFFontCMapRepository* PDFFontCMapRepository::getInstance()
{
static PDFFontCMapRepository repository;
return &repository;
}
void PDFFontCMapRepository::saveToFile(const QString& fileName) const
{
QFile file(fileName);
if (file.open(QFile::WriteOnly | QFile::Truncate))
{
size_t size = m_cmaps.size();
{
QDataStream stream(&file);
stream << size;
for (const auto& item : m_cmaps)
{
stream << item.first;
stream << item.second;
}
}
file.close();
}
}
bool PDFFontCMapRepository::loadFromFile(const QString& fileName)
{
QFile file(fileName);
if (file.open(QFile::ReadOnly))
{
{
QDataStream stream(&file);
size_t size = 0;
stream >> size;
for (size_t i = 0; i < size; ++i)
{
QByteArray key;
QByteArray value;
stream >> key;
stream >> value;
m_cmaps[qMove(key)] = qMove(value);
}
}
file.close();
return true;
}
return false;
}
PDFFontCMapRepository::PDFFontCMapRepository()
{
}
PDFReal PDFType0Font::getGlyphAdvance(CID cid) const
{
auto it = m_advances.find(cid);
if (it != m_advances.cend())
{
return it->second;
}
return m_defaultAdvance;
}
PDFType3Font::PDFType3Font(FontDescriptor fontDescriptor,
int firstCharacterIndex,
int lastCharacterIndex,
QMatrix fontMatrix,
std::map<int, QByteArray>&& characterContentStreams,
std::vector<double>&& widths,
const PDFObject& resources,
PDFFontCMap toUnicode) :
PDFFont(qMove(fontDescriptor)),
m_firstCharacterIndex(firstCharacterIndex),
m_lastCharacterIndex(lastCharacterIndex),
m_fontMatrix(fontMatrix),
m_characterContentStreams(qMove(characterContentStreams)),
m_widths(qMove(widths)),
m_resources(resources),
m_toUnicode(qMove(toUnicode))
{
}
FontType PDFType3Font::getFontType() const
{
return FontType::Type3;
}
void PDFType3Font::dumpFontToTreeItem(QTreeWidgetItem* item) const
{
new QTreeWidgetItem(item, { PDFTranslationContext::tr("Character count"), QString::number(m_characterContentStreams.size()) });
}
double PDFType3Font::getWidth(int characterIndex) const
{
if (characterIndex >= m_firstCharacterIndex && characterIndex <= m_lastCharacterIndex)
{
size_t index = characterIndex - m_firstCharacterIndex;
if (index < m_widths.size())
{
return m_widths[index];
}
}
return 0.0;
}
const QByteArray* PDFType3Font::getContentStream(int characterIndex) const
{
auto it = m_characterContentStreams.find(characterIndex);
if (it != m_characterContentStreams.cend())
{
return &it->second;
}
return nullptr;
}
void PDFRealizedType3FontImpl::fillTextSequence(const QByteArray& byteArray, TextSequence& textSequence, PDFRenderErrorReporter* reporter)
{
Q_ASSERT(dynamic_cast<const PDFType3Font*>(m_parentFont.get()));
const PDFType3Font* parentFont = static_cast<const PDFType3Font*>(m_parentFont.get());
textSequence.items.reserve(byteArray.size());
for (int i = 0, characterCount = byteArray.size(); i < characterCount; ++i)
{
int index = static_cast<uint8_t>(byteArray[i]);
const QByteArray* contentStream = parentFont->getContentStream(index);
QChar character = parentFont->getUnicode(index);
const double width = parentFont->getWidth(index);
if (contentStream)
{
textSequence.items.emplace_back(contentStream, character, width);
}
else
{
// Report error, and add advance, if we have it
reporter->reportRenderError(RenderErrorType::Warning, PDFTranslationContext::tr("Content stream for type 3 font character code '%1' not found.").arg(index));
if (width > 0.0)
{
textSequence.items.emplace_back(width);
}
}
}
}
bool PDFRealizedType3FontImpl::isHorizontalWritingSystem() const
{
return true;
}
CharacterInfos PDFRealizedType3FontImpl::getCharacterInfos() const
{
CharacterInfos result;
Q_ASSERT(dynamic_cast<const PDFType3Font*>(m_parentFont.get()));
const PDFType3Font* parentFont = static_cast<const PDFType3Font*>(m_parentFont.get());
for (const auto& contentStreamItem : parentFont->getContentStreams())
{
CharacterInfo info;
info.gid = contentStreamItem.first;
info.character = parentFont->getUnicode(contentStreamItem.first);
result.emplace_back(qMove(info));
}
return result;
}
} // namespace pdf