| // Copyright 2016 PDFium Authors. All rights reserved. |
| // Use of this source code is governed by a BSD-style license that can be |
| // found in the LICENSE file. |
| |
| // Original code copyright 2014 Foxit Software Inc. http://www.foxitsoftware.com |
| |
| #include "core/fpdfapi/font/cpdf_font.h" |
| |
| #include <limits> |
| #include <memory> |
| #include <utility> |
| #include <vector> |
| |
| #include "core/fpdfapi/cpdf_modulemgr.h" |
| #include "core/fpdfapi/font/cpdf_fontencoding.h" |
| #include "core/fpdfapi/font/cpdf_truetypefont.h" |
| #include "core/fpdfapi/font/cpdf_type1font.h" |
| #include "core/fpdfapi/font/cpdf_type3font.h" |
| #include "core/fpdfapi/page/cpdf_docpagedata.h" |
| #include "core/fpdfapi/page/cpdf_pagemodule.h" |
| #include "core/fpdfapi/parser/cpdf_array.h" |
| #include "core/fpdfapi/parser/cpdf_dictionary.h" |
| #include "core/fpdfapi/parser/cpdf_document.h" |
| #include "core/fpdfapi/parser/cpdf_name.h" |
| #include "core/fpdfapi/parser/cpdf_stream_acc.h" |
| #include "core/fxcrt/fx_memory.h" |
| #include "core/fxge/fx_freetype.h" |
| #include "third_party/base/logging.h" |
| #include "third_party/base/ptr_util.h" |
| #include "third_party/base/stl_util.h" |
| |
| namespace { |
| |
| const uint8_t kChineseFontNames[][5] = {{0xCB, 0xCE, 0xCC, 0xE5, 0x00}, |
| {0xBF, 0xAC, 0xCC, 0xE5, 0x00}, |
| {0xBA, 0xDA, 0xCC, 0xE5, 0x00}, |
| {0xB7, 0xC2, 0xCB, 0xCE, 0x00}, |
| {0xD0, 0xC2, 0xCB, 0xCE, 0x00}}; |
| |
| void GetPredefinedEncoding(const CFX_ByteString& value, int* basemap) { |
| if (value == "WinAnsiEncoding") |
| *basemap = PDFFONT_ENCODING_WINANSI; |
| else if (value == "MacRomanEncoding") |
| *basemap = PDFFONT_ENCODING_MACROMAN; |
| else if (value == "MacExpertEncoding") |
| *basemap = PDFFONT_ENCODING_MACEXPERT; |
| else if (value == "PDFDocEncoding") |
| *basemap = PDFFONT_ENCODING_PDFDOC; |
| } |
| |
| } // namespace |
| |
| CPDF_Font::CPDF_Font() |
| : m_pFontFile(nullptr), |
| m_pFontDict(nullptr), |
| m_bToUnicodeLoaded(false), |
| m_Flags(0), |
| m_StemV(0), |
| m_Ascent(0), |
| m_Descent(0), |
| m_ItalicAngle(0) {} |
| |
| CPDF_Font::~CPDF_Font() { |
| if (m_pFontFile) { |
| auto* pPageData = m_pDocument->GetPageData(); |
| if (pPageData) { |
| pPageData->MaybePurgeFontFileStreamAcc( |
| m_pFontFile->GetStream()->AsStream()); |
| } |
| } |
| } |
| |
| bool CPDF_Font::IsType1Font() const { |
| return false; |
| } |
| |
| bool CPDF_Font::IsTrueTypeFont() const { |
| return false; |
| } |
| |
| bool CPDF_Font::IsType3Font() const { |
| return false; |
| } |
| |
| bool CPDF_Font::IsCIDFont() const { |
| return false; |
| } |
| |
| const CPDF_Type1Font* CPDF_Font::AsType1Font() const { |
| return nullptr; |
| } |
| |
| CPDF_Type1Font* CPDF_Font::AsType1Font() { |
| return nullptr; |
| } |
| |
| const CPDF_TrueTypeFont* CPDF_Font::AsTrueTypeFont() const { |
| return nullptr; |
| } |
| |
| CPDF_TrueTypeFont* CPDF_Font::AsTrueTypeFont() { |
| return nullptr; |
| } |
| |
| const CPDF_Type3Font* CPDF_Font::AsType3Font() const { |
| return nullptr; |
| } |
| |
| CPDF_Type3Font* CPDF_Font::AsType3Font() { |
| return nullptr; |
| } |
| |
| const CPDF_CIDFont* CPDF_Font::AsCIDFont() const { |
| return nullptr; |
| } |
| |
| CPDF_CIDFont* CPDF_Font::AsCIDFont() { |
| return nullptr; |
| } |
| |
| bool CPDF_Font::IsUnicodeCompatible() const { |
| return false; |
| } |
| |
| int CPDF_Font::CountChar(const char* pString, int size) const { |
| return size; |
| } |
| |
| int CPDF_Font::GlyphFromCharCodeExt(uint32_t charcode) { |
| return GlyphFromCharCode(charcode, nullptr); |
| } |
| |
| bool CPDF_Font::IsVertWriting() const { |
| const CPDF_CIDFont* pCIDFont = AsCIDFont(); |
| return pCIDFont ? pCIDFont->IsVertWriting() : m_Font.IsVertical(); |
| } |
| |
| int CPDF_Font::AppendChar(char* buf, uint32_t charcode) const { |
| *buf = static_cast<char>(charcode); |
| return 1; |
| } |
| |
| void CPDF_Font::AppendChar(CFX_ByteString* str, uint32_t charcode) const { |
| char buf[4]; |
| int len = AppendChar(buf, charcode); |
| *str += CFX_ByteStringC(buf, len); |
| } |
| |
| CFX_WideString CPDF_Font::UnicodeFromCharCode(uint32_t charcode) const { |
| if (!m_bToUnicodeLoaded) |
| LoadUnicodeMap(); |
| |
| return m_pToUnicodeMap ? m_pToUnicodeMap->Lookup(charcode) : CFX_WideString(); |
| } |
| |
| uint32_t CPDF_Font::CharCodeFromUnicode(wchar_t unicode) const { |
| if (!m_bToUnicodeLoaded) |
| LoadUnicodeMap(); |
| |
| return m_pToUnicodeMap ? m_pToUnicodeMap->ReverseLookup(unicode) : 0; |
| } |
| |
| bool CPDF_Font::HasFontWidths() const { |
| return true; |
| } |
| |
| void CPDF_Font::LoadFontDescriptor(CPDF_Dictionary* pFontDesc) { |
| m_Flags = pFontDesc->GetIntegerFor("Flags", FXFONT_NONSYMBOLIC); |
| int ItalicAngle = 0; |
| bool bExistItalicAngle = false; |
| if (pFontDesc->KeyExist("ItalicAngle")) { |
| ItalicAngle = pFontDesc->GetIntegerFor("ItalicAngle"); |
| bExistItalicAngle = true; |
| } |
| if (ItalicAngle < 0) { |
| m_Flags |= FXFONT_ITALIC; |
| m_ItalicAngle = ItalicAngle; |
| } |
| bool bExistStemV = false; |
| if (pFontDesc->KeyExist("StemV")) { |
| m_StemV = pFontDesc->GetIntegerFor("StemV"); |
| bExistStemV = true; |
| } |
| bool bExistAscent = false; |
| if (pFontDesc->KeyExist("Ascent")) { |
| m_Ascent = pFontDesc->GetIntegerFor("Ascent"); |
| bExistAscent = true; |
| } |
| bool bExistDescent = false; |
| if (pFontDesc->KeyExist("Descent")) { |
| m_Descent = pFontDesc->GetIntegerFor("Descent"); |
| bExistDescent = true; |
| } |
| bool bExistCapHeight = false; |
| if (pFontDesc->KeyExist("CapHeight")) |
| bExistCapHeight = true; |
| if (bExistItalicAngle && bExistAscent && bExistCapHeight && bExistDescent && |
| bExistStemV) { |
| m_Flags |= FXFONT_USEEXTERNATTR; |
| } |
| if (m_Descent > 10) |
| m_Descent = -m_Descent; |
| CPDF_Array* pBBox = pFontDesc->GetArrayFor("FontBBox"); |
| if (pBBox) { |
| m_FontBBox.left = pBBox->GetIntegerAt(0); |
| m_FontBBox.bottom = pBBox->GetIntegerAt(1); |
| m_FontBBox.right = pBBox->GetIntegerAt(2); |
| m_FontBBox.top = pBBox->GetIntegerAt(3); |
| } |
| |
| CPDF_Stream* pFontFile = pFontDesc->GetStreamFor("FontFile"); |
| if (!pFontFile) |
| pFontFile = pFontDesc->GetStreamFor("FontFile2"); |
| if (!pFontFile) |
| pFontFile = pFontDesc->GetStreamFor("FontFile3"); |
| if (!pFontFile) |
| return; |
| |
| m_pFontFile = m_pDocument->LoadFontFile(pFontFile); |
| if (!m_pFontFile) |
| return; |
| |
| const uint8_t* pFontData = m_pFontFile->GetData(); |
| uint32_t dwFontSize = m_pFontFile->GetSize(); |
| if (!m_Font.LoadEmbedded(pFontData, dwFontSize)) { |
| m_pDocument->GetPageData()->MaybePurgeFontFileStreamAcc( |
| m_pFontFile->GetStream()->AsStream()); |
| m_pFontFile = nullptr; |
| } |
| } |
| |
| void CPDF_Font::CheckFontMetrics() { |
| if (m_FontBBox.top == 0 && m_FontBBox.bottom == 0 && m_FontBBox.left == 0 && |
| m_FontBBox.right == 0) { |
| FXFT_Face face = m_Font.GetFace(); |
| if (face) { |
| m_FontBBox.left = TT2PDF(FXFT_Get_Face_xMin(face), face); |
| m_FontBBox.bottom = TT2PDF(FXFT_Get_Face_yMin(face), face); |
| m_FontBBox.right = TT2PDF(FXFT_Get_Face_xMax(face), face); |
| m_FontBBox.top = TT2PDF(FXFT_Get_Face_yMax(face), face); |
| m_Ascent = TT2PDF(FXFT_Get_Face_Ascender(face), face); |
| m_Descent = TT2PDF(FXFT_Get_Face_Descender(face), face); |
| } else { |
| bool bFirst = true; |
| for (int i = 0; i < 256; i++) { |
| FX_RECT rect = GetCharBBox(i); |
| if (rect.left == rect.right) { |
| continue; |
| } |
| if (bFirst) { |
| m_FontBBox = rect; |
| bFirst = false; |
| } else { |
| if (m_FontBBox.top < rect.top) { |
| m_FontBBox.top = rect.top; |
| } |
| if (m_FontBBox.right < rect.right) { |
| m_FontBBox.right = rect.right; |
| } |
| if (m_FontBBox.left > rect.left) { |
| m_FontBBox.left = rect.left; |
| } |
| if (m_FontBBox.bottom > rect.bottom) { |
| m_FontBBox.bottom = rect.bottom; |
| } |
| } |
| } |
| } |
| } |
| if (m_Ascent == 0 && m_Descent == 0) { |
| FX_RECT rect = GetCharBBox('A'); |
| m_Ascent = rect.bottom == rect.top ? m_FontBBox.top : rect.top; |
| rect = GetCharBBox('g'); |
| m_Descent = rect.bottom == rect.top ? m_FontBBox.bottom : rect.bottom; |
| } |
| } |
| |
| void CPDF_Font::LoadUnicodeMap() const { |
| m_bToUnicodeLoaded = true; |
| CPDF_Stream* pStream = m_pFontDict->GetStreamFor("ToUnicode"); |
| if (!pStream) { |
| return; |
| } |
| m_pToUnicodeMap = pdfium::MakeUnique<CPDF_ToUnicodeMap>(); |
| m_pToUnicodeMap->Load(pStream); |
| } |
| |
| int CPDF_Font::GetStringWidth(const char* pString, int size) { |
| int offset = 0; |
| int width = 0; |
| while (offset < size) { |
| uint32_t charcode = GetNextChar(pString, size, offset); |
| width += GetCharWidthF(charcode); |
| } |
| return width; |
| } |
| |
| // static |
| CPDF_Font* CPDF_Font::GetStockFont(CPDF_Document* pDoc, |
| const CFX_ByteStringC& name) { |
| CFX_ByteString fontname(name); |
| int font_id = PDF_GetStandardFontName(&fontname); |
| if (font_id < 0) |
| return nullptr; |
| |
| CPDF_FontGlobals* pFontGlobals = |
| CPDF_ModuleMgr::Get()->GetPageModule()->GetFontGlobals(); |
| CPDF_Font* pFont = pFontGlobals->Find(pDoc, font_id); |
| if (pFont) |
| return pFont; |
| |
| CPDF_Dictionary* pDict = new CPDF_Dictionary(pDoc->GetByteStringPool()); |
| pDict->SetNewFor<CPDF_Name>("Type", "Font"); |
| pDict->SetNewFor<CPDF_Name>("Subtype", "Type1"); |
| pDict->SetNewFor<CPDF_Name>("BaseFont", fontname); |
| pDict->SetNewFor<CPDF_Name>("Encoding", "WinAnsiEncoding"); |
| return pFontGlobals->Set(pDoc, font_id, CPDF_Font::Create(nullptr, pDict)); |
| } |
| |
| std::unique_ptr<CPDF_Font> CPDF_Font::Create(CPDF_Document* pDoc, |
| CPDF_Dictionary* pFontDict) { |
| CFX_ByteString type = pFontDict->GetStringFor("Subtype"); |
| std::unique_ptr<CPDF_Font> pFont; |
| if (type == "TrueType") { |
| CFX_ByteString tag = pFontDict->GetStringFor("BaseFont").Left(4); |
| for (size_t i = 0; i < FX_ArraySize(kChineseFontNames); ++i) { |
| if (tag == CFX_ByteString(kChineseFontNames[i], 4)) { |
| CPDF_Dictionary* pFontDesc = pFontDict->GetDictFor("FontDescriptor"); |
| if (!pFontDesc || !pFontDesc->KeyExist("FontFile2")) |
| pFont = pdfium::MakeUnique<CPDF_CIDFont>(); |
| break; |
| } |
| } |
| if (!pFont) |
| pFont = pdfium::MakeUnique<CPDF_TrueTypeFont>(); |
| } else if (type == "Type3") { |
| pFont = pdfium::MakeUnique<CPDF_Type3Font>(); |
| } else if (type == "Type0") { |
| pFont = pdfium::MakeUnique<CPDF_CIDFont>(); |
| } else { |
| pFont = pdfium::MakeUnique<CPDF_Type1Font>(); |
| } |
| pFont->m_pFontDict = pFontDict; |
| pFont->m_pDocument = pDoc; |
| pFont->m_BaseFont = pFontDict->GetStringFor("BaseFont"); |
| return pFont->Load() ? std::move(pFont) : nullptr; |
| } |
| |
| uint32_t CPDF_Font::GetNextChar(const char* pString, |
| int nStrLen, |
| int& offset) const { |
| if (offset < 0 || nStrLen < 1) { |
| return 0; |
| } |
| uint8_t ch = offset < nStrLen ? pString[offset++] : pString[nStrLen - 1]; |
| return static_cast<uint32_t>(ch); |
| } |
| |
| void CPDF_Font::LoadPDFEncoding(CPDF_Object* pEncoding, |
| int& iBaseEncoding, |
| std::vector<CFX_ByteString>* pCharNames, |
| bool bEmbedded, |
| bool bTrueType) { |
| if (!pEncoding) { |
| if (m_BaseFont == "Symbol") { |
| iBaseEncoding = bTrueType ? PDFFONT_ENCODING_MS_SYMBOL |
| : PDFFONT_ENCODING_ADOBE_SYMBOL; |
| } else if (!bEmbedded && iBaseEncoding == PDFFONT_ENCODING_BUILTIN) { |
| iBaseEncoding = PDFFONT_ENCODING_WINANSI; |
| } |
| return; |
| } |
| if (pEncoding->IsName()) { |
| if (iBaseEncoding == PDFFONT_ENCODING_ADOBE_SYMBOL || |
| iBaseEncoding == PDFFONT_ENCODING_ZAPFDINGBATS) { |
| return; |
| } |
| if ((m_Flags & FXFONT_SYMBOLIC) && m_BaseFont == "Symbol") { |
| if (!bTrueType) |
| iBaseEncoding = PDFFONT_ENCODING_ADOBE_SYMBOL; |
| return; |
| } |
| CFX_ByteString bsEncoding = pEncoding->GetString(); |
| if (bsEncoding.Compare("MacExpertEncoding") == 0) { |
| bsEncoding = "WinAnsiEncoding"; |
| } |
| GetPredefinedEncoding(bsEncoding, &iBaseEncoding); |
| return; |
| } |
| |
| CPDF_Dictionary* pDict = pEncoding->AsDictionary(); |
| if (!pDict) |
| return; |
| |
| if (iBaseEncoding != PDFFONT_ENCODING_ADOBE_SYMBOL && |
| iBaseEncoding != PDFFONT_ENCODING_ZAPFDINGBATS) { |
| CFX_ByteString bsEncoding = pDict->GetStringFor("BaseEncoding"); |
| if (bsEncoding.Compare("MacExpertEncoding") == 0 && bTrueType) { |
| bsEncoding = "WinAnsiEncoding"; |
| } |
| GetPredefinedEncoding(bsEncoding, &iBaseEncoding); |
| } |
| if ((!bEmbedded || bTrueType) && iBaseEncoding == PDFFONT_ENCODING_BUILTIN) |
| iBaseEncoding = PDFFONT_ENCODING_STANDARD; |
| |
| CPDF_Array* pDiffs = pDict->GetArrayFor("Differences"); |
| if (!pDiffs) |
| return; |
| |
| pCharNames->resize(256); |
| uint32_t cur_code = 0; |
| for (uint32_t i = 0; i < pDiffs->GetCount(); i++) { |
| CPDF_Object* pElement = pDiffs->GetDirectObjectAt(i); |
| if (!pElement) |
| continue; |
| |
| if (CPDF_Name* pName = pElement->AsName()) { |
| if (cur_code < 256) |
| (*pCharNames)[cur_code] = pName->GetString(); |
| cur_code++; |
| } else { |
| cur_code = pElement->GetInteger(); |
| } |
| } |
| } |
| |
| bool CPDF_Font::IsStandardFont() const { |
| if (!IsType1Font()) |
| return false; |
| if (m_pFontFile) |
| return false; |
| if (AsType1Font()->GetBase14Font() < 0) |
| return false; |
| return true; |
| } |
| |
| const char* CPDF_Font::GetAdobeCharName( |
| int iBaseEncoding, |
| const std::vector<CFX_ByteString>& charnames, |
| int charcode) { |
| if (charcode < 0 || charcode >= 256) { |
| NOTREACHED(); |
| return nullptr; |
| } |
| |
| if (!charnames.empty() && !charnames[charcode].IsEmpty()) |
| return charnames[charcode].c_str(); |
| |
| const char* name = nullptr; |
| if (iBaseEncoding) |
| name = PDF_CharNameFromPredefinedCharSet(iBaseEncoding, charcode); |
| return name && name[0] ? name : nullptr; |
| } |
| |
| uint32_t CPDF_Font::FallbackFontFromCharcode(uint32_t charcode) { |
| if (m_FontFallbacks.empty()) { |
| m_FontFallbacks.push_back(pdfium::MakeUnique<CFX_Font>()); |
| m_FontFallbacks[0]->LoadSubst("Arial", IsTrueTypeFont(), m_Flags, |
| m_StemV * 5, m_ItalicAngle, 0, |
| IsVertWriting()); |
| } |
| return 0; |
| } |
| |
| int CPDF_Font::FallbackGlyphFromCharcode(int fallbackFont, uint32_t charcode) { |
| if (!pdfium::IndexInBounds(m_FontFallbacks, fallbackFont)) |
| return -1; |
| |
| CFX_WideString str = UnicodeFromCharCode(charcode); |
| uint32_t unicode = !str.IsEmpty() ? str.GetAt(0) : charcode; |
| int glyph = |
| FXFT_Get_Char_Index(m_FontFallbacks[fallbackFont]->GetFace(), unicode); |
| if (glyph == 0) |
| return -1; |
| |
| return glyph; |
| } |
| |
| CFX_Font* CPDF_Font::GetFontFallback(int position) { |
| if (position < 0 || static_cast<size_t>(position) >= m_FontFallbacks.size()) |
| return nullptr; |
| return m_FontFallbacks[position].get(); |
| } |
| |
| // static |
| int CPDF_Font::TT2PDF(int m, FXFT_Face face) { |
| int upm = FXFT_Get_Face_UnitsPerEM(face); |
| if (upm == 0) |
| return m; |
| |
| return static_cast<int>( |
| pdfium::clamp((m * 1000.0 + upm / 2) / upm, |
| static_cast<double>(std::numeric_limits<int>::min()), |
| static_cast<double>(std::numeric_limits<int>::max()))); |
| } |
| |
| // static |
| bool CPDF_Font::FT_UseTTCharmap(FXFT_Face face, |
| int platform_id, |
| int encoding_id) { |
| auto** pCharMap = FXFT_Get_Face_Charmaps(face); |
| for (int i = 0; i < FXFT_Get_Face_CharmapCount(face); i++) { |
| if (FXFT_Get_Charmap_PlatformID(pCharMap[i]) == platform_id && |
| FXFT_Get_Charmap_EncodingID(pCharMap[i]) == encoding_id) { |
| FXFT_Set_Charmap(face, pCharMap[i]); |
| return true; |
| } |
| } |
| return false; |
| } |