From 10cfea1fdafc8fcf1edd60bc783e9db9ef6229c0 Mon Sep 17 00:00:00 2001 From: Dan Sinclair Date: Mon, 16 Nov 2015 13:09:00 -0500 Subject: Merge to XFA: Reland "Cleanup some numeric code."" This reverts commit 0569ab0b11b723d9bca4ddd642b0cf8828c4bdd1. This changes the various comparisons of char >= '0' && char <= '9' and char < '0' || char > '9' to use std::isdigit checks. It also cleans up a handful of hex to digit conversions to call one common method. R=tsepez@chromium.org Review URL: https://codereview.chromium.org/1449873003 . (cherry picked from commit 3f148915d12f54a946a0c0bf526162b79c39d650) Review URL: https://codereview.chromium.org/1452673002 . --- core/src/fpdfapi/fpdf_font/fpdf_font.cpp | 51 ++----- core/src/fpdfapi/fpdf_font/fpdf_font_cid.cpp | 44 ++---- .../src/fpdfapi/fpdf_page/fpdf_page_parser_old.cpp | 64 +++----- .../src/fpdfapi/fpdf_parser/fpdf_parser_decode.cpp | 32 ++-- .../src/fpdfapi/fpdf_parser/fpdf_parser_parser.cpp | 163 ++++++++++----------- .../fpdfapi/fpdf_parser/fpdf_parser_utility.cpp | 17 +-- core/src/fpdftext/fpdf_text.cpp | 8 +- core/src/fpdftext/fpdf_text_int.cpp | 12 +- core/src/fxcodec/codec/fx_codec.cpp | 49 ++++--- core/src/fxcrt/fx_basic_bstring.cpp | 9 +- core/src/fxcrt/fx_basic_gcc.cpp | 48 ++++-- core/src/fxcrt/fx_basic_util.cpp | 18 +-- core/src/fxcrt/fx_basic_wstring.cpp | 14 +- core/src/fxcrt/fx_extension.cpp | 21 +-- core/src/fxcrt/fx_extension_unittest.cpp | 24 +++ core/src/fxcrt/fx_xml_parser.cpp | 8 +- 16 files changed, 264 insertions(+), 318 deletions(-) create mode 100644 core/src/fxcrt/fx_extension_unittest.cpp (limited to 'core/src') diff --git a/core/src/fpdfapi/fpdf_font/fpdf_font.cpp b/core/src/fpdfapi/fpdf_font/fpdf_font.cpp index 9f83da062c..64d6cdce92 100644 --- a/core/src/fpdfapi/fpdf_font/fpdf_font.cpp +++ b/core/src/fpdfapi/fpdf_font/fpdf_font.cpp @@ -11,6 +11,7 @@ #include "core/include/fpdfapi/fpdf_page.h" #include "core/include/fpdfapi/fpdf_pageobj.h" #include "core/include/fpdfapi/fpdf_resource.h" +#include "core/include/fxcrt/fx_ext.h" #include "core/include/fxge/fx_freetype.h" #if _FXM_PLATFORM_ == _FXM_PLATFORM_APPLE_ @@ -518,32 +519,19 @@ FX_DWORD CPDF_ToUnicodeMap::ReverseLookup(FX_WCHAR unicode) { FX_DWORD CPDF_ToUnicodeMap::StringToCode(const CFX_ByteStringC& str) { const FX_CHAR* buf = str.GetCStr(); int len = str.GetLength(); - if (len == 0) { + if (len == 0) return 0; - } + int result = 0; if (buf[0] == '<') { - for (int i = 1; i < len; i++) { - int digit; - if (buf[i] >= '0' && buf[i] <= '9') { - digit = buf[i] - '0'; - } else if (buf[i] >= 'a' && buf[i] <= 'f') { - digit = buf[i] - 'a' + 10; - } else if (buf[i] >= 'A' && buf[i] <= 'F') { - digit = buf[i] - 'A' + 10; - } else { - break; - } - result = result * 16 + digit; - } + for (int i = 1; i < len && std::isxdigit(buf[i]); ++i) + result = result * 16 + FXSYS_toHexDigit(buf[i]); return result; } - for (int i = 0; i < len; i++) { - if (buf[i] < '0' || buf[i] > '9') { - break; - } - result = result * 10 + buf[i] - '0'; - } + + for (int i = 0; i < len && std::isdigit(buf[i]); ++i) + result = result * 10 + FXSYS_toDecimalDigit(buf[i]); + return result; } static CFX_WideString StringDataAdd(CFX_WideString str) { @@ -570,26 +558,15 @@ CFX_WideString CPDF_ToUnicodeMap::StringToWideString( const CFX_ByteStringC& str) { const FX_CHAR* buf = str.GetCStr(); int len = str.GetLength(); - if (len == 0) { + if (len == 0) return CFX_WideString(); - } + CFX_WideString result; if (buf[0] == '<') { int byte_pos = 0; FX_WCHAR ch = 0; - for (int i = 1; i < len; i++) { - int digit; - if (buf[i] >= '0' && buf[i] <= '9') { - digit = buf[i] - '0'; - } else if (buf[i] >= 'a' && buf[i] <= 'f') { - digit = buf[i] - 'a' + 10; - } else if (buf[i] >= 'A' && buf[i] <= 'F') { - digit = buf[i] - 'A' + 10; - } else { - break; - } - ch = ch * 16 + digit; - + for (int i = 1; i < len && std::isxdigit(buf[i]); ++i) { + ch = ch * 16 + FXSYS_toHexDigit(buf[i]); byte_pos++; if (byte_pos == 4) { result += ch; @@ -599,8 +576,6 @@ CFX_WideString CPDF_ToUnicodeMap::StringToWideString( } return result; } - if (buf[0] == '(') { - } return result; } void CPDF_ToUnicodeMap::Load(CPDF_Stream* pStream) { diff --git a/core/src/fpdfapi/fpdf_font/fpdf_font_cid.cpp b/core/src/fpdfapi/fpdf_font/fpdf_font_cid.cpp index 65b4b781b1..d4c71085a1 100644 --- a/core/src/fpdfapi/fpdf_font/fpdf_font_cid.cpp +++ b/core/src/fpdfapi/fpdf_font/fpdf_font_cid.cpp @@ -9,6 +9,7 @@ #include "core/include/fpdfapi/fpdf_module.h" #include "core/include/fpdfapi/fpdf_page.h" #include "core/include/fpdfapi/fpdf_resource.h" +#include "core/include/fxcrt/fx_ext.h" #include "core/include/fxge/fx_freetype.h" #include "core/include/fxge/fx_ge.h" #include "core/src/fpdfapi/fpdf_cmaps/cmap_int.h" @@ -699,30 +700,17 @@ void CPDF_CMapParser::ParseWord(const CFX_ByteStringC& word) { m_LastWord = word; } +// Static. FX_DWORD CPDF_CMapParser::CMap_GetCode(const CFX_ByteStringC& word) { int num = 0; if (word.GetAt(0) == '<') { - for (int i = 1; i < word.GetLength(); i++) { - uint8_t digit = word.GetAt(i); - if (digit >= '0' && digit <= '9') { - digit = digit - '0'; - } else if (digit >= 'a' && digit <= 'f') { - digit = digit - 'a' + 10; - } else if (digit >= 'A' && digit <= 'F') { - digit = digit - 'A' + 10; - } else { - return num; - } - num = num * 16 + digit; - } - } else { - for (int i = 0; i < word.GetLength(); i++) { - if (word.GetAt(i) < '0' || word.GetAt(i) > '9') { - return num; - } - num = num * 10 + word.GetAt(i) - '0'; - } + for (int i = 1; i < word.GetLength() && std::isxdigit(word.GetAt(i)); ++i) + num = num * 16 + FXSYS_toHexDigit(word.GetAt(i)); + return num; } + + for (int i = 0; i < word.GetLength() && std::isdigit(word.GetAt(i)); ++i) + num = num * 10 + FXSYS_toDecimalDigit(word.GetAt(i)); return num; } @@ -746,13 +734,7 @@ bool CPDF_CMapParser::CMap_GetCodeRange(CMap_CodeRange& range, for (i = 0; i < range.m_CharSize; ++i) { uint8_t digit1 = first.GetAt(i * 2 + 1); uint8_t digit2 = first.GetAt(i * 2 + 2); - uint8_t byte = (digit1 >= '0' && digit1 <= '9') - ? (digit1 - '0') - : ((digit1 & 0xdf) - 'A' + 10); - byte = byte * 16 + ((digit2 >= '0' && digit2 <= '9') - ? (digit2 - '0') - : ((digit2 & 0xdf) - 'A' + 10)); - range.m_Lower[i] = byte; + range.m_Lower[i] = FXSYS_toHexDigit(digit1) * 16 + FXSYS_toHexDigit(digit2); } FX_DWORD size = second.GetLength(); @@ -763,13 +745,7 @@ bool CPDF_CMapParser::CMap_GetCodeRange(CMap_CodeRange& range, uint8_t digit2 = ((FX_DWORD)i * 2 + 2 < size) ? second.GetAt((FX_STRSIZE)i * 2 + 2) : '0'; - uint8_t byte = (digit1 >= '0' && digit1 <= '9') - ? (digit1 - '0') - : ((digit1 & 0xdf) - 'A' + 10); - byte = byte * 16 + ((digit2 >= '0' && digit2 <= '9') - ? (digit2 - '0') - : ((digit2 & 0xdf) - 'A' + 10)); - range.m_Upper[i] = byte; + range.m_Upper[i] = FXSYS_toHexDigit(digit1) * 16 + FXSYS_toHexDigit(digit2); } return true; } diff --git a/core/src/fpdfapi/fpdf_page/fpdf_page_parser_old.cpp b/core/src/fpdfapi/fpdf_page/fpdf_page_parser_old.cpp index 8d155e787b..a1aa686248 100644 --- a/core/src/fpdfapi/fpdf_page/fpdf_page_parser_old.cpp +++ b/core/src/fpdfapi/fpdf_page/fpdf_page_parser_old.cpp @@ -11,6 +11,7 @@ #include "core/include/fpdfapi/fpdf_module.h" #include "core/include/fpdfapi/fpdf_page.h" #include "core/include/fxcodec/fx_codec.h" +#include "core/include/fxcrt/fx_ext.h" namespace { @@ -791,7 +792,7 @@ CFX_ByteString CPDF_StreamParser::ReadString() { break; case 1: if (ch >= '0' && ch <= '7') { - iEscCode = ch - '0'; + iEscCode = FXSYS_toDecimalDigit(ch); status = 2; break; } @@ -816,7 +817,7 @@ CFX_ByteString CPDF_StreamParser::ReadString() { break; case 2: if (ch >= '0' && ch <= '7') { - iEscCode = iEscCode * 8 + ch - '0'; + iEscCode = iEscCode * 8 + FXSYS_toDecimalDigit(ch); status = 3; } else { buf.AppendChar(iEscCode); @@ -826,7 +827,7 @@ CFX_ByteString CPDF_StreamParser::ReadString() { break; case 3: if (ch >= '0' && ch <= '7') { - iEscCode = iEscCode * 8 + ch - '0'; + iEscCode = iEscCode * 8 + FXSYS_toDecimalDigit(ch); buf.AppendChar(iEscCode); status = 0; } else { @@ -859,50 +860,33 @@ CFX_ByteString CPDF_StreamParser::ReadHexString() { if (!PositionIsInBounds()) return CFX_ByteString(); - int ch = m_pBuf[m_Pos++]; CFX_ByteTextBuf buf; - FX_BOOL bFirst = TRUE; + bool bFirst = true; int code = 0; - while (1) { - if (ch == '>') { - break; - } - if (ch >= '0' && ch <= '9') { - if (bFirst) { - code = (ch - '0') * 16; - } else { - code += ch - '0'; - buf.AppendChar((char)code); - } - bFirst = !bFirst; - } else if (ch >= 'A' && ch <= 'F') { - if (bFirst) { - code = (ch - 'A' + 10) * 16; - } else { - code += ch - 'A' + 10; - buf.AppendChar((char)code); - } - bFirst = !bFirst; - } else if (ch >= 'a' && ch <= 'f') { - if (bFirst) { - code = (ch - 'a' + 10) * 16; - } else { - code += ch - 'a' + 10; - buf.AppendChar((char)code); - } - bFirst = !bFirst; - } - if (!PositionIsInBounds()) + while (PositionIsInBounds()) { + int ch = m_pBuf[m_Pos++]; + + if (ch == '>') break; - ch = m_pBuf[m_Pos++]; + if (!std::isxdigit(ch)) + continue; + + int val = FXSYS_toHexDigit(ch); + if (bFirst) { + code = val * 16; + } else { + code += val; + buf.AppendByte((uint8_t)code); + } + bFirst = !bFirst; } - if (!bFirst) { + if (!bFirst) buf.AppendChar((char)code); - } - if (buf.GetLength() > MAX_STRING_LENGTH) { + + if (buf.GetLength() > MAX_STRING_LENGTH) return CFX_ByteString(buf.GetBuffer(), MAX_STRING_LENGTH); - } + return buf.GetByteString(); } diff --git a/core/src/fpdfapi/fpdf_parser/fpdf_parser_decode.cpp b/core/src/fpdfapi/fpdf_parser/fpdf_parser_decode.cpp index 588ab5dff6..57d1971889 100644 --- a/core/src/fpdfapi/fpdf_parser/fpdf_parser_decode.cpp +++ b/core/src/fpdfapi/fpdf_parser/fpdf_parser_decode.cpp @@ -9,6 +9,7 @@ #include "core/include/fpdfapi/fpdf_module.h" #include "core/include/fpdfapi/fpdf_parser.h" #include "core/include/fxcodec/fx_codec.h" +#include "core/include/fxcrt/fx_ext.h" #define _STREAM_MAX_SIZE_ 20 * 1024 * 1024 @@ -129,37 +130,32 @@ FX_DWORD HexDecode(const uint8_t* src_buf, } dest_buf = FX_Alloc(uint8_t, i / 2 + 1); dest_size = 0; - FX_BOOL bFirstDigit = TRUE; + bool bFirst = true; for (i = 0; i < src_size; i++) { uint8_t ch = src_buf[i]; if (PDFCharIsLineEnding(ch) || ch == ' ' || ch == '\t') continue; - int digit; - if (ch <= '9' && ch >= '0') { - digit = ch - '0'; - } else if (ch <= 'f' && ch >= 'a') { - digit = ch - 'a' + 10; - } else if (ch <= 'F' && ch >= 'A') { - digit = ch - 'A' + 10; - } else if (ch == '>') { - i++; + if (ch == '>') { + ++i; break; - } else { - continue; } - if (bFirstDigit) { + if (!std::isxdigit(ch)) + continue; + + int digit = FXSYS_toHexDigit(ch); + if (bFirst) dest_buf[dest_size] = digit * 16; - } else { + else dest_buf[dest_size++] += digit; - } - bFirstDigit = !bFirstDigit; + + bFirst = !bFirst; } - if (!bFirstDigit) { + if (!bFirst) dest_size++; - } return i; } + FX_DWORD RunLengthDecode(const uint8_t* src_buf, FX_DWORD src_size, uint8_t*& dest_buf, diff --git a/core/src/fpdfapi/fpdf_parser/fpdf_parser_parser.cpp b/core/src/fpdfapi/fpdf_parser/fpdf_parser_parser.cpp index 543f8f6323..9318403766 100644 --- a/core/src/fpdfapi/fpdf_parser/fpdf_parser_parser.cpp +++ b/core/src/fpdfapi/fpdf_parser/fpdf_parser_parser.cpp @@ -13,6 +13,7 @@ #include "core/include/fpdfapi/fpdf_module.h" #include "core/include/fpdfapi/fpdf_page.h" #include "core/include/fpdfapi/fpdf_parser.h" +#include "core/include/fxcrt/fx_ext.h" #include "core/include/fxcrt/fx_safe_types.h" #include "core/src/fpdfapi/fpdf_page/pageint.h" #include "third_party/base/nonstd_unique_ptr.h" @@ -164,85 +165,83 @@ FX_DWORD CPDF_Parser::StartParse(IFX_FileRead* pFileAccess, m_bXRefStream = FALSE; m_LastXRefOffset = 0; m_bOwnFileRead = bOwnFileRead; + int32_t offset = GetHeaderOffset(pFileAccess); if (offset == -1) { - if (bOwnFileRead && pFileAccess) { + if (bOwnFileRead && pFileAccess) pFileAccess->Release(); - } return PDFPARSE_ERROR_FORMAT; } m_Syntax.InitParser(pFileAccess, offset); + uint8_t ch; - if (!m_Syntax.GetCharAt(5, ch)) { + if (!m_Syntax.GetCharAt(5, ch)) return PDFPARSE_ERROR_FORMAT; - } - if (ch >= '0' && ch <= '9') { - m_FileVersion = (ch - '0') * 10; - } - if (!m_Syntax.GetCharAt(7, ch)) { + if (std::isdigit(ch)) + m_FileVersion = FXSYS_toDecimalDigit(ch) * 10; + + if (!m_Syntax.GetCharAt(7, ch)) return PDFPARSE_ERROR_FORMAT; - } - if (ch >= '0' && ch <= '9') { - m_FileVersion += ch - '0'; - } - if (m_Syntax.m_FileLen < m_Syntax.m_HeaderOffset + 9) { + if (std::isdigit(ch)) + m_FileVersion += FXSYS_toDecimalDigit(ch); + + if (m_Syntax.m_FileLen < m_Syntax.m_HeaderOffset + 9) return PDFPARSE_ERROR_FORMAT; - } + m_Syntax.RestorePos(m_Syntax.m_FileLen - m_Syntax.m_HeaderOffset - 9); - if (!bReParse) { + if (!bReParse) m_pDocument = new CPDF_Document(this); - } + FX_BOOL bXRefRebuilt = FALSE; if (m_Syntax.SearchWord(FX_BSTRC("startxref"), TRUE, FALSE, 4096)) { FX_FILESIZE startxref_offset = m_Syntax.SavePos(); void* pResult = FXSYS_bsearch(&startxref_offset, m_SortedOffset.GetData(), m_SortedOffset.GetSize(), sizeof(FX_FILESIZE), CompareFileSize); - if (pResult == NULL) { + if (!pResult) m_SortedOffset.Add(startxref_offset); - } + m_Syntax.GetKeyword(); FX_BOOL bNumber; CFX_ByteString xrefpos_str = m_Syntax.GetNextWord(bNumber); - if (!bNumber) { + if (!bNumber) return PDFPARSE_ERROR_FORMAT; - } + m_LastXRefOffset = (FX_FILESIZE)FXSYS_atoi64(xrefpos_str); if (!LoadAllCrossRefV4(m_LastXRefOffset) && !LoadAllCrossRefV5(m_LastXRefOffset)) { - if (!RebuildCrossRef()) { + if (!RebuildCrossRef()) return PDFPARSE_ERROR_FORMAT; - } + bXRefRebuilt = TRUE; m_LastXRefOffset = 0; } } else { - if (!RebuildCrossRef()) { + if (!RebuildCrossRef()) return PDFPARSE_ERROR_FORMAT; - } + bXRefRebuilt = TRUE; } FX_DWORD dwRet = SetEncryptHandler(); - if (dwRet != PDFPARSE_ERROR_SUCCESS) { + if (dwRet != PDFPARSE_ERROR_SUCCESS) return dwRet; - } + m_pDocument->LoadDoc(); - if (m_pDocument->GetRoot() == NULL || m_pDocument->GetPageCount() == 0) { - if (bXRefRebuilt) { + if (!m_pDocument->GetRoot() || m_pDocument->GetPageCount() == 0) { + if (bXRefRebuilt) return PDFPARSE_ERROR_FORMAT; - } + ReleaseEncryptHandler(); - if (!RebuildCrossRef()) { + if (!RebuildCrossRef()) return PDFPARSE_ERROR_FORMAT; - } + dwRet = SetEncryptHandler(); - if (dwRet != PDFPARSE_ERROR_SUCCESS) { + if (dwRet != PDFPARSE_ERROR_SUCCESS) return dwRet; - } + m_pDocument->LoadDoc(); - if (m_pDocument->GetRoot() == NULL) { + if (!m_pDocument->GetRoot()) return PDFPARSE_ERROR_FORMAT; - } } FXSYS_qsort(m_SortedOffset.GetData(), m_SortedOffset.GetSize(), sizeof(FX_FILESIZE), CompareFileSize); @@ -251,13 +250,12 @@ FX_DWORD CPDF_Parser::StartParse(IFX_FileRead* pFileAccess, ReleaseEncryptHandler(); RebuildCrossRef(); RootObjNum = GetRootObjNum(); - if (RootObjNum == 0) { + if (RootObjNum == 0) return PDFPARSE_ERROR_FORMAT; - } + dwRet = SetEncryptHandler(); - if (dwRet != PDFPARSE_ERROR_SUCCESS) { + if (dwRet != PDFPARSE_ERROR_SUCCESS) return dwRet; - } } if (m_pSecurityHandler && !m_pSecurityHandler->IsMetadataEncrypted()) { CPDF_Reference* pMetadata = @@ -461,9 +459,8 @@ FX_BOOL CPDF_Parser::LoadLinearizedCrossRefV4(FX_FILESIZE pos, int32_t offset = FXSYS_atoi(pEntry); if (offset == 0) { for (int32_t c = 0; c < 10; c++) { - if (pEntry[c] < '0' || pEntry[c] > '9') { + if (!std::isdigit(pEntry[c])) return FALSE; - } } } m_CrossRef.SetAtGrow(objnum, offset); @@ -562,9 +559,8 @@ bool CPDF_Parser::LoadCrossRefV4(FX_FILESIZE pos, FX_FILESIZE offset = (FX_FILESIZE)FXSYS_atoi64(pEntry); if (offset == 0) { for (int32_t c = 0; c < 10; c++) { - if (pEntry[c] < '0' || pEntry[c] > '9') { + if (!std::isdigit(pEntry[c])) return false; - } } } m_CrossRef.SetAtGrow(objnum, offset); @@ -632,28 +628,32 @@ FX_BOOL CPDF_Parser::RebuildCrossRef() { uint8_t byte = buffer[i]; switch (status) { case 0: - if (PDFCharIsWhitespace(byte)) { + if (PDFCharIsWhitespace(byte)) status = 1; - } - if (byte <= '9' && byte >= '0') { + + if (std::isdigit(byte)) { --i; status = 1; } + if (byte == '%') { inside_index = 0; status = 9; } + if (byte == '(') { status = 10; depth = 1; } + if (byte == '<') { inside_index = 1; status = 11; } - if (byte == '\\') { + + if (byte == '\\') status = 13; - } + if (byte == 't') { status = 7; inside_index = 1; @@ -662,10 +662,10 @@ FX_BOOL CPDF_Parser::RebuildCrossRef() { case 1: if (PDFCharIsWhitespace(byte)) { break; - } else if (byte <= '9' && byte >= '0') { + } else if (std::isdigit(byte)) { start_pos = pos + i; status = 2; - objnum = byte - '0'; + objnum = FXSYS_toDecimalDigit(byte); } else if (byte == 't') { status = 7; inside_index = 1; @@ -678,8 +678,8 @@ FX_BOOL CPDF_Parser::RebuildCrossRef() { } break; case 2: - if (byte <= '9' && byte >= '0') { - objnum = objnum * 10 + byte - '0'; + if (std::isdigit(byte)) { + objnum = objnum * 10 + FXSYS_toDecimalDigit(byte); break; } else if (PDFCharIsWhitespace(byte)) { status = 3; @@ -690,10 +690,10 @@ FX_BOOL CPDF_Parser::RebuildCrossRef() { } break; case 3: - if (byte <= '9' && byte >= '0') { + if (std::isdigit(byte)) { start_pos1 = pos + i; status = 4; - gennum = byte - '0'; + gennum = FXSYS_toDecimalDigit(byte); } else if (PDFCharIsWhitespace(byte)) { break; } else if (byte == 't') { @@ -705,8 +705,8 @@ FX_BOOL CPDF_Parser::RebuildCrossRef() { } break; case 4: - if (byte <= '9' && byte >= '0') { - gennum = gennum * 10 + byte - '0'; + if (std::isdigit(byte)) { + gennum = gennum * 10 + FXSYS_toDecimalDigit(byte); break; } else if (PDFCharIsWhitespace(byte)) { status = 5; @@ -721,9 +721,9 @@ FX_BOOL CPDF_Parser::RebuildCrossRef() { inside_index = 1; } else if (PDFCharIsWhitespace(byte)) { break; - } else if (byte <= '9' && byte >= '0') { + } else if (std::isdigit(byte)) { objnum = gennum; - gennum = byte - '0'; + gennum = FXSYS_toDecimalDigit(byte); start_pos = start_pos1; start_pos1 = pos + i; status = 4; @@ -1858,7 +1858,7 @@ CFX_ByteString CPDF_SyntaxParser::ReadString() { break; case 1: if (ch >= '0' && ch <= '7') { - iEscCode = ch - '0'; + iEscCode = FXSYS_toDecimalDigit(ch); status = 2; break; } @@ -1883,7 +1883,7 @@ CFX_ByteString CPDF_SyntaxParser::ReadString() { break; case 2: if (ch >= '0' && ch <= '7') { - iEscCode = iEscCode * 8 + ch - '0'; + iEscCode = iEscCode * 8 + FXSYS_toDecimalDigit(ch); status = 3; } else { buf.AppendChar(iEscCode); @@ -1893,7 +1893,7 @@ CFX_ByteString CPDF_SyntaxParser::ReadString() { break; case 3: if (ch >= '0' && ch <= '7') { - iEscCode = iEscCode * 8 + ch - '0'; + iEscCode = iEscCode * 8 + FXSYS_toDecimalDigit(ch); buf.AppendChar(iEscCode); status = 0; } else { @@ -1918,48 +1918,33 @@ CFX_ByteString CPDF_SyntaxParser::ReadString() { } CFX_ByteString CPDF_SyntaxParser::ReadHexString() { uint8_t ch; - if (!GetNextChar(ch)) { + if (!GetNextChar(ch)) return CFX_ByteString(); - } + CFX_BinaryBuf buf; - FX_BOOL bFirst = TRUE; + bool bFirst = true; uint8_t code = 0; while (1) { - if (ch == '>') { + if (ch == '>') break; - } - if (ch >= '0' && ch <= '9') { - if (bFirst) { - code = (ch - '0') * 16; - } else { - code += ch - '0'; - buf.AppendByte((uint8_t)code); - } - bFirst = !bFirst; - } else if (ch >= 'A' && ch <= 'F') { - if (bFirst) { - code = (ch - 'A' + 10) * 16; - } else { - code += ch - 'A' + 10; - buf.AppendByte((uint8_t)code); - } - bFirst = !bFirst; - } else if (ch >= 'a' && ch <= 'f') { + + if (std::isxdigit(ch)) { + int val = FXSYS_toHexDigit(ch); if (bFirst) { - code = (ch - 'a' + 10) * 16; + code = val * 16; } else { - code += ch - 'a' + 10; + code += val; buf.AppendByte((uint8_t)code); } bFirst = !bFirst; } - if (!GetNextChar(ch)) { + + if (!GetNextChar(ch)) break; - } } - if (!bFirst) { + if (!bFirst) buf.AppendByte((uint8_t)code); - } + return buf.GetByteString(); } void CPDF_SyntaxParser::ToNextLine() { diff --git a/core/src/fpdfapi/fpdf_parser/fpdf_parser_utility.cpp b/core/src/fpdfapi/fpdf_parser/fpdf_parser_utility.cpp index 89a5deb62d..9729bab942 100644 --- a/core/src/fpdfapi/fpdf_parser/fpdf_parser_utility.cpp +++ b/core/src/fpdfapi/fpdf_parser/fpdf_parser_utility.cpp @@ -5,6 +5,7 @@ // Original code copyright 2014 Foxit Software Inc. http://www.foxitsoftware.com #include "core/include/fpdfapi/fpdf_parser.h" +#include "core/include/fxcrt/fx_ext.h" // Indexed by 8-bit character code, contains either: // 'W' - for whitespace: NUL, TAB, CR, LF, FF, 0x80, 0xff @@ -279,18 +280,7 @@ FX_BOOL CPDF_SimpleParser::FindTagParam(const CFX_ByteStringC& token, } return FALSE; } -static int _hex2dec(char ch) { - if (ch >= '0' && ch <= '9') { - return ch - '0'; - } - if (ch >= 'a' && ch <= 'f') { - return ch - 'a' + 10; - } - if (ch >= 'A' && ch <= 'F') { - return ch - 'A' + 10; - } - return 0; -} + CFX_ByteString PDF_NameDecode(const CFX_ByteStringC& bstr) { int size = bstr.GetLength(); const FX_CHAR* pSrc = bstr.GetCStr(); @@ -302,7 +292,8 @@ CFX_ByteString PDF_NameDecode(const CFX_ByteStringC& bstr) { FX_CHAR* pDest = pDestStart; for (int i = 0; i < size; i++) { if (pSrc[i] == '#' && i < size - 2) { - *pDest++ = _hex2dec(pSrc[i + 1]) * 16 + _hex2dec(pSrc[i + 2]); + *pDest++ = + FXSYS_toHexDigit(pSrc[i + 1]) * 16 + FXSYS_toHexDigit(pSrc[i + 2]); i += 2; } else { *pDest++ = pSrc[i]; diff --git a/core/src/fpdftext/fpdf_text.cpp b/core/src/fpdftext/fpdf_text.cpp index 55ab483722..0b7a849c75 100644 --- a/core/src/fpdftext/fpdf_text.cpp +++ b/core/src/fpdftext/fpdf_text.cpp @@ -14,6 +14,9 @@ #include "third_party/base/nonstd_unique_ptr.h" #include "txtproc.h" +#include +#include + CFX_ByteString CharFromUnicodeAlt(FX_WCHAR unicode, int destcp, const FX_CHAR* defchar) { @@ -436,10 +439,9 @@ void NormalizeString(CFX_WideString& str) { static FX_BOOL IsNumber(CFX_WideString& str) { for (int i = 0; i < str.GetLength(); i++) { FX_WCHAR ch = str[i]; - if ((ch < '0' || ch > '9') && ch != '-' && ch != '+' && ch != '.' && - ch != ' ') { + // TODO(dsinclair): --.+ +.-- should probably not be a number. + if (!std::iswdigit(ch) && ch != '-' && ch != '+' && ch != '.' && ch != ' ') return FALSE; - } } return TRUE; } diff --git a/core/src/fpdftext/fpdf_text_int.cpp b/core/src/fpdftext/fpdf_text_int.cpp index f02ca41e1a..0754757940 100644 --- a/core/src/fpdftext/fpdf_text_int.cpp +++ b/core/src/fpdftext/fpdf_text_int.cpp @@ -4,7 +4,8 @@ // Original code copyright 2014 Foxit Software Inc. http://www.foxitsoftware.com -#include +#include +#include #include #include "core/include/fpdfapi/fpdf_module.h" @@ -2446,8 +2447,8 @@ void CPDF_TextPageFind::ExtractFindWhat(const CFX_WideString& findwhat) { FX_BOOL CPDF_TextPageFind::IsMatchWholeWord(const CFX_WideString& csPageText, int startPos, int endPos) { - int char_left = 0; - int char_right = 0; + FX_WCHAR char_left = 0; + FX_WCHAR char_right = 0; int char_count = endPos - startPos + 1; if (char_count < 1) { return FALSE; @@ -2463,12 +2464,11 @@ FX_BOOL CPDF_TextPageFind::IsMatchWholeWord(const CFX_WideString& csPageText, } if ((char_left > 'A' && char_left < 'a') || (char_left > 'a' && char_left < 'z') || - (char_left > 0xfb00 && char_left < 0xfb06) || - (char_left >= '0' && char_left <= '9') || + (char_left > 0xfb00 && char_left < 0xfb06) || std::iswdigit(char_left) || (char_right > 'A' && char_right < 'a') || (char_right > 'a' && char_right < 'z') || (char_right > 0xfb00 && char_right < 0xfb06) || - (char_right >= '0' && char_right <= '9')) { + std::iswdigit(char_right)) { return FALSE; } if (!(('A' > char_left || char_left > 'Z') && diff --git a/core/src/fxcodec/codec/fx_codec.cpp b/core/src/fxcodec/codec/fx_codec.cpp index 7b6aacea2d..b357f8ac08 100644 --- a/core/src/fxcodec/codec/fx_codec.cpp +++ b/core/src/fxcodec/codec/fx_codec.cpp @@ -1,3 +1,4 @@ + // Copyright 2014 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. @@ -9,6 +10,7 @@ #include #include "codec_int.h" +#include "core/include/fxcrt/fx_ext.h" #include "core/include/fxcrt/fx_safe_types.h" #include "third_party/base/logging.h" @@ -151,6 +153,20 @@ FX_BOOL CCodec_BasicModule::RunLengthEncode(const uint8_t* src_buf, FX_DWORD& dest_size) { return FALSE; } + +#define EXPONENT_DETECT(ptr) \ + for (;; ptr++) { \ + if (!std::isdigit(*ptr)) { \ + if (endptr) \ + *endptr = (char*)ptr; \ + break; \ + } else { \ + exp_ret *= 10; \ + exp_ret += FXSYS_toDecimalDigit(*ptr); \ + continue; \ + } \ + } + extern "C" double FXstrtod(const char* nptr, char** endptr) { double ret = 0.0; const char* ptr = nptr; @@ -161,20 +177,20 @@ extern "C" double FXstrtod(const char* nptr, char** endptr) { return 0.0; } for (;; ptr++) { - if (!e_number && !e_point && (*ptr == '\t' || *ptr == ' ')) { + if (!e_number && !e_point && (*ptr == '\t' || *ptr == ' ')) continue; - } - if (*ptr >= '0' && *ptr <= '9') { - if (!e_number) { + + if (std::isdigit(*ptr)) { + if (!e_number) e_number = 1; - } + if (!e_point) { ret *= 10; - ret += (*ptr - '0'); + ret += FXSYS_toDecimalDigit(*ptr); } else { fra_count++; fra_ret *= 10; - fra_ret += (*ptr - '0'); + fra_ret += FXSYS_toDecimalDigit(*ptr); } continue; } @@ -192,29 +208,17 @@ extern "C" double FXstrtod(const char* nptr, char** endptr) { } } if (e_number && (*ptr == 'e' || *ptr == 'E')) { -#define EXPONENT_DETECT(ptr) \ - for (;; ptr++) { \ - if (*ptr < '0' || *ptr > '9') { \ - if (endptr) \ - *endptr = (char*)ptr; \ - break; \ - } else { \ - exp_ret *= 10; \ - exp_ret += (*ptr - '0'); \ - continue; \ - } \ - } exp_ptr = ptr++; if (*ptr == '+' || *ptr == '-') { exp_sig = (*ptr++ == '+') ? 1 : -1; - if (*ptr < '0' || *ptr > '9') { + if (!std::isdigit(*ptr)) { if (endptr) { *endptr = (char*)exp_ptr; } break; } EXPONENT_DETECT(ptr); - } else if (*ptr >= '0' && *ptr <= '9') { + } else if (std::isdigit(*ptr)) { EXPONENT_DETECT(ptr); } else { if (endptr) { @@ -222,7 +226,6 @@ extern "C" double FXstrtod(const char* nptr, char** endptr) { } break; } -#undef EXPONENT_DETECT break; } if (ptr != nptr && !e_number) { @@ -251,6 +254,8 @@ extern "C" double FXstrtod(const char* nptr, char** endptr) { } return is_negative ? -ret : ret; } +#undef EXPONENT_DETECT + FX_BOOL CCodec_BasicModule::A85Encode(const uint8_t* src_buf, FX_DWORD src_size, uint8_t*& dest_buf, diff --git a/core/src/fxcrt/fx_basic_bstring.cpp b/core/src/fxcrt/fx_basic_bstring.cpp index 9d64fbe139..574e57a9a6 100644 --- a/core/src/fxcrt/fx_basic_bstring.cpp +++ b/core/src/fxcrt/fx_basic_bstring.cpp @@ -5,6 +5,7 @@ // Original code copyright 2014 Foxit Software Inc. http://www.foxitsoftware.com #include // For offsetof(). +#include #include "core/include/fxcrt/fx_basic.h" #include "third_party/base/numerics/safe_math.h" @@ -493,8 +494,8 @@ void CFX_ByteString::FormatV(const FX_CHAR* lpszFormat, va_list argList) { } if (nWidth == 0) { nWidth = FXSYS_atoi(lpsz); - for (; (*lpsz) >= '0' && (*lpsz) <= '9'; lpsz++) - ; + while (std::isdigit(*lpsz)) + lpsz++; } if (nWidth < 0 || nWidth > 128 * 1024) { lpszFormat = "Bad width"; @@ -509,8 +510,8 @@ void CFX_ByteString::FormatV(const FX_CHAR* lpszFormat, va_list argList) { lpsz++; } else { nPrecision = FXSYS_atoi(lpsz); - for (; (*lpsz) >= '0' && (*lpsz) <= '9'; lpsz++) - ; + while (std::isdigit(*lpsz)) + lpsz++; } } if (nPrecision < 0 || nPrecision > 128 * 1024) { diff --git a/core/src/fxcrt/fx_basic_gcc.cpp b/core/src/fxcrt/fx_basic_gcc.cpp index f8b0c7ac78..c352ee3f81 100644 --- a/core/src/fxcrt/fx_basic_gcc.cpp +++ b/core/src/fxcrt/fx_basic_gcc.cpp @@ -5,29 +5,49 @@ // Original code copyright 2014 Foxit Software Inc. http://www.foxitsoftware.com #include +#include +#include #include "core/include/fxcrt/fx_ext.h" #include "core/include/fxcrt/fx_string.h" -template -T FXSYS_StrToInt(STR_T str) { +template +T FXSYS_StrToInt(const FX_CHAR* str) { FX_BOOL neg = FALSE; - if (str == NULL) { + if (!str) return 0; - } + if (*str == '-') { neg = TRUE; str++; } T num = 0; - while (*str) { - if ((*str) < '0' || (*str) > '9') { + while (*str && std::isdigit(*str)) { + if (num > (std::numeric_limits::max() - 9) / 10) break; - } - if (num > (std::numeric_limits::max() - 9) / 10) { + + num = num * 10 + FXSYS_toDecimalDigit(*str); + str++; + } + return neg ? -num : num; +} + +template +T FXSYS_StrToInt(const FX_WCHAR* str) { + FX_BOOL neg = FALSE; + if (!str) + return 0; + + if (*str == '-') { + neg = TRUE; + str++; + } + T num = 0; + while (*str && std::iswdigit(*str)) { + if (num > (std::numeric_limits::max() - 9) / 10) break; - } - num = num * 10 + (*str) - '0'; + + num = num * 10 + FXSYS_toDecimalDigitWide(*str); str++; } return neg ? -num : num; @@ -71,16 +91,16 @@ STR_T FXSYS_IntToStr(T value, STR_T string, int radix) { extern "C" { #endif int32_t FXSYS_atoi(const FX_CHAR* str) { - return FXSYS_StrToInt(str); + return FXSYS_StrToInt(str); } int32_t FXSYS_wtoi(const FX_WCHAR* str) { - return FXSYS_StrToInt(str); + return FXSYS_StrToInt(str); } int64_t FXSYS_atoi64(const FX_CHAR* str) { - return FXSYS_StrToInt(str); + return FXSYS_StrToInt(str); } int64_t FXSYS_wtoi64(const FX_WCHAR* str) { - return FXSYS_StrToInt(str); + return FXSYS_StrToInt(str); } const FX_CHAR* FXSYS_i64toa(int64_t value, FX_CHAR* str, int radix) { return FXSYS_IntToStr(value, str, radix); diff --git a/core/src/fxcrt/fx_basic_util.cpp b/core/src/fxcrt/fx_basic_util.cpp index 3e9d6169cd..b4c7064da2 100644 --- a/core/src/fxcrt/fx_basic_util.cpp +++ b/core/src/fxcrt/fx_basic_util.cpp @@ -5,6 +5,9 @@ // Original code copyright 2014 Foxit Software Inc. http://www.foxitsoftware.com #include "core/include/fxcrt/fx_basic.h" +#include "core/include/fxcrt/fx_ext.h" + +#include #if _FXM_PLATFORM_ != _FXM_PLATFORM_WINDOWS_ #include @@ -102,14 +105,11 @@ void FX_atonum(const CFX_ByteStringC& strc, FX_BOOL& bInteger, void* pData) { bNegative = TRUE; cc++; } - while (cc < len) { - if (str[cc] < '0' || str[cc] > '9') { + while (cc < len && std::isdigit(str[cc])) { + // TODO(dsinclair): This is not the right way to handle overflow. + integer = integer * 10 + FXSYS_toDecimalDigit(str[cc]); + if (integer < 0) break; - } - integer = integer * 10 + str[cc] - '0'; - if (integer < 0) { - break; - } cc++; } if (bNegative) { @@ -146,7 +146,7 @@ FX_FLOAT FX_atof(const CFX_ByteStringC& strc) { if (str[cc] == '.') { break; } - value = value * 10 + str[cc] - '0'; + value = value * 10 + FXSYS_toDecimalDigit(str[cc]); cc++; } static const FX_FLOAT fraction_scales[] = { @@ -157,7 +157,7 @@ FX_FLOAT FX_atof(const CFX_ByteStringC& strc) { if (cc < len && str[cc] == '.') { cc++; while (cc < len) { - value += fraction_scales[scale] * (str[cc] - '0'); + value += fraction_scales[scale] * FXSYS_toDecimalDigit(str[cc]); scale++; if (scale == sizeof fraction_scales / sizeof(FX_FLOAT)) { break; diff --git a/core/src/fxcrt/fx_basic_wstring.cpp b/core/src/fxcrt/fx_basic_wstring.cpp index 131672da2b..220ffbd57d 100644 --- a/core/src/fxcrt/fx_basic_wstring.cpp +++ b/core/src/fxcrt/fx_basic_wstring.cpp @@ -5,8 +5,10 @@ // Original code copyright 2014 Foxit Software Inc. http://www.foxitsoftware.com #include // For offsetof(). +#include #include "core/include/fxcrt/fx_basic.h" +#include "core/include/fxcrt/fx_ext.h" #include "third_party/base/numerics/safe_math.h" // static @@ -765,8 +767,8 @@ void CFX_WideString::FormatV(const FX_WCHAR* lpszFormat, va_list argList) { } if (nWidth == 0) { nWidth = FXSYS_wtoi(lpsz); - for (; *lpsz != 0 && (*lpsz) <= '9' && (*lpsz) >= '0'; lpsz++) - ; + while (std::iswdigit(*lpsz)) + ++lpsz; } if (nWidth < 0 || nWidth > 128 * 1024) { lpszFormat = L"Bad width"; @@ -781,8 +783,8 @@ void CFX_WideString::FormatV(const FX_WCHAR* lpszFormat, va_list argList) { lpsz++; } else { nPrecision = FXSYS_wtoi(lpsz); - for (; *lpsz != 0 && (*lpsz) >= '0' && (*lpsz) <= '9'; lpsz++) - ; + while (std::iswdigit(*lpsz)) + ++lpsz; } } if (nPrecision < 0 || nPrecision > 128 * 1024) { @@ -968,7 +970,7 @@ FX_FLOAT FX_wtof(const FX_WCHAR* str, int len) { if (str[cc] == '.') { break; } - integer = integer * 10 + str[cc] - '0'; + integer = integer * 10 + FXSYS_toDecimalDigitWide(str[cc]); cc++; } FX_FLOAT fraction = 0; @@ -976,7 +978,7 @@ FX_FLOAT FX_wtof(const FX_WCHAR* str, int len) { cc++; FX_FLOAT scale = 0.1f; while (cc < len) { - fraction += scale * (str[cc] - '0'); + fraction += scale * FXSYS_toDecimalDigitWide(str[cc]); scale *= 0.1f; cc++; } diff --git a/core/src/fxcrt/fx_extension.cpp b/core/src/fxcrt/fx_extension.cpp index dd80b2284a..3761192a8d 100644 --- a/core/src/fxcrt/fx_extension.cpp +++ b/core/src/fxcrt/fx_extension.cpp @@ -63,9 +63,7 @@ IFX_MemoryStream* FX_CreateMemoryStream(uint8_t* pBuffer, IFX_MemoryStream* FX_CreateMemoryStream(FX_BOOL bConsecutive) { return new CFX_MemoryStream(bConsecutive); } -#ifdef __cplusplus -extern "C" { -#endif + FX_FLOAT FXSYS_tan(FX_FLOAT a) { return (FX_FLOAT)tan(a); } @@ -202,12 +200,7 @@ FX_DWORD FX_HashCode_String_GetW(const FX_WCHAR* pStr, } return dwHashCode; } -#ifdef __cplusplus -} -#endif -#ifdef __cplusplus -extern "C" { -#endif + void* FX_Random_MT_Start(FX_DWORD dwSeed) { FX_LPMTRANDOMCONTEXT pContext = FX_Alloc(FX_MTRANDOMCONTEXT, 1); pContext->mt[0] = dwSeed; @@ -310,12 +303,7 @@ void FX_Random_GenerateCrypto(FX_DWORD* pBuffer, int32_t iCount) { FX_Random_GenerateBase(pBuffer, iCount); #endif } -#ifdef __cplusplus -} -#endif -#ifdef __cplusplus -extern "C" { -#endif + void FX_GUID_CreateV4(FX_LPGUID pGUID) { #if (_FX_OS_ == _FX_WIN32_DESKTOP_ || _FX_OS_ == _FX_WIN32_MOBILE_ || \ _FX_OS_ == _FX_WIN64_) @@ -348,6 +336,3 @@ void FX_GUID_ToString(FX_LPCGUID pGUID, } bsStr.ReleaseBuffer(bSeparator ? 36 : 32); } -#ifdef __cplusplus -} -#endif diff --git a/core/src/fxcrt/fx_extension_unittest.cpp b/core/src/fxcrt/fx_extension_unittest.cpp new file mode 100644 index 0000000000..b12dcbe07f --- /dev/null +++ b/core/src/fxcrt/fx_extension_unittest.cpp @@ -0,0 +1,24 @@ +// Copyright 2015 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. + +#include "testing/gtest/include/gtest/gtest.h" + +#include "core/include/fxcrt/fx_ext.h" + +TEST(fxcrt, FXSYS_toHexDigit) { + EXPECT_EQ(10, FXSYS_toHexDigit('a')); + EXPECT_EQ(10, FXSYS_toHexDigit('A')); + EXPECT_EQ(7, FXSYS_toHexDigit('7')); + EXPECT_EQ(0, FXSYS_toHexDigit('i')); +} + +TEST(fxcrt, FXSYS_toDecimalDigit) { + EXPECT_EQ(7, FXSYS_toDecimalDigit('7')); + EXPECT_EQ(0, FXSYS_toDecimalDigit('a')); +} + +TEST(fxcrt, FXSYS_toDecimalDigitWide) { + EXPECT_EQ(7, FXSYS_toDecimalDigitWide(L'7')); + EXPECT_EQ(0, FXSYS_toDecimalDigitWide(L'a')); +} diff --git a/core/src/fxcrt/fx_xml_parser.cpp b/core/src/fxcrt/fx_xml_parser.cpp index 429bc38289..2d3ff6631e 100644 --- a/core/src/fxcrt/fx_xml_parser.cpp +++ b/core/src/fxcrt/fx_xml_parser.cpp @@ -6,6 +6,7 @@ #include "xml_int.h" +#include "core/include/fxcrt/fx_ext.h" #include "core/include/fxcrt/fx_xml.h" CXML_Parser::~CXML_Parser() { @@ -228,9 +229,8 @@ FX_DWORD CXML_Parser::GetCharRef() { iState = 10; break; } - if (g_FXCRT_XML_IsDigital(ch)) { - code = code * 10 + ch - '0'; - } + if (g_FXCRT_XML_IsDigital(ch)) + code = code * 10 + FXSYS_toDecimalDigit(ch); break; case 4: m_dwIndex++; @@ -242,7 +242,7 @@ FX_DWORD CXML_Parser::GetCharRef() { g_FXCRT_XML_ByteTypes[ch] & FXCRTM_XML_CHARTYPE_HexChar; if (nHex) { if (nHex == FXCRTM_XML_CHARTYPE_HexDigital) { - code = (code << 4) + ch - '0'; + code = (code << 4) + FXSYS_toDecimalDigit(ch); } else if (nHex == FXCRTM_XML_CHARTYPE_HexLowerLetter) { code = (code << 4) + ch - 87; } else { -- cgit v1.2.3