From 9bd18183ba8210c91d71c3060146235750a4c71c Mon Sep 17 00:00:00 2001 From: Jun Fang Date: Fri, 25 Sep 2015 20:32:46 -0700 Subject: Fix the issue that pdfium swallows 'fi' or 'ff' in some pdf files Pdfium swallows 'fi' or 'ff' in some tested files because it doesn't load the embedded font file correctly. The root cause is that there is incorrect keyword like 'ngendstream' in the stream of the embedded font file. Pdfium tries to find another correct keyword but uses wrong offset rather than accumulated offset. BUG=524043 R=thestig@chromium.org, tsepez@chromium.org Review URL: https://codereview.chromium.org/1307353005 . --- .../src/fpdfapi/fpdf_parser/fpdf_parser_parser.cpp | 56 +++++++++------------- 1 file changed, 23 insertions(+), 33 deletions(-) (limited to 'core/src/fpdfapi/fpdf_parser') diff --git a/core/src/fpdfapi/fpdf_parser/fpdf_parser_parser.cpp b/core/src/fpdfapi/fpdf_parser/fpdf_parser_parser.cpp index 7482f0b8e4..cd923f3e22 100644 --- a/core/src/fpdfapi/fpdf_parser/fpdf_parser_parser.cpp +++ b/core/src/fpdfapi/fpdf_parser/fpdf_parser_parser.cpp @@ -2455,8 +2455,8 @@ CPDF_Stream* CPDF_SyntaxParser::ReadStream(CPDF_Dictionary* pDict, len = pLenObj->GetInteger(); } // Check whether end of line markers follow the keyword 'stream'. - unsigned int numMarkers = ReadEOLMarkers(m_Pos); - m_Pos += numMarkers; + // The stream starts after end of line markers. + m_Pos += ReadEOLMarkers(m_Pos); FX_FILESIZE streamStartPos = m_Pos; if (pContext) { pContext->m_DataStart = streamStartPos; @@ -2467,19 +2467,18 @@ CPDF_Stream* CPDF_SyntaxParser::ReadStream(CPDF_Dictionary* pDict, objnum == (FX_DWORD)m_MetadataObjnum ? nullptr : m_pCryptoHandler; if (!pCryptoHandler) { FX_BOOL bSearchForKeyword = TRUE; - unsigned int prevMarkers = 0; - unsigned int nextMarkers = 0; if (len >= 0) { pdfium::base::CheckedNumeric pos = m_Pos; pos += len; if (pos.IsValid() && pos.ValueOrDie() < m_FileLen) { m_Pos = pos.ValueOrDie(); } - prevMarkers = ReadEOLMarkers(m_Pos); + m_Pos += ReadEOLMarkers(m_Pos); + FXSYS_memset(m_WordBuffer, 0, ENDSTREAM_LEN + 1); GetNextWord(); - nextMarkers = ReadEOLMarkers(m_Pos); - if (m_WordSize == ENDSTREAM_LEN && prevMarkers != 0 && nextMarkers != 0 && - FXSYS_memcmp(m_WordBuffer, "endstream", ENDSTREAM_LEN) == 0) { + if (FXSYS_memcmp(m_WordBuffer, "endstream", ENDSTREAM_LEN) == 0 && + IsWholeWord(m_Pos - ENDSTREAM_LEN, m_FileLen, + FX_BSTRC("endstream").GetPtr(), ENDSTREAM_LEN, TRUE)) { bSearchForKeyword = FALSE; } } @@ -2494,22 +2493,12 @@ CPDF_Stream* CPDF_SyntaxParser::ReadStream(CPDF_Dictionary* pDict, // Can't find any "endstream". break; } - prevMarkers = ReadEOLMarkers(streamStartPos + endStreamOffset - 1); - nextMarkers = - ReadEOLMarkers(streamStartPos + endStreamOffset + ENDSTREAM_LEN); - if (prevMarkers != 0 && nextMarkers != 0) { + if (IsWholeWord(m_Pos - ENDSTREAM_LEN, m_FileLen, + FX_BSTRC("endstream").GetPtr(), ENDSTREAM_LEN, TRUE)) { // Stop searching when the keyword "endstream" is found. + endStreamOffset = m_Pos - streamStartPos - ENDSTREAM_LEN; break; - } else { - unsigned char ch = 0x00; - GetCharAt(streamStartPos + endStreamOffset + ENDSTREAM_LEN, ch); - if (ch == 0x09 || ch == 0x20) { - //"endstream" is treated as a keyword - // when it is followed by a tab or whitespace - break; - } } - m_Pos += ENDSTREAM_LEN; } m_Pos = streamStartPos; FX_FILESIZE endObjOffset = 0; @@ -2519,14 +2508,12 @@ CPDF_Stream* CPDF_SyntaxParser::ReadStream(CPDF_Dictionary* pDict, // Can't find any "endobj". break; } - prevMarkers = ReadEOLMarkers(streamStartPos + endObjOffset - 1); - nextMarkers = - ReadEOLMarkers(streamStartPos + endObjOffset + ENDOBJ_LEN); - if (prevMarkers != 0 && nextMarkers != 0) { + if (IsWholeWord(m_Pos - ENDOBJ_LEN, m_FileLen, + FX_BSTRC("endobj").GetPtr(), ENDOBJ_LEN, TRUE)) { // Stop searching when the keyword "endobj" is found. + endObjOffset = m_Pos - streamStartPos - ENDOBJ_LEN; break; } - m_Pos += ENDOBJ_LEN; } if (endStreamOffset < 0 && endObjOffset < 0) { // Can't find "endstream" or "endobj". @@ -2542,7 +2529,7 @@ CPDF_Stream* CPDF_SyntaxParser::ReadStream(CPDF_Dictionary* pDict, endStreamOffset = endObjOffset; } len = endStreamOffset; - numMarkers = ReadEOLMarkers(streamStartPos + endStreamOffset - 2); + int numMarkers = ReadEOLMarkers(streamStartPos + endStreamOffset - 2); if (numMarkers == 2) { len -= 2; } else { @@ -2579,8 +2566,9 @@ CPDF_Stream* CPDF_SyntaxParser::ReadStream(CPDF_Dictionary* pDict, pContext->m_DataEnd = pContext->m_DataStart + len; } streamStartPos = m_Pos; + FXSYS_memset(m_WordBuffer, 0, ENDOBJ_LEN + 1); GetNextWord(); - numMarkers = ReadEOLMarkers(m_Pos); + int numMarkers = ReadEOLMarkers(m_Pos); if (m_WordSize == ENDOBJ_LEN && numMarkers != 0 && FXSYS_memcmp(m_WordBuffer, "endobj", ENDOBJ_LEN) == 0) { m_Pos = streamStartPos; @@ -2611,7 +2599,8 @@ int32_t CPDF_SyntaxParser::GetDirectNum() { FX_BOOL CPDF_SyntaxParser::IsWholeWord(FX_FILESIZE startpos, FX_FILESIZE limit, const uint8_t* tag, - FX_DWORD taglen) { + FX_DWORD taglen, + FX_BOOL checkKeyword) { uint8_t type = PDF_CharType[tag[0]]; FX_BOOL bCheckLeft = type != 'D' && type != 'W'; type = PDF_CharType[tag[taglen - 1]]; @@ -2620,13 +2609,13 @@ FX_BOOL CPDF_SyntaxParser::IsWholeWord(FX_FILESIZE startpos, if (bCheckRight && startpos + (int32_t)taglen <= limit && GetCharAt(startpos + (int32_t)taglen, ch)) { uint8_t type = PDF_CharType[ch]; - if (type == 'N' || type == 'R') { + if (type == 'N' || type == 'R' || (checkKeyword && type == 'D')) { return FALSE; } } if (bCheckLeft && startpos > 0 && GetCharAt(startpos - 1, ch)) { uint8_t type = PDF_CharType[ch]; - if (type == 'N' || type == 'R') { + if (type == 'N' || type == 'R' || (checkKeyword && type == 'D')) { return FALSE; } } @@ -2682,7 +2671,8 @@ FX_BOOL CPDF_SyntaxParser::SearchWord(const CFX_ByteStringC& tag, } } FX_FILESIZE startpos = bForward ? pos - taglen + 1 : pos; - if (!bWholeWord || IsWholeWord(startpos, limit, tag.GetPtr(), taglen)) { + if (!bWholeWord || + IsWholeWord(startpos, limit, tag.GetPtr(), taglen, FALSE)) { m_Pos = startpos; return TRUE; } @@ -2739,7 +2729,7 @@ int32_t CPDF_SyntaxParser::SearchMultiWord(const CFX_ByteStringC& tags, if (pPatterns[i].m_Offset == pPatterns[i].m_Len) { if (!bWholeWord || IsWholeWord(pos - pPatterns[i].m_Len, limit, pPatterns[i].m_pTag, - pPatterns[i].m_Len)) { + pPatterns[i].m_Len, FALSE)) { found = i; goto end; } else { -- cgit v1.2.3