From dcc0723e70cb815630d2043abcd8f4a50b7199b0 Mon Sep 17 00:00:00 2001
From: Adrian-George Bostan <adrg@epistack.com>
Date: Sun, 28 Apr 2019 15:26:50 +0300
Subject: [PATCH] Search xref objects with tolerance both to the left and right
 (#447)

* Search xref objects with tolerance both to the left and right. Try searching xref to the left only if not found to the right.
---
 pdf/core/io.go     |  4 +++
 pdf/core/parser.go | 62 +++++++++++++++++++++++-----------------------
 2 files changed, 35 insertions(+), 31 deletions(-)

diff --git a/pdf/core/io.go b/pdf/core/io.go
index 882d5df7..bf23ed50 100644
--- a/pdf/core/io.go
+++ b/pdf/core/io.go
@@ -83,6 +83,10 @@ func (parser *PdfParser) GetFileOffset() int64 {
 
 // SetFileOffset sets the file to an offset position and resets buffer.
 func (parser *PdfParser) SetFileOffset(offset int64) {
+	if offset < 0 {
+		offset = 0
+	}
+
 	parser.rs.Seek(offset, io.SeekStart)
 	parser.reader = bufio.NewReader(parser.rs)
 }
diff --git a/pdf/core/parser.go b/pdf/core/parser.go
index c5a24021..bf8a1c1d 100755
--- a/pdf/core/parser.go
+++ b/pdf/core/parser.go
@@ -1103,43 +1103,43 @@ func (parser *PdfParser) parseXrefStream(xstm *PdfObjectInteger) (*PdfObjectDict
 	return trailerDict, nil
 }
 
-// Parse xref table at the current file position.  Can either be a
-// standard xref table, or an xref stream.
+// Parse xref table at the current file position. Can either be a standard xref
+// table, or an xref stream.
 func (parser *PdfParser) parseXref() (*PdfObjectDictionary, error) {
-	var err error
-	var trailerDict *PdfObjectDictionary
-
-	// Points to xref table or xref stream object?
-	bb, _ := parser.reader.Peek(20)
-	if reIndirectObject.MatchString(string(bb)) {
-		common.Log.Trace("xref points to an object.  Probably xref object")
-		common.Log.Trace("starting with \"%s\"", string(bb))
-		trailerDict, err = parser.parseXrefStream(nil)
-		if err != nil {
-			return nil, err
+	// Search xrefs within 20 bytes of the current location. If the first
+	// iteration of the loop is unable to find a match, peek another 20 bytes
+	// left of the current location, add them to the previously read buffer
+	// and try again.
+	const bufLen = 20
+	bb, _ := parser.reader.Peek(bufLen)
+	for i := 0; i < 2; i++ {
+		if reIndirectObject.Match(bb) {
+			common.Log.Trace("xref points to an object. Probably xref object")
+			common.Log.Debug("starting with \"%s\"", string(bb))
+			return parser.parseXrefStream(nil)
 		}
-	} else if reXrefTable.MatchString(string(bb)) {
-		common.Log.Trace("Standard xref section table!")
-		var err error
-		trailerDict, err = parser.parseXrefTable()
-		if err != nil {
-			return nil, err
-		}
-	} else {
-		common.Log.Debug("Warning: Unable to find xref table or stream. Repair attempted: Looking for earliest xref from bottom.")
-		err := parser.repairSeekXrefMarker()
-		if err != nil {
-			common.Log.Debug("Repair failed - %v", err)
-			return nil, err
+		if reXrefTable.Match(bb) {
+			common.Log.Trace("Standard xref section table!")
+			return parser.parseXrefTable()
 		}
 
-		trailerDict, err = parser.parseXrefTable()
-		if err != nil {
-			return nil, err
-		}
+		// xref match failed. Peek 20 bytes to the left of the current offset,
+		// append them to the previously read buffer and try again. Reset to the
+		// original offset after reading.
+		offset := parser.GetFileOffset()
+		parser.SetFileOffset(offset - bufLen)
+		defer parser.SetFileOffset(offset)
+
+		lbb, _ := parser.reader.Peek(bufLen)
+		bb = append(lbb, bb...)
 	}
 
-	return trailerDict, err
+	common.Log.Debug("Warning: Unable to find xref table or stream. Repair attempted: Looking for earliest xref from bottom.")
+	if err := parser.repairSeekXrefMarker(); err != nil {
+		common.Log.Debug("Repair failed - %v", err)
+		return nil, err
+	}
+	return parser.parseXrefTable()
 }
 
 // Look for EOF marker and seek to its beginning.