ABI change: fix use of off_t, size_t, and integer types

Significantly improve the code's use of off_t for file offsets, size_t
for memory sizes, and integer types in cases where there has to be
compatibility with external interfaces.  Rework sections of the code
that would have prevented qpdf from working on files larger than 2 (or
maybe 4) GB in size.

13 files changed, 43 insertions, 32 deletions

diff --git a/include/qpdf/Buffer.hh b/include/qpdf/Buffer.hh
index 31db93e2..234d6ec0 100644
--- a/include/qpdf/Buffer.hh
+++ b/include/qpdf/Buffer.hh
@@ -9,6 +9,7 @@
 #define __BUFFER_HH__
 
 #include <qpdf/DLL.h>
+#include <cstring>              // for size_t
 
 class Buffer
 {
@@ -19,12 +20,12 @@ class Buffer
     // Create a Buffer object whose memory is owned by the class and
     // will be freed when the Buffer object is destroyed.
     QPDF_DLL
-    Buffer(unsigned long size);
+    Buffer(size_t size);
 
     // Create a Buffer object whose memory is owned by the caller and
     // will not be freed when the Buffer is destroyed.
     QPDF_DLL
-    Buffer(unsigned char* buf, unsigned long size);
+    Buffer(unsigned char* buf, size_t size);
 
     QPDF_DLL
     Buffer(Buffer const&);
@@ -33,19 +34,19 @@ class Buffer
     QPDF_DLL
     ~Buffer();
     QPDF_DLL
-    unsigned long getSize() const;
+    size_t getSize() const;
     QPDF_DLL
     unsigned char const* getBuffer() const;
     QPDF_DLL
     unsigned char* getBuffer();
 
   private:
-    void init(unsigned long size, unsigned char* buf, bool own_memory);
+    void init(size_t size, unsigned char* buf, bool own_memory);
     void copy(Buffer const&);
     void destroy();
 
     bool own_memory;
-    unsigned long size;
+    size_t size;
     unsigned char* buf;
 };
 
diff --git a/include/qpdf/Pipeline.hh b/include/qpdf/Pipeline.hh
index 796bd0eb..4bab715a 100644
--- a/include/qpdf/Pipeline.hh
+++ b/include/qpdf/Pipeline.hh
@@ -46,7 +46,7 @@ class Pipeline
     // and then, if they are not end-of-line pipelines, call
     // getNext()->write or getNext()->finish.
     QPDF_DLL
-    virtual void write(unsigned char* data, int len) = 0;
+    virtual void write(unsigned char* data, size_t len) = 0;
     QPDF_DLL
     virtual void finish() = 0;
 
diff --git a/include/qpdf/Pl_Buffer.hh b/include/qpdf/Pl_Buffer.hh
index a85a595d..c8cdc92b 100644
--- a/include/qpdf/Pl_Buffer.hh
+++ b/include/qpdf/Pl_Buffer.hh
@@ -32,7 +32,7 @@ class Pl_Buffer: public Pipeline
     QPDF_DLL
     virtual ~Pl_Buffer();
     QPDF_DLL
-    virtual void write(unsigned char*, int);
+    virtual void write(unsigned char*, size_t);
     QPDF_DLL
     virtual void finish();
 
diff --git a/include/qpdf/Pl_Count.hh b/include/qpdf/Pl_Count.hh
index c2543f21..e5b8939f 100644
--- a/include/qpdf/Pl_Count.hh
+++ b/include/qpdf/Pl_Count.hh
@@ -12,6 +12,7 @@
 // calling finish().
 
 #include <qpdf/Pipeline.hh>
+#include <qpdf/Types.h>
 
 class Pl_Count: public Pipeline
 {
@@ -21,19 +22,19 @@ class Pl_Count: public Pipeline
     QPDF_DLL
     virtual ~Pl_Count();
     QPDF_DLL
-    virtual void write(unsigned char*, int);
+    virtual void write(unsigned char*, size_t);
     QPDF_DLL
     virtual void finish();
     // Returns the number of bytes written
     QPDF_DLL
-    int getCount() const;
+    off_t getCount() const;
     // Returns the last character written, or '\0' if no characters
     // have been written (in which case getCount() returns 0)
     QPDF_DLL
     unsigned char getLastChar() const;
 
   private:
-    int count;
+    off_t count;
     unsigned char last_char;
 };
 
diff --git a/include/qpdf/Pl_Discard.hh b/include/qpdf/Pl_Discard.hh
index c08afd88..6531506a 100644
--- a/include/qpdf/Pl_Discard.hh
+++ b/include/qpdf/Pl_Discard.hh
@@ -24,7 +24,7 @@ class Pl_Discard: public Pipeline
     QPDF_DLL
     virtual ~Pl_Discard();
     QPDF_DLL
-    virtual void write(unsigned char*, int);
+    virtual void write(unsigned char*, size_t);
     QPDF_DLL
     virtual void finish();
 };
diff --git a/include/qpdf/Pl_Flate.hh b/include/qpdf/Pl_Flate.hh
index 5244377e..fac9d467 100644
--- a/include/qpdf/Pl_Flate.hh
+++ b/include/qpdf/Pl_Flate.hh
@@ -24,7 +24,7 @@ class Pl_Flate: public Pipeline
     virtual ~Pl_Flate();
 
     QPDF_DLL
-    virtual void write(unsigned char* data, int len);
+    virtual void write(unsigned char* data, size_t len);
     QPDF_DLL
     virtual void finish();
 
diff --git a/include/qpdf/Pl_StdioFile.hh b/include/qpdf/Pl_StdioFile.hh
index b40f43bb..fc7c58c1 100644
--- a/include/qpdf/Pl_StdioFile.hh
+++ b/include/qpdf/Pl_StdioFile.hh
@@ -29,7 +29,7 @@ class Pl_StdioFile: public Pipeline
     virtual ~Pl_StdioFile();
 
     QPDF_DLL
-    virtual void write(unsigned char* buf, int len);
+    virtual void write(unsigned char* buf, size_t len);
     QPDF_DLL
     virtual void finish();
 
diff --git a/include/qpdf/QPDF.hh b/include/qpdf/QPDF.hh
index 6a9728b1..ee9dba87 100644
--- a/include/qpdf/QPDF.hh
+++ b/include/qpdf/QPDF.hh
@@ -388,7 +388,7 @@ class QPDF
 	virtual off_t tell() = 0;
 	virtual void seek(off_t offset, int whence) = 0;
 	virtual void rewind() = 0;
-	virtual size_t read(char* buffer, int length) = 0;
+	virtual size_t read(char* buffer, size_t length) = 0;
 	virtual void unreadCh(char ch) = 0;
 
       protected:
@@ -405,7 +405,7 @@ class QPDF
 	virtual off_t tell();
 	virtual void seek(off_t offset, int whence);
 	virtual void rewind();
-	virtual size_t read(char* buffer, int length);
+	virtual size_t read(char* buffer, size_t length);
 	virtual void unreadCh(char ch);
 
       private:
@@ -428,7 +428,7 @@ class QPDF
 	virtual off_t tell();
 	virtual void seek(off_t offset, int whence);
 	virtual void rewind();
-	virtual size_t read(char* buffer, int length);
+	virtual size_t read(char* buffer, size_t length);
 	virtual void unreadCh(char ch);
 
       private:
@@ -490,7 +490,7 @@ class QPDF
 	PointerHolder<InputSource> input, int objid, int generation,
 	bool in_object_stream,
 	bool in_array, bool in_dictionary);
-    int recoverStreamLength(
+    size_t recoverStreamLength(
 	PointerHolder<InputSource> input, int objid, int generation,
 	off_t stream_offset);
     QPDFTokenizer::Token readToken(PointerHolder<InputSource>);
diff --git a/include/qpdf/QPDFObjectHandle.hh b/include/qpdf/QPDFObjectHandle.hh
index 6213ff10..f2ed22b7 100644
--- a/include/qpdf/QPDFObjectHandle.hh
+++ b/include/qpdf/QPDFObjectHandle.hh
@@ -327,7 +327,7 @@ class QPDFObjectHandle
 	// object must be dictionary object
 	static QPDFObjectHandle newStream(
 	    QPDF* qpdf, int objid, int generation,
-	    QPDFObjectHandle stream_dict, off_t offset, int length)
+	    QPDFObjectHandle stream_dict, off_t offset, size_t length)
 	{
 	    return QPDFObjectHandle::newStream(
 		qpdf, objid, generation, stream_dict, offset, length);
@@ -371,7 +371,7 @@ class QPDFObjectHandle
     static QPDFObjectHandle newIndirect(QPDF*, int objid, int generation);
     static QPDFObjectHandle newStream(
 	QPDF* qpdf, int objid, int generation,
-	QPDFObjectHandle stream_dict, off_t offset, int length);
+	QPDFObjectHandle stream_dict, off_t offset, size_t length);
 
     void assertInitialized() const;
     void assertType(char const* type_name, bool istype);
diff --git a/include/qpdf/QPDFWriter.hh b/include/qpdf/QPDFWriter.hh
index a8a51791..6a1c23cf 100644
--- a/include/qpdf/QPDFWriter.hh
+++ b/include/qpdf/QPDFWriter.hh
@@ -21,6 +21,7 @@
 
 #include <qpdf/DLL.h>
 #include <qpdf/Constants.h>
+#include <qpdf/Types.h>
 
 #include <qpdf/QPDFXRefEntry.hh>
 
@@ -219,7 +220,7 @@ class QPDFWriter
     void writePad(int nspaces);
     void assignCompressedObjectNumbers(int objid);
     void enqueueObject(QPDFObjectHandle object);
-    void writeObjectStreamOffsets(std::vector<int>& offsets, int first_obj);
+    void writeObjectStreamOffsets(std::vector<off_t>& offsets, int first_obj);
     void writeObjectStream(QPDFObjectHandle object);
     void writeObject(QPDFObjectHandle object, int object_stream_index = -1);
     void writeTrailer(trailer_e which, int size,
@@ -229,7 +230,7 @@ class QPDFWriter
     void unparseObject(QPDFObjectHandle object, int level,
 		       unsigned int flags,
 		       // for stream dictionaries
-		       int stream_length, bool compress);
+		       size_t stream_length, bool compress);
     void unparseChild(QPDFObjectHandle child, int level, int flags);
     void initializeSpecialStreams();
     void preserveObjectStreams();
@@ -266,8 +267,8 @@ class QPDFWriter
 		       int prev,
 		       bool suppress_offsets,
 		       int hint_id,
-		       int hint_offset,
-		       int hint_length);
+		       off_t hint_offset,
+		       off_t hint_length);
     int writeXRefStream(int objid, int max_id, int max_offset,
 			trailer_e which, int first, int last, int size);
     int writeXRefStream(int objid, int max_id, int max_offset,
@@ -275,8 +276,8 @@ class QPDFWriter
 			// for linearization
 			int prev,
 			int hint_id,
-			int hint_offset,
-			int hint_length,
+			off_t hint_offset,
+			off_t hint_length,
 			bool skip_compression);
     int calculateXrefStreamPadding(int xref_bytes);
 
@@ -295,7 +296,7 @@ class QPDFWriter
     // stack items are of type Pl_Buffer, the buffer is retrieved.
     void popPipelineStack(PointerHolder<Buffer>* bp = 0);
 
-    void adjustAESStreamLength(unsigned long& length);
+    void adjustAESStreamLength(size_t& length);
     void pushEncryptionFilter();
     void pushDiscardFilter();
 
@@ -336,7 +337,7 @@ class QPDFWriter
     std::map<int, size_t> lengths;
     int next_objid;
     int cur_stream_length_id;
-    unsigned long cur_stream_length;
+    size_t cur_stream_length;
     bool added_newline;
     int max_ostream_index;
     std::set<int> normalized_streams;
diff --git a/include/qpdf/QPDFXRefEntry.hh b/include/qpdf/QPDFXRefEntry.hh
index 35843970..c362cc97 100644
--- a/include/qpdf/QPDFXRefEntry.hh
+++ b/include/qpdf/QPDFXRefEntry.hh
@@ -23,12 +23,12 @@ class QPDFXRefEntry
     QPDF_DLL
     QPDFXRefEntry();
     QPDF_DLL
-    QPDFXRefEntry(int type, int field1, int field2);
+    QPDFXRefEntry(int type, off_t field1, int field2);
 
     QPDF_DLL
     int getType() const;
     QPDF_DLL
-    int getOffset() const;	// only for type 1
+    off_t getOffset() const;	// only for type 1
     QPDF_DLL
     int getObjStreamNumber() const; // only for type 2
     QPDF_DLL
@@ -36,7 +36,7 @@ class QPDFXRefEntry
 
   private:
     int type;
-    int field1;
+    off_t field1;
     int field2;
 };
 
diff --git a/include/qpdf/QUtil.hh b/include/qpdf/QUtil.hh
index 7f5a38df..37307101 100644
--- a/include/qpdf/QUtil.hh
+++ b/include/qpdf/QUtil.hh
@@ -9,6 +9,7 @@
 #define __QUTIL_HH__
 
 #include <qpdf/DLL.h>
+#include <qpdf/Types.h>
 #include <string>
 #include <list>
 #include <stdexcept>
@@ -20,7 +21,7 @@ namespace QUtil
     // This is a collection of useful utility functions that don't
     // really go anywhere else.
     QPDF_DLL
-    std::string int_to_string(int, int length = 0);
+    std::string int_to_string(long long, int length = 0);
     QPDF_DLL
     std::string double_to_string(double, int decimal_places = 0);
 
@@ -44,6 +45,12 @@ namespace QUtil
     QPDF_DLL
     FILE* fopen_wrapper(std::string const&, FILE*);
 
+    // Wrap around off_t versions of fseek and ftell if available
+    QPDF_DLL
+    int fseek_off_t(FILE* stream, off_t offset, int whence);
+    QPDF_DLL
+    off_t ftell_off_t(FILE* stream);
+
     QPDF_DLL
     char* copy_string(std::string const&);
 
diff --git a/include/qpdf/qpdf-c.h b/include/qpdf/qpdf-c.h
index 168da154..77af4741 100644
--- a/include/qpdf/qpdf-c.h
+++ b/include/qpdf/qpdf-c.h
@@ -71,6 +71,7 @@
 
 #include <qpdf/DLL.h>
 #include <qpdf/Constants.h>
+#include <qpdf/Types.h>
 
 #ifdef __cplusplus
 extern "C" {
@@ -300,7 +301,7 @@ extern "C" {
      * if a subsequent call to qpdf_init_write or
      * qpdf_init_write_memory is called. */
     QPDF_DLL
-    unsigned long qpdf_get_buffer_length(qpdf_data qpdf);
+    size_t qpdf_get_buffer_length(qpdf_data qpdf);
     QPDF_DLL
     unsigned char const* qpdf_get_buffer(qpdf_data qpdf);