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.

TODO

@@ -1,35 +1,12 @@
 Next
 ====
 
- * Get rid of off_t.  size_t is okay.  Use autoconf to figure out what
+*** ABI changes have been made.  build.mk has been updated.
-   type to use for offsets.
-
- * Get rid of int/size_t/off_t inconsistencies.  MSVC 2010 can find
-   these if you add /w14267 to the compilation.  We might want to do
-   this by default.  The easiest way to fix this on Windows is to
-   modify msvc.mk to add this to both cl /c lines and run
-
-   make 2>&1 | tee build.log
-
- * Deal with portability issues from gcc 4.7.  See portability.patch
-   from debian package.
 
  * Do a Windows 64-bit build.  MSVC 2010 Professional x86_64 verified
    to build and test cleanly in 2.3.1.  Hopefully the next release
    will include 64-bit binary distributions and external libraries.
 
- * Provide an option to copy encryption parameters from another file.
-   This would make it possible to decrypt a file, manually work with
-   it, and then re-encrypt it using the original encryption parameters
-   including a possibly unknown owner password.
-
- * See if I can support the new encryption formats mentioned in the
-   open bug on sourceforge.  Check other sourceforge bugs.
-
-
-Better 64-bit support on Windows
-================================
-
  * Building 64-bit libraries with MSVC works with existing build.sh as
    long as the x86_64 version of the compiler is in the path.
    Currently this generates 32-bit mingw and 64-bit msvc.  We need to
@@ -38,16 +15,13 @@ Better 64-bit support on Windows
    only been verified with MSVC 2010 so far; we still need to get it
    working with 64-bit mingw.
 
- * Get rid of int/size_t/off_t inconsistencies.  MSVC 2010 can find
+ * Provide an option to copy encryption parameters from another file.
-   these if you add /w14267 to the compilation.  We might want to do
+   This would make it possible to decrypt a file, manually work with
-   this by default.  The easiest way to fix this on Windows is to
+   it, and then re-encrypt it using the original encryption parameters
-   modify msvc.mk to add this to both cl /c lines and run
+   including a possibly unknown owner password.
 
-   make 2>&1 | tee build.log
+ * See if I can support the new encryption formats mentioned in the
-
+   open bug on sourceforge.  Check other sourceforge bugs.
-   Then, from emacs, compile with command cat build.log.
-
-   This will probably require ABI changes, but it seems worth doing.
 
 
 General

configure.ac

@@ -37,8 +37,29 @@ if test "$BUILD_INTERNAL_LIBS" = "0"; then
    AC_SEARCH_LIBS(pcre_compile,pcre,,[MISSING_PCRE=1; MISSING_ANY=1])
 fi
 
+AC_SYS_LARGEFILE
+AC_FUNC_FSEEKO
 AC_TYPE_UINT16_T
 AC_TYPE_UINT32_T
+
+AC_MSG_CHECKING(for whether printf supports %ll)
+AC_COMPILE_IFELSE([AC_LANG_SOURCE([[
+#include <stdio.h>
+#include <string.h>
+#include <sys/types.h>
+int
+main()
+{
+    long long a = 160591605916059ll;
+    char t[50];
+    sprintf(t, "%lld", a);
+}
+]])],[qpdf_PRINTF_LL=yes],[qpdf_PRINTF_LL=no])
+AC_MSG_RESULT($qpdf_PRINTF_LL)
+if test "$qpdf_PRINTF_LL" = "yes"; then
+   AC_DEFINE([HAVE_PRINTF_LL], [1], [Whether printf supports %ll])
+fi
+
 AC_CHECK_FUNCS(random)
 
 # Check if LD supports linker scripts, and define conditional

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;
 };
 

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;
 

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();
 

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;
 };
 

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();
 };

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();
 

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();
 

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>);

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);

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,
+		       off_t hint_offset,
-		       int hint_length);
+		       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,
+			off_t hint_offset,
-			int hint_length,
+			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;

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;
 };
 

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&);
 

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);
 

libqpdf/BitWriter.cc

@@ -12,7 +12,7 @@ BitWriter::BitWriter(Pipeline* pl) :
 }
 
 void
-BitWriter::writeBits(unsigned long val, int bits)
+BitWriter::writeBits(unsigned long val, unsigned int bits)
 {
     write_bits(this->ch, this->bit_offset, val, bits, this->pl);
 }

libqpdf/Buffer.cc

@@ -7,12 +7,12 @@ Buffer::Buffer()
     init(0, 0, true);
 }
 
-Buffer::Buffer(unsigned long size)
+Buffer::Buffer(size_t size)
 {
     init(size, 0, true);
 }
 
-Buffer::Buffer(unsigned char* buf, unsigned long size)
+Buffer::Buffer(unsigned char* buf, size_t size)
 {
     init(size, buf, false);
 }
@@ -36,7 +36,7 @@ Buffer::~Buffer()
 }
 
 void
-Buffer::init(unsigned long size, unsigned char* buf, bool own_memory)
+Buffer::init(size_t size, unsigned char* buf, bool own_memory)
 {
     this->own_memory = own_memory;
     this->size = size;
@@ -75,7 +75,7 @@ Buffer::destroy()
     this->buf = 0;
 }
 
-unsigned long
+size_t
 Buffer::getSize() const
 {
     return this->size;

libqpdf/MD5.cc

@@ -308,7 +308,7 @@ void MD5::reset()
 
 void MD5::encodeString(char const* str)
 {
-    unsigned int len = strlen(str);
+    unsigned int len = (unsigned int)strlen(str);
 
     update((unsigned char *)str, len);
     final();
@@ -316,7 +316,7 @@ void MD5::encodeString(char const* str)
 
 void MD5::appendString(char const* input_string)
 {
-    update((unsigned char *)input_string, strlen(input_string));
+    update((unsigned char *)input_string, (unsigned int) strlen(input_string));
 }
 
 void MD5::encodeDataIncrementally(char const* data, int len)
@@ -332,7 +332,7 @@ void MD5::encodeFile(char const *filename, int up_to_size)
 	std::string("MD5: open ") + filename,
 	fopen(filename, "rb"));
 
-    int len;
+    size_t len;
     int so_far = 0;
     int to_try = 1024;
     do
@@ -344,7 +344,7 @@ void MD5::encodeFile(char const *filename, int up_to_size)
 	len = fread(buffer, 1, to_try, file);
 	if (len > 0)
 	{
-	    update(buffer, len);
+	    update(buffer, (unsigned int) len);
 	    so_far += len;
 	    if ((up_to_size >= 0) && (so_far >= up_to_size))
 	    {

libqpdf/PCRE.cc

@@ -166,7 +166,7 @@ PCRE::match(char const* subject, int options, int startoffset, int size)
 {
     if (size == -1)
     {
-	size = strlen(subject);
+	size = (int) strlen(subject);
     }
 
     Match result(this->nbackrefs, subject);

libqpdf/Pl_AES_PDF.cc

@@ -60,9 +60,9 @@ Pl_AES_PDF::useStaticIV()
 }
 
 void
-Pl_AES_PDF::write(unsigned char* data, int len)
+Pl_AES_PDF::write(unsigned char* data, size_t len)
 {
-    unsigned int bytes_left = len;
+    size_t bytes_left = len;
     unsigned char* p = data;
 
     while (bytes_left > 0)
@@ -72,8 +72,8 @@ Pl_AES_PDF::write(unsigned char* data, int len)
 	    flush(false);
 	}
 
-	unsigned int available = this->buf_size - this->offset;
+	size_t available = this->buf_size - this->offset;
-	int bytes = (bytes_left < available ? bytes_left : available);
+	size_t bytes = (bytes_left < available ? bytes_left : available);
 	bytes_left -= bytes;
 	std::memcpy(this->inbuf + this->offset, p, bytes);
 	this->offset += bytes;
@@ -93,7 +93,7 @@ Pl_AES_PDF::finish()
 	// Pad as described in section 3.5.1 of version 1.7 of the PDF
 	// specification, including providing an entire block of padding
 	// if the input was a multiple of 16 bytes.
-	unsigned char pad = this->buf_size - this->offset;
+	unsigned char pad = (unsigned char) (this->buf_size - this->offset);
 	memset(this->inbuf + this->offset, pad, pad);
 	this->offset = this->buf_size;
 	flush(false);

libqpdf/Pl_ASCII85Decoder.cc

@@ -16,13 +16,13 @@ Pl_ASCII85Decoder::~Pl_ASCII85Decoder()
 }
 
 void
-Pl_ASCII85Decoder::write(unsigned char* buf, int len)
+Pl_ASCII85Decoder::write(unsigned char* buf, size_t len)
 {
     if (eod > 1)
     {
 	return;
     }
-    for (int i = 0; i < len; ++i)
+    for (size_t i = 0; i < len; ++i)
     {
 	if (eod > 1)
 	{

libqpdf/Pl_ASCIIHexDecoder.cc

@@ -17,13 +17,13 @@ Pl_ASCIIHexDecoder::~Pl_ASCIIHexDecoder()
 }
 
 void
-Pl_ASCIIHexDecoder::write(unsigned char* buf, int len)
+Pl_ASCIIHexDecoder::write(unsigned char* buf, size_t len)
 {
     if (this->eod)
     {
 	return;
     }
-    for (int i = 0; i < len; ++i)
+    for (size_t i = 0; i < len; ++i)
     {
 	char ch = toupper(buf[i]);
 	switch (ch)

libqpdf/Pl_Buffer.cc

@@ -15,7 +15,7 @@ Pl_Buffer::~Pl_Buffer()
 }
 
 void
-Pl_Buffer::write(unsigned char* buf, int len)
+Pl_Buffer::write(unsigned char* buf, size_t len)
 {
     Buffer* b = new Buffer(len);
     memcpy(b->getBuffer(), buf, len);

libqpdf/Pl_Count.cc

@@ -12,7 +12,7 @@ Pl_Count::~Pl_Count()
 }
 
 void
-Pl_Count::write(unsigned char* buf, int len)
+Pl_Count::write(unsigned char* buf, size_t len)
 {
     if (len)
     {
@@ -28,7 +28,7 @@ Pl_Count::finish()
     getNext()->finish();
 }
 
-int
+off_t
 Pl_Count::getCount() const
 {
     return this->count;

libqpdf/Pl_Discard.cc

@@ -12,7 +12,7 @@ Pl_Discard::~Pl_Discard()
 }
 
 void
-Pl_Discard::write(unsigned char* buf, int len)
+Pl_Discard::write(unsigned char* buf, size_t len)
 {
 }
 

libqpdf/Pl_Flate.cc

@@ -40,7 +40,7 @@ Pl_Flate::~Pl_Flate()
 }
 
 void
-Pl_Flate::write(unsigned char* data, int len)
+Pl_Flate::write(unsigned char* data, size_t len)
 {
     if (this->outbuf == 0)
     {
@@ -48,7 +48,19 @@ Pl_Flate::write(unsigned char* data, int len)
 	    this->identifier +
 	    ": Pl_Flate: write() called after finish() called");
     }
-    handleData(data, len, Z_NO_FLUSH);
+
+    // Write in chunks in case len is too big to fit in an int.
+    // Assume int is at least 32 bits.
+    static size_t const max_bytes = 1 << 30;
+    size_t bytes_left = len;
+    unsigned char* buf = data;
+    while (bytes_left > 0)
+    {
+	size_t bytes = (bytes_left >= max_bytes ? max_bytes : bytes_left);
+        handleData(buf, (int)bytes, Z_NO_FLUSH);
+	bytes_left -= bytes;
+        buf += bytes;
+    }
 }
 
 void

libqpdf/Pl_LZWDecoder.cc

@@ -25,9 +25,9 @@ Pl_LZWDecoder::~Pl_LZWDecoder()
 }
 
 void
-Pl_LZWDecoder::write(unsigned char* bytes, int len)
+Pl_LZWDecoder::write(unsigned char* bytes, size_t len)
 {
-    for (int i = 0; i < len; ++i)
+    for (size_t i = 0; i < len; ++i)
     {
 	this->buf[next++] = bytes[i];
 	if (this->next == 3)
@@ -131,7 +131,7 @@ Pl_LZWDecoder::addToTable(unsigned char next)
 	assert(idx < table.size());
 	Buffer& b = table[idx];
 	last_data = b.getBuffer();
-	last_size = b.getSize();
+	last_size = (unsigned int) b.getSize();
     }
 
     Buffer entry(1 + last_size);
@@ -170,7 +170,7 @@ Pl_LZWDecoder::handleCode(int code)
 	    // be what we read last plus the first character of what
 	    // we're reading now.
 	    unsigned char next = '\0';
-	    unsigned int table_size = table.size();
+	    unsigned int table_size = (unsigned int) table.size();
 	    if (code < 256)
 	    {
 		// just read < 256; last time's next was code
@@ -178,7 +178,7 @@ Pl_LZWDecoder::handleCode(int code)
 	    }
 	    else if (code > 257)
 	    {
-		unsigned int idx = code - 258;
+		size_t idx = code - 258;
 		if (idx > table_size)
 		{
 		    throw std::runtime_error("LZWDecoder: bad code received");

libqpdf/Pl_MD5.cc

@@ -12,14 +12,27 @@ Pl_MD5::~Pl_MD5()
 }
 
 void
-Pl_MD5::write(unsigned char* buf, int len)
+Pl_MD5::write(unsigned char* buf, size_t len)
 {
     if (! this->in_progress)
     {
 	this->md5.reset();
 	this->in_progress = true;
     }
-    this->md5.encodeDataIncrementally((char*) buf, len);
+
+    // Write in chunks in case len is too big to fit in an int.
+    // Assume int is at least 32 bits.
+    static size_t const max_bytes = 1 << 30;
+    size_t bytes_left = len;
+    unsigned char* data = buf;
+    while (bytes_left > 0)
+    {
+	size_t bytes = (bytes_left >= max_bytes ? max_bytes : bytes_left);
+        this->md5.encodeDataIncrementally((char*) data, (int)bytes);
+	bytes_left -= bytes;
+        data += bytes;
+    }
+
     this->getNext()->write(buf, len);
 }
 

libqpdf/Pl_PNGFilter.cc

@@ -29,10 +29,10 @@ Pl_PNGFilter::~Pl_PNGFilter()
 }
 
 void
-Pl_PNGFilter::write(unsigned char* data, int len)
+Pl_PNGFilter::write(unsigned char* data, size_t len)
 {
-    int left = this->incoming - this->pos;
+    size_t left = this->incoming - this->pos;
-    unsigned int offset = 0;
+    size_t offset = 0;
     while (len >= left)
     {
 	// finish off current row

libqpdf/Pl_QPDFTokenizer.cc

@@ -22,7 +22,7 @@ Pl_QPDFTokenizer::~Pl_QPDFTokenizer()
 }
 
 void
-Pl_QPDFTokenizer::writeNext(char const* buf, int len)
+Pl_QPDFTokenizer::writeNext(char const* buf, size_t len)
 {
     if (len)
     {
@@ -159,10 +159,10 @@ Pl_QPDFTokenizer::checkUnread()
 }
 
 void
-Pl_QPDFTokenizer::write(unsigned char* buf, int len)
+Pl_QPDFTokenizer::write(unsigned char* buf, size_t len)
 {
     checkUnread();
-    for (int i = 0; i < len; ++i)
+    for (size_t i = 0; i < len; ++i)
     {
 	processChar(buf[i]);
 	checkUnread();

libqpdf/Pl_RC4.cc

@@ -3,7 +3,7 @@
 
 Pl_RC4::Pl_RC4(char const* identifier, Pipeline* next,
 	       unsigned char const* key_data, int key_len,
-	       int out_bufsize) :
+	       size_t out_bufsize) :
     Pipeline(identifier, next),
     out_bufsize(out_bufsize),
     rc4(key_data, key_len)
@@ -21,7 +21,7 @@ Pl_RC4::~Pl_RC4()
 }
 
 void
-Pl_RC4::write(unsigned char* data, int len)
+Pl_RC4::write(unsigned char* data, size_t len)
 {
     if (this->outbuf == 0)
     {
@@ -30,14 +30,15 @@ Pl_RC4::write(unsigned char* data, int len)
 	    ": Pl_RC4: write() called after finish() called");
     }
 
-    int bytes_left = len;
+    size_t bytes_left = len;
     unsigned char* p = data;
 
     while (bytes_left > 0)
     {
-	int bytes = (bytes_left < this->out_bufsize ? bytes_left : out_bufsize);
+	size_t bytes =
+            (bytes_left < this->out_bufsize ? bytes_left : out_bufsize);
 	bytes_left -= bytes;
-	rc4.process(p, bytes, outbuf);
+	rc4.process(p, (int)bytes, outbuf);
 	p += bytes;
 	getNext()->write(outbuf, bytes);
     }

libqpdf/Pl_StdioFile.cc

@@ -14,7 +14,7 @@ Pl_StdioFile::~Pl_StdioFile()
 }
 
 void
-Pl_StdioFile::write(unsigned char* buf, int len)
+Pl_StdioFile::write(unsigned char* buf, size_t len)
 {
     size_t so_far = 0;
     while (len > 0)

libqpdf/QPDF.cc

@@ -117,7 +117,7 @@ QPDF::FileInputSource::getName() const
 off_t
 QPDF::FileInputSource::tell()
 {
-    return ftell(this->file);
+    return QUtil::ftell_off_t(this->file);
 }
 
 void
@@ -126,7 +126,7 @@ QPDF::FileInputSource::seek(off_t offset, int whence)
     QUtil::os_wrapper(std::string("seek to ") + this->filename + ", offset " +
 		      QUtil::int_to_string(offset) + " (" +
 		      QUtil::int_to_string(whence) + ")",
-		      fseek(this->file, offset, whence));
+		      QUtil::fseek_off_t(this->file, offset, whence));
 }
 
 void
@@ -136,9 +136,9 @@ QPDF::FileInputSource::rewind()
 }
 
 size_t
-QPDF::FileInputSource::read(char* buffer, int length)
+QPDF::FileInputSource::read(char* buffer, size_t length)
 {
-    this->last_offset = ftell(this->file);
+    this->last_offset = QUtil::ftell_off_t(this->file);
     size_t len = fread(buffer, 1, length, this->file);
     if ((len == 0) && ferror(this->file))
     {
@@ -197,7 +197,7 @@ QPDF::BufferInputSource::seek(off_t offset, int whence)
 	break;
 
       case SEEK_END:
-	this->cur_offset = this->buf->getSize() + offset;
+	this->cur_offset = (off_t)this->buf->getSize() + offset;
 	break;
 
       case SEEK_CUR:
@@ -218,9 +218,9 @@ QPDF::BufferInputSource::rewind()
 }
 
 size_t
-QPDF::BufferInputSource::read(char* buffer, int length)
+QPDF::BufferInputSource::read(char* buffer, size_t length)
 {
-    off_t end_pos = this->buf->getSize();
+    off_t end_pos = (off_t) this->buf->getSize();
     if (this->cur_offset >= end_pos)
     {
 	this->last_offset = end_pos;
@@ -228,7 +228,7 @@ QPDF::BufferInputSource::read(char* buffer, int length)
     }
 
     this->last_offset = this->cur_offset;
-    size_t len = std::min((int)(end_pos - this->cur_offset), length);
+    size_t len = std::min((size_t)(end_pos - this->cur_offset), length);
     memcpy(buffer, buf->getBuffer() + this->cur_offset, len);
     this->cur_offset += len;
     return len;
@@ -432,7 +432,7 @@ QPDF::parse(char const* password)
 	    throw QPDFExc(qpdf_e_damaged_pdf, this->file->getName(), "", 0,
 			  "can't find startxref");
 	}
-	off_t xref_offset = atoi(m2.getMatch(1).c_str());
+	off_t xref_offset = atol(m2.getMatch(1).c_str());
 	read_xref(xref_offset);
     }
     catch (QPDFExc& e)
@@ -878,10 +878,10 @@ QPDF::processXRefStream(off_t xref_offset, QPDFObjectHandle& xref_obj)
 	entry_size += W[i];
     }
 
-    int expected_size = entry_size * num_entries;
+    size_t expected_size = entry_size * num_entries;
 
     PointerHolder<Buffer> bp = xref_obj.getStreamData();
-    int actual_size = bp->getSize();
+    size_t actual_size = bp->getSize();
 
     if (expected_size != actual_size)
     {
@@ -1396,7 +1396,7 @@ QPDF::readObjectInternal(PointerHolder<InputSource> input,
 		// objects since resolving a previously unresolved
 		// indirect object will change file position.
 		off_t stream_offset = input->tell();
-		int length = 0;
+		size_t length = 0;
 
 		try
 		{
@@ -1419,7 +1419,7 @@ QPDF::readObjectInternal(PointerHolder<InputSource> input,
 		    }
 
 		    length = length_obj.getIntValue();
-		    input->seek(stream_offset + length, SEEK_SET);
+		    input->seek(stream_offset + (off_t)length, SEEK_SET);
 		    if (! (readToken(input) ==
 			   QPDFTokenizer::Token(
 			       QPDFTokenizer::tt_word, "endstream")))
@@ -1457,7 +1457,7 @@ QPDF::readObjectInternal(PointerHolder<InputSource> input,
     return object;
 }
 
-int
+size_t
 QPDF::recoverStreamLength(PointerHolder<InputSource> input,
 			  int objid, int generation, off_t stream_offset)
 {
@@ -1474,7 +1474,7 @@ QPDF::recoverStreamLength(PointerHolder<InputSource> input,
     input->seek(stream_offset, SEEK_SET);
     std::string last_line;
     off_t last_line_offset = 0;
-    int length = 0;
+    size_t length = 0;
     while (input->tell() < eof)
     {
 	std::string line = input->readLine();

libqpdf/QPDFObjectHandle.cc

@@ -604,7 +604,7 @@ QPDFObjectHandle::newDictionary(
 QPDFObjectHandle
 QPDFObjectHandle::newStream(QPDF* qpdf, int objid, int generation,
 			    QPDFObjectHandle stream_dict,
-			    off_t offset, int length)
+			    off_t offset, size_t length)
 {
     return QPDFObjectHandle(new QPDF_Stream(
 				qpdf, objid, generation,

libqpdf/QPDFTokenizer.cc

@@ -175,7 +175,7 @@ QPDFTokenizer::presentCharacter(char ch)
 	    string_ignoring_newline = false;
 	}
 
-	unsigned int bs_num_count = strlen(bs_num_register);
+	size_t bs_num_count = strlen(bs_num_register);
 	bool ch_is_octal = ((ch >= '0') && (ch <= '7'));
 	if ((bs_num_count == 3) || ((bs_num_count > 0) && (! ch_is_octal)))
 	{

libqpdf/QPDFWriter.cc

@@ -651,7 +651,7 @@ QPDFWriter::popPipelineStack(PointerHolder<Buffer>* bp)
 }
 
 void
-QPDFWriter::adjustAESStreamLength(unsigned long& length)
+QPDFWriter::adjustAESStreamLength(size_t& length)
 {
     if (this->encrypted && (! this->cur_data_key.empty()) &&
 	this->encrypt_use_aes)
@@ -679,7 +679,7 @@ QPDFWriter::pushEncryptionFilter()
 	{
 	    p = new Pl_RC4("rc4 stream encryption", this->pipeline,
 			   (unsigned char*) this->cur_data_key.c_str(),
-			   this->cur_data_key.length());
+			   (int)this->cur_data_key.length());
 	}
 	pushPipeline(p);
     }
@@ -879,9 +879,9 @@ QPDFWriter::writeTrailer(trailer_e which, int size, bool xref_stream, int prev)
 		if (which == t_lin_first)
 		{
 		    writeString(" /Prev ");
-		    int pos = this->pipeline->getCount();
+		    off_t pos = this->pipeline->getCount();
 		    writeString(QUtil::int_to_string(prev));
-		    int nspaces = pos + 11 - this->pipeline->getCount();
+		    int nspaces = (int)(pos - this->pipeline->getCount() + 11);
 		    assert(nspaces >= 0);
 		    writePad(nspaces);
 		}
@@ -926,7 +926,8 @@ QPDFWriter::unparseObject(QPDFObjectHandle object, int level,
 
 void
 QPDFWriter::unparseObject(QPDFObjectHandle object, int level,
-			  unsigned int flags, int stream_length, bool compress)
+			  unsigned int flags, size_t stream_length,
+                          bool compress)
 {
     unsigned int child_flags = flags & ~f_stream;
 
@@ -1137,16 +1138,16 @@ QPDFWriter::unparseObject(QPDFObjectHandle object, int level,
 		Buffer* buf = bufpl.getBuffer();
 		val = QPDF_String(
 		    std::string((char*)buf->getBuffer(),
-				(size_t)buf->getSize())).unparse(true);
+				buf->getSize())).unparse(true);
 		delete buf;
 	    }
 	    else
 	    {
 		char* tmp = QUtil::copy_string(val);
-		unsigned int vlen = val.length();
+		size_t vlen = val.length();
 		RC4 rc4((unsigned char const*)this->cur_data_key.c_str(),
-			this->cur_data_key.length());
+			(int)this->cur_data_key.length());
-		rc4.process((unsigned char*)tmp, vlen);
+		rc4.process((unsigned char*)tmp, (int)vlen);
 		val = QPDF_String(std::string(tmp, vlen)).unparse();
 		delete [] tmp;
 	    }
@@ -1164,7 +1165,7 @@ QPDFWriter::unparseObject(QPDFObjectHandle object, int level,
 }
 
 void
-QPDFWriter::writeObjectStreamOffsets(std::vector<int>& offsets,
+QPDFWriter::writeObjectStreamOffsets(std::vector<off_t>& offsets,
 				     int first_obj)
 {
     for (unsigned int i = 0; i < offsets.size(); ++i)
@@ -1190,8 +1191,8 @@ QPDFWriter::writeObjectStream(QPDFObjectHandle object)
     int old_id = object.getObjectID();
     int new_id = obj_renumber[old_id];
 
-    std::vector<int> offsets;
+    std::vector<off_t> offsets;
-    int first = 0;
+    off_t first = 0;
 
     // Generate stream itself.  We have to do this in two passes so we
     // can calculate offsets in the first pass.
@@ -1209,7 +1210,7 @@ QPDFWriter::writeObjectStream(QPDFObjectHandle object)
 	    // Adjust offsets to skip over comment before first object
 
 	    first = offsets[0];
-	    for (std::vector<int>::iterator iter = offsets.begin();
+	    for (std::vector<off_t>::iterator iter = offsets.begin();
 		 iter != offsets.end(); ++iter)
 	    {
 		*iter -= first;
@@ -1280,7 +1281,7 @@ QPDFWriter::writeObjectStream(QPDFObjectHandle object)
     writeStringQDF("\n ");
     writeString(" /Type /ObjStm");
     writeStringQDF("\n ");
-    unsigned long length = stream_buffer->getSize();
+    size_t length = stream_buffer->getSize();
     adjustAESStreamLength(length);
     writeString(" /Length " + QUtil::int_to_string(length));
     writeStringQDF("\n ");
@@ -1523,8 +1524,9 @@ QPDFWriter::generateObjectStreams()
     // This code doesn't do anything with /Extends.
 
     std::vector<int> const& eligible = this->pdf.getCompressibleObjects();
-    unsigned int n_object_streams = (eligible.size() + 99) / 100;
+    unsigned int n_object_streams =
-    unsigned int n_per = eligible.size() / n_object_streams;
+        (unsigned int)((eligible.size() + 99) / 100);
+    unsigned int n_per = (unsigned int)(eligible.size() / n_object_streams);
     if (n_per * n_object_streams < eligible.size())
     {
 	++n_per;
@@ -1777,7 +1779,7 @@ QPDFWriter::writeHintStream(int hint_id)
     openObject(hint_id);
     setDataKey(hint_id);
 
-    unsigned long hlen = hint_buffer->getSize();
+    size_t hlen = hint_buffer->getSize();
 
     writeString("<< /Filter /FlateDecode /S ");
     writeString(QUtil::int_to_string(S));
@@ -1817,13 +1819,13 @@ QPDFWriter::writeXRefTable(trailer_e which, int first, int last, int size)
 int
 QPDFWriter::writeXRefTable(trailer_e which, int first, int last, int size,
 			   int prev, bool suppress_offsets,
-			   int hint_id, int hint_offset, int hint_length)
+			   int hint_id, off_t hint_offset, off_t hint_length)
 {
     writeString("xref\n");
     writeString(QUtil::int_to_string(first));
     writeString(" ");
     writeString(QUtil::int_to_string(last - first + 1));
-    int space_before_zero = this->pipeline->getCount();
+    off_t space_before_zero = this->pipeline->getCount();
     writeString("\n");
     for (int i = first; i <= last; ++i)
     {
@@ -1865,10 +1867,10 @@ int
 QPDFWriter::writeXRefStream(int xref_id, int max_id, int max_offset,
 			    trailer_e which, int first, int last, int size,
 			    int prev, int hint_id,
-			    int hint_offset, int hint_length,
+			    off_t hint_offset, off_t hint_length,
 			    bool skip_compression)
 {
-    int xref_offset = this->pipeline->getCount();
+    off_t xref_offset = this->pipeline->getCount();
     int space_before_zero = xref_offset - 1;
 
     // field 1 contains offsets and object stream identifiers
@@ -2021,7 +2023,8 @@ QPDFWriter::writeLinearized()
     //
 
     // Second half objects
-    int second_half_uncompressed = part7.size() + part8.size() + part9.size();
+    int second_half_uncompressed =
+        (int)(part7.size() + part8.size() + part9.size());
     int second_half_first_obj = 1;
     int after_second_half = 1 + second_half_uncompressed;
     this->next_objid = after_second_half;
@@ -2077,13 +2080,13 @@ QPDFWriter::writeLinearized()
 
     int part4_end_marker = part4.back().getObjectID();
     int part6_end_marker = part6.back().getObjectID();
-    int space_before_zero = 0;
+    off_t space_before_zero = 0;
-    int file_size = 0;
+    off_t file_size = 0;
-    int part6_end_offset = 0;
+    off_t part6_end_offset = 0;
-    int first_half_max_obj_offset = 0;
+    off_t first_half_max_obj_offset = 0;
-    int second_xref_offset = 0;
+    off_t second_xref_offset = 0;
-    int first_xref_end = 0;
+    off_t first_xref_end = 0;
-    int second_xref_end = 0;
+    off_t second_xref_end = 0;
 
     this->next_objid = part4_first_obj;
     enqueuePart(part4);
@@ -2097,7 +2100,7 @@ QPDFWriter::writeLinearized()
     enqueuePart(part9);
     assert(this->next_objid == after_second_half);
 
-    int hint_length = 0;
+    off_t hint_length = 0;
     PointerHolder<Buffer> hint_buffer;
 
     // Write file in two passes.  Part numbers refer to PDF spec 1.4.
@@ -2118,14 +2121,14 @@ QPDFWriter::writeLinearized()
 	// space if all numerical values in the parameter dictionary
 	// are 10 digits long plus a few extra characters for safety.
 
-	int pos = this->pipeline->getCount();
+	off_t pos = this->pipeline->getCount();
 	openObject(lindict_id);
 	writeString("<<");
 	if (pass == 2)
 	{
 	    std::vector<QPDFObjectHandle> const& pages = pdf.getAllPages();
 	    int first_page_object = obj_renumber[pages[0].getObjectID()];
-	    int npages = pages.size();
+	    int npages = (int)pages.size();
 
 	    writeString(" /Linearized 1 /L ");
 	    writeString(QUtil::int_to_string(file_size + hint_length));
@@ -2147,15 +2150,15 @@ QPDFWriter::writeLinearized()
 	writeString(" >>");
 	closeObject(lindict_id);
 	static int const pad = 150;
-	int spaces = (pos + pad - this->pipeline->getCount());
+	int spaces = (pos - this->pipeline->getCount() + pad);
 	assert(spaces >= 0);
 	writePad(spaces);
 	writeString("\n");
 
 	// Part 3: first page cross reference table and trailer.
 
-	int first_xref_offset = this->pipeline->getCount();
+	off_t first_xref_offset = this->pipeline->getCount();
-	int hint_offset = 0;
+	off_t hint_offset = 0;
 	if (pass == 2)
 	{
 	    hint_offset = this->xref[hint_id].getOffset();
@@ -2183,7 +2186,7 @@ QPDFWriter::writeLinearized()
 			    hint_length + second_xref_offset,
 			    hint_id, hint_offset, hint_length,
 			    (pass == 1));
-	    int endpos = this->pipeline->getCount();
+	    off_t endpos = this->pipeline->getCount();
 	    if (pass == 1)
 	    {
 		// Pad so we have enough room for the real xref
@@ -2260,7 +2263,7 @@ QPDFWriter::writeLinearized()
 				t_lin_second, 0, second_half_end,
 				second_trailer_size,
 				0, 0, 0, 0, (pass == 1));
-	    int endpos = this->pipeline->getCount();
+	    off_t endpos = this->pipeline->getCount();
 
 	    if (pass == 1)
 	    {
@@ -2274,14 +2277,14 @@ QPDFWriter::writeLinearized()
 	    else
 	    {
 		// Make the file size the same.
-		int pos = this->pipeline->getCount();
+		off_t pos = this->pipeline->getCount();
 		writePad(second_xref_end + hint_length - 1 - pos);
 		writeString("\n");
 
 		// If this assertion fails, maybe we didn't have
 		// enough padding above.
 		assert(this->pipeline->getCount() ==
-		       second_xref_end + hint_length);
+		       (off_t)(second_xref_end + hint_length));
 	    }
 	}
 	else
@@ -2309,7 +2312,7 @@ QPDFWriter::writeLinearized()
 	    activatePipelineStack();
 	    writeHintStream(hint_id);
 	    popPipelineStack(&hint_buffer);
-	    hint_length = hint_buffer->getSize();
+	    hint_length = (off_t)hint_buffer->getSize();
 
 	    // Restore hint offset
 	    this->xref[hint_id] = QPDFXRefEntry(1, hint_offset, 0);

libqpdf/QPDFXRefEntry.cc

@@ -9,7 +9,7 @@ QPDFXRefEntry::QPDFXRefEntry() :
 {
 }
 
-QPDFXRefEntry::QPDFXRefEntry(int type, int field1, int field2) :
+QPDFXRefEntry::QPDFXRefEntry(int type, off_t field1, int field2) :
     type(type),
     field1(field1),
     field2(field2)
@@ -27,7 +27,7 @@ QPDFXRefEntry::getType() const
     return this->type;
 }
 
-int
+off_t
 QPDFXRefEntry::getOffset() const
 {
     if (this->type != 1)
@@ -46,7 +46,7 @@ QPDFXRefEntry::getObjStreamNumber() const
 	throw std::logic_error(
 	    "getObjStreamNumber called for xref entry of type != 2");
     }
-    return this->field1;
+    return (int) this->field1;
 }
 
 int

libqpdf/QPDF_Array.cc

@@ -37,7 +37,7 @@ QPDF_Array::unparse()
 int
 QPDF_Array::getNItems() const
 {
-    return this->items.size();
+    return (int)this->items.size();
 }
 
 QPDFObjectHandle

libqpdf/QPDF_Stream.cc

@@ -22,7 +22,7 @@ std::map<std::string, std::string> QPDF_Stream::filter_abbreviations;
 
 QPDF_Stream::QPDF_Stream(QPDF* qpdf, int objid, int generation,
 			 QPDFObjectHandle stream_dict,
-			 off_t offset, int length) :
+			 off_t offset, size_t length) :
     qpdf(qpdf),
     objid(objid),
     generation(generation),
@@ -379,8 +379,8 @@ QPDF_Stream::pipeStreamData(Pipeline* pipeline, bool filter,
 	Pl_Count count("stream provider count", pipeline);
 	this->stream_provider->provideStreamData(
 	    this->objid, this->generation, &count);
-	size_t actual_length = count.getCount();
+	off_t actual_length = count.getCount();
-	size_t desired_length =
+	off_t desired_length =
 	    this->stream_dict.getKey("/Length").getIntValue();
 	if (actual_length == desired_length)
 	{
@@ -446,5 +446,5 @@ QPDF_Stream::replaceFilterData(QPDFObjectHandle const& filter,
     this->stream_dict.replaceOrRemoveKey("/Filter", filter);
     this->stream_dict.replaceOrRemoveKey("/DecodeParms", decode_parms);
     this->stream_dict.replaceKey("/Length",
-				 QPDFObjectHandle::newInteger(length));
+				 QPDFObjectHandle::newInteger((int)length));
 }

libqpdf/QPDF_String.cc

@@ -157,7 +157,7 @@ std::string
 QPDF_String::getUTF8Val() const
 {
     std::string result;
-    unsigned int len = this->val.length();
+    size_t len = this->val.length();
     if ((len >= 2) && (len % 2 == 0) &&
 	(this->val[0] == '\xfe') && (this->val[1] == '\xff'))
     {

libqpdf/QPDF_encryption.cc

@@ -29,7 +29,7 @@ static unsigned int const key_bytes = 32;
 void
 pad_or_truncate_password(std::string const& password, char k1[key_bytes])
 {
-    int password_bytes = std::min((size_t) key_bytes, password.length());
+    int password_bytes = std::min(key_bytes, (unsigned int)password.length());
     int pad_bytes = key_bytes - password_bytes;
     memcpy(k1, password.c_str(), password_bytes);
     memcpy(k1 + password_bytes, padding_string, pad_bytes);
@@ -121,7 +121,7 @@ QPDF::compute_data_key(std::string const& encryption_key,
     }
 
     MD5 md5;
-    md5.encodeDataIncrementally(result.c_str(), result.length());
+    md5.encodeDataIncrementally(result.c_str(), (int)result.length());
     MD5::Digest digest;
     md5.digest(digest);
     return std::string((char*) digest,
@@ -144,7 +144,7 @@ QPDF::compute_encryption_key(
     pbytes[2] = (char) ((data.P >> 16) & 0xff);
     pbytes[3] = (char) ((data.P >> 24) & 0xff);
     md5.encodeDataIncrementally(pbytes, 4);
-    md5.encodeDataIncrementally(data.id1.c_str(), data.id1.length());
+    md5.encodeDataIncrementally(data.id1.c_str(), (int)data.id1.length());
     if ((data.R >= 4) && (! data.encrypt_metadata))
     {
 	char bytes[4];
@@ -218,7 +218,7 @@ compute_U_value_R3(std::string const& user_password,
     MD5 md5;
     md5.encodeDataIncrementally(
 	pad_or_truncate_password("").c_str(), key_bytes);
-    md5.encodeDataIncrementally(data.id1.c_str(), data.id1.length());
+    md5.encodeDataIncrementally(data.id1.c_str(), (int)data.id1.length());
     MD5::Digest digest;
     md5.digest(digest);
     iterate_rc4(digest, sizeof(MD5::Digest),
@@ -583,16 +583,16 @@ QPDF::decryptString(std::string& str, int objid, int generation)
 	    pl.write((unsigned char*)str.c_str(), str.length());
 	    pl.finish();
 	    PointerHolder<Buffer> buf = bufpl.getBuffer();
-	    str = std::string((char*)buf->getBuffer(), (size_t)buf->getSize());
+	    str = std::string((char*)buf->getBuffer(), buf->getSize());
 	}
 	else
 	{
 	    QTC::TC("qpdf", "QPDF_encryption rc4 decode string");
-	    unsigned int vlen = str.length();
+	    unsigned int vlen = (int)str.length();
 	    // Using PointerHolder guarantees that tmp will
 	    // be freed even if rc4.process throws an exception.
 	    PointerHolder<char> tmp(true, QUtil::copy_string(str));
-	    RC4 rc4((unsigned char const*)key.c_str(), key.length());
+	    RC4 rc4((unsigned char const*)key.c_str(), (int)key.length());
 	    rc4.process((unsigned char*)tmp.getPointer(), vlen);
 	    str = std::string(tmp.getPointer(), vlen);
 	}
@@ -704,7 +704,7 @@ QPDF::decryptStream(Pipeline*& pipeline, int objid, int generation,
     {
 	QTC::TC("qpdf", "QPDF_encryption rc4 decode stream");
 	pipeline = new Pl_RC4("RC4 stream decryption", pipeline,
-			      (unsigned char*) key.c_str(), key.length());
+			      (unsigned char*) key.c_str(), (int)key.length());
     }
     heap.push_back(pipeline);
 }

libqpdf/QPDF_linearization.cc

@@ -289,7 +289,7 @@ QPDF::readLinearizationData()
     PointerHolder<Buffer> hbp = pb.getBuffer();
     Buffer* hb = hbp.getPointer();
     unsigned char const* h_buf = hb->getBuffer();
-    int h_size = hb->getSize();
+    int h_size = (int)hb->getSize();
 
     readHPageOffset(BitStream(h_buf, h_size));
 
@@ -345,7 +345,7 @@ QPDF::readHintStream(Pipeline& pl, off_t offset, size_t length)
     {
 	QTC::TC("qpdf", "QPDF hint table length direct");
     }
-    off_t computed_end = offset + length;
+    off_t computed_end = offset + (off_t)length;
     if ((computed_end < min_end_offset) ||
 	(computed_end > max_end_offset))
     {
@@ -488,7 +488,7 @@ QPDF::checkLinearizationInternal()
     }
 
     // N: number of pages
-    int npages = pages.size();
+    int npages = (int)pages.size();
     if (p.npages != npages)
     {
 	// Not tested in the test suite
@@ -736,7 +736,7 @@ QPDF::checkHPageOffset(std::list<std::string>& errors,
     // under a page's /Resources dictionary in with shared objects
     // even when they are private.
 
-    unsigned int npages = pages.size();
+    unsigned int npages = (unsigned int)pages.size();
     int table_offset = adjusted_offset(
 	this->page_offset_hints.first_page_offset);
     ObjGen first_page_og(pages[0].getObjectID(), pages[0].getGeneration());
@@ -1425,7 +1425,7 @@ QPDF::calculateLinearizationData(std::map<int, int> const& object_stream_data)
 	    pages.push_back(getUncompressedObject(*iter, object_stream_data));
 	}
     }
-    unsigned int npages = pages.size();
+    unsigned int npages = (unsigned int)pages.size();
 
     // We will be initializing some values of the computed hint
     // tables.  Specifically, we can initialize any items that deal
@@ -1495,7 +1495,7 @@ QPDF::calculateLinearizationData(std::map<int, int> const& object_stream_data)
     // in garbage values for all the shared object identifiers on the
     // first page.
 
-    this->c_page_offset_data.entries[0].nobjects = this->part6.size();
+    this->c_page_offset_data.entries[0].nobjects = (int)this->part6.size();
 
     // Part 7: other pages' private objects
 
@@ -1646,9 +1646,11 @@ QPDF::calculateLinearizationData(std::map<int, int> const& object_stream_data)
 
     // Make sure we got everything exactly once.
 
-    unsigned int num_placed = this->part4.size() + this->part6.size() +
+    unsigned int num_placed =
-	this->part7.size() + this->part8.size() + this->part9.size();
+        (unsigned int)(this->part4.size() + this->part6.size() +
-    unsigned int num_wanted = this->object_to_obj_users.size();
+                       this->part7.size() + this->part8.size() +
+                       this->part9.size());
+    unsigned int num_wanted = (unsigned int)this->object_to_obj_users.size();
     if (num_placed != num_wanted)
     {
 	throw std::logic_error(
@@ -1672,10 +1674,11 @@ QPDF::calculateLinearizationData(std::map<int, int> const& object_stream_data)
     // can map from object number only without regards to generation.
     std::map<int, int> obj_to_index;
 
-    this->c_shared_object_data.nshared_first_page = this->part6.size();
+    this->c_shared_object_data.nshared_first_page =
+        (unsigned int)this->part6.size();
     this->c_shared_object_data.nshared_total =
 	this->c_shared_object_data.nshared_first_page +
-	this->part8.size();
+	(unsigned int) this->part8.size();
 
     std::vector<CHSharedObjectEntry>& shared =
 	this->c_shared_object_data.entries;
@@ -1684,7 +1687,7 @@ QPDF::calculateLinearizationData(std::map<int, int> const& object_stream_data)
     {
 	QPDFObjectHandle& oh = *iter;
 	int obj = oh.getObjectID();
-	obj_to_index[obj] = shared.size();
+	obj_to_index[obj] = (int)shared.size();
 	shared.push_back(CHSharedObjectEntry(obj));
     }
     QTC::TC("qpdf", "QPDF lin part 8 empty", this->part8.empty() ? 1 : 0);
@@ -1698,7 +1701,7 @@ QPDF::calculateLinearizationData(std::map<int, int> const& object_stream_data)
 	{
 	    QPDFObjectHandle& oh = *iter;
 	    int obj = oh.getObjectID();
-	    obj_to_index[obj] = shared.size();
+	    obj_to_index[obj] = (int)shared.size();
 	    shared.push_back(CHSharedObjectEntry(obj));
 	}
     }
@@ -1814,7 +1817,7 @@ QPDF::calculateHPageOffset(
     // values.
 
     std::vector<QPDFObjectHandle> const& pages = getAllPages();
-    unsigned int npages = pages.size();
+    unsigned int npages = (unsigned int)pages.size();
     CHPageOffset& cph = this->c_page_offset_data;
     std::vector<CHPageOffsetEntry>& cphe = cph.entries;
 
@@ -2019,7 +2022,7 @@ QPDF::writeHPageOffset(BitWriter& w)
     w.writeBits(t.nbits_shared_numerator, 16);		    // 12
     w.writeBits(t.shared_denominator, 16);		    // 13
 
-    unsigned int nitems = getAllPages().size();
+    unsigned int nitems = (unsigned int)getAllPages().size();
     std::vector<HPageOffsetEntry>& entries = t.entries;
 
     write_vector_int(w, nitems, entries,
@@ -2110,12 +2113,12 @@ QPDF::generateHintStream(std::map<int, QPDFXRefEntry> const& xref,
     BitWriter w(&c);
 
     writeHPageOffset(w);
-    S = c.getCount();
+    S = (int)c.getCount();
     writeHSharedObject(w);
     O = 0;
     if (this->outline_hints.nobjects > 0)
     {
-	O = c.getCount();
+	O = (int)c.getCount();
 	writeHGeneric(w, this->outline_hints);
     }
     c.finish();

libqpdf/QUtil.cc

@@ -1,4 +1,5 @@
 #include <qpdf/QUtil.hh>
+#include <qpdf/qpdf-config.h>
 #include <stdio.h>
 #include <errno.h>
 #include <ctype.h>
@@ -14,7 +15,7 @@
 #endif
 
 std::string
-QUtil::int_to_string(int num, int fullpad)
+QUtil::int_to_string(long long num, int fullpad)
 {
     // This routine will need to be recompiled if an int can be longer than
     // 49 digits.
@@ -28,14 +29,20 @@ QUtil::int_to_string(int num, int fullpad)
 			       "limit");
     }
 
+#ifdef HAVE_PRINTF_LL
+# define PRINTF_LL "ll"
+#else
+# define PRINTF_LL "l"
+#endif
     if (fullpad)
     {
-	sprintf(t, "%0*d", fullpad, num);
+	sprintf(t, "%0*" PRINTF_LL "d", fullpad, num);
     }
     else
     {
-	sprintf(t, "%d", num);
+	sprintf(t, "%" PRINTF_LL "d", num);
     }
+#undef PRINTF_LL
 
     return std::string(t);
 }
@@ -101,6 +108,26 @@ QUtil::fopen_wrapper(std::string const& description, FILE* f)
     return f;
 }
 
+int
+QUtil::fseek_off_t(FILE* stream, off_t offset, int whence)
+{
+#if HAVE_FSEEKO
+    return fseeko(stream, offset, whence);
+#else
+    return fseek(stream, offset, whence);
+#endif
+}
+
+off_t
+QUtil::ftell_off_t(FILE* stream)
+{
+#if HAVE_FSEEKO
+    return ftello(stream);
+#else
+    return ftell(stream);
+#endif
+}
+
 char*
 QUtil::copy_string(std::string const& str)
 {

libqpdf/RC4.cc

@@ -15,7 +15,7 @@ RC4::RC4(unsigned char const* key_data, int key_len)
 {
     if (key_len == -1)
     {
-	key_len = strlen((char*)key_data);
+	key_len = (int)strlen((char*)key_data);
     }
 
     for (int i = 0; i < 256; ++i)

libqpdf/bits.icc

@@ -95,7 +95,7 @@ read_bits(unsigned char const*& p, unsigned int& bit_offset,
 #ifdef BITS_WRITE
 static void
 write_bits(unsigned char& ch, unsigned int& bit_offset,
-	   unsigned long val, unsigned bits, Pipeline* pipeline)
+	   unsigned long val, unsigned int bits, Pipeline* pipeline)
 {
     if (bits > 32)
     {

libqpdf/qpdf-c.cc

@@ -462,10 +462,10 @@ static void qpdf_get_buffer_internal(qpdf_data qpdf)
     }
 }
 
-unsigned long qpdf_get_buffer_length(qpdf_data qpdf)
+size_t qpdf_get_buffer_length(qpdf_data qpdf)
 {
     qpdf_get_buffer_internal(qpdf);
-    unsigned long result = 0L;
+    size_t result = 0;
     if (qpdf->output_buffer)
     {
 	result = qpdf->output_buffer->getSize();

libqpdf/qpdf/BitWriter.hh

@@ -15,7 +15,7 @@ class BitWriter
     QPDF_DLL
     BitWriter(Pipeline* pl);
     QPDF_DLL
-    void writeBits(unsigned long val, int bits);
+    void writeBits(unsigned long val, unsigned int bits);
     // Force any partial byte to be written to the pipeline.
     QPDF_DLL
     void flush();

libqpdf/qpdf/Pl_AES_PDF.hh

@@ -22,7 +22,7 @@ class Pl_AES_PDF: public Pipeline
     virtual ~Pl_AES_PDF();
 
     QPDF_DLL
-    virtual void write(unsigned char* data, int len);
+    virtual void write(unsigned char* data, size_t len);
     QPDF_DLL
     virtual void finish();
 
@@ -43,7 +43,7 @@ class Pl_AES_PDF: public Pipeline
     bool encrypt;
     bool cbc_mode;
     bool first;
-    unsigned int offset;
+    size_t offset;              // offset into memory buffer
     unsigned char key[key_size];
     uint32_t rk[key_size + 28];
     unsigned char inbuf[buf_size];

libqpdf/qpdf/Pl_ASCII85Decoder.hh

@@ -11,7 +11,7 @@ class Pl_ASCII85Decoder: public Pipeline
     QPDF_DLL
     virtual ~Pl_ASCII85Decoder();
     QPDF_DLL
-    virtual void write(unsigned char* buf, int len);
+    virtual void write(unsigned char* buf, size_t len);
     QPDF_DLL
     virtual void finish();
 
@@ -19,8 +19,8 @@ class Pl_ASCII85Decoder: public Pipeline
     void flush();
 
     char inbuf[5];
-    int pos;
+    size_t pos;
-    int eod;
+    size_t eod;
 };
 
 #endif // __PL_ASCII85DECODER_HH__

libqpdf/qpdf/Pl_ASCIIHexDecoder.hh

@@ -11,7 +11,7 @@ class Pl_ASCIIHexDecoder: public Pipeline
     QPDF_DLL
     virtual ~Pl_ASCIIHexDecoder();
     QPDF_DLL
-    virtual void write(unsigned char* buf, int len);
+    virtual void write(unsigned char* buf, size_t len);
     QPDF_DLL
     virtual void finish();
 
@@ -19,7 +19,7 @@ class Pl_ASCIIHexDecoder: public Pipeline
     void flush();
 
     char inbuf[3];
-    int pos;
+    size_t pos;
     bool eod;
 };
 

libqpdf/qpdf/Pl_LZWDecoder.hh

@@ -15,7 +15,7 @@ class Pl_LZWDecoder: public Pipeline
     QPDF_DLL
     virtual ~Pl_LZWDecoder();
     QPDF_DLL
-    virtual void write(unsigned char* buf, int len);
+    virtual void write(unsigned char* buf, size_t len);
     QPDF_DLL
     virtual void finish();
 

libqpdf/qpdf/Pl_MD5.hh

@@ -20,7 +20,7 @@ class Pl_MD5: public Pipeline
     QPDF_DLL
     virtual ~Pl_MD5();
     QPDF_DLL
-    virtual void write(unsigned char*, int);
+    virtual void write(unsigned char*, size_t);
     QPDF_DLL
     virtual void finish();
     QPDF_DLL

libqpdf/qpdf/Pl_PNGFilter.hh

@@ -30,7 +30,7 @@ class Pl_PNGFilter: public Pipeline
     virtual ~Pl_PNGFilter();
 
     QPDF_DLL
-    virtual void write(unsigned char* data, int len);
+    virtual void write(unsigned char* data, size_t len);
     QPDF_DLL
     virtual void finish();
 
@@ -45,8 +45,8 @@ class Pl_PNGFilter: public Pipeline
     unsigned char* prev_row;
     unsigned char* buf1;
     unsigned char* buf2;
-    int pos;
+    size_t pos;
-    int incoming;
+    size_t incoming;
 };
 
 #endif // __PL_PNGFILTER_HH__

libqpdf/qpdf/Pl_QPDFTokenizer.hh

@@ -18,13 +18,13 @@ class Pl_QPDFTokenizer: public Pipeline
   public:
     Pl_QPDFTokenizer(char const* identifier, Pipeline* next);
     virtual ~Pl_QPDFTokenizer();
-    virtual void write(unsigned char* buf, int len);
+    virtual void write(unsigned char* buf, size_t len);
     virtual void finish();
 
   private:
     void processChar(char ch);
     void checkUnread();
-    void writeNext(char const*, int len);
+    void writeNext(char const*, size_t len);
     void writeToken(QPDFTokenizer::Token&);
 
     QPDFTokenizer tokenizer;

libqpdf/qpdf/Pl_RC4.hh

@@ -14,18 +14,18 @@ class Pl_RC4: public Pipeline
     QPDF_DLL
     Pl_RC4(char const* identifier, Pipeline* next,
 	   unsigned char const* key_data, int key_len = -1,
-	   int out_bufsize = def_bufsize);
+	   size_t out_bufsize = def_bufsize);
     QPDF_DLL
     virtual ~Pl_RC4();
 
     QPDF_DLL
-    virtual void write(unsigned char* data, int len);
+    virtual void write(unsigned char* data, size_t len);
     QPDF_DLL
     virtual void finish();
 
   private:
     unsigned char* outbuf;
-    int out_bufsize;
+    size_t out_bufsize;
     RC4 rc4;
 };
 

libqpdf/qpdf/QPDF_Stream.hh

@@ -13,7 +13,7 @@ class QPDF_Stream: public QPDFObject
   public:
     QPDF_Stream(QPDF*, int objid, int generation,
 		QPDFObjectHandle stream_dict,
-		off_t offset, int length);
+		off_t offset, size_t length);
     virtual ~QPDF_Stream();
     virtual std::string unparse();
     QPDFObjectHandle getDict() const;
@@ -51,7 +51,7 @@ class QPDF_Stream: public QPDFObject
     int generation;
     QPDFObjectHandle stream_dict;
     off_t offset;
-    int length;
+    size_t length;
     PointerHolder<Buffer> stream_data;
     PointerHolder<QPDFObjectHandle::StreamDataProvider> stream_provider;
 };

libtests/aes.cc

@@ -46,7 +46,7 @@ int main(int argc, char* argv[])
 	usage();
     }
 
-    unsigned int hexkeylen = strlen(hexkey);
+    unsigned int hexkeylen = (unsigned int)strlen(hexkey);
     unsigned int keylen = hexkeylen / 2;
     if (keylen != Pl_AES_PDF::key_size)
     {
@@ -93,7 +93,7 @@ int main(int argc, char* argv[])
     bool done = false;
     while (! done)
     {
-	int len = fread(buf, 1, sizeof(buf), infile);
+	size_t len = fread(buf, 1, sizeof(buf), infile);
 	if (len <= 0)
 	{
 	    done = true;

libtests/ascii85.cc

@@ -15,7 +15,7 @@ int main()
 	bool done = false;
 	while (! done)
 	{
-	    int len = fread(buf, 1, sizeof(buf), stdin);
+	    size_t len = fread(buf, 1, sizeof(buf), stdin);
 	    if (len <= 0)
 	    {
 		done = true;

libtests/bits.cc

@@ -46,7 +46,7 @@ print_buffer(Pl_Buffer* bp)
     bp->finish();
     Buffer* b = bp->getBuffer();
     unsigned char const* p = b->getBuffer();
-    unsigned long l = b->getSize();
+    size_t l = b->getSize();
     for (unsigned long i = 0; i < l; ++i)
     {
 	printf("%02x%s", (unsigned int)(p[i]),

libtests/hex.cc

@@ -15,7 +15,7 @@ int main()
 	bool done = false;
 	while (! done)
 	{
-	    int len = fread(buf, 1, sizeof(buf), stdin);
+	    size_t len = fread(buf, 1, sizeof(buf), stdin);
 	    if (len <= 0)
 	    {
 		done = true;

libtests/lzw.cc

@@ -38,7 +38,7 @@ int main(int argc, char* argv[])
 	bool done = false;
 	while (! done)
 	{
-	    int len = fread(buf, 1, sizeof(buf), infile);
+	    size_t len = fread(buf, 1, sizeof(buf), infile);
 	    if (len <= 0)
 	    {
 		done = true;

libtests/md5.cc

@@ -54,7 +54,7 @@ int main(int, char*[])
 	    bool done = false;
 	    while (! done)
 	    {
-		int len = fread(buf, 1, sizeof(buf), f);
+		size_t len = fread(buf, 1, sizeof(buf), f);
 		if (len <= 0)
 		{
 		    done = true;

libtests/rc4.cc

@@ -17,8 +17,8 @@ int main(int argc, char* argv[])
     char* hexkey = argv[1];
     char* infilename = argv[2];
     char* outfilename = argv[3];
-    int hexkeylen = strlen(hexkey);
+    unsigned int hexkeylen = (unsigned int)strlen(hexkey);
-    int keylen = hexkeylen / 2;
+    unsigned int keylen = hexkeylen / 2;
     unsigned char* key = new unsigned char[keylen + 1];
     key[keylen] = '\0';
 
@@ -56,7 +56,7 @@ int main(int argc, char* argv[])
     bool done = false;
     while (! done)
     {
-	int len = fread(buf, 1, sizeof(buf), infile);
+	size_t len = fread(buf, 1, sizeof(buf), infile);
 	if (len <= 0)
 	{
 	    done = true;

make/msvc.mk

@@ -19,6 +19,16 @@ endef
 CFLAGS := $(filter-out -g,$(CFLAGS))
 CXXFLAGS := $(filter-out -g,$(CXXFLAGS))
 
+# /WX makes all warnings errors.
+CFLAGS += /WX
+CXXFLAGS += /WX
+
+# /w14267 makes warning 4267 a level 1 warning.  This warning reports
+# potential issues between size_t, off_t, and non-compatible integer
+# types.
+CFLAGS += /w14267
+CXXFLAGS += /w14267
+
 clean::
 	$(RM) *.pdb
 
@@ -35,7 +45,7 @@ endef
 #                       1   2
 # Usage: $(call c_compile,src,includes)
 define c_compile
-	cl /nologo /O2 /Zi /Gy /EHsc /MD $(CPPFLAGS) $(CXXFLAGS) \
+	cl /nologo /O2 /Zi /Gy /EHsc /MD $(CPPFLAGS) $(CFLAGS) \
 		$(foreach I,$(2),-I$(I)) \
 		/c $(1) /Fo$(call c_src_to_obj,$(1))
 endef

qpdf/test_driver.cc

@@ -74,7 +74,7 @@ void runtest(int n, char const* filename)
 	FILE* f = QUtil::fopen_wrapper(std::string("open ") + filename,
 				       fopen(filename, "rb"));
 	fseek(f, 0, SEEK_END);
-	size_t size = (size_t) ftell(f);
+	size_t size = (size_t) QUtil::ftell_off_t(f);
 	fseek(f, 0, SEEK_SET);
 	file_buf = PointerHolder<char>(true, new char[size]);
 	char* buf_p = file_buf.getPointer();

zlib-flate/zlib-flate.cc

@@ -70,7 +70,7 @@ int main(int argc, char* argv[])
 	bool done = false;
 	while (! done)
 	{
-	    int len = fread(buf, 1, sizeof(buf), stdin);
+	    size_t len = fread(buf, 1, sizeof(buf), stdin);
 	    if (len <= 0)
 	    {
 		done = true;