Fix object map to use unique file hash instead of pointer values (Issue #125)

CHANGES.md

@@ -9,6 +9,8 @@ v1.5.3 - YYYY-MM-DD
 - Fixed decryption of UTF-16 strings (Issue #42)
 - Fixed decryption of PDF files with large permission values.
 - Fixed support for EncryptMetadata key in the encryption dictionary.
+- Fixed `pdfioObjCopy` and `pdfioPageCopy` to properly identify the source PDF
+  file being used (Issue #125)
 
 
 v1.5.2 - 2025-04-12

doc/pdfio.3

@@ -1,4 +1,4 @@
-.TH pdfio 3 "pdf read/write library" "2025-04-12" "pdf read/write library"
+.TH pdfio 3 "pdf read/write library" "2025-04-24" "pdf read/write library"
 .SH NAME
 pdfio \- pdf read/write library
 .SH Introduction
@@ -325,7 +325,7 @@ where the five arguments to the function are the filename ("myinputfile.pdf"), a
     }
 .fi
 .PP
-The error callback is called for both errors and warnings and accepts the pdfio_file_t pointer, a message string, and the callback pointer value, for example:
+The error callback is called for both errors and warnings and accepts the pdfio_file_t pointer, a message string, and the callback pointer value. It returns true to continue processing the file or false to stop, for example:
 .nf
 
     bool
@@ -335,12 +335,15 @@ The error callback is called for both errors and warnings and accepts the pdfio_
     
       fprintf(stderr, "%s: %s\\n", pdfioFileGetName(pdf), message);
     
-      // Return false to treat warnings as errors
+      // Return true for warning messages (continue) and false for errors (stop)
-      return (false);
+      return (!strncmp(message, "WARNING:", 8));
     }
 .fi
 .PP
 The default error callback (NULL) does the equivalent of the above.
+.PP
+Note: Many errors are unrecoverable, so PDFio ignores the return value from the error callback and always stops processing the PDF file. Warning messages start with the prefix "WARNING:" while errors have no prefix.
+
 .PP
 Each PDF file contains one or more pages. The pdfioFileGetNumPages function returns the number of pages in the file while the pdfioFileGetPage function gets the specified page in the PDF file:
 .nf
@@ -2477,7 +2480,7 @@ PDFIO_ENCRYPTION_RC4_128
 .TP 5
 PDFIO_ENCRYPTION_RC4_40
 .br
-40-bit RC4 encryption (PDF 1.3)
+40-bit RC4 encryption (PDF 1.3, reading only)
 .SS pdfio_filter_e
 Compression/decompression filters for streams
 .TP 5
@@ -2947,6 +2950,8 @@ bool  pdfioArrayRemove (
     size_t n
 );
 .fi
+.PP
+
 .SS pdfioContentClip
 Clip output to the current path.
 .PP
@@ -3083,6 +3088,8 @@ bool  pdfioContentPathEnd (
     pdfio_stream_t *st
 );
 .fi
+.PP
+
 .SS pdfioContentPathLineTo
 Add a straight line to the current path.
 .PP
@@ -3423,6 +3430,8 @@ double  pdfioContentTextMeasure (
 This function measures the given text string "s" and returns its width based
 on "size". The text string must always use the UTF-8 (Unicode) encoding but
 any control characters (such as newlines) are ignored.
+
+
 .SS pdfioContentTextMoveLine
 Move to the next line and offset.
 .PP
@@ -3451,6 +3460,8 @@ bool  pdfioContentTextNewLine (
     pdfio_stream_t *st
 );
 .fi
+.PP
+
 .SS pdfioContentTextNewLineShow
 Move to the next line and show text.
 .PP
@@ -3468,6 +3479,8 @@ This function moves to the next line and then shows some text with optional
 word and character spacing in a PDF content stream. The "unicode" argument
 specifies that the current font maps to full Unicode.  The "s" argument
 specifies a UTF-8 encoded string.
+
+
 .SS pdfioContentTextNewLineShowf
 Show formatted text.
 .PP
@@ -3486,6 +3499,8 @@ This function moves to the next line and shows some formatted text with
 optional word and character spacing in a PDF content stream. The "unicode"
 argument specifies that the current font maps to full Unicode.  The "format"
 argument specifies a UTF-8 encoded \fBprintf\fR-style format string.
+
+
 .SS pdfioContentTextShow
 Show text.
 .PP
@@ -3536,6 +3551,8 @@ bool  pdfioDictClear (
     const char *key
 );
 .fi
+.PP
+
 .SS pdfioDictCopy
 Copy a dictionary to a PDF file.
 .PP
@@ -3608,6 +3625,8 @@ const char * pdfioDictGetKey (
     size_t n
 );
 .fi
+.PP
+
 .SS pdfioDictGetName
 Get a key name value from a dictionary.
 .PP
@@ -3625,6 +3644,8 @@ size_t  pdfioDictGetNumPairs (
     pdfio_dict_t *dict
 );
 .fi
+.PP
+
 .SS pdfioDictGetNumber
 Get a key number value from a dictionary.
 .PP
@@ -3697,6 +3718,8 @@ function "cb":
 
 The iteration continues as long as the callback returns \fBtrue\fR or all keys
 have been iterated.
+
+
 .SS pdfioDictSetArray
 Set a key array in a dictionary.
 .PP
@@ -3852,8 +3875,9 @@ CropBox for pages in the PDF file - if \fBNULL\fR then a default "Universal" siz
 of 8.27x11in (the intersection of US Letter and ISO A4) is used.
 .PP
 The "error_cb" and "error_cbdata" arguments specify an error handler callback
-and its data pointer - if \fBNULL\fR the default error handler is used that
+and its data pointer - if \fBNULL\fR then the default error handler is used that
-writes error messages to \fBstderr\fR.
+writes error messages to \fBstderr\fR.  The error handler callback should return
+\fBtrue\fR to continue writing the PDF file or \fBfalse\fR to stop.
 .SS pdfioFileCreateArrayObj
 Create a new object in a PDF file containing an array.
 .PP
@@ -4011,6 +4035,8 @@ pdfio_obj_t * pdfioFileCreateNameObj (
 .PP
 This function creates a new object with a name value in a PDF file.
 You must call \fIpdfioObjClose\fR to write the object to the file.
+
+
 .SS pdfioFileCreateNumberObj
 Create a new object in a PDF file containing a number.
 .PP
@@ -4023,6 +4049,8 @@ pdfio_obj_t * pdfioFileCreateNumberObj (
 .PP
 This function creates a new object with a number value in a PDF file.
 You must call \fIpdfioObjClose\fR to write the object to the file.
+
+
 .SS pdfioFileCreateObj
 Create a new object in a PDF file.
 .PP
@@ -4071,8 +4099,9 @@ CropBox for pages in the PDF file - if \fBNULL\fR then a default "Universal" siz
 of 8.27x11in (the intersection of US Letter and ISO A4) is used.
 .PP
 The "error_cb" and "error_cbdata" arguments specify an error handler callback
-and its data pointer - if \fBNULL\fR the default error handler is used that
+and its data pointer - if \fBNULL\fR then the default error handler is used that
-writes error messages to \fBstderr\fR.
+writes error messages to \fBstderr\fR.  The error handler callback should return
+\fBtrue\fR to continue writing the PDF file or \fBfalse\fR to stop.
 .PP
 .IP 5
 \fINote\fR: Files created using this API are slightly larger than those
@@ -4101,6 +4130,8 @@ pdfio_obj_t * pdfioFileCreateStringObj (
 .PP
 This function creates a new object with a string value in a PDF file.
 You must call \fIpdfioObjClose\fR to write the object to the file.
+
+
 .SS pdfioFileCreateTemporary
 
 .PP
@@ -4294,8 +4325,18 @@ cancel the open.  If \fBNULL\fR is specified for the callback function and the
 PDF file requires a password, the open will always fail.
 .PP
 The "error_cb" and "error_cbdata" arguments specify an error handler callback
-and its data pointer - if \fBNULL\fR the default error handler is used that
+and its data pointer - if \fBNULL\fR then the default error handler is used that
-writes error messages to \fBstderr\fR.
+writes error messages to \fBstderr\fR.  The error handler callback should return
+\fBtrue\fR to continue reading the PDF file or \fBfalse\fR to stop.
+.PP
+.IP 5
+Note: Error messages starting with "WARNING:" are actually warning
+.IP 5
+messages - the callback should normally return \fBtrue\fR to allow PDFio to
+.IP 5
+try to resolve the issue.  In addition, some errors are unrecoverable and
+.IP 5
+ignore the return value of the error callback.
 .SS pdfioFileSetAuthor
 Set the author for a PDF file.
 .PP
@@ -4474,6 +4515,8 @@ const char * pdfioObjGetName (
     pdfio_obj_t *obj
 );
 .fi
+.PP
+
 .SS pdfioObjGetNumber
 Get the object's number.
 .PP
@@ -4677,8 +4720,9 @@ bool  pdfioStreamPrintf (
 .fi
 .PP
 This function writes a formatted string to a stream.  In addition to the
-standard \fBprintf\fR format characters, you can use "%N" to format a PDF name
+standard \fBprintf\fR format characters, you can use "%H" to format a HTML/XML
-value ("/Name") and "%S" to format a PDF string ("(String)") value.
+string value, "%N" to format a PDF name value ("/Name"), and "%S" to format
+a PDF string ("(String)") value.
 .SS pdfioStreamPutChar
 Write a single character to a stream.
 .PP

doc/pdfio.html

@@ -728,7 +728,7 @@ password_cb(<span class="reserved">void</span> *data, <span class="reserved">con
   <span class="reserved">return</span> (<span class="string">&quot;Password42&quot;</span>);
 }
 </code></pre>
-<p>The error callback is called for both errors and warnings and accepts the <code>pdfio_file_t</code> pointer, a message string, and the callback pointer value, for example:</p>
+<p>The error callback is called for both errors and warnings and accepts the <code>pdfio_file_t</code> pointer, a message string, and the callback pointer value. It returns <code>true</code> to continue processing the file or <code>false</code> to stop, for example:</p>
 <pre><code class="language-c"><span class="reserved">bool</span>
 error_cb(pdfio_file_t *pdf, <span class="reserved">const</span> <span class="reserved">char</span> *message, <span class="reserved">void</span> *data)
 {
@@ -736,11 +736,14 @@ error_cb(pdfio_file_t *pdf, <span class="reserved">const</span> <span class="res
 
   fprintf(stderr, <span class="string">&quot;%s: %s\n&quot;</span>, pdfioFileGetName(pdf), message);
 
-  <span class="comment">// Return false to treat warnings as errors</span>
+  <span class="comment">// Return true for warning messages (continue) and false for errors (stop)</span>
-  <span class="reserved">return</span> (<span class="reserved">false</span>);
+  <span class="reserved">return</span> (!strncmp(message, <span class="string">&quot;WARNING:&quot;</span>, <span class="number">8</span>));
 }
 </code></pre>
 <p>The default error callback (<code>NULL</code>) does the equivalent of the above.</p>
+<blockquote>
+<p>Note: Many errors are unrecoverable, so PDFio ignores the return value from the error callback and always stops processing the PDF file. Warning messages start with the prefix &quot;WARNING:&quot; while errors have no prefix.</p>
+</blockquote>
 <p>Each PDF file contains one or more pages. The <a href="#pdfioFileGetNumPages"><code>pdfioFileGetNumPages</code></a> function returns the number of pages in the file while the <a href="#pdfioFileGetPage"><code>pdfioFileGetPage</code></a> function gets the specified page in the PDF file:</p>
 <pre><code class="language-c">pdfio_file_t *pdf;   <span class="comment">// PDF file</span>
 size_t       i;      <span class="comment">// Looping var</span>
@@ -2746,7 +2749,7 @@ size_t pdfioArrayGetSize(<a href="#pdfio_array_t">pdfio_array_t</a> *a);</p>
 </tbody></table>
 <h4 class="returnvalue">Return Value</h4>
 <p class="description">Value type</p>
-<h3 class="function"><a id="pdfioArrayRemove">pdfioArrayRemove</a></h3>
+<h3 class="function"><span class="info">&#160;PDFio v1.4&#160;</span><a id="pdfioArrayRemove">pdfioArrayRemove</a></h3>
 <p class="description">Remove an array entry.</p>
 <p class="code">
 <span class="reserved">bool</span> pdfioArrayRemove(<a href="#pdfio_array_t">pdfio_array_t</a> *a, size_t n);</p>
@@ -2951,7 +2954,7 @@ using the <a href="#pdfioPageDictAddImage"><code>pdfioPageDictAddImage</code></a
 </tbody></table>
 <h4 class="returnvalue">Return Value</h4>
 <p class="description"><code>true</code> on success, <code>false</code> on failure</p>
-<h3 class="function"><a id="pdfioContentPathEnd">pdfioContentPathEnd</a></h3>
+<h3 class="function"><span class="info">&#160;PDFio v1.1&#160;</span><a id="pdfioContentPathEnd">pdfioContentPathEnd</a></h3>
 <p class="description">Clear the current path.</p>
 <p class="code">
 <span class="reserved">bool</span> pdfioContentPathEnd(<a href="#pdfio_stream_t">pdfio_stream_t</a> *st);</p>
@@ -3441,7 +3444,7 @@ are 0, a solid line is drawn.</p>
 </tbody></table>
 <h4 class="returnvalue">Return Value</h4>
 <p class="description"><code>true</code> on success, <code>false</code> on failure</p>
-<h3 class="function"><a id="pdfioContentTextMeasure">pdfioContentTextMeasure</a></h3>
+<h3 class="function"><span class="info">&#160;PDFio v1.2&#160;</span><a id="pdfioContentTextMeasure">pdfioContentTextMeasure</a></h3>
 <p class="description">Measure a text string and return its width.</p>
 <p class="code">
 <span class="reserved">double</span> pdfioContentTextMeasure(<a href="#pdfio_obj_t">pdfio_obj_t</a> *font, <span class="reserved">const</span> <span class="reserved">char</span> *s, <span class="reserved">double</span> size);</p>
@@ -3459,7 +3462,9 @@ are 0, a solid line is drawn.</p>
 <h4 class="discussion">Discussion</h4>
 <p class="discussion">This function measures the given text string &quot;s&quot; and returns its width based
 on &quot;size&quot;. The text string must always use the UTF-8 (Unicode) encoding but
-any control characters (such as newlines) are ignored.</p>
+any control characters (such as newlines) are ignored.
+
+</p>
 <h3 class="function"><a id="pdfioContentTextMoveLine">pdfioContentTextMoveLine</a></h3>
 <p class="description">Move to the next line and offset.</p>
 <p class="code">
@@ -3490,7 +3495,7 @@ any control characters (such as newlines) are ignored.</p>
 </tbody></table>
 <h4 class="returnvalue">Return Value</h4>
 <p class="description"><code>true</code> on success, <code>false</code> on failure</p>
-<h3 class="function"><a id="pdfioContentTextNewLine">pdfioContentTextNewLine</a></h3>
+<h3 class="function"><span class="info">&#160;PDFio v1.2&#160;</span><a id="pdfioContentTextNewLine">pdfioContentTextNewLine</a></h3>
 <p class="description">Move to the next line.</p>
 <p class="code">
 <span class="reserved">bool</span> pdfioContentTextNewLine(<a href="#pdfio_stream_t">pdfio_stream_t</a> *st);</p>
@@ -3501,7 +3506,7 @@ any control characters (such as newlines) are ignored.</p>
 </tbody></table>
 <h4 class="returnvalue">Return Value</h4>
 <p class="description"><code>true</code> on success, <code>false</code> on failure</p>
-<h3 class="function"><a id="pdfioContentTextNewLineShow">pdfioContentTextNewLineShow</a></h3>
+<h3 class="function"><span class="info">&#160;PDFio v1.2&#160;</span><a id="pdfioContentTextNewLineShow">pdfioContentTextNewLineShow</a></h3>
 <p class="description">Move to the next line and show text.</p>
 <p class="code">
 <span class="reserved">bool</span> pdfioContentTextNewLineShow(<a href="#pdfio_stream_t">pdfio_stream_t</a> *st, <span class="reserved">double</span> ws, <span class="reserved">double</span> cs, <span class="reserved">bool</span> unicode, <span class="reserved">const</span> <span class="reserved">char</span> *s);</p>
@@ -3524,8 +3529,10 @@ any control characters (such as newlines) are ignored.</p>
 <p class="discussion">This function moves to the next line and then shows some text with optional
 word and character spacing in a PDF content stream. The &quot;unicode&quot; argument
 specifies that the current font maps to full Unicode.  The &quot;s&quot; argument
-specifies a UTF-8 encoded string.</p>
+specifies a UTF-8 encoded string.
-<h3 class="function"><a id="pdfioContentTextNewLineShowf">pdfioContentTextNewLineShowf</a></h3>
+
+</p>
+<h3 class="function"><span class="info">&#160;PDFio v1.2&#160;</span><a id="pdfioContentTextNewLineShowf">pdfioContentTextNewLineShowf</a></h3>
 <p class="description">Show formatted text.</p>
 <p class="code">
 <span class="reserved">bool</span> pdfioContentTextNewLineShowf(<a href="#pdfio_stream_t">pdfio_stream_t</a> *st, <span class="reserved">double</span> ws, <span class="reserved">double</span> cs, <span class="reserved">bool</span> unicode, <span class="reserved">const</span> <span class="reserved">char</span> *format, ...);</p>
@@ -3550,7 +3557,9 @@ specifies a UTF-8 encoded string.</p>
 <p class="discussion">This function moves to the next line and shows some formatted text with
 optional word and character spacing in a PDF content stream. The &quot;unicode&quot;
 argument specifies that the current font maps to full Unicode.  The &quot;format&quot;
-argument specifies a UTF-8 encoded <code>printf</code>-style format string.</p>
+argument specifies a UTF-8 encoded <code>printf</code>-style format string.
+
+</p>
 <h3 class="function"><a id="pdfioContentTextShow">pdfioContentTextShow</a></h3>
 <p class="description">Show text.</p>
 <p class="code">
@@ -3613,7 +3622,7 @@ argument specifies an array of UTF-8 encoded strings.</p>
 <p class="discussion">This function shows some formatted text in a PDF content stream. The
 &quot;unicode&quot; argument specifies that the current font maps to full Unicode.
 The &quot;format&quot; argument specifies a UTF-8 encoded <code>printf</code>-style format string.</p>
-<h3 class="function"><a id="pdfioDictClear">pdfioDictClear</a></h3>
+<h3 class="function"><span class="info">&#160;PDFio v1.4&#160;</span><a id="pdfioDictClear">pdfioDictClear</a></h3>
 <p class="description">Remove a key/value pair from a dictionary.</p>
 <p class="code">
 <span class="reserved">bool</span> pdfioDictClear(<a href="#pdfio_dict_t">pdfio_dict_t</a> *dict, <span class="reserved">const</span> <span class="reserved">char</span> *key);</p>
@@ -3717,7 +3726,7 @@ time_t pdfioDictGetDate(<a href="#pdfio_dict_t">pdfio_dict_t</a> *dict, <span cl
 </tbody></table>
 <h4 class="returnvalue">Return Value</h4>
 <p class="description">Value</p>
-<h3 class="function"><a id="pdfioDictGetKey">pdfioDictGetKey</a></h3>
+<h3 class="function"><span class="info">&#160;PDFio v1.4&#160;</span><a id="pdfioDictGetKey">pdfioDictGetKey</a></h3>
 <p class="description">Get the key for the specified pair.</p>
 <p class="code">
 <span class="reserved">const</span> <span class="reserved">char</span> *pdfioDictGetKey(<a href="#pdfio_dict_t">pdfio_dict_t</a> *dict, size_t n);</p>
@@ -3743,7 +3752,7 @@ time_t pdfioDictGetDate(<a href="#pdfio_dict_t">pdfio_dict_t</a> *dict, <span cl
 </tbody></table>
 <h4 class="returnvalue">Return Value</h4>
 <p class="description">Value</p>
-<h3 class="function"><a id="pdfioDictGetNumPairs">pdfioDictGetNumPairs</a></h3>
+<h3 class="function"><span class="info">&#160;PDFio v1.4&#160;</span><a id="pdfioDictGetNumPairs">pdfioDictGetNumPairs</a></h3>
 <p class="description">Get the number of key/value pairs in a dictionary.</p>
 <p class="code">
 size_t pdfioDictGetNumPairs(<a href="#pdfio_dict_t">pdfio_dict_t</a> *dict);</p>
@@ -3821,7 +3830,7 @@ size_t pdfioDictGetNumPairs(<a href="#pdfio_dict_t">pdfio_dict_t</a> *dict);</p>
 </tbody></table>
 <h4 class="returnvalue">Return Value</h4>
 <p class="description">Value type</p>
-<h3 class="function"><a id="pdfioDictIterateKeys">pdfioDictIterateKeys</a></h3>
+<h3 class="function"><span class="info">&#160;PDFio v1.1&#160;</span><a id="pdfioDictIterateKeys">pdfioDictIterateKeys</a></h3>
 <p class="description">Iterate the keys in a dictionary.</p>
 <p class="code">
 <span class="reserved">void</span> pdfioDictIterateKeys(<a href="#pdfio_dict_t">pdfio_dict_t</a> *dict, <a href="#pdfio_dict_cb_t">pdfio_dict_cb_t</a> cb, <span class="reserved">void</span> *cb_data);</p>
@@ -3848,7 +3857,9 @@ my_dict_cb(pdfio_dict_t *dict, const char *key, void *cb_data)
 </pre>
 
 The iteration continues as long as the callback returns <code>true</code> or all keys
-have been iterated.</p>
+have been iterated.
+
+</p>
 <h3 class="function"><a id="pdfioDictSetArray">pdfioDictSetArray</a></h3>
 <p class="description">Set a key array in a dictionary.</p>
 <p class="code">
@@ -4076,8 +4087,9 @@ CropBox for pages in the PDF file - if <code>NULL</code> then a default &quot;Un
 of 8.27x11in (the intersection of US Letter and ISO A4) is used.<br>
 <br>
 The &quot;error_cb&quot; and &quot;error_cbdata&quot; arguments specify an error handler callback
-and its data pointer - if <code>NULL</code> the default error handler is used that
+and its data pointer - if <code>NULL</code> then the default error handler is used that
-writes error messages to <code>stderr</code>.</p>
+writes error messages to <code>stderr</code>.  The error handler callback should return
+<code>true</code> to continue writing the PDF file or <code>false</code> to stop.</p>
 <h3 class="function"><a id="pdfioFileCreateArrayObj">pdfioFileCreateArrayObj</a></h3>
 <p class="description">Create a new object in a PDF file containing an array.</p>
 <p class="code">
@@ -4260,7 +4272,7 @@ image on the page.<br>
 Note: Currently PNG support is limited to grayscale, RGB, or indexed files
 without interlacing or alpha.  Transparency (masking) based on color/index
 is supported.</blockquote>
-<h3 class="function"><a id="pdfioFileCreateNameObj">pdfioFileCreateNameObj</a></h3>
+<h3 class="function"><span class="info">&#160;PDFio v1.4&#160;</span><a id="pdfioFileCreateNameObj">pdfioFileCreateNameObj</a></h3>
 <p class="description">Create a new object in a PDF file containing a name.</p>
 <p class="code">
 <a href="#pdfio_obj_t">pdfio_obj_t</a> *pdfioFileCreateNameObj(<a href="#pdfio_file_t">pdfio_file_t</a> *pdf, <span class="reserved">const</span> <span class="reserved">char</span> *name);</p>
@@ -4275,8 +4287,10 @@ is supported.</blockquote>
 <p class="description">New object</p>
 <h4 class="discussion">Discussion</h4>
 <p class="discussion">This function creates a new object with a name value in a PDF file.
-You must call <a href="#pdfioObjClose"><code>pdfioObjClose</code></a> to write the object to the file.</p>
+You must call <a href="#pdfioObjClose"><code>pdfioObjClose</code></a> to write the object to the file.
-<h3 class="function"><a id="pdfioFileCreateNumberObj">pdfioFileCreateNumberObj</a></h3>
+
+</p>
+<h3 class="function"><span class="info">&#160;PDFio v1.2&#160;</span><a id="pdfioFileCreateNumberObj">pdfioFileCreateNumberObj</a></h3>
 <p class="description">Create a new object in a PDF file containing a number.</p>
 <p class="code">
 <a href="#pdfio_obj_t">pdfio_obj_t</a> *pdfioFileCreateNumberObj(<a href="#pdfio_file_t">pdfio_file_t</a> *pdf, <span class="reserved">double</span> number);</p>
@@ -4291,7 +4305,9 @@ You must call <a href="#pdfioObjClose"><code>pdfioObjClose</code></a> to write t
 <p class="description">New object</p>
 <h4 class="discussion">Discussion</h4>
 <p class="discussion">This function creates a new object with a number value in a PDF file.
-You must call <a href="#pdfioObjClose"><code>pdfioObjClose</code></a> to write the object to the file.</p>
+You must call <a href="#pdfioObjClose"><code>pdfioObjClose</code></a> to write the object to the file.
+
+</p>
 <h3 class="function"><a id="pdfioFileCreateObj">pdfioFileCreateObj</a></h3>
 <p class="description">Create a new object in a PDF file.</p>
 <p class="code">
@@ -4352,8 +4368,9 @@ CropBox for pages in the PDF file - if <code>NULL</code> then a default &quot;Un
 of 8.27x11in (the intersection of US Letter and ISO A4) is used.<br>
 <br>
 The &quot;error_cb&quot; and &quot;error_cbdata&quot; arguments specify an error handler callback
-and its data pointer - if <code>NULL</code> the default error handler is used that
+and its data pointer - if <code>NULL</code> then the default error handler is used that
-writes error messages to <code>stderr</code>.<br>
+writes error messages to <code>stderr</code>.  The error handler callback should return
+<code>true</code> to continue writing the PDF file or <code>false</code> to stop.<br>
 <br>
 </p><blockquote>
 <em>Note</em>: Files created using this API are slightly larger than those
@@ -4372,7 +4389,7 @@ stored as indirect object references.</blockquote>
 </tbody></table>
 <h4 class="returnvalue">Return Value</h4>
 <p class="description">Contents stream</p>
-<h3 class="function"><a id="pdfioFileCreateStringObj">pdfioFileCreateStringObj</a></h3>
+<h3 class="function"><span class="info">&#160;PDFio v1.2&#160;</span><a id="pdfioFileCreateStringObj">pdfioFileCreateStringObj</a></h3>
 <p class="description">Create a new object in a PDF file containing a string.</p>
 <p class="code">
 <a href="#pdfio_obj_t">pdfio_obj_t</a> *pdfioFileCreateStringObj(<a href="#pdfio_file_t">pdfio_file_t</a> *pdf, <span class="reserved">const</span> <span class="reserved">char</span> *string);</p>
@@ -4387,7 +4404,9 @@ stored as indirect object references.</blockquote>
 <p class="description">New object</p>
 <h4 class="discussion">Discussion</h4>
 <p class="discussion">This function creates a new object with a string value in a PDF file.
-You must call <a href="#pdfioObjClose"><code>pdfioObjClose</code></a> to write the object to the file.</p>
+You must call <a href="#pdfioObjClose"><code>pdfioObjClose</code></a> to write the object to the file.
+
+</p>
 <h3 class="function"><a id="pdfioFileCreateTemporary">pdfioFileCreateTemporary</a></h3>
 <p class="description"></p>
 <p class="code">
@@ -4670,8 +4689,15 @@ cancel the open.  If <code>NULL</code> is specified for the callback function an
 PDF file requires a password, the open will always fail.<br>
 <br>
 The &quot;error_cb&quot; and &quot;error_cbdata&quot; arguments specify an error handler callback
-and its data pointer - if <code>NULL</code> the default error handler is used that
+and its data pointer - if <code>NULL</code> then the default error handler is used that
-writes error messages to <code>stderr</code>.</p>
+writes error messages to <code>stderr</code>.  The error handler callback should return
+<code>true</code> to continue reading the PDF file or <code>false</code> to stop.<br>
+<br>
+</p><blockquote>
+Note: Error messages starting with &quot;WARNING:&quot; are actually warning
+messages - the callback should normally return <code>true</code> to allow PDFio to
+try to resolve the issue.  In addition, some errors are unrecoverable and
+ignore the return value of the error callback.</blockquote>
 <h3 class="function"><a id="pdfioFileSetAuthor">pdfioFileSetAuthor</a></h3>
 <p class="description">Set the author for a PDF file.</p>
 <p class="code">
@@ -4892,7 +4918,7 @@ size_t pdfioObjGetLength(<a href="#pdfio_obj_t">pdfio_obj_t</a> *obj);</p>
 </tbody></table>
 <h4 class="returnvalue">Return Value</h4>
 <p class="description">Length in bytes or <code>0</code> for none</p>
-<h3 class="function"><a id="pdfioObjGetName">pdfioObjGetName</a></h3>
+<h3 class="function"><span class="info">&#160;PDFio v1.4&#160;</span><a id="pdfioObjGetName">pdfioObjGetName</a></h3>
 <p class="description">Get the name value associated with an object.</p>
 <p class="code">
 <span class="reserved">const</span> <span class="reserved">char</span> *pdfioObjGetName(<a href="#pdfio_obj_t">pdfio_obj_t</a> *obj);</p>
@@ -5168,8 +5194,9 @@ ssize_t pdfioStreamPeek(<a href="#pdfio_stream_t">pdfio_stream_t</a> *st, <span
 <p class="description"><code>true</code> on success, <code>false</code> on failure</p>
 <h4 class="discussion">Discussion</h4>
 <p class="discussion">This function writes a formatted string to a stream.  In addition to the
-standard <code>printf</code> format characters, you can use &quot;%N&quot; to format a PDF name
+standard <code>printf</code> format characters, you can use &quot;%H&quot; to format a HTML/XML
-value (&quot;/Name&quot;) and &quot;%S&quot; to format a PDF string (&quot;(String)&quot;) value.</p>
+string value, &quot;%N&quot; to format a PDF name value (&quot;/Name&quot;), and &quot;%S&quot; to format
+a PDF string (&quot;(String)&quot;) value.</p>
 <h3 class="function"><a id="pdfioStreamPutChar">pdfioStreamPutChar</a></h3>
 <p class="description">Write a single character to a stream.</p>
 <p class="code">
@@ -5403,7 +5430,7 @@ typedef enum <a href="#pdfio_valtype_e">pdfio_valtype_e</a> pdfio_valtype_t;
 <tr><th>PDFIO_ENCRYPTION_AES_128 </th><td class="description">128-bit AES encryption (PDF 1.6)</td></tr>
 <tr><th>PDFIO_ENCRYPTION_NONE </th><td class="description">No encryption</td></tr>
 <tr><th>PDFIO_ENCRYPTION_RC4_128 </th><td class="description">128-bit RC4 encryption (PDF 1.4)</td></tr>
-<tr><th>PDFIO_ENCRYPTION_RC4_40 </th><td class="description">40-bit RC4 encryption (PDF 1.3)</td></tr>
+<tr><th>PDFIO_ENCRYPTION_RC4_40 </th><td class="description">40-bit RC4 encryption (PDF 1.3, reading only)</td></tr>
 </tbody></table>
 <h3 class="enumeration"><a id="pdfio_filter_e">pdfio_filter_e</a></h3>
 <p class="description">Compression/decompression filters for streams</p>

doc/pdfio.md

@@ -343,8 +343,8 @@ password_cb(void *data, const char *filename)
 ```
 
 The error callback is called for both errors and warnings and accepts the
-`pdfio_file_t` pointer, a message string, and the callback pointer value, for
+`pdfio_file_t` pointer, a message string, and the callback pointer value. It
-example:
+returns `true` to continue processing the file or `false` to stop, for example:
 
 ```c
 bool
@@ -354,13 +354,17 @@ error_cb(pdfio_file_t *pdf, const char *message, void *data)
 
   fprintf(stderr, "%s: %s\n", pdfioFileGetName(pdf), message);
 
-  // Return false to treat warnings as errors
+  // Return true for warning messages (continue) and false for errors (stop)
-  return (false);
+  return (!strncmp(message, "WARNING:", 8));
 }
 ```
 
 The default error callback (`NULL`) does the equivalent of the above.
 
+> Note: Many errors are unrecoverable, so PDFio ignores the return value from
+> the error callback and always stops processing the PDF file.  Warning messages
+> start with the prefix "WARNING:" while errors have no prefix.
+
 Each PDF file contains one or more pages.  The [`pdfioFileGetNumPages`](@@)
 function returns the number of pages in the file while the
 [`pdfioFileGetPage`](@@) function gets the specified page in the PDF file:

examples/pdfiomerge.c

@@ -8,8 +8,8 @@
 //
 // Usage:
 //
-//   ./pdfmerge [-o OUTPUT.pdf] INPUT.pdf [... INPUT.pdf]
+//   ./pdfiomerge [-o OUTPUT.pdf] INPUT.pdf [... INPUT.pdf]
-//   ./pdfmerge INPUT.pdf [... INPUT.pdf] >OUTPUT.pdf
+//   ./pdfiomerge INPUT.pdf [... INPUT.pdf] >OUTPUT.pdf
 //
 
 #include <pdfio.h>
@@ -47,7 +47,7 @@ main(int  argc,				// I - Number of command-line arguments
     }
     else if (!strncmp(argv[i], "--", 2))
     {
-      fprintf(stderr, "pdfmerge: Unknown option '%s'.\n", argv[i]);
+      fprintf(stderr, "pdfiomerge: Unknown option '%s'.\n", argv[i]);
       return (usage(stderr));
     }
     else if (argv[i][0] == '-')
@@ -59,14 +59,14 @@ main(int  argc,				// I - Number of command-line arguments
           case 'o' : // -o OUTPUT.pdf
               if (outpdf)
               {
-                fputs("pdfmerge: Only one output file can be specified.\n", stderr);
+                fputs("pdfiomerge: Only one output file can be specified.\n", stderr);
                 return (usage(stderr));
               }
 
               i ++;
               if (i >= argc)
               {
-                fputs("pdfmerge: Missing output filename after '-o'.\n", stderr);
+                fputs("pdfiomerge: Missing output filename after '-o'.\n", stderr);
                 return (usage(stderr));
               }
 
@@ -75,7 +75,7 @@ main(int  argc,				// I - Number of command-line arguments
               break;
 
           default :
-              fprintf(stderr, "pdfmerge: Unknown option '-%c'.\n", *opt);
+              fprintf(stderr, "pdfiomerge: Unknown option '-%c'.\n", *opt);
               return (usage(stderr));
         }
       }

pdfio-array.c

@@ -640,6 +640,8 @@ _pdfioArrayRead(pdfio_file_t   *pdf,	// I - PDF file
 //
 // 'pdfioArrayRemove()' - Remove an array entry.
 //
+// @since PDFio v1.4@
+//
 
 bool					// O - `true` on success, `false` otherwise
 pdfioArrayRemove(pdfio_array_t *a,	// I - Array

pdfio-common.c

@@ -47,7 +47,7 @@ _pdfioFileConsume(pdfio_file_t *pdf,	// I - PDF file
 // `false` to halt.
 //
 
-bool					// O - `false` to stop
+bool					// O - `false` to stop, `true` to continue
 _pdfioFileDefaultError(
     pdfio_file_t *pdf,			// I - PDF file
     const char   *message,		// I - Error message
@@ -57,7 +57,7 @@ _pdfioFileDefaultError(
 
   fprintf(stderr, "%s: %s\n", pdf->filename, message);
 
-  return (false);
+  return (!strncmp(message, "WARNING:", 8));
 }
 
 
@@ -134,19 +134,20 @@ _pdfioFileGetChar(pdfio_file_t *pdf)	// I - PDF file
 bool					// O - `true` on success, `false` on error
 _pdfioFileGets(pdfio_file_t *pdf,	// I - PDF file
                char         *buffer,	// I - Line buffer
-	       size_t       bufsize)	// I - Size of line buffer
+	       size_t       bufsize,	// I - Size of line buffer
+	       bool         discard)	// I - OK to discard excess line chars?
 {
   bool	eol = false;			// End of line?
   char	*bufptr = buffer,		// Pointer into buffer
 	*bufend = buffer + bufsize - 1;	// Pointer to end of buffer
 
 
-  PDFIO_DEBUG("_pdfioFileGets(pdf=%p, buffer=%p, bufsize=%lu) bufpos=%ld, buffer=%p, bufptr=%p, bufend=%p, offset=%lu\n", pdf, buffer, (unsigned long)bufsize, (long)pdf->bufpos, pdf->buffer, pdf->bufptr, pdf->bufend, (unsigned long)(pdf->bufpos + (pdf->bufptr - pdf->buffer)));
+  PDFIO_DEBUG("_pdfioFileGets(pdf=%p, buffer=%p, bufsize=%lu, discard=%s) bufpos=%ld, buffer=%p, bufptr=%p, bufend=%p, offset=%lu\n", pdf, buffer, (unsigned long)bufsize, discard ? "true" : "false", (long)pdf->bufpos, pdf->buffer, pdf->bufptr, pdf->bufend, (unsigned long)(pdf->bufpos + (pdf->bufptr - pdf->buffer)));
 
   while (!eol)
   {
     // If there are characters ready in the buffer, use them...
-    while (!eol && pdf->bufptr < pdf->bufend && bufptr < bufend)
+    while (!eol && pdf->bufptr < pdf->bufend)
     {
       char ch = *(pdf->bufptr++);	// Next character in buffer
 
@@ -168,8 +169,10 @@ _pdfioFileGets(pdfio_file_t *pdf,	// I - PDF file
 	    pdf->bufptr ++;
 	}
       }
-      else
+      else if (bufptr < bufend)
         *bufptr++ = ch;
+      else if (!discard)
+        break;
     }
 
     // Fill the read buffer as needed...
@@ -424,7 +427,7 @@ off_t					// O - Offset from beginning of file
 _pdfioFileTell(pdfio_file_t *pdf)	// I - PDF file
 {
   if (pdf->bufptr)
-    return (pdf->bufpos + (pdf->bufptr - pdf->buffer));
+    return (pdf->bufpos + (off_t)(pdf->bufptr - pdf->buffer));
   else
     return (pdf->bufpos);
 }
@@ -452,7 +455,7 @@ _pdfioFileWrite(pdfio_file_t *pdf,	// I - PDF file
       if (!write_buffer(pdf, buffer, bytes))
         return (false);
 
-      pdf->bufpos += bytes;
+      pdf->bufpos += (off_t)bytes;
 
       return (true);
     }
@@ -478,7 +481,7 @@ fill_buffer(pdfio_file_t *pdf)		// I - PDF file
 
   // Advance current position in file as needed...
   if (pdf->bufend)
-    pdf->bufpos += pdf->bufend - pdf->buffer;
+    pdf->bufpos += (off_t)(pdf->bufend - pdf->buffer);
 
   // Try reading from the file...
   if ((bytes = read_buffer(pdf, pdf->buffer, sizeof(pdf->buffer))) <= 0)

pdfio-content.c

@@ -632,6 +632,8 @@ pdfioContentPathCurve23(
 //
 // 'pdfioContentPathEnd()' - Clear the current path.
 //
+// @since PDFio v1.1@
+//
 
 bool					// O - `true` on success, `false` on failure
 pdfioContentPathEnd(pdfio_stream_t *st)	// I - Stream
@@ -1111,6 +1113,8 @@ pdfioContentTextEnd(pdfio_stream_t *st)	// I - Stream
 // on "size". The text string must always use the UTF-8 (Unicode) encoding but
 // any control characters (such as newlines) are ignored.
 //
+// @since PDFio v1.2@
+//
 
 double					// O - Width
 pdfioContentTextMeasure(
@@ -1267,6 +1271,8 @@ pdfioContentTextMoveTo(
 //
 // 'pdfioContentTextNewLine()' - Move to the next line.
 //
+// @since PDFio v1.2@
+//
 
 bool					// O - `true` on success, `false` on failure
 pdfioContentTextNewLine(
@@ -1298,6 +1304,8 @@ pdfioContentTextNextLine(
 // specifies that the current font maps to full Unicode.  The "s" argument
 // specifies a UTF-8 encoded string.
 //
+// @since PDFio v1.2@
+//
 
 bool					// O - `true` on success, `false` on failure
 pdfioContentTextNewLineShow(
@@ -1346,6 +1354,8 @@ pdfioContentTextNewLineShow(
 // argument specifies that the current font maps to full Unicode.  The "format"
 // argument specifies a UTF-8 encoded `printf`-style format string.
 //
+// @since PDFio v1.2@
+//
 
 bool					// O - `true` on success, `false` on failure
 pdfioContentTextNewLineShowf(

pdfio-dict.c

@@ -20,6 +20,8 @@ static int	compare_pairs(_pdfio_pair_t *a, _pdfio_pair_t *b);
 //
 // 'pdfioDictClear()' - Remove a key/value pair from a dictionary.
 //
+// @since PDFio v1.4@
+//
 
 bool					// O - `true` if cleared, `false` otherwise
 pdfioDictClear(pdfio_dict_t *dict,	// I - Dictionary
@@ -345,6 +347,8 @@ pdfioDictGetDict(pdfio_dict_t *dict,	// I - Dictionary
 //
 // 'pdfioDictGetKey()' - Get the key for the specified pair.
 //
+// @since PDFio v1.4@
+//
 
 const char *				// O - Key for specified pair
 pdfioDictGetKey(pdfio_dict_t *dict,	// I - Dictionary
@@ -375,6 +379,8 @@ pdfioDictGetName(pdfio_dict_t *dict,	// I - Dictionary
 //
 // 'pdfioDictGetNumPairs()' - Get the number of key/value pairs in a dictionary.
 //
+// @since PDFio v1.4@
+//
 
 size_t					// O - Number of pairs
 pdfioDictGetNumPairs(pdfio_dict_t *dict)// I - Dictionary
@@ -562,6 +568,8 @@ _pdfioDictGetValue(pdfio_dict_t *dict,	// I - Dictionary
 // The iteration continues as long as the callback returns `true` or all keys
 // have been iterated.
 //
+// @since PDFio v1.1@
+//
 
 void
 pdfioDictIterateKeys(
@@ -635,9 +643,11 @@ _pdfioDictRead(pdfio_file_t   *pdf,	// I - PDF file
     {
       // Issue 118: Discard duplicate key/value pairs, in the future this will
       // be a warning message...
-      _pdfioFileError(pdf, "WARNING: Discarding value for duplicate dictionary key '%s'.", key + 1);
       _pdfioValueDelete(&value);
-      continue;
+      if (_pdfioFileError(pdf, "WARNING: Discarding value for duplicate dictionary key '%s'.", key + 1))
+        continue;
+      else
+        break;
     }
     else if (!_pdfioDictSetValue(dict, pdfioStringCreate(pdf, key + 1), &value))
       break;

pdfio-file.c

@@ -61,9 +61,10 @@ _pdfioFileAddMappedObj(
   pdf->num_objmaps ++;
 
   map->obj        = dst_obj;
-  map->src_pdf    = src_obj->pdf;
   map->src_number = src_obj->number;
 
+  memcpy(map->src_id, src_obj->pdf->file_id, sizeof(map->src_id));
+
   // Sort as needed...
   if (pdf->num_objmaps > 1 && compare_objmaps(map, pdf->objmaps + pdf->num_objmaps - 2) < 0)
     qsort(pdf->objmaps, pdf->num_objmaps, sizeof(_pdfio_objmap_t), (int (*)(const void *, const void *))compare_objmaps);
@@ -181,8 +182,9 @@ pdfioFileClose(pdfio_file_t *pdf)	// I - PDF file
 // of 8.27x11in (the intersection of US Letter and ISO A4) is used.
 //
 // The "error_cb" and "error_cbdata" arguments specify an error handler callback
-// and its data pointer - if `NULL` the default error handler is used that
+// and its data pointer - if `NULL` then the default error handler is used that
-// writes error messages to `stderr`.
+// writes error messages to `stderr`.  The error handler callback should return
+// `true` to continue writing the PDF file or `false` to stop.
 //
 
 pdfio_file_t *				// O - PDF file or `NULL` on error
@@ -260,6 +262,8 @@ pdfioFileCreateArrayObj(
 // This function creates a new object with a name value in a PDF file.
 // You must call @link pdfioObjClose@ to write the object to the file.
 //
+// @since PDFio v1.4@
+//
 
 pdfio_obj_t *				// O - New object
 pdfioFileCreateNameObj(
@@ -289,6 +293,8 @@ pdfioFileCreateNameObj(
 // This function creates a new object with a number value in a PDF file.
 // You must call @link pdfioObjClose@ to write the object to the file.
 //
+// @since PDFio v1.2@
+//
 
 pdfio_obj_t *				// O - New object
 pdfioFileCreateNumberObj(
@@ -415,8 +421,9 @@ _pdfioFileCreateObj(
 // of 8.27x11in (the intersection of US Letter and ISO A4) is used.
 //
 // The "error_cb" and "error_cbdata" arguments specify an error handler callback
-// and its data pointer - if `NULL` the default error handler is used that
+// and its data pointer - if `NULL` then the default error handler is used that
-// writes error messages to `stderr`.
+// writes error messages to `stderr`.  The error handler callback should return
+// `true` to continue writing the PDF file or `false` to stop.
 //
 // > *Note*: Files created using this API are slightly larger than those
 // > created using the @link pdfioFileCreate@ function since stream lengths are
@@ -518,6 +525,8 @@ pdfioFileCreatePage(pdfio_file_t *pdf,	// I - PDF file
 // This function creates a new object with a string value in a PDF file.
 // You must call @link pdfioObjClose@ to write the object to the file.
 //
+// @since PDFio v1.2@
+//
 
 pdfio_obj_t *				// O - New object
 pdfioFileCreateStringObj(
@@ -669,7 +678,7 @@ _pdfioFileFindMappedObj(
     return (NULL);
 
   // Otherwise search for a match...
-  key.src_pdf    = src_pdf;
+  memcpy(key.src_id, src_pdf->file_id, sizeof(key.src_id));
   key.src_number = src_number;
 
   if ((match = (_pdfio_objmap_t *)bsearch(&key, pdf->objmaps, pdf->num_objmaps, sizeof(_pdfio_objmap_t), (int (*)(const void *, const void *))compare_objmaps)) != NULL)
@@ -987,8 +996,14 @@ pdfioFileGetVersion(
 // PDF file requires a password, the open will always fail.
 //
 // The "error_cb" and "error_cbdata" arguments specify an error handler callback
-// and its data pointer - if `NULL` the default error handler is used that
+// and its data pointer - if `NULL` then the default error handler is used that
-// writes error messages to `stderr`.
+// writes error messages to `stderr`.  The error handler callback should return
+// `true` to continue reading the PDF file or `false` to stop.
+//
+// > Note: Error messages starting with "WARNING:" are actually warning
+// > messages - the callback should normally return `true` to allow PDFio to
+// > try to resolve the issue.  In addition, some errors are unrecoverable and
+// > ignore the return value of the error callback.
 //
 
 pdfio_file_t *				// O - PDF file
@@ -1006,6 +1021,10 @@ pdfioFileOpen(
 		*end;			// End of line
   ssize_t	bytes;			// Bytes read
   off_t		xref_offset;		// Offset to xref table
+  time_t	curtime;		// Creation date/time
+  unsigned char	*id_value;		// ID value
+  size_t	id_valuelen;		// Length of ID value
+  _pdfio_sha256_t ctx;			// Hashing context
 
 
   PDFIO_DEBUG("pdfioFileOpen(filename=\"%s\", password_cb=%p, password_cbdata=%p, error_cb=%p, error_cbdata=%p)\n", filename, (void *)password_cb, (void *)password_cbdata, (void *)error_cb, (void *)error_cbdata);
@@ -1049,7 +1068,7 @@ pdfioFileOpen(
   }
 
   // Read the header from the first line...
-  if (!_pdfioFileGets(pdf, line, sizeof(line)))
+  if (!_pdfioFileGets(pdf, line, sizeof(line), true))
     goto error;
 
   if ((strncmp(line, "%PDF-1.", 7) && strncmp(line, "%PDF-2.", 7)) || !isdigit(line[7] & 255))
@@ -1063,7 +1082,7 @@ pdfioFileOpen(
   pdf->version = strdup(line + 5);
 
   // Grab the last 1k of the file to find the start of the xref table...
-  if (_pdfioFileSeek(pdf, -1024, SEEK_END) < 0)
+  if (_pdfioFileSeek(pdf, 1 - sizeof(line), SEEK_END) < 0)
   {
     _pdfioFileError(pdf, "Unable to read startxref data.");
     goto error;
@@ -1075,28 +1094,48 @@ pdfioFileOpen(
     goto error;
   }
 
+  PDFIO_DEBUG("pdfioOpen: Read %d bytes at end of file.\n", (int)bytes);
+
   line[bytes] = '\0';
   end = line + bytes - 9;
 
   for (ptr = line; ptr < end; ptr ++)
   {
-    if (!memcmp(ptr, "startxref", 9))
+    if (!strncmp(ptr, "startxref", 9) && !strstr(ptr + 9, "startxref") && strtol(ptr + 9, NULL, 10) > 0)
       break;
   }
 
   if (ptr >= end)
   {
-    _pdfioFileError(pdf, "Unable to find start of xref table.");
+    if (!_pdfioFileError(pdf, "WARNING: Unable to find start of cross-reference table, will attempt to rebuild."))
-    goto error;
+      goto error;
-  }
 
-  xref_offset = (off_t)strtol(ptr + 9, NULL, 10);
-
-  if (!load_xref(pdf, xref_offset, password_cb, password_cbdata))
-  {
     if (!repair_xref(pdf, password_cb, password_cbdata))
       goto error;
   }
+  else
+  {
+    PDFIO_DEBUG("pdfioFileOpen: line=%p,ptr=%p(\"%s\")\n", line, ptr, ptr);
+
+    xref_offset = (off_t)strtol(ptr + 9, NULL, 10);
+
+    PDFIO_DEBUG("pdfioFileOpen: xref_offset=%lu\n", (unsigned long)xref_offset);
+
+    if (!load_xref(pdf, xref_offset, password_cb, password_cbdata))
+      goto error;
+  }
+
+  // Create the unique file identifier string for the object map...
+  curtime = pdfioFileGetCreationDate(pdf);
+
+  _pdfioCryptoSHA256Init(&ctx);
+  _pdfioCryptoSHA256Append(&ctx, (uint8_t *)pdf->filename, strlen(pdf->filename));
+  _pdfioCryptoSHA256Append(&ctx, (uint8_t *)&curtime, sizeof(curtime));
+  if ((id_value = pdfioArrayGetBinary(pdf->id_array, 0, &id_valuelen)) != NULL)
+    _pdfioCryptoSHA256Append(&ctx, id_value, id_valuelen);
+  if ((id_value = pdfioArrayGetBinary(pdf->id_array, 1, &id_valuelen)) != NULL)
+    _pdfioCryptoSHA256Append(&ctx, id_value, id_valuelen);
+  _pdfioCryptoSHA256Finish(&ctx, pdf->file_id);
 
   return (pdf);
 
@@ -1352,10 +1391,11 @@ static int				// O - Result of comparison
 compare_objmaps(_pdfio_objmap_t *a,	// I - First object map
                 _pdfio_objmap_t *b)	// I - Second object map
 {
-  if (a->src_pdf < b->src_pdf)
+  int	ret = memcmp(a->src_id, b->src_id, sizeof(a->src_id));
-    return (-1);
+					// Result of comparison
-  else if (a->src_pdf > b->src_pdf)
+
-    return (1);
+  if (ret)
+    return (ret);
   else if (a->src_number < b->src_number)
     return (-1);
   else if (a->src_number > b->src_number)
@@ -1384,6 +1424,8 @@ create_common(
   pdfio_file_t	*pdf;			// New PDF file
   pdfio_dict_t	*dict;			// Dictionary
   unsigned char	id_value[16];		// File ID value
+  time_t	curtime;		// Creation date/time
+  _pdfio_sha256_t ctx;			// Hashing context
 
 
   PDFIO_DEBUG("create_common(filename=\"%s\", fd=%d, output_cb=%p, output_cbdata=%p, version=\"%s\", media_box=%p, crop_box=%p, error_cb=%p, error_cbdata=%p)\n", filename, fd, (void *)output_cb, (void *)output_cbdata, version, (void *)media_box, (void *)crop_box, (void *)error_cb, (void *)error_cbdata);
@@ -1474,7 +1516,9 @@ create_common(
   if ((dict = pdfioDictCreate(pdf)) == NULL)
     goto error;
 
-  pdfioDictSetDate(dict, "CreationDate", time(NULL));
+  curtime = time(NULL);
+
+  pdfioDictSetDate(dict, "CreationDate", curtime);
   pdfioDictSetString(dict, "Producer", "pdfio/" PDFIO_VERSION);
 
   if ((pdf->info_obj = pdfioFileCreateObj(pdf, dict)) == NULL)
@@ -1499,6 +1543,14 @@ create_common(
     pdfioArrayAppendBinary(pdf->id_array, id_value, sizeof(id_value));
   }
 
+  // Create the unique file identifier string for the object map...
+  _pdfioCryptoSHA256Init(&ctx);
+  _pdfioCryptoSHA256Append(&ctx, (uint8_t *)pdf->filename, strlen(pdf->filename));
+  _pdfioCryptoSHA256Append(&ctx, (uint8_t *)&curtime, sizeof(curtime));
+  _pdfioCryptoSHA256Append(&ctx, id_value, sizeof(id_value));
+  _pdfioCryptoSHA256Append(&ctx, id_value, sizeof(id_value));
+  _pdfioCryptoSHA256Finish(&ctx, pdf->file_id);
+
   return (pdf);
 
   // Common error handling code...
@@ -1677,7 +1729,10 @@ load_pages(pdfio_file_t *pdf,		// I - PDF file
   }
 
   if ((type = pdfioDictGetName(dict, "Type")) == NULL || (strcmp(type, "Pages") && strcmp(type, "Page")))
-    return (false);
+  {
+    if (!_pdfioFileError(pdf, "WARNING: No Type value for pages object."))
+      return (false);
+  }
 
   // If there is a Kids array, then this is a parent node and we have to look
   // at the child objects...
@@ -1744,31 +1799,32 @@ load_xref(
   int		generation;		// Generation number
   _pdfio_token_t tb;			// Token buffer/stack
   off_t		line_offset;		// Offset to start of line
+  pdfio_obj_t	*pages_obj;		// Pages object
 
 
   while (!done)
   {
     if (_pdfioFileSeek(pdf, xref_offset, SEEK_SET) != xref_offset)
     {
-      _pdfioFileError(pdf, "Unable to seek to start of xref table.");
+      PDFIO_DEBUG("load_xref: Unable to seek to %lu.\n", (unsigned long)xref_offset);
-      return (false);
+      goto repair;
     }
 
     do
     {
       line_offset = _pdfioFileTell(pdf);
 
-      if (!_pdfioFileGets(pdf, line, sizeof(line)))
+      if (!_pdfioFileGets(pdf, line, sizeof(line), true))
       {
-	_pdfioFileError(pdf, "Unable to read start of xref table.");
+	PDFIO_DEBUG("load_xref: Unable to read line at offset %lu.\n", (unsigned long)line_offset);
-	return (false);
+        goto repair;
       }
     }
     while (!line[0]);
 
     PDFIO_DEBUG("load_xref: line_offset=%lu, line='%s'\n", (unsigned long)line_offset, line);
 
-    if (isdigit(line[0] & 255) && strlen(line) > 4 && (!strcmp(line + strlen(line) - 4, " obj") || ((ptr = strstr(line, " obj")) != NULL && ptr[4] == '<')))
+    if (isdigit(line[0] & 255) && strlen(line) > 4 && (!strcmp(line + strlen(line) - 4, " obj") || ((ptr = strstr(line, " obj")) != NULL && (ptr[4] == '<' || isspace(ptr[4])))))
     {
       // Cross-reference stream
       pdfio_obj_t	*obj;		// Object
@@ -1790,14 +1846,14 @@ load_xref(
 
       if ((number = strtoimax(line, &ptr, 10)) < 1)
       {
-	_pdfioFileError(pdf, "Bad xref table header '%s'.", line);
+        PDFIO_DEBUG("load_xref: Unable to scan object number.\n");
-	return (false);
+        goto repair;
       }
 
       if ((generation = (int)strtol(ptr, &ptr, 10)) < 0 || (generation > 65535 && number != 0))
       {
-	_pdfioFileError(pdf, "Bad xref table header '%s'.", line);
+        PDFIO_DEBUG("load_xref: Unable to scan generation number (%u).\n", (unsigned)generation);
-	return (false);
+        goto repair;
       }
 
       while (isspace(*ptr & 255))
@@ -1805,14 +1861,14 @@ load_xref(
 
       if (strncmp(ptr, "obj", 3))
       {
-	_pdfioFileError(pdf, "Bad xref table header '%s'.", line);
+        PDFIO_DEBUG("load_xref: No 'obj' after object number and generation (saw '%s').\n", ptr);
-	return (false);
+        goto repair;
       }
 
       if (_pdfioFileSeek(pdf, line_offset + (off_t)(ptr + 3 - line), SEEK_SET) < 0)
       {
-        _pdfioFileError(pdf, "Unable to seek to xref object %lu %u.", (unsigned long)number, (unsigned)generation);
+        PDFIO_DEBUG("load_xref: Unable to seek to start of cross-reference object dictionary.\n");
-        return (false);
+        goto repair;
       }
 
       PDFIO_DEBUG("load_xref: Loading object %lu %u.\n", (unsigned long)number, (unsigned)generation);
@@ -1827,21 +1883,21 @@ load_xref(
 
       if (!_pdfioValueRead(pdf, obj, &tb, &trailer, 0))
       {
-        _pdfioFileError(pdf, "Unable to read cross-reference stream dictionary.");
+        PDFIO_DEBUG("load_xref: Unable to read cross-reference object dictionary.\n");
-        return (false);
+        goto repair;
       }
       else if (trailer.type != PDFIO_VALTYPE_DICT)
       {
-	_pdfioFileError(pdf, "Cross-reference stream does not have a dictionary.");
+        PDFIO_DEBUG("load_xref: Expected dictionary for cross-reference object (type=%d).", trailer.type);
-	return (false);
+        goto repair;
       }
 
       obj->value = trailer;
 
       if (!_pdfioTokenGet(&tb, line, sizeof(line)) || strcmp(line, "stream"))
       {
-        _pdfioFileError(pdf, "Unable to get stream after xref dictionary.");
+        PDFIO_DEBUG("load_xref: No stream token after dictionary (got '%s').\n", line);
-        return (false);
+        goto repair;
       }
 
       PDFIO_DEBUG("load_xref: tb.bufptr=%p, tb.bufend=%p, tb.bufptr[0]=0x%02x, tb.bufptr[0]=0x%02x\n", tb.bufptr, tb.bufend, tb.bufptr[0], tb.bufptr[1]);
@@ -1859,8 +1915,8 @@ load_xref(
 
       if ((w_array = pdfioDictGetArray(trailer.value.dict, "W")) == NULL)
       {
-	_pdfioFileError(pdf, "Cross-reference stream does not have required W key.");
+        PDFIO_DEBUG("load_xref: Missing W array in cross-reference objection dictionary.\n");
-	return (false);
+        goto repair;
       }
 
       w[0]    = (size_t)pdfioArrayGetNumber(w_array, 0);
@@ -1870,16 +1926,18 @@ load_xref(
       w_2     = w[0];
       w_3     = w[0] + w[1];
 
-      if (w[1] == 0 || w[2] > 4 || w[0] > sizeof(buffer) || w[1] > sizeof(buffer) || w[2] > sizeof(buffer) || w_total > sizeof(buffer))
+      PDFIO_DEBUG("W=[%u %u %u], w_total=%u\n", (unsigned)w[0], (unsigned)w[1], (unsigned)w[2], (unsigned)w_total);
+
+      if (pdfioArrayGetSize(w_array) > 3 || w[1] == 0 || w[2] > 4 || w[0] > sizeof(buffer) || w[1] > sizeof(buffer) || w[2] > sizeof(buffer) || w_total > sizeof(buffer))
       {
-	_pdfioFileError(pdf, "Cross-reference stream has invalid W key [%u %u %u].", (unsigned)w[0], (unsigned)w[1], (unsigned)w[2]);
+        PDFIO_DEBUG("load_xref: Bad W array in cross-reference objection dictionary.\n");
-	return (false);
+        goto repair;
       }
 
       if ((st = pdfioObjOpenStream(obj, true)) == NULL)
       {
-	_pdfioFileError(pdf, "Unable to open cross-reference stream.");
+        PDFIO_DEBUG("load_xref: Unable to open cross-reference stream.\n");
-	return (false);
+        goto repair;
       }
 
       for (index_n = 0; index_n < index_count; index_n += 2)
@@ -1899,7 +1957,20 @@ load_xref(
 	{
 	  count --;
 
-	  PDFIO_DEBUG("load_xref: number=%u %02X%02X%02X%02X%02X\n", (unsigned)number, buffer[0], buffer[1], buffer[2], buffer[3], buffer[4]);
+#ifdef DEBUG
+          if (w_total > 5)
+	    PDFIO_DEBUG("load_xref: number=%u %02X%02X%02X%02X%02X...\n", (unsigned)number, buffer[0], buffer[1], buffer[2], buffer[3], buffer[4]);
+	  else if (w_total == 5)
+	    PDFIO_DEBUG("load_xref: number=%u %02X%02X%02X%02X%02X\n", (unsigned)number, buffer[0], buffer[1], buffer[2], buffer[3], buffer[4]);
+	  else if (w_total == 4)
+	    PDFIO_DEBUG("load_xref: number=%u %02X%02X%02X%02X\n", (unsigned)number, buffer[0], buffer[1], buffer[2], buffer[3]);
+	  else if (w_total == 3)
+	    PDFIO_DEBUG("load_xref: number=%u %02X%02X%02X\n", (unsigned)number, buffer[0], buffer[1], buffer[2]);
+	  else if (w_total == 2)
+	    PDFIO_DEBUG("load_xref: number=%u %02X%02X\n", (unsigned)number, buffer[0], buffer[1]);
+	  else
+	    PDFIO_DEBUG("load_xref: number=%u %02X\n", (unsigned)number, buffer[0]);
+#endif // DEBUG
 
 	  // Check whether this is an object definition...
 	  if (w[0] > 0)
@@ -1981,6 +2052,7 @@ load_xref(
 	      else
 	      {
 		_pdfioFileError(pdf, "Too many object streams.");
+		pdfioStreamClose(st);
 	        return (false);
 	      }
 	    }
@@ -1989,7 +2061,10 @@ load_xref(
 	  {
 	    // Add this object...
 	    if (!add_obj(pdf, (size_t)number, (unsigned short)generation, (off_t)offset))
+	    {
+	      pdfioStreamClose(st);
 	      return (false);
+	    }
 	  }
 
 	  number ++;
@@ -2037,7 +2112,7 @@ load_xref(
 					// Offset of current line
 
       PDFIO_DEBUG("load_xref: Reading xref table starting at offset %lu\n", (unsigned long)trailer_offset);
-      while (_pdfioFileGets(pdf, line, sizeof(line)))
+      while (_pdfioFileGets(pdf, line, sizeof(line), false))
       {
         PDFIO_DEBUG("load_xref: '%s' at offset %lu\n", line, (unsigned long)trailer_offset);
 
@@ -2062,8 +2137,8 @@ load_xref(
 
 	if (sscanf(line, "%jd%jd", &number, &num_objects) != 2)
 	{
-	  _pdfioFileError(pdf, "Malformed xref table section '%s'.", line);
+	  PDFIO_DEBUG("load_xref: Unable to scan START COUNT from line.\n");
-	  return (false);
+	  goto repair;
 	}
 
 	// Read this group of objects...
@@ -2071,41 +2146,45 @@ load_xref(
 	{
 	  // Read a line from the file and validate it...
 	  if (_pdfioFileRead(pdf, line, 20) != 20)
-	    return (false);
+	  {
+	    PDFIO_DEBUG("load_xref: Unable to read 20 byte xref record.\n");
+	    goto repair;
+	  }
 
 	  line[20] = '\0';
 
-	  if (strcmp(line + 18, "\r\n") && strcmp(line + 18, " \n") && strcmp(line + 18, " \r"))
+	  if (strcmp(line + 18, "\r\n") && strcmp(line + 18, "\r\r") && strcmp(line + 18, " \n") && strcmp(line + 18, " \r"))
 	  {
-	    _pdfioFileError(pdf, "Malformed xref table entry '%s'.", line);
+	    PDFIO_DEBUG("load_xref: Bad end-of-line <%02X%02X>\n", line[18], line[19]);
-	    return (false);
+	    goto repair;
 	  }
+
 	  line[18] = '\0';
 
 	  // Parse the line
 	  if ((offset = strtoimax(line, &ptr, 10)) < 0)
 	  {
-	    _pdfioFileError(pdf, "Malformed xref table entry '%s'.", line);
+	    PDFIO_DEBUG("load_xref: Unable to scan offset.\n");
-	    return (false);
+	    goto repair;
 	  }
 
 	  if ((generation = (int)strtol(ptr, &ptr, 10)) < 0 || (generation > 65535 && offset != 0))
 	  {
-	    _pdfioFileError(pdf, "Malformed xref table entry '%s'.", line);
+	    PDFIO_DEBUG("load_xref: Unable to scan generation (%u).\n", (unsigned)generation);
-	    return (false);
+	    goto repair;
 	  }
 
 	  if (*ptr != ' ')
 	  {
-	    _pdfioFileError(pdf, "Malformed xref table entry '%s'.", line);
+	    PDFIO_DEBUG("load_xref: Missing space before type.\n");
-	    return (false);
+	    goto repair;
 	  }
 
 	  ptr ++;
 	  if (*ptr != 'f' && *ptr != 'n')
 	  {
-	    _pdfioFileError(pdf, "Malformed xref table entry '%s'.", line);
+	    PDFIO_DEBUG("load_xref: Bad type '%c'.\n", *ptr);
-	    return (false);
+	    goto repair;
 	  }
 
 	  if (*ptr == 'f')
@@ -2124,21 +2203,21 @@ load_xref(
 
       if (strncmp(line, "trailer", 7))
       {
-	_pdfioFileError(pdf, "Missing trailer.");
+        PDFIO_DEBUG("load_xref: No trailer after xref table.\n");
-	return (false);
+	goto repair;
       }
 
       _pdfioTokenInit(&tb, pdf, (_pdfio_tconsume_cb_t)_pdfioFileConsume, (_pdfio_tpeek_cb_t)_pdfioFilePeek, pdf);
 
       if (!_pdfioValueRead(pdf, NULL, &tb, &trailer, 0))
       {
-	_pdfioFileError(pdf, "Unable to read trailer dictionary.");
+        PDFIO_DEBUG("load_xref: Unable to read trailer dictionary.\n");
-	return (false);
+	goto repair;
       }
       else if (trailer.type != PDFIO_VALTYPE_DICT)
       {
-	_pdfioFileError(pdf, "Trailer is not a dictionary.");
+        PDFIO_DEBUG("load_xref: Trailer not a dictionary (type=%d).\n", trailer.type);
-	return (false);
+	goto repair;
       }
 
       PDFIO_DEBUG("load_xref: Got trailer dict.\n");
@@ -2160,8 +2239,7 @@ load_xref(
     }
     else
     {
-      _pdfioFileError(pdf, "Bad xref table header '%s'.", line);
+      goto repair;
-      return (false);
     }
 
     PDFIO_DEBUG("load_xref: Contents of trailer dictionary:\n");
@@ -2190,13 +2268,31 @@ load_xref(
 
   if ((pdf->root_obj = pdfioDictGetObj(pdf->trailer_dict, "Root")) == NULL)
   {
-    _pdfioFileError(pdf, "Missing Root object.");
+    PDFIO_DEBUG("load_xref: Missing Root object.\n");
-    return (false);
+    goto repair;
   }
 
   PDFIO_DEBUG("load_xref: Root=%p(%lu)\n", pdf->root_obj, (unsigned long)pdf->root_obj->number);
 
-  return (load_pages(pdf, pdfioDictGetObj(pdfioObjGetDict(pdf->root_obj), "Pages"), 0));
+  if ((pages_obj = pdfioDictGetObj(pdfioObjGetDict(pdf->root_obj), "Pages")) == NULL)
+  {
+    PDFIO_DEBUG("load_xref: Missing Pages object.\n");
+    goto repair;
+  }
+
+  PDFIO_DEBUG("load_xref: Pages=%p(%lu)\n", pdf->root_obj, (unsigned long)pdf->root_obj->number);
+
+  return (load_pages(pdf, pages_obj, 0));
+
+  // If we get here the cross-reference table is busted - try repairing if the
+  // error callback says to proceed...
+
+  repair:
+
+  if (_pdfioFileError(pdf, "WARNING: Cross-reference is damaged, will attempt to rebuild."))
+    return (repair_xref(pdf, password_cb, password_data));
+  else
+    return (false);
 }
 
 
@@ -2210,7 +2306,7 @@ repair_xref(
     pdfio_password_cb_t password_cb,	// I - Password callback or `NULL` for none
     void                *password_data)	// I - Password callback data, if any
 {
-  char		line[16384],		// Line from file
+  char		line[1024],		// Line from file
 		*ptr;			// Pointer into line
   off_t		line_offset;		// Offset in file
   intmax_t	number;			// Object number
@@ -2218,16 +2314,22 @@ repair_xref(
   size_t	i;			// Looping var
   size_t	num_sobjs = 0;		// Number of object streams
   pdfio_obj_t	*sobjs[16384];		// Object streams to load
+  pdfio_dict_t	*backup_trailer = NULL;	// Backup trailer dictionary
+  pdfio_obj_t	*pages_obj;		// Pages object
 
 
-  // Let caller know something is wrong...
+  // Clear trailer data...
-  _pdfioFileError(pdf, "WARNING: Cross-reference table is damaged, attempting to rebuild.");
+  pdf->trailer_dict = NULL;
+  pdf->root_obj     = NULL;
+  pdf->info_obj     = NULL;
+  pdf->pages_obj    = NULL;
+  pdf->encrypt_obj  = NULL;
 
-  // Read from the beginning of the file, looking for
+  // Read from the beginning of the file, looking for objects...
   if ((line_offset = _pdfioFileSeek(pdf, 0, SEEK_SET)) < 0)
     return (false);
 
-  while (_pdfioFileGets(pdf, line, sizeof(line)))
+  while (_pdfioFileGets(pdf, line, sizeof(line), true))
   {
     // See if this is the start of an object...
     if (line[0] >= '1' && line[0] <= '9')
@@ -2244,43 +2346,75 @@ repair_xref(
 	  pdfio_obj_t	*obj;		// Object
 	  _pdfio_token_t tb;		// Token buffer/stack
 
-          PDFIO_DEBUG("OBJECT %ld %d at offset %ld\n", (long)number, generation, (long)line_offset);
+          PDFIO_DEBUG("repair_xref: OBJECT %ld %d at offset %ld\n", (long)number, generation, (long)line_offset);
 
-	  if ((obj = add_obj(pdf, (size_t)number, (unsigned short)generation, line_offset)) == NULL)
+          if ((obj = pdfioFileFindObj(pdf, (size_t)number)) != NULL)
+          {
+            obj->offset = line_offset;
+          }
+          else if ((obj = add_obj(pdf, (size_t)number, (unsigned short)generation, line_offset)) == NULL)
 	  {
 	    _pdfioFileError(pdf, "Unable to allocate memory for object.");
 	    return (false);
 	  }
 
+          if (ptr[3])
+          {
+	    // Probably the start of the object dictionary, rewind the file so
+	    // we can read it...
+	    _pdfioFileSeek(pdf, line_offset + (ptr - line + 3), SEEK_SET);
+	  }
+
 	  _pdfioTokenInit(&tb, pdf, (_pdfio_tconsume_cb_t)_pdfioFileConsume, (_pdfio_tpeek_cb_t)_pdfioFilePeek, pdf);
 
 	  if (!_pdfioValueRead(pdf, obj, &tb, &obj->value, 0))
 	  {
-	    _pdfioFileError(pdf, "Unable to read cross-reference stream dictionary.");
+	    if (!_pdfioFileError(pdf, "WARNING: Unable to read object dictionary/value."))
-	    return (false);
+	      return (false);
+	    else
+	      continue;
 	  }
 
-	  if (_pdfioTokenGet(&tb, line, sizeof(line)) && strcmp(line, "stream"))
+	  if (_pdfioTokenGet(&tb, line, sizeof(line)))
 	  {
 	    const char *type = pdfioObjGetType(obj);
 					// Object type
 
 	    _pdfioTokenFlush(&tb);
-	    obj->stream_offset = _pdfioFileTell(pdf);
 
-	    if (type && !strcmp(type, "ObjStm") && num_sobjs < (sizeof(sobjs) / sizeof(sobjs[0])))
+            if (type && !strcmp(line, "stream"))
-	    {
-	      sobjs[num_sobjs] = obj;
-	      num_sobjs ++;
-	    }
-
-            if (type && !strcmp(type, "XRef") && !pdf->trailer_dict)
             {
-              // Save the trailer dictionary...
+              // Possible object or XRef stream...
-	      pdf->trailer_dict = pdfioObjGetDict(obj);
+	      obj->stream_offset = _pdfioFileTell(pdf);
-	      pdf->encrypt_obj  = pdfioDictGetObj(pdf->trailer_dict, "Encrypt");
+
-	      pdf->id_array     = pdfioDictGetArray(pdf->trailer_dict, "ID");
+	      if (!strcmp(type, "ObjStm") && num_sobjs < (sizeof(sobjs) / sizeof(sobjs[0])))
-            }
+	      {
+	        PDFIO_DEBUG("repair_xref: Object stream...\n");
+		sobjs[num_sobjs] = obj;
+		num_sobjs ++;
+	      }
+
+	      if (!strcmp(type, "XRef") && !pdf->trailer_dict)
+	      {
+		// Save the trailer dictionary...
+	        PDFIO_DEBUG("repair_xref: XRef stream...\n");
+		pdf->trailer_dict = pdfioObjGetDict(obj);
+		pdf->encrypt_obj  = pdfioDictGetObj(pdf->trailer_dict, "Encrypt");
+		pdf->id_array     = pdfioDictGetArray(pdf->trailer_dict, "ID");
+	      }
+	    }
+	    else if (type && !strcmp(line, "endobj"))
+	    {
+	      // Possible catalog or pages object...
+	      if (!strcmp(type, "Catalog"))
+	      {
+	        PDFIO_DEBUG("repair_xref: Catalog (root) object...\n");
+	        if (!backup_trailer)
+	          backup_trailer = pdfioDictCreate(pdf);
+
+	        pdfioDictSetObj(backup_trailer, "Root", obj);
+	      }
+	    }
 	  }
 	}
       }
@@ -2291,6 +2425,8 @@ repair_xref(
       _pdfio_token_t tb;		// Token buffer/stack
       _pdfio_value_t trailer;		// Trailer
 
+      PDFIO_DEBUG("repair_xref: line=\"%s\"\n", line);
+
       if (line[7])
       {
 	// Probably the start of the trailer dictionary, rewind the file so
@@ -2298,7 +2434,7 @@ repair_xref(
 	_pdfioFileSeek(pdf, line_offset + 7, SEEK_SET);
       }
 
-      PDFIO_DEBUG("TRAILER at offset %ld\n", (long)line_offset);
+      PDFIO_DEBUG("repair_xref: TRAILER at offset %ld\n", (long)line_offset);
 
       _pdfioTokenInit(&tb, pdf, (_pdfio_tconsume_cb_t)_pdfioFileConsume, (_pdfio_tpeek_cb_t)_pdfioFilePeek, pdf);
       if (!_pdfioValueRead(pdf, NULL, &tb, &trailer, 0))
@@ -2314,10 +2450,12 @@ repair_xref(
 
       _pdfioTokenFlush(&tb);
 
-      if (!pdf->trailer_dict)
+      if (_pdfioDictGetValue(trailer.value.dict, "Root"))
       {
 	// Save the trailer dictionary and grab the root (catalog) and info
 	// objects...
+	PDFIO_DEBUG("repair_xref: Using this trailer dictionary.\n");
+
 	pdf->trailer_dict = trailer.value.dict;
 	pdf->encrypt_obj  = pdfioDictGetObj(pdf->trailer_dict, "Encrypt");
 	pdf->id_array     = pdfioDictGetArray(pdf->trailer_dict, "ID");
@@ -2328,11 +2466,18 @@ repair_xref(
     line_offset = _pdfioFileTell(pdf);
   }
 
+  PDFIO_DEBUG("repair_xref: Stopped at line_offset=%lu\n", (unsigned long)line_offset);
+
+  if (!pdf->trailer_dict && backup_trailer)
+    pdf->trailer_dict = backup_trailer;
+
   // If the trailer contains an Encrypt key, try unlocking the file...
   if (pdf->encrypt_obj && !_pdfioCryptoUnlock(pdf, password_cb, password_data))
     return (false);
 
   // Load any stream objects...
+  PDFIO_DEBUG("repair_xref: Found %lu stream objects.\n", (unsigned long)num_sobjs);
+
   for (i = 0; i < num_sobjs; i ++)
   {
     if (!load_obj_stream(sobjs[i]))
@@ -2351,8 +2496,16 @@ repair_xref(
 
   PDFIO_DEBUG("repair_xref: Root=%p(%lu)\n", pdf->root_obj, (unsigned long)pdf->root_obj->number);
 
+  if ((pages_obj = pdfioDictGetObj(pdfioObjGetDict(pdf->root_obj), "Pages")) == NULL)
+  {
+    _pdfioFileError(pdf, "Missing Pages object.");
+    return (false);
+  }
+
+  PDFIO_DEBUG("repair_xref: Pages=%p(%lu)\n", pages_obj, (unsigned long)pages_obj->number);
+
   // Load pages...
-  return (load_pages(pdf, pdfioDictGetObj(pdfioObjGetDict(pdf->root_obj), "Pages"), 0));
+  return (load_pages(pdf, pages_obj, 0));
 }
 
 
@@ -2502,12 +2655,16 @@ write_trailer(pdfio_file_t *pdf)	// I - PDF file
             buffer[2] = (obj->offset >> 8) & 255;
             buffer[3] = obj->offset & 255;
             break;
+#ifdef _WIN32
+	default :
+#endif // _WIN32
         case 4 :
             buffer[1] = (obj->offset >> 24) & 255;
             buffer[2] = (obj->offset >> 16) & 255;
             buffer[3] = (obj->offset >> 8) & 255;
             buffer[4] = obj->offset & 255;
             break;
+#ifndef _WIN32 // Windows off_t is 32-bits?!?
         case 5 :
             buffer[1] = (obj->offset >> 32) & 255;
             buffer[2] = (obj->offset >> 24) & 255;
@@ -2542,6 +2699,7 @@ write_trailer(pdfio_file_t *pdf)	// I - PDF file
             buffer[7] = (obj->offset >> 8) & 255;
             buffer[8] = obj->offset & 255;
             break;
+#endif // !_WIN32
       }
 
       if (!pdfioStreamWrite(xref_st, buffer, offsize + 2))

pdfio-object.c

@@ -79,6 +79,10 @@ pdfioObjCopy(pdfio_file_t *pdf,		// I - PDF file
   if (srcobj->value.type == PDFIO_VALTYPE_NONE)
     _pdfioObjLoad(srcobj);
 
+  // See if we have already mapped this object...
+  if ((dstobj = _pdfioFileFindMappedObj(pdf, srcobj->pdf, srcobj->number)) != NULL)
+    return (dstobj);			// Yes, return that one...
+
   // Create the new object...
   if ((dstobj = _pdfioFileCreateObj(pdf, srcobj->pdf, NULL)) == NULL)
     return (NULL);
@@ -141,6 +145,7 @@ pdfioObjCreateStream(
     pdfio_obj_t    *obj,		// I - Object
     pdfio_filter_t filter)		// I - Type of compression to apply
 {
+  pdfio_stream_t *st;			// Stream
   pdfio_obj_t	*length_obj = NULL;	// Length object, if any
 
 
@@ -194,11 +199,13 @@ pdfioObjCreateStream(
   if (!_pdfioFilePuts(obj->pdf, "stream\n"))
     return (NULL);
 
-  obj->stream_offset    = _pdfioFileTell(obj->pdf);
+  obj->stream_offset = _pdfioFileTell(obj->pdf);
-  obj->pdf->current_obj = obj;
 
   // Return the new stream...
-  return (_pdfioStreamCreate(obj, length_obj, 0, filter));
+  if ((st = _pdfioStreamCreate(obj, length_obj, 0, filter)) != NULL)
+    obj->pdf->current_obj = obj;
+
+  return (st);
 }
 
 
@@ -330,6 +337,8 @@ pdfioObjGetLength(pdfio_obj_t *obj)	// I - Object
 //
 // 'pdfioObjGetName()' - Get the name value associated with an object.
 //
+// @since PDFio v1.4@
+//
 
 const char *				// O - Dictionary or `NULL` on error
 pdfioObjGetName(pdfio_obj_t *obj)	// I - Object
@@ -532,6 +541,9 @@ pdfio_stream_t *			// O - Stream or `NULL` on error
 pdfioObjOpenStream(pdfio_obj_t *obj,	// I - Object
                    bool        decode)	// I - Decode/decompress data?
 {
+  pdfio_stream_t	*st;		// Stream
+
+
   // Range check input...
   if (!obj)
     return (NULL);
@@ -554,9 +566,10 @@ pdfioObjOpenStream(pdfio_obj_t *obj,	// I - Object
     return (NULL);
 
   // Open the stream...
-  obj->pdf->current_obj = obj;
+  if ((st = _pdfioStreamOpen(obj, decode)) != NULL)
+    obj->pdf->current_obj = obj;
 
-  return (_pdfioStreamOpen(obj, decode));
+  return (st);
 }
 
 

pdfio-private.h

@@ -10,7 +10,7 @@
 #ifndef PDFIO_PRIVATE_H
 #  define PDFIO_PRIVATE_H
 #  ifdef _WIN32
-#    define _CRT_SECURE_NO_WARNINGS	// Disable bogus VS warnings/errors...
+#    define _CRT_SECURE_NO_WARNINGS 1	// Disable bogus VS warnings/errors...
 #  endif // _WIN32
 #  include "pdfio.h"
 #  include <stdarg.h>
@@ -28,16 +28,16 @@
 #    define access	_access		// Map standard POSIX/C99 names
 #    define close	_close
 #    define fileno	_fileno
-#    define lseek	_lseek
+#    define lseek(f,o,w) (off_t)_lseek((f),(long)(o),(w))
 #    define mkdir(d,p)	_mkdir(d)
 #    define open	_open
-#    define read	_read
+#    define read(f,b,s)	_read((f),(b),(unsigned)(s))
 #    define rmdir	_rmdir
 #    define snprintf	_snprintf
 #    define strdup	_strdup
 #    define unlink	_unlink
 #    define vsnprintf	_vsnprintf
-#    define write	_write
+#    define write(f,b,s) _write((f),(b),(unsigned)(s))
 #    ifndef F_OK
 #      define F_OK	00		// POSIX parameters/flags
 #      define W_OK	02
@@ -221,7 +221,7 @@ struct _pdfio_dict_s			// Dictionary
 typedef struct _pdfio_objmap_s		// PDF object map
 {
   pdfio_obj_t	*obj;			// Object for this file
-  pdfio_file_t	*src_pdf;		// Source PDF file
+  unsigned char	src_id[32];		// Source PDF file file identifier
   size_t	src_number;		// Source object number
 } _pdfio_objmap_t;
 
@@ -236,6 +236,7 @@ typedef struct _pdfio_strbuf_s		// PDF string buffer
 struct _pdfio_file_s			// PDF file structure
 {
   char		*filename;		// Filename
+  unsigned char	file_id[32];		// File identifier bytes
   struct lconv	*loc;			// Locale data
   char		*version;		// Version number
   pdfio_rect_t	media_box,		// Default MediaBox value
@@ -384,7 +385,7 @@ extern bool		_pdfioFileError(pdfio_file_t *pdf, const char *format, ...) _PDFIO_
 extern pdfio_obj_t	*_pdfioFileFindMappedObj(pdfio_file_t *pdf, pdfio_file_t *src_pdf, size_t src_number) _PDFIO_INTERNAL;
 extern bool		_pdfioFileFlush(pdfio_file_t *pdf) _PDFIO_INTERNAL;
 extern int		_pdfioFileGetChar(pdfio_file_t *pdf) _PDFIO_INTERNAL;
-extern bool		_pdfioFileGets(pdfio_file_t *pdf, char *buffer, size_t bufsize) _PDFIO_INTERNAL;
+extern bool		_pdfioFileGets(pdfio_file_t *pdf, char *buffer, size_t bufsize, bool discard) _PDFIO_INTERNAL;
 extern ssize_t		_pdfioFilePeek(pdfio_file_t *pdf, void *buffer, size_t bytes) _PDFIO_INTERNAL;
 extern bool		_pdfioFilePrintf(pdfio_file_t *pdf, const char *format, ...) _PDFIO_INTERNAL;
 extern bool		_pdfioFilePuts(pdfio_file_t *pdf, const char *s) _PDFIO_INTERNAL;

pdfio-stream.c

@@ -259,7 +259,7 @@ _pdfioStreamCreate(
     {
       colors = 1;
     }
-    else if (colors < 0 || colors > 4)
+    else if (colors < 0 || colors > 32)
     {
       _pdfioFileError(st->pdf, "Unsupported Colors value %d.", colors);
       free(st);
@@ -270,7 +270,7 @@ _pdfioStreamCreate(
     {
       columns = 1;
     }
-    else if (columns < 0)
+    else if (columns < 0 || columns > 65536)
     {
       _pdfioFileError(st->pdf, "Unsupported Columns value %d.", columns);
       free(st);
@@ -532,7 +532,7 @@ _pdfioStreamOpen(pdfio_obj_t *obj,	// I - Object
       {
         colors = 1;
       }
-      else if (colors < 0 || colors > 4)
+      else if (colors < 0 || colors > 32)
       {
         _pdfioFileError(st->pdf, "Unsupported Colors value %d.", colors);
 	goto error;
@@ -542,7 +542,7 @@ _pdfioStreamOpen(pdfio_obj_t *obj,	// I - Object
       {
         columns = 1;
       }
-      else if (columns < 0)
+      else if (columns < 0 || columns > 65536)
       {
         _pdfioFileError(st->pdf, "Unsupported Columns value %d.", columns);
 	goto error;
@@ -697,8 +697,9 @@ pdfioStreamPeek(pdfio_stream_t *st,	// I - Stream
 // 'pdfioStreamPrintf()' - Write a formatted string to a stream.
 //
 // This function writes a formatted string to a stream.  In addition to the
-// standard `printf` format characters, you can use "%N" to format a PDF name
+// standard `printf` format characters, you can use "%H" to format a HTML/XML
-// value ("/Name") and "%S" to format a PDF string ("(String)") value.
+// string value, "%N" to format a PDF name value ("/Name"), and "%S" to format
+// a PDF string ("(String)") value.
 //
 
 bool					// O - `true` on success, `false` on failure

pdfio-string.c

@@ -486,6 +486,59 @@ _pdfio_vsnprintf(pdfio_file_t *pdf,	// I - PDF file
 	    }
 	    break;
 
+	case 'H' : // XML/HTML string
+	    if ((s = va_arg(ap, char *)) == NULL)
+	      s = "(null)";
+
+            // Loop through the literal string...
+            while (*s)
+            {
+              // Escape special characters
+              if (*s == '&')
+              {
+                // &
+                if (bufptr < bufend)
+                {
+                  _pdfio_strlcpy(bufptr, "&amp;", (size_t)(bufend - bufptr + 1));
+                  bufptr += strlen(bufptr);
+                }
+
+                bytes += 5;
+              }
+              else if (*s == '<')
+              {
+                // &lt;
+                if (bufptr < bufend)
+                {
+                  _pdfio_strlcpy(bufptr, "&lt;", (size_t)(bufend - bufptr + 1));
+                  bufptr += strlen(bufptr);
+                }
+
+                bytes += 4;
+              }
+              else if (*s == '>')
+              {
+                // &gt;
+                if (bufptr < bufend)
+                {
+                  _pdfio_strlcpy(bufptr, "&gt;", (size_t)(bufend - bufptr + 1));
+                  bufptr += strlen(bufptr);
+                }
+
+                bytes += 4;
+              }
+              else
+              {
+                // Literal character...
+                if (bufptr < bufend)
+                  *bufptr++ = *s;
+                bytes ++;
+              }
+
+              s ++;
+            }
+	    break;
+
 	case 'S' : // PDF string
 	    if ((s = va_arg(ap, char *)) == NULL)
 	      s = "(null)";

pdfio.h

@@ -46,7 +46,7 @@ extern "C" {
 //
 
 #  if _WIN32
-typedef __int64 ssize_t;		// POSIX type not present on Windows... @private@
+typedef __int64 ssize_t;		// POSIX type not present on Windows @private@
 #  endif // _WIN32
 
 typedef struct _pdfio_array_s pdfio_array_t;
@@ -62,7 +62,7 @@ typedef bool (*pdfio_error_cb_t)(pdfio_file_t *pdf, const char *message, void *d
 typedef enum pdfio_encryption_e		// PDF encryption modes
 {
   PDFIO_ENCRYPTION_NONE = 0,		// No encryption
-  PDFIO_ENCRYPTION_RC4_40,		// 40-bit RC4 encryption (PDF 1.3)
+  PDFIO_ENCRYPTION_RC4_40,		// 40-bit RC4 encryption (PDF 1.3, reading only)
   PDFIO_ENCRYPTION_RC4_128,		// 128-bit RC4 encryption (PDF 1.4)
   PDFIO_ENCRYPTION_AES_128,		// 128-bit AES encryption (PDF 1.6)
   PDFIO_ENCRYPTION_AES_256		// 256-bit AES encryption (PDF 2.0) @exclude all@