setOutputStreams

git-svn-id: svn+q:///qpdf/trunk@1035 71b93d88-0707-0410-a8cf-f5a4172ac649

setOutputStreams
git-svn-id: svn+q:///qpdf/trunk@1035 71b93d88-0707-0410-a8cf-f5a4172ac649
Jay Berkenbilt
1 parent 9f444ffe
Showing 11 changed files with 276 additions and 45 deletions
ChangeLog
TODO
include/qpdf/QPDF.hh
libqpdf/QPDF.cc
libqpdf/QPDF_linearization.cc
manual/qpdf-manual.xml
qpdf/qtest/qpdf.test
qpdf/qtest/qpdf/linearized-and-warnings-1.out
qpdf/qtest/qpdf/linearized-and-warnings-2.out
qpdf/qtest/qpdf/linearized-and-warnings.pdf
qpdf/test_driver.cc
 2010-10-01  Jay Berkenbilt  <ejb@ql.org>
+	* include/qpdf/QPDF.hh: Add setOutputStreams method to allow
+	redirection of library-generated output/error to alternative
+	streams.
+
 	* include/qpdf/QPDF.hh: Add processMemoryFile method for
 	processing a PDF file from a memory buffer instead of a file.
-General
-=======
+2.2.1
+=====
+
+ * Remove cerr and cout from QPDF*.cc
  * QPDF::checkLinearization writes things to std::cout, which makes it
    hard for GUIs that want to display the result.  Go through all
@@ -12,6 +14,10 @@ General
    rethrow if needed.  The old version should call the ostream version
    with std::cout.
+
+General
+=======
+
  * In general, take a fresh look at private methods to see which, if
    any, should be protected.
@@ -12,6 +12,7 @@
 #include <string>
 #include <map>
 #include <list>
+#include <iostream>
 #include <qpdf/DLL.h>
@@ -61,6 +62,21 @@ class QPDF
     // Parameter settings
+    // By default, warning messages are issued to std::cerr and output
+    // messages printed by certain check calls are issued to
+    // std::cout.  This method allows you to specify alternative
+    // streams for this purpose.  Note that no normal QPDF operations
+    // generate output to std::cout, so for applications that just
+    // wish to avoid creating output and don't call any check
+    // functions, calling setSuppressWarnings(true) is sufficient.
+    // Applications that wish to present check or warning information
+    // to users may replace the output and error streams to capture
+    // the output and errors for other use.  A null value for either
+    // stream will cause QPDF to use std::cout or std::cerr as
+    // appropriate.
+    QPDF_DLL
+    void setOutputStreams(std::ostream* out_stream, std::ostream* err_stream);
+
     // If true, ignore any cross-reference streams in a hybrid file
     // (one that contains both cross-reference streams and
     // cross-reference tables).  This can be useful for testing to
@@ -68,7 +84,8 @@ class QPDF
     QPDF_DLL
     void setIgnoreXRefStreams(bool);
-    // By default, any warnings are issued to stderr as they are
+    // By default, any warnings are issued to std::cerr or the error
+    // stream specified in a call to setOutputStreams as they are
     // encountered.  If this is called with a true value, reporting of
     // warnings is suppressed.  You may still retrieve warnings by
     // calling getWarnings.
@@ -88,7 +105,7 @@ class QPDF
     // clear the list.  This method may be called even if processFile
     // throws an exception.  Note that if setSuppressWarnings was not
     // called or was called with a false value, any warnings retrieved
-    // here will have already been issued to stderr.
+    // here will have already been output.
     QPDF_DLL
     std::vector<QPDFExc> getWarnings();
@@ -204,12 +221,14 @@ class QPDF
     // Performs various sanity checks on a linearized file.  Return
     // true if no errors or warnings.  Otherwise, return false and
-    // output errors and warnings to stdout.
+    // output errors and warnings to std::cout or the output stream
+    // specified in a call to setOutputStreams.
     QPDF_DLL
     bool checkLinearization();
     // Calls checkLinearization() and, if possible, prints normalized
-    // contents of some of the hints tables to stdout.  Normalization
+    // contents of some of the hints tables to std::cout or the output
+    // stream specified in a call to setOutputStreams.  Normalization
     // includes adding min values to delta values and adjusting
     // offsets based on the location and size of the primary hint
     // stream.
@@ -803,6 +822,8 @@ class QPDF
     bool encryption_initialized;
     bool ignore_xref_streams;
     bool suppress_warnings;
+    std::ostream* out_stream;
+    std::ostream* err_stream;
     bool attempt_recovery;
     int encryption_V;
     bool encrypt_metadata;
@@ -267,6 +267,8 @@ QPDF::QPDF() :
     encryption_initialized(false),
     ignore_xref_streams(false),
     suppress_warnings(false),
+    out_stream(&std::cout),
+    err_stream(&std::cerr),
     attempt_recovery(true),
     encryption_V(0),
     encrypt_metadata(true),
@@ -333,6 +335,13 @@ QPDF::setIgnoreXRefStreams(bool val)
 }
 void
+QPDF::setOutputStreams(std::ostream* out, std::ostream* err)
+{
+    this->out_stream = out ? out : &std::cout;
+    this->err_stream = err ? err : &std::cerr;
+}
+
+void
 QPDF::setSuppressWarnings(bool val)
 {
     this->suppress_warnings = val;
@@ -449,7 +458,8 @@ QPDF::warn(QPDFExc const&amp; e)
     this->warnings.push_back(e);
     if (! this->suppress_warnings)
     {
-	std::cerr << "WARNING: " << this->warnings.back().what() << std::endl;
+	*err_stream << "WARNING: "
+		    << this->warnings.back().what() << std::endl;
     }
 }
@@ -1045,16 +1055,16 @@ QPDF::showXRefTable()
     {
 	ObjGen const& og = (*iter).first;
 	QPDFXRefEntry const& entry = (*iter).second;
-	std::cout << og.obj << "/" << og.gen << ": ";
+	*out_stream << og.obj << "/" << og.gen << ": ";
 	switch (entry.getType())
 	{
 	  case 1:
-	    std::cout << "uncompressed; offset = " << entry.getOffset();
+	    *out_stream << "uncompressed; offset = " << entry.getOffset();
 	    break;
 	  case 2:
-	    std::cout << "compressed; stream = " << entry.getObjStreamNumber()
-		      << ", index = " << entry.getObjStreamIndex();
+	    *out_stream << "compressed; stream = " << entry.getObjStreamNumber()
+			<< ", index = " << entry.getObjStreamIndex();
 	    break;
 	  default:
@@ -1062,7 +1072,7 @@ QPDF::showXRefTable()
 				   " showing xref_table");
 	    break;
 	}
-	std::cout << std::endl;
+	*out_stream << std::endl;
     }
 }
@@ -64,7 +64,7 @@ QPDF::checkLinearization()
     }
     catch (QPDFExc& e)
     {
-	std::cout << e.what() << std::endl;
+	*out_stream << e.what() << std::endl;
     }
     return result;
 }
@@ -351,9 +351,9 @@ QPDF::readHintStream(Pipeline&amp; pl, off_t offset, size_t length)
     if ((computed_end < min_end_offset) ||
 	(computed_end > max_end_offset))
     {
-	std::cout << "expected = " << computed_end
-		  << "; actual = " << min_end_offset << ".."
-		  << max_end_offset << std::endl;
+	*out_stream << "expected = " << computed_end
+		    << "; actual = " << min_end_offset << ".."
+		    << max_end_offset << std::endl;
 	throw QPDFExc(qpdf_e_damaged_pdf, this->file->getName(),
 		      "linearization dictionary",
 		      this->file->getLastOffset(),
@@ -617,7 +617,7 @@ QPDF::checkLinearizationInternal()
 	for (std::list<std::string>::iterator iter = errors.begin();
 	     iter != errors.end(); ++iter)
 	{
-	    std::cout << "ERROR: " << (*iter) << std::endl;
+	    *out_stream << "ERROR: " << (*iter) << std::endl;
 	}
     }
@@ -627,7 +627,7 @@ QPDF::checkLinearizationInternal()
 	for (std::list<std::string>::iterator iter = warnings.begin();
 	     iter != warnings.end(); ++iter)
 	{
-	    std::cout << "WARNING: " << (*iter) << std::endl;
+	    *out_stream << "WARNING: " << (*iter) << std::endl;
 	}
     }
@@ -1008,17 +1008,17 @@ QPDF::showLinearizationData()
     }
     catch (QPDFExc& e)
     {
-	std::cout << e.what() << std::endl;
+	*out_stream << e.what() << std::endl;
     }
 }
 void
 QPDF::dumpLinearizationDataInternal()
 {
-    std::cout << this->file->getName() << ": linearization data:" << std::endl
-	      << std::endl;
+    *out_stream << this->file->getName() << ": linearization data:" << std::endl
+		<< std::endl;
-    std::cout
+    *out_stream
 	<< "file_size: " << this->linp.file_size << std::endl
 	<< "first_page_object: " << this->linp.first_page_object << std::endl
 	<< "first_page_end: " << this->linp.first_page_end << std::endl
@@ -1029,19 +1029,19 @@ QPDF::dumpLinearizationDataInternal()
 	<< "H_length: " << this->linp.H_length << std::endl
 	<< std::endl;
-    std::cout << "Page Offsets Hint Table" << std::endl
-	      << std::endl;
+    *out_stream << "Page Offsets Hint Table" << std::endl
+		<< std::endl;
     dumpHPageOffset();
-    std::cout << std::endl
-	      << "Shared Objects Hint Table" << std::endl
-	      << std::endl;
+    *out_stream << std::endl
+		<< "Shared Objects Hint Table" << std::endl
+		<< std::endl;
     dumpHSharedObject();
     if (this->outline_hints.nobjects > 0)
     {
-	std::cout << std::endl
-		  << "Outlines Hint Table" << std::endl
-		  << std::endl;
+	*out_stream << std::endl
+		    << "Outlines Hint Table" << std::endl
+		    << std::endl;
 	dumpHGeneric(this->outline_hints);
     }
 }
@@ -1064,7 +1064,7 @@ void
 QPDF::dumpHPageOffset()
 {
     HPageOffset& t = this->page_offset_hints;
-    std::cout
+    *out_stream
 	<< "min_nobjects: " << t.min_nobjects
 	<< std::endl
 	<< "first_page_offset: " << adjusted_offset(t.first_page_offset)
@@ -1095,7 +1095,7 @@ QPDF::dumpHPageOffset()
     for (int i1 = 0; i1 < this->linp.npages; ++i1)
     {
 	HPageOffsetEntry& pe = t.entries[i1];
-	std::cout
+	*out_stream
 	    << "Page " << i1 << ":" << std::endl
 	    << "  nobjects: " << pe.delta_nobjects + t.min_nobjects
 	    << std::endl
@@ -1109,10 +1109,10 @@ QPDF::dumpHPageOffset()
 	    << "  nshared_objects: " << pe.nshared_objects << std::endl;
 	for (int i2 = 0; i2 < pe.nshared_objects; ++i2)
 	{
-	    std::cout << "    identifier " << i2 << ": "
-		      << pe.shared_identifiers[i2] << std::endl;
-	    std::cout << "    numerator " << i2 << ": "
-		      << pe.shared_numerators[i2] << std::endl;
+	    *out_stream << "    identifier " << i2 << ": "
+			<< pe.shared_identifiers[i2] << std::endl;
+	    *out_stream << "    numerator " << i2 << ": "
+			<< pe.shared_numerators[i2] << std::endl;
 	}
     }
 }
@@ -1121,7 +1121,7 @@ void
 QPDF::dumpHSharedObject()
 {
     HSharedObject& t = this->shared_object_hints;
-    std::cout
+    *out_stream
 	<< "first_shared_obj: " << t.first_shared_obj
 	<< std::endl
 	<< "first_shared_offset: " << adjusted_offset(t.first_shared_offset)
@@ -1140,19 +1140,19 @@ QPDF::dumpHSharedObject()
     for (int i = 0; i < t.nshared_total; ++i)
     {
 	HSharedObjectEntry& se = t.entries[i];
-	std::cout << "Shared Object " << i << ":" << std::endl;
-	std::cout << "  group length: "
-		  << se.delta_group_length + t.min_group_length << std::endl;
+	*out_stream << "Shared Object " << i << ":" << std::endl;
+	*out_stream << "  group length: "
+		    << se.delta_group_length + t.min_group_length << std::endl;
 	// PDF spec says signature present nobjects_minus_one are
 	// always 0, so print them only if they have a non-zero value.
 	if (se.signature_present)
 	{
-	    std::cout << "  signature present" << std::endl;
+	    *out_stream << "  signature present" << std::endl;
 	}
 	if (se.nobjects_minus_one != 0)
 	{
-	    std::cout << "  nobjects: "
-		      << se.nobjects_minus_one + 1 << std::endl;
+	    *out_stream << "  nobjects: "
+			<< se.nobjects_minus_one + 1 << std::endl;
 	}
     }
 }
@@ -1160,7 +1160,7 @@ QPDF::dumpHSharedObject()
 void
 QPDF::dumpHGeneric(HGeneric& t)
 {
-    std::cout
+    *out_stream
 	<< "first_object: " << t.first_object
 	<< std::endl
 	<< "first_object_offset: " << adjusted_offset(t.first_object_offset)
@@ -2075,6 +2075,61 @@ print &quot;\n&quot;;
   </para>
   <variablelist>
    <varlistentry>
+    <term>2.2.1: XXX</term>
+    <listitem>
+     <itemizedlist>
+      <listitem>
+       <para>
+        Add new method <classname>QPDF::processMemoryFile</classname>
+        for operating on PDF files that are loaded into memory rather
+        than in a file on disk.
+       </para>
+      </listitem>
+      <listitem>
+       <para>
+        Add new method <classname>QPDF::setOutputStreams</classname>
+        to replace <varname>std::cout</varname> and
+        <varname>std::cerr</varname> with other streams for generation
+        of diagnostic messages and error messages.  This can be useful
+        for GUIs or other applications that want to capture any output
+        generated by the library to present to the user in some other
+        way.  Note that QPDF does not write to
+        <varname>std::cout</varname> (or the specified output stream)
+        except where explicitly mentioned in
+        <filename>QPDF.hh</filename>, and that the only use of the
+        error stream is for warnings.  Note also that output of
+        warnings is suppressed when
+        <literal>setSuppressWarnings(true)</literal> is called.
+       </para>
+      </listitem>
+      <listitem>
+       <para>
+        Give a warning but otherwise ignore empty PDF objects by
+        treating them as null.  Empty object are not permitted by the
+        PDF specification but have been known to appear in some actual
+        PDF files.
+       </para>
+      </listitem>
+      <listitem>
+       <para>
+        Handle inline image filter abbreviations when the appear as
+        stream filter abbreviations.  The PDF specification does not
+        allow use of stream filter abbreviations in this way, but
+        Adobe Reader and some other PDF readers accept them since they
+        sometimes appear incorrectly in actual PDF files.
+       </para>
+      </listitem>
+      <listitem>
+       <para>
+        Implement miscellaneous enhancements to
+        <classname>PointerHolder</classname> and
+        <classname>Buffer</classname> to support other changes.
+       </para>
+      </listitem>
+     </itemizedlist>
+    </listitem>
+   </varlistentry>
+   <varlistentry>
     <term>2.2.0: August 14, 2010</term>
     <listitem>
      <itemizedlist>
@@ -111,7 +111,7 @@ $td-&gt;runtest(&quot;new stream&quot;,
 show_ntests();
 # ----------
 $td->notify("--- Miscellaneous Tests ---");
-$n_tests += 26;
+$n_tests += 28;
 $td->runtest("qpdf version",
 	     {$td->COMMAND => "qpdf --version"},
@@ -245,6 +245,18 @@ $td-&gt;runtest(&quot;empty object&quot;,
 	      $td->EXIT_STATUS => 3},
 	     $td->NORMALIZE_NEWLINES);
+$td->runtest("error/output redirection to null",
+	     {$td->COMMAND => "test_driver 12 linearized-and-warnings.pdf"},
+	     {$td->FILE => "linearized-and-warnings-1.out",
+	      $td->EXIT_STATUS => 0},
+	     $td->NORMALIZE_NEWLINES);
+
+$td->runtest("error/output redirection to strings",
+	     {$td->COMMAND => "test_driver 13 linearized-and-warnings.pdf"},
+	     {$td->FILE => "linearized-and-warnings-2.out",
+	      $td->EXIT_STATUS => 0},
+	     $td->NORMALIZE_NEWLINES);
+
 show_ntests();
 # ----------
 $td->notify("--- Error Condition Tests ---");
+WARNING: linearized-and-warnings.pdf (object 2 0, file position 1117): empty object treated as null
+linearized-and-warnings.pdf: linearization data:
+
+file_size: 1310
+first_page_object: 6
+first_page_end: 1044
+npages: 1
+xref_zero_offset: 1132
+first_page: 0
+H_offset: 528
+H_length: 118
+
+Page Offsets Hint Table
+
+min_nobjects: 4
+first_page_offset: 646
+nbits_delta_nobjects: 0
+min_page_length: 398
+nbits_delta_page_length: 0
+min_content_offset: 0
+nbits_delta_content_offset: 0
+min_content_length: 398
+nbits_delta_content_length: 0
+nbits_nshared_objects: 0
+nbits_shared_identifier: 3
+nbits_shared_numerator: 0
+shared_denominator: 4
+Page 0:
+  nobjects: 4
+  length: 398
+  content_offset: 0
+  content_length: 398
+  nshared_objects: 0
+
+Shared Objects Hint Table
+
+first_shared_obj: 0
+first_shared_offset: 0
+nshared_first_page: 4
+nshared_total: 4
+nbits_nobjects: 0
+min_group_length: 30
+nbits_delta_group_length: 7
+Shared Object 0:
+  group length: 143
+Shared Object 1:
+  group length: 118
+Shared Object 2:
+  group length: 30
+Shared Object 3:
+  group length: 107
+test 12 done
+---output---
+linearized-and-warnings.pdf: linearization data:
+
+file_size: 1310
+first_page_object: 6
+first_page_end: 1044
+npages: 1
+xref_zero_offset: 1132
+first_page: 0
+H_offset: 528
+H_length: 118
+
+Page Offsets Hint Table
+
+min_nobjects: 4
+first_page_offset: 646
+nbits_delta_nobjects: 0
+min_page_length: 398
+nbits_delta_page_length: 0
+min_content_offset: 0
+nbits_delta_content_offset: 0
+min_content_length: 398
+nbits_delta_content_length: 0
+nbits_nshared_objects: 0
+nbits_shared_identifier: 3
+nbits_shared_numerator: 0
+shared_denominator: 4
+Page 0:
+  nobjects: 4
+  length: 398
+  content_offset: 0
+  content_length: 398
+  nshared_objects: 0
+
+Shared Objects Hint Table
+
+first_shared_obj: 0
+first_shared_offset: 0
+nshared_first_page: 4
+nshared_total: 4
+nbits_nobjects: 0
+min_group_length: 30
+nbits_delta_group_length: 7
+Shared Object 0:
+  group length: 143
+Shared Object 1:
+  group length: 118
+Shared Object 2:
+  group length: 30
+Shared Object 3:
+  group length: 107
+---error---
+WARNING: linearized-and-warnings.pdf (object 2 0, file position 1117): empty object treated as null
+test 13 done
@@ -10,6 +10,7 @@
 #include <qpdf/Pl_Flate.hh>
 #include <qpdf/QPDFWriter.hh>
 #include <iostream>
+#include <sstream>
 #include <stdio.h>
 #include <string.h>
 #include <stdlib.h>
@@ -502,6 +503,22 @@ void runtest(int n, char const* filename)
 	    std::cout << "raw stream data okay" << std::endl;
 	}
     }
+    else if (n == 12)
+    {
+	pdf.setOutputStreams(0, 0);
+	pdf.showLinearizationData();
+    }
+    else if (n == 13)
+    {
+	std::ostringstream out;
+	std::ostringstream err;
+	pdf.setOutputStreams(&out, &err);
+	pdf.showLinearizationData();
+	std::cout << "---output---" << std::endl
+		  << out.str()
+		  << "---error---" << std::endl
+		  << err.str();
+    }
     else
     {
 	throw std::runtime_error(std::string("invalid test ") +