Convert many more errors to warnings

Jay Berkenbilt
1 parent a4fd4b91
Showing 32 changed files with 205 additions and 70 deletions
examples/qtest/bookmarks.test
examples/qtest/npages.test
libqpdf/QPDF.cc
qpdf/qpdf.testcov
qpdf/qtest/qpdf.test
qpdf/qtest/qpdf/bad1-recover.out
qpdf/qtest/qpdf/bad1.out
qpdf/qtest/qpdf/bad22-recover.out
qpdf/qtest/qpdf/bad22.out
qpdf/qtest/qpdf/bad23-recover.out
qpdf/qtest/qpdf/bad23.out
qpdf/qtest/qpdf/bad24-recover.out
qpdf/qtest/qpdf/bad24.out
qpdf/qtest/qpdf/bad25.out
qpdf/qtest/qpdf/bad26.out
qpdf/qtest/qpdf/bad27.out
qpdf/qtest/qpdf/bad32.out
qpdf/qtest/qpdf/bad34.out
qpdf/qtest/qpdf/bad35-recover.out
qpdf/qtest/qpdf/bad35.out
@@ -29,8 +29,8 @@ $td-&gt;runtest(&quot;no bookmarks&quot;,
 $td->runtest("bad",
 	     {$td->COMMAND => "pdf-bookmarks 3.pdf"},
-	     {$td->STRING => "pdf-bookmarks processing file 3.pdf: " .
-		  "3.pdf: not a PDF file\n",
+	     {$td->REGEXP => "pdf-bookmarks processing file 3.pdf: " .
+		  ".*unable to find trailer.*",
 		  $td->EXIT_STATUS => 2},
 	     $td->NORMALIZE_NEWLINES);
@@ -16,7 +16,7 @@ $td-&gt;runtest(&quot;normal&quot;,
 $td->runtest("error",
 	     {$td->COMMAND => "pdf-npages bad"},
-	     {$td->STRING => "pdf-npages: bad: not a PDF file\n",
+	     {$td->REGEXP => "pdf-npages: bad: unable to find trailer.*",
 	      $td->EXIT_STATUS => 2},
 	     $td->NORMALIZE_NEWLINES);
@@ -205,7 +205,7 @@ QPDF::getWarnings()
 void
 QPDF::parse(char const* password)
 {
-    PCRE header_re("\\A((?s).*?)%PDF-(1.\d+)\b");
+    PCRE header_re("\\A((?s).*?)%PDF-(\\d+.\d+)\b");
     PCRE eof_re("(?s:startxref\\s+(\\d+)\\s+%%EOF\\b)");
     if (password)
@@ -233,16 +233,17 @@ QPDF::parse(char const* password)
             this->file = new OffsetInputSource(this->file, global_offset);
         }
 	this->pdf_version = m1.getMatch(2);
-	if (atof(this->pdf_version.c_str()) < 1.2)
-	{
-	    this->tokenizer.allowPoundAnywhereInName();
-	}
     }
     else
     {
 	QTC::TC("qpdf", "QPDF not a pdf file");
-	throw QPDFExc(qpdf_e_damaged_pdf, this->file->getName(),
-		      "", 0, "not a PDF file");
+	warn(QPDFExc(qpdf_e_damaged_pdf, this->file->getName(),
+                     "", 0, "can't find PDF header"));
+        this->pdf_version = "1.0";
+    }
+    if (atof(this->pdf_version.c_str()) < 1.2)
+    {
+        this->tokenizer.allowPoundAnywhereInName();
     }
     // PDF spec says %%EOF must be found within the last 1024 bytes of
@@ -1152,7 +1153,7 @@ QPDF::readObject(PointerHolder&lt;InputSource&gt; input,
             {
                 if (this->attempt_recovery)
                 {
-                    // may throw an exception
+                    warn(e);
                     length = recoverStreamLength(
                         input, objid, generation, stream_offset);
                 }
@@ -1288,9 +1289,9 @@ QPDF::recoverStreamLength(PointerHolder&lt;InputSource&gt; input,
     if (length == 0)
     {
-	throw QPDFExc(qpdf_e_damaged_pdf, input->getName(),
-		      this->last_object_description, stream_offset,
-		      "unable to recover stream data");
+        warn(QPDFExc(qpdf_e_damaged_pdf, input->getName(),
+                     this->last_object_description, stream_offset,
+                     "unable to recover stream data; treating stream as empty"));
     }
     QTC::TC("qpdf", "QPDF recovered stream length");
@@ -1309,6 +1310,10 @@ QPDF::readObjectAtOffset(bool try_recovery,
 			 int exp_objid, int exp_generation,
 			 int& objid, int& generation)
 {
+    if (! this->attempt_recovery)
+    {
+        try_recovery = false;
+    }
     setLastObjectDescription(description, exp_objid, exp_generation);
     // Special case: if offset is 0, just return null.  Some PDF
@@ -1363,16 +1368,27 @@ QPDF::readObjectAtOffset(bool try_recovery,
 	    (! ((objid == exp_objid) && (generation == exp_generation))))
 	{
 	    QTC::TC("qpdf", "QPDF err wrong objid/generation");
-	    throw QPDFExc(qpdf_e_damaged_pdf, this->file->getName(),
-			  this->last_object_description, offset,
-			  std::string("expected ") +
-			  QUtil::int_to_string(exp_objid) + " " +
-			  QUtil::int_to_string(exp_generation) + " obj");
+	    QPDFExc e(qpdf_e_damaged_pdf, this->file->getName(),
+                      this->last_object_description, offset,
+                      std::string("expected ") +
+                      QUtil::int_to_string(exp_objid) + " " +
+                      QUtil::int_to_string(exp_generation) + " obj");
+            if (try_recovery)
+            {
+                // Will be retried below
+                throw e;
+            }
+            else
+            {
+                // We can try reading the object anyway even if the ID
+                // doesn't match.
+                warn(e);
+            }
 	}
     }
     catch (QPDFExc& e)
     {
-	if ((exp_objid >= 0) && try_recovery && this->attempt_recovery)
+	if ((exp_objid >= 0) && try_recovery)
 	{
 	    // Try again after reconstructing xref table
 	    reconstruct_xref(e);
@@ -1496,31 +1512,42 @@ QPDF::resolve(int objid, int generation)
 	}
 	QPDFXRefEntry const& entry = this->xref_table[og];
-	switch (entry.getType())
-	{
-	  case 1:
-	    {
-		qpdf_offset_t offset = entry.getOffset();
-		// Object stored in cache by readObjectAtOffset
-		int aobjid;
-		int ageneration;
-		QPDFObjectHandle oh =
-		    readObjectAtOffset(true, offset, "", objid, generation,
-				       aobjid, ageneration);
-	    }
-	    break;
+        try
+        {
+            switch (entry.getType())
+            {
+              case 1:
+                {
+                    qpdf_offset_t offset = entry.getOffset();
+                    // Object stored in cache by readObjectAtOffset
+                    int aobjid;
+                    int ageneration;
+                    QPDFObjectHandle oh =
+                        readObjectAtOffset(true, offset, "", objid, generation,
+                                           aobjid, ageneration);
+                }
+                break;
-	  case 2:
-	    resolveObjectsInStream(entry.getObjStreamNumber());
-	    break;
+              case 2:
+                resolveObjectsInStream(entry.getObjStreamNumber());
+                break;
-	  default:
-	    throw QPDFExc(qpdf_e_damaged_pdf, this->file->getName(), "", 0,
-			  "object " +
-			  QUtil::int_to_string(objid) + "/" +
-			  QUtil::int_to_string(generation) +
-			  " has unexpected xref entry type");
-	}
+              default:
+                throw QPDFExc(qpdf_e_damaged_pdf, this->file->getName(), "", 0,
+                              "object " +
+                              QUtil::int_to_string(objid) + "/" +
+                              QUtil::int_to_string(generation) +
+                              " has unexpected xref entry type");
+            }
+        }
+        catch (QPDFExc& e)
+        {
+            QTC::TC("qpdf", "QPDF resolve failure to null");
+            warn(e);
+            QPDFObjectHandle oh = QPDFObjectHandle::newNull();
+            this->obj_cache[og] =
+                ObjCache(QPDFObjectHandle::ObjAccessor::getObject(oh), -1, -1);
+        }
     }
     return this->obj_cache[og].object;
@@ -278,3 +278,4 @@ QPDF recursion loop in resolve 0
 QPDFObjectHandle treat word as string 0
 QPDFObjectHandle found fake 1
 QPDFObjectHandle no val for last key 0
+QPDF resolve failure to null 0
@@ -220,7 +220,7 @@ $td-&gt;runtest(&quot;C API: qpdf version&quot;,
 # Files to reproduce various bugs
 foreach my $d (
-    ["51", "resolve loop", 2],
+    ["51", "resolve loop", 3],
     ["99", "object 0", 2],
     ["99b", "object 0", 2],
     ["100", "xref reconstruction loop", 2],
@@ -228,7 +228,7 @@ foreach my $d (
     ["117", "other infinite loop", 2],
     ["118", "other infinite loop", 2],
     ["119", "other infinite loop", 3],
-    ["120", "other infinite loop", 2],
+    ["120", "other infinite loop", 3],
     )
 {
     my ($n, $description, $exit_status) = @$d;
@@ -464,7 +464,7 @@ $td-&gt;runtest(&quot;EOF terminating literal tokens&quot;,
              $td->NORMALIZE_NEWLINES);
 $td->runtest("EOF reading token",
              {$td->COMMAND => "qpdf --check eof-reading-token.pdf"},
-             {$td->FILE => "eof-reading-token.out", $td->EXIT_STATUS => 2},
+             {$td->FILE => "eof-reading-token.out", $td->EXIT_STATUS => 3},
              $td->NORMALIZE_NEWLINES);
 $td->runtest("extra header text",
              {$td->COMMAND => "test_driver 32 minimal.pdf"},
@@ -794,9 +794,12 @@ $n_tests += @badfiles + 3;
 # neither Acrobat nor other PDF viewers really care.  Tests 12 and 28
 # have error conditions that used to be fatal but are now considered
 # non-fatal.
-my %badtest_overrides = (6 => 0, 12 => 0, 13 => 0,
-                         14 => 0, 15 => 0, 17 => 0,
-                         28 => 0, 30 => 0, 31 => 0, 36 => 0);
+my %badtest_overrides = ();
+for(6, 12..15, 17, 22..28, 30..32, 34, 36)
+{
+    $badtest_overrides{$_} = 0;
+}
+
 for (my $i = 1; $i <= scalar(@badfiles); ++$i)
 {
     my $status = $badtest_overrides{$i};
@@ -835,7 +838,7 @@ $n_tests += @badfiles + 8;
 # though in some cases it may.  Acrobat Reader would not be able to
 # recover any of these files any better.
 my %recover_failures = ();
-for (1, 7, 16, 18..21, 24, 29, 35)
+for (1, 7, 16, 18..21, 29, 35)
 {
     $recover_failures{$_} = 1;
 }
-bad1.pdf: not a PDF file
+WARNING: bad1.pdf: can't find PDF header
+WARNING: bad1.pdf: file is damaged
+WARNING: bad1.pdf: can't find startxref
+WARNING: bad1.pdf: Attempting to reconstruct cross-reference table
+bad1.pdf: unable to find trailer dictionary while recovering damaged file
-bad1.pdf: not a PDF file
+WARNING: bad1.pdf: can't find PDF header
+bad1.pdf: can't find startxref
+WARNING: bad22.pdf (object 4 0, file position 314): stream dictionary lacks /Length key
 WARNING: bad22.pdf (object 4 0, file position 341): attempting to recover stream length
 /QTest is indirect and has type stream (10)
 /QTest is a stream.  Dictionary: << /Qength 44 >>
-bad22.pdf (object 4 0, file position 314): stream dictionary lacks /Length key
+WARNING: bad22.pdf (object 4 0, file position 314): stream dictionary lacks /Length key
+/QTest is implicit
+/QTest is indirect and has type null (2)
+/QTest is null
+unparse: 4 0 R
+unparseResolved: null
+test 0 done
+WARNING: bad23.pdf (object 4 0, file position 314): /Length key in stream dictionary is not an integer
 WARNING: bad23.pdf (object 4 0, file position 341): attempting to recover stream length
 /QTest is indirect and has type stream (10)
 /QTest is a stream.  Dictionary: << /Length () >>
-bad23.pdf (object 4 0, file position 314): /Length key in stream dictionary is not an integer
+WARNING: bad23.pdf (object 4 0, file position 314): /Length key in stream dictionary is not an integer
+/QTest is implicit
+/QTest is indirect and has type null (2)
+/QTest is null
+unparse: 4 0 R
+unparseResolved: null
+test 0 done
+WARNING: bad24.pdf (object 4 0, file position 385): expected endstream
 WARNING: bad24.pdf (object 4 0, file position 341): attempting to recover stream length
-bad24.pdf (object 4 0, file position 341): unable to recover stream data
+WARNING: bad24.pdf (object 4 0, file position 341): unable to recover stream data; treating stream as empty
+WARNING: bad24.pdf (object 4 0, file position 778): EOF while reading token
+/QTest is implicit
+/QTest is indirect and has type null (2)
+/QTest is null
+unparse: 4 0 R
+unparseResolved: null
+test 1 done
-bad24.pdf (object 4 0, file position 385): expected endstream
+WARNING: bad24.pdf (object 4 0, file position 385): expected endstream
+/QTest is implicit
+/QTest is indirect and has type null (2)
+/QTest is null
+unparse: 4 0 R
+unparseResolved: null
+test 0 done
-bad25.pdf (object 4 0, file position 307): expected n n obj
+WARNING: bad25.pdf (object 4 0, file position 307): expected n n obj
+/QTest is implicit
+/QTest is indirect and has type null (2)
+/QTest is null
+unparse: 4 0 R
+unparseResolved: null
+test 0 done
-bad26.pdf (object 4 0, file position 307): expected n n obj
+WARNING: bad26.pdf (object 4 0, file position 307): expected n n obj
+/QTest is implicit
+/QTest is indirect and has type null (2)
+/QTest is null
+unparse: 4 0 R
+unparseResolved: null
+test 0 done
-bad27.pdf (object 4 0, file position 307): expected n n obj
+WARNING: bad27.pdf (object 4 0, file position 307): expected n n obj
+/QTest is implicit
+/QTest is indirect and has type null (2)
+/QTest is null
+unparse: 4 0 R
+unparseResolved: null
+test 0 done
-bad32.pdf (object 4 0, file position 307): expected 4 0 obj
+WARNING: bad32.pdf (object 4 0, file position 307): expected 4 0 obj
+/QTest is implicit
+/QTest is indirect and has type null (2)
+/QTest is null
+unparse: 4 0 R
+unparseResolved: null
+test 0 done
-bad34.pdf (object 4 0, file position 322): expected n n obj
+WARNING: bad34.pdf (object 4 0, file position 322): expected n n obj
+/QTest is implicit
+/QTest is indirect and has type null (2)
+/QTest is null
+unparse: 4 0 R
+unparseResolved: null
+test 0 done
-bad35.pdf (object 1 0, file position 521): supposed object stream 1 has wrong type
+WARNING: bad35.pdf (object 1 0, file position 521): supposed object stream 1 has wrong type
+operation for Dictionary object attempted on object of wrong type
-bad35.pdf (object 1 0, file position 521): supposed object stream 1 has wrong type
+WARNING: bad35.pdf (object 1 0, file position 521): supposed object stream 1 has wrong type
+operation for Dictionary object attempted on object of wrong type
-error: bad1.pdf: not a PDF file
+warning: bad1.pdf: can't find PDF header
   code: 5
   file: bad1.pdf
   pos : 0
-  text: not a PDF file
+  text: can't find PDF header
+warning: bad1.pdf: file is damaged
+  code: 5
+  file: bad1.pdf
+  pos : 0
+  text: file is damaged
+warning: bad1.pdf: can't find startxref
+  code: 5
+  file: bad1.pdf
+  pos : 0
+  text: can't find startxref
+warning: bad1.pdf: Attempting to reconstruct cross-reference table
+  code: 5
+  file: bad1.pdf
+  pos : 0
+  text: Attempting to reconstruct cross-reference table
+error: bad1.pdf: unable to find trailer dictionary while recovering damaged file
+  code: 5
+  file: bad1.pdf
+  pos : 0
+  text: unable to find trailer dictionary while recovering damaged file
@@ -2,4 +2,4 @@ checking eof-reading-token.pdf
 PDF Version: 1.3
 File is not encrypted
 File is not linearized
-object stream 12 (file position 5): EOF while reading token
+WARNING: object stream 12 (file position 5): EOF while reading token
 WARNING: heifer.pdf: file is damaged
 WARNING: heifer.pdf (file position 92741): xref not found
 WARNING: heifer.pdf: Attempting to reconstruct cross-reference table
+WARNING: heifer.pdf (object 2 0, file position 2165): expected endstream
 WARNING: heifer.pdf (object 2 0, file position 51): attempting to recover stream length
 qpdf: operation succeeded with warnings; resulting file may have some problems
@@ -7,6 +7,8 @@ WARNING: issue-100.pdf (file position 289): unknown token while reading object; 
 WARNING: issue-100.pdf (file position 294): unknown token while reading object; treating as string
 WARNING: issue-100.pdf (file position 297): unknown token while reading object; treating as string
 WARNING: issue-100.pdf (file position 304): unknown token while reading object; treating as string
+WARNING: issue-100.pdf (file position 308): unexpected )
+WARNING: issue-100.pdf (object 5 0, file position 418): /Length key in stream dictionary is not an integer
 WARNING: issue-100.pdf (object 5 0, file position 489): attempting to recover stream length
 WARNING: issue-100.pdf (trailer, file position 953): expected dictionary key but found non-name object; inserting key /QPDFFake1
 WARNING: issue-100.pdf (trailer, file position 953): dictionary ended prematurely; using null as value for last key
@@ -3,15 +3,22 @@ WARNING: issue-101.pdf (file position 3526): xref not found
 WARNING: issue-101.pdf: Attempting to reconstruct cross-reference table
 WARNING: issue-101.pdf (file position 1242): expected dictionary key but found non-name object; inserting key /QPDFFake1
 WARNING: issue-101.pdf (file position 1242): dictionary ended prematurely; using null as value for last key
+WARNING: issue-101.pdf (object 5 0, file position 1438): /Length key in stream dictionary is not an integer
 WARNING: issue-101.pdf (object 5 0, file position 1509): attempting to recover stream length
+WARNING: issue-101.pdf (trailer, file position 2026): /Length key in stream dictionary is not an integer
 WARNING: issue-101.pdf (trailer, file position 2097): attempting to recover stream length
 WARNING: issue-101.pdf (trailer, file position 2928): unknown token while reading object; treating as string
 WARNING: issue-101.pdf (trailer, file position 2930): unknown token while reading object; treating as string
 WARNING: issue-101.pdf (trailer, file position 2928): expected dictionary key but found non-name object; inserting key /QPDFFake1
 WARNING: issue-101.pdf (trailer, file position 2928): expected dictionary key but found non-name object; inserting key /QPDFFake2
 WARNING: issue-101.pdf (trailer, file position 2928): expected dictionary key but found non-name object; inserting key /QPDFFake3
+WARNING: issue-101.pdf (trailer, file position 2925): /Length key in stream dictionary is not an integer
 WARNING: issue-101.pdf (trailer, file position 2996): attempting to recover stream length
+WARNING: issue-101.pdf (trailer, file position 3339): /Length key in stream dictionary is not an integer
 WARNING: issue-101.pdf (trailer, file position 3410): attempting to recover stream length
+WARNING: issue-101.pdf (trailer, file position 3560): /Length key in stream dictionary is not an integer
 WARNING: issue-101.pdf (trailer, file position 3631): attempting to recover stream length
+WARNING: issue-101.pdf (trailer, file position 4113): /Length key in stream dictionary is not an integer
 WARNING: issue-101.pdf (trailer, file position 4184): attempting to recover stream length
-issue-101.pdf (trailer, file position 4184): unable to recover stream data
+WARNING: issue-101.pdf (trailer, file position 4184): unable to recover stream data; treating stream as empty
+issue-101.pdf: unable to find trailer dictionary while recovering damaged file
@@ -2,5 +2,6 @@ WARNING: issue-117.pdf: file is damaged
 WARNING: issue-117.pdf: can't find startxref
 WARNING: issue-117.pdf: Attempting to reconstruct cross-reference table
 WARNING: issue-117.pdf (file position 66): loop detected resolving object 2 0
+WARNING: issue-117.pdf (object 2 0, file position 22): /Length key in stream dictionary is not an integer
 WARNING: issue-117.pdf (object 2 0, file position 67): attempting to recover stream length
 attempt to make a stream into a direct object
 WARNING: issue-118.pdf (file position 732): loop detected resolving object 2 0
-issue-118.pdf (xref stream: object 8 0, file position 732): supposed object stream 2 is not a stream
+WARNING: issue-118.pdf (xref stream: object 8 0, file position 732): supposed object stream 2 is not a stream
+operation for Dictionary object attempted on object of wrong type
 WARNING: issue-120.pdf (file position 85): loop detected resolving object 3 0
-issue-120.pdf (object 6 0, file position 85): supposed object stream 3 is not a stream
+WARNING: issue-120.pdf (object 6 0, file position 85): supposed object stream 3 is not a stream
+qpdf: operation succeeded with warnings; resulting file may have some problems
@@ -2,5 +2,8 @@ WARNING: issue-51.pdf: reported number of objects (0) inconsistent with actual n
 WARNING: issue-51.pdf (object 7 0, file position 553): expected endobj
 WARNING: issue-51.pdf (object 1 0, file position 359): expected endobj
 WARNING: issue-51.pdf (file position 70): loop detected resolving object 2 0
+WARNING: issue-51.pdf (object 2 0, file position 26): /Length key in stream dictionary is not an integer
 WARNING: issue-51.pdf (object 2 0, file position 71): attempting to recover stream length
-issue-51.pdf (object 2 0, file position 71): unable to recover stream data
+WARNING: issue-51.pdf (object 2 0, file position 71): unable to recover stream data; treating stream as empty
+WARNING: issue-51.pdf (object 2 0, file position 977): EOF while reading token
+qpdf: operation succeeded with warnings; resulting file may have some problems
@@ -2,5 +2,6 @@ checking linearization-bounds-1.pdf
 PDF Version: 1.3
 File is not encrypted
 File is linearized
+WARNING: linearization-bounds-1.pdf (linearization hint stream: object 62 0, file position 1001182): EOF while reading token
 WARNING: linearization-bounds-1.pdf (linearization hint stream: object 62 0, file position 1183): attempting to recover stream length
 linearization-bounds-1.pdf (linearization hint table, file position 1183): /S (shared object) offset is out of bounds
@@ -2,5 +2,6 @@ checking linearization-bounds-2.pdf
 PDF Version: 1.3
 File is not encrypted
 File is linearized
+WARNING: linearization-bounds-2.pdf (linearization hint stream: object 62 0, file position 1282): expected endstream
 WARNING: linearization-bounds-2.pdf (linearization hint stream: object 62 0, file position 1183): attempting to recover stream length
 linearization-bounds-2.pdf (linearization hint table, file position 1183): /S (shared object) offset is out of bounds
@@ -2,5 +2,6 @@ checking linearization-large-vector-alloc.pdf
 PDF Version: 1.3
 File is not encrypted
 File is linearized
+WARNING: linearization-large-vector-alloc.pdf (linearization hint stream: object 62 0, file position 1282): expected endstream
 WARNING: linearization-large-vector-alloc.pdf (linearization hint stream: object 62 0, file position 1183): attempting to recover stream length
 overflow reading bit stream