Add QPDFObjectHandle::parseContentStream method

This method allows parsing of the PDF objects in a content stream or
array of content streams.

7 files changed, 186 insertions, 2 deletions

diff --git a/qpdf/qpdf.testcov b/qpdf/qpdf.testcov
index a0578f28..b09e966c 100644
--- a/qpdf/qpdf.testcov
+++ b/qpdf/qpdf.testcov
@@ -236,7 +236,7 @@ QPDFWriter copy use_aes 1
 QPDFObjectHandle indirect without context 0
 QPDFObjectHandle trailing data in parse 0
 qpdf pages encryption password 0
-QPDF_Tokenizer EOF reading token 0
+QPDF_Tokenizer EOF reading token 1
 QPDF_Tokenizer EOF reading appendable token 0
 QPDFWriter extra header text no newline 0
 QPDFWriter extra header text add newline 0
@@ -259,3 +259,5 @@ QPDFWriter remove Crypt 0
 qpdf-c called qpdf_get_pdf_extension_level 0
 qpdf-c called qpdf_set_r5_encryption_parameters 0
 qpdf-c called qpdf_set_r6_encryption_parameters 0
+QPDFObjectHandle EOF in inline image 0
+QPDFObjectHandle inline image token 0
diff --git a/qpdf/qtest/qpdf.test b/qpdf/qtest/qpdf.test
index bf62ceea..8d2b5cfc 100644
--- a/qpdf/qtest/qpdf.test
+++ b/qpdf/qtest/qpdf.test
@@ -199,7 +199,7 @@ $td->runtest("remove page we don't have",
 show_ntests();
 # ----------
 $td->notify("--- Miscellaneous Tests ---");
-$n_tests += 57;
+$n_tests += 59;
 
 $td->runtest("qpdf version",
 	     {$td->COMMAND => "qpdf --version"},
@@ -468,6 +468,16 @@ $td->runtest("check file with leading junk",
              {$td->COMMAND => "qpdf --check leading-junk.pdf"},
              {$td->FILE => "leading-junk.out", $td->EXIT_STATUS => 0},
              $td->NORMALIZE_NEWLINES);
+$td->runtest("EOF inside inline image",
+             {$td->COMMAND => "test_driver 37 eof-in-inline-image.pdf"},
+             {$td->FILE => "eof-in-inline-image.out",
+              $td->EXIT_STATUS => 2},
+             $td->NORMALIZE_NEWLINES);
+$td->runtest("tokenize content streams",
+             {$td->COMMAND => "test_driver 37 tokenize-content-streams.pdf"},
+             {$td->FILE => "tokenize-content-streams.out",
+              $td->EXIT_STATUS => 0},
+             $td->NORMALIZE_NEWLINES);
 
 show_ntests();
 # ----------
diff --git a/qpdf/qtest/qpdf/eof-in-inline-image.out b/qpdf/qtest/qpdf/eof-in-inline-image.out
new file mode 100644
index 00000000..8ac365c4
--- /dev/null
+++ b/qpdf/qtest/qpdf/eof-in-inline-image.out
@@ -0,0 +1,25 @@
+BT
+/F1
+24
+Tf
+72
+720
+Td
+(Potato)
+Tj
+ET
+BI
+/CS
+/G
+/W
+1
+/H
+1
+/BPC
+8
+/F
+/Fl
+/DP
+<< /Columns 1 /Predictor 15 >>
+ID
+content stream object 4 0 (stream data, file position 139): EOF found while reading inline image
diff --git a/qpdf/qtest/qpdf/eof-in-inline-image.pdf b/qpdf/qtest/qpdf/eof-in-inline-image.pdf
new file mode 100644
index 00000000..e970b77d
--- /dev/null
+++ b/qpdf/qtest/qpdf/eof-in-inline-image.pdf
diff --git a/qpdf/qtest/qpdf/tokenize-content-streams.out b/qpdf/qtest/qpdf/tokenize-content-streams.out
new file mode 100644
index 00000000..9bc933dc
--- /dev/null
+++ b/qpdf/qtest/qpdf/tokenize-content-streams.out
@@ -0,0 +1,95 @@
+BT
+/F1
+24
+Tf
+72
+720
+Td
+(Potato)
+Tj
+ET
+-EOF-
+0.1
+0
+0
+0.1
+0
+0
+cm
+q
+0
+1.1999
+-1.1999
+0
+121.19
+150.009
+cm
+BI
+/CS
+/G
+/W
+1
+/H
+1
+/BPC
+8
+/F
+/Fl
+/DP
+<< /Columns 1 /Predictor 15 >>
+ID
+inline image: 789c63fc0f0001030101
+EI
+Q
+q
+0
+35.997
+-128.389
+0
+431.964
+7269.02
+cm
+BI
+/CS
+/G
+/W
+30
+/H
+107
+/BPC
+8
+/F
+/Fl
+/DP
+<< /Columns 30 /Predictor 15 >>
+ID
+inline image: 789cedd1a11100300800b1b2ffd06503148283bc8dfcf8af2a306ee352eff2e06318638c31c63b3801627b620a
+EI
+Q
+q
+0
+38.3968
+-93.5922
+0
+431.964
+7567.79
+cm
+BI
+/CS
+/G
+/W
+32
+/H
+78
+/BPC
+8
+/F
+/Fl
+/DP
+<< /Columns 32 /Predictor 15 >>
+ID
+inline image: 789c63fccf801f308e2a185530aa60882a20203faa605401890a0643aa1e5530aa6054010d140000bdd03c13
+EI
+Q
+-EOF-
+test 37 done
diff --git a/qpdf/qtest/qpdf/tokenize-content-streams.pdf b/qpdf/qtest/qpdf/tokenize-content-streams.pdf
new file mode 100644
index 00000000..ea97a6e2
--- /dev/null
+++ b/qpdf/qtest/qpdf/tokenize-content-streams.pdf
diff --git a/qpdf/test_driver.cc b/qpdf/test_driver.cc
index 48017908..cd6aa991 100644
--- a/qpdf/test_driver.cc
+++ b/qpdf/test_driver.cc
@@ -58,6 +58,45 @@ class Provider: public QPDFObjectHandle::StreamDataProvider
     bool bad_length;
 };
 
+class ParserCallbacks: public QPDFObjectHandle::ParserCallbacks
+{
+  public:
+    virtual ~ParserCallbacks()
+    {
+    }
+
+    virtual void handleObject(QPDFObjectHandle);
+    virtual void handleEOF();
+};
+
+void
+ParserCallbacks::handleObject(QPDFObjectHandle obj)
+{
+    if (obj.isInlineImage())
+    {
+        std::string val = obj.getInlineImageValue();
+        std::cout << "inline image: ";
+        char buf[3];
+        buf[2] = '\0';
+        for (size_t i = 0; i < val.length(); ++i)
+        {
+            sprintf(buf, "%02x", (unsigned char)(val[i]));
+            std::cout << buf;
+        }
+        std::cout << std::endl;
+    }
+    else
+    {
+        std::cout << obj.unparse() << std::endl;
+    }
+}
+
+void
+ParserCallbacks::handleEOF()
+{
+    std::cout << "-EOF-" << std::endl;
+}
+
 static std::string getPageContents(QPDFObjectHandle page)
 {
     PointerHolder<Buffer> b1 =
@@ -1245,6 +1284,19 @@ void runtest(int n, char const* filename1, char const* arg2)
             }
         }
     }
+    else if (n == 37)
+    {
+        // Parse content streams of all pages
+        std::vector<QPDFObjectHandle> pages = pdf.getAllPages();
+        for (std::vector<QPDFObjectHandle>::iterator iter = pages.begin();
+             iter != pages.end(); ++iter)
+        {
+            QPDFObjectHandle page = *iter;
+            QPDFObjectHandle contents = page.getKey("/Contents");
+            ParserCallbacks cb;
+            QPDFObjectHandle::parseContentStream(contents, &cb);
+        }
+    }
     else
     {
 	throw std::runtime_error(std::string("invalid test ") +