Merge pull request #1384 from m-holger/ostream

Enhance error checking of object stream object ids and offsets

Merge pull request #1384 from m-holger/ostream
Enhance error checking of object stream object ids and offsets
m-holger · GitHub
2 parents 52e63412 f06209ca
Showing 6 changed files with 45 additions and 3 deletions
include/qpdf/QPDFXRefEntry.hh
libqpdf/QPDF.cc
manual/release-notes.rst
qpdf/qpdf.testcov
qpdf/qtest/qpdf/issue-143.out
qpdf/qtest/qpdf/issue-143.pdf
@@ -62,6 +62,9 @@ class QPDFXRefEntry
   private:
     // This class does not use the Members pattern to avoid a memory allocation for every one of
     // these. A lot of these get created.
+
+    // The layout can be changed to reduce the size from 24 to 16 bytes. However, this would have a
+    // definite runtime cost.
     int type{0};
     qpdf_offset_t field1{0};
     int field2{0};
@@ -25,6 +25,7 @@
 #include <qpdf/Util.hh>
  
 using namespace qpdf;
+using namespace std::literals;
  
 // This must be a fixed value. This API returns a const reference to it, and the C API relies on its
 // being static as well.
@@ -1967,6 +1968,7 @@ QPDF::resolveObjectsInStream(int obj_stream_number)
             (m->file->getName() + " object stream " + std::to_string(obj_stream_number)),
             bp.get()));
  
+    long long last_offset = -1;
     for (int i = 0; i < n; ++i) {
         QPDFTokenizer::Token tnum = readToken(*input);
         QPDFTokenizer::Token toffset = readToken(*input);
@@ -1980,9 +1982,7 @@ QPDF::resolveObjectsInStream(int obj_stream_number)
  
         int num = QUtil::string_to_int(tnum.getValue().c_str());
         long long offset = QUtil::string_to_int(toffset.getValue().c_str());
-        if (num > m->xref_table_max_id) {
-            continue;
-        }
+
         if (num == obj_stream_number) {
             QTC::TC("qpdf", "QPDF ignore self-referential object stream");
             warn(damagedPDF(
@@ -1992,6 +1992,33 @@ QPDF::resolveObjectsInStream(int obj_stream_number)
                 "object stream claims to contain itself"));
             continue;
         }
+
+        if (num < 1) {
+            QTC::TC("qpdf", "QPDF object stream contains id < 1");
+            warn(damagedPDF(
+                *input,
+                "object "s + std::to_string(num) + " 0, offset " + std::to_string(offset),
+                0,
+                "object id is invalid"s));
+            continue;
+        }
+
+        if (offset <= last_offset) {
+            QTC::TC("qpdf", "QPDF object stream offsets not increasing");
+            warn(damagedPDF(
+                *input,
+                "object "s + std::to_string(num) + " 0, offset " + std::to_string(offset),
+                0,
+                "offset is invalid (must be larger than previous offset " +
+                    std::to_string(last_offset) + ")"));
+            continue;
+        }
+        last_offset = offset;
+
+        if (num > m->xref_table_max_id) {
+            continue;
+        }
+
         offsets[num] = toI(offset + first);
     }
  
@@ -81,6 +81,12 @@ more detail.
       copies the file ``check`` to ``out.pdf`` but may in future check ``out.pdf``.
       Use ``qpdf ./check out.pdf`` or ``qpdf -- check out.pdf`` instead.
  
+
+  - Bug fixes
+
+    - In object streams ignore objects with invalid offset. Report objects with invalid
+      id or offset.
+
   - Library Enhancements
  
     - ``QPDFObjectHandle`` supports move construction/assignment.
@@ -4,7 +4,9 @@ QPDF err wrong objid/generation 0
 QPDF check objid 1
 QPDF check generation 1
 QPDF check obj 1
+QPDF object stream offsets not increasing 0
 QPDF ignore self-referential object stream 0
+QPDF object stream contains id < 1 0
 QPDF hint table length indirect 0
 QPDF hint table length direct 0
 QPDF P absent in lindict 0
@@ -14,5 +14,9 @@ WARNING: issue-143.pdf (object 1 0, offset 24): expected dictionary key but foun
 WARNING: issue-143.pdf (object 1 0, offset 21): stream dictionary lacks /Length key
 WARNING: issue-143.pdf (object 1 0, offset 84): attempting to recover stream length
 WARNING: issue-143.pdf (object 1 0, offset 84): recovered stream length: 606
+WARNING: issue-143.pdf object stream 1 (object 0 0, offset 0): object id is invalid
+WARNING: issue-143.pdf object stream 1 (object 0 0, offset 0): object id is invalid
+WARNING: issue-143.pdf object stream 1 (object 6 0, offset 0): offset is invalid (must be larger than previous offset 0)
+WARNING: issue-143.pdf object stream 1 (object 0 0, offset 0): object id is invalid
 WARNING: issue-143.pdf object stream 1 (object 2 0, offset 33): expected dictionary key but found non-name object; inserting key /QPDFFake1
 qpdf: issue-143.pdf: unable to find page tree