Add private method QPDF::validateStreamLineEnd

m-holger
1 parent ef866d68
Showing 2 changed files with 46 additions and 39 deletions
include/qpdf/QPDF.hh
libqpdf/QPDF.cc
@@ -1009,6 +1009,7 @@ class QPDF
     QPDFObjectHandle readTrailer();
     QPDFObjectHandle readObject(std::string const& description, QPDFObjGen og);
     void readStream(QPDFObjectHandle& object, QPDFObjGen og, qpdf_offset_t offset);
+    void validateStreamLineEnd(QPDFObjectHandle& object, QPDFObjGen og, qpdf_offset_t offset);
     QPDFObjectHandle readObjectInStream(std::shared_ptr<InputSource>, QPDFObjGen og);
     size_t recoverStreamLength(
         std::shared_ptr<InputSource> input, QPDFObjGen const& og, qpdf_offset_t stream_offset);
@@ -1324,6 +1324,47 @@ QPDF::readObject(std::string const&amp; description, QPDFObjGen og)
 void
 QPDF::readStream(QPDFObjectHandle& object, QPDFObjGen og, qpdf_offset_t offset)
 {
+    validateStreamLineEnd(object, og, offset);
+
+    // Must get offset before accessing any additional objects since resolving a previously
+    // unresolved indirect object will change file position.
+    qpdf_offset_t stream_offset = m->file->tell();
+    size_t length = 0;
+
+    try {
+        auto length_obj = object.getKey("/Length");
+
+        if (!length_obj.isInteger()) {
+            if (length_obj.isNull()) {
+                QTC::TC("qpdf", "QPDF stream without length");
+                throw damagedPDF(offset, "stream dictionary lacks /Length key");
+            }
+            QTC::TC("qpdf", "QPDF stream length not integer");
+            throw damagedPDF(offset, "/Length key in stream dictionary is not an integer");
+        }
+
+        length = toS(length_obj.getUIntValue());
+        // Seek in two steps to avoid potential integer overflow
+        m->file->seek(stream_offset, SEEK_SET);
+        m->file->seek(toO(length), SEEK_CUR);
+        if (!readToken(m->file).isWord("endstream")) {
+            QTC::TC("qpdf", "QPDF missing endstream");
+            throw damagedPDF("expected endstream");
+        }
+    } catch (QPDFExc& e) {
+        if (m->attempt_recovery) {
+            warn(e);
+            length = recoverStreamLength(m->file, og, stream_offset);
+        } else {
+            throw;
+        }
+    }
+    object = newIndirect(og, QPDF_Stream::create(this, og, object, stream_offset, length));
+}
+
+void
+QPDF::validateStreamLineEnd(QPDFObjectHandle& object, QPDFObjGen og, qpdf_offset_t offset)
+{
     // The PDF specification states that the word "stream" should be followed by either a carriage
     // return and a newline or by a newline alone.  It specifically disallowed following it by a
     // carriage return alone since, in that case, there would be no way to tell whether the NL in a
@@ -1336,12 +1377,12 @@ QPDF::readStream(QPDFObjectHandle&amp; object, QPDFObjGen og, qpdf_offset_t offset)
         if (m->file->read(&ch, 1) == 0) {
             // A premature EOF here will result in some other problem that will get reported at
             // another time.
-            break;
+            return;
         }
         if (ch == '\n') {
             // ready to read stream data
             QTC::TC("qpdf", "QPDF stream with NL only");
-            break;
+            return;
         }
         if (ch == '\r') {
             // Read another character
@@ -1358,52 +1399,17 @@ QPDF::readStream(QPDFObjectHandle&amp; object, QPDFObjGen og, qpdf_offset_t offset)
                         m->file->tell(), "stream keyword followed by carriage return only"));
                 }
             }
-            break;
+            return;
         }
         if (!QUtil::is_space(ch)) {
             QTC::TC("qpdf", "QPDF stream without newline");
             m->file->unreadCh(ch);
             warn(damagedPDF(
                 m->file->tell(), "stream keyword not followed by proper line terminator"));
-            break;
+            return;
         }
         warn(damagedPDF(m->file->tell(), "stream keyword followed by extraneous whitespace"));
     }
-
-    // Must get offset before accessing any additional objects since resolving a previously
-    // unresolved indirect object will change file position.
-    qpdf_offset_t stream_offset = m->file->tell();
-    size_t length = 0;
-
-    try {
-        auto length_obj = object.getKey("/Length");
-
-        if (!length_obj.isInteger()) {
-            if (length_obj.isNull()) {
-                QTC::TC("qpdf", "QPDF stream without length");
-                throw damagedPDF(offset, "stream dictionary lacks /Length key");
-            }
-            QTC::TC("qpdf", "QPDF stream length not integer");
-            throw damagedPDF(offset, "/Length key in stream dictionary is not an integer");
-        }
-
-        length = toS(length_obj.getUIntValue());
-        // Seek in two steps to avoid potential integer overflow
-        m->file->seek(stream_offset, SEEK_SET);
-        m->file->seek(toO(length), SEEK_CUR);
-        if (!readToken(m->file).isWord("endstream")) {
-            QTC::TC("qpdf", "QPDF missing endstream");
-            throw damagedPDF("expected endstream");
-        }
-    } catch (QPDFExc& e) {
-        if (m->attempt_recovery) {
-            warn(e);
-            length = recoverStreamLength(m->file, og, stream_offset);
-        } else {
-            throw;
-        }
-    }
-    object = newIndirect(og, QPDF_Stream::create(this, og, object, stream_offset, length));
 }
 QPDFObjectHandle