Allow some badchars in the JS code

peepdf/JSAnalysis.py

@@ -190,13 +190,18 @@ def isJavascript(content):
     results = 0
     length = len(content)
     smallScriptLength = 100
+    badChars = 0
 
     if re.findall(reJSscript, content, re.DOTALL | re.IGNORECASE):
         return True
 
     for char in content:
         if (ord(char) < 32 and char not in ['\n', '\r', '\t', '\f', '\x00']) or ord(char) >= 127:
-            return False
+            badChars += 1
+
+    # More than 20% of the content are bad chars
+    if badChars >= len(content) * 0.2:
+        return False
 
     for string in jsStrings:
         cont = content.count(string)

tests/test_pee.py

@@ -25,6 +25,26 @@
 import peepdf.main
 import pytest
 
+def test_js_detect():
+    p = peepdf.PDFCore.PDFParser()
+    r, f = p.parse(
+        "tests/files/js_in_pdf.js", forceMode=True,
+        looseMode=True, manualAnalysis=False
+    )
+    assert not r
+
+    for version in xrange(f.updates + 1):
+        for obj in f.body[version].objects.values():
+            if isinstance(obj, peepdf.PDFCore.PDFIndirectObject):
+                o = obj.getObject()
+                if isinstance(o, peepdf.PDFCore.PDFStream):
+                    stream = o.decodedStream
+                    isJS = peepdf.JSAnalysis.isJavascript(stream)
+                    if "function docOpened()" in stream:
+                        assert isJS
+                    else:
+                        assert not isJS
+
 def test_whitespace_after_opening():
     p = peepdf.PDFCore.PDFParser()
     r, f = p.parse(