Make sure to dereference MediaBox in /Pages

CHANGELOG.md

@@ -22,6 +22,7 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/).
 - `RecursionError` when corrupt PDF specifies a recursive /Pages object ([#998](https://github.com/pdfminer/pdfminer.six/pull/998))
 - `TypeError` when corrupt PDF specifies text-positioning operators with invalid values ([#1000](https://github.com/pdfminer/pdfminer.six/pull/1000))
 - inline image parsing fails when stream data contains "EI\n" ([#1008](https://github.com/pdfminer/pdfminer.six/issues/1008))
+- `TypeError` when MediaBox is an indirect object reference ([#1004](https://github.com/pdfminer/pdfminer.six/issues/1004))
 
 ### Removed
 

pdfminer/pdfpage.py

@@ -68,7 +68,8 @@ def __init__(
             self.attrs.get("Resources", dict()),
         )
         mediabox_params: List[Any] = [
-            resolve1(mediabox_param) for mediabox_param in self.attrs["MediaBox"]
+            resolve1(mediabox_param)
+            for mediabox_param in resolve1(self.attrs["MediaBox"])
         ]
         self.mediabox = parse_rect(resolve1(mediabox_params))
         self.cropbox = self.mediabox

tests/test_tools_pdf2txt.py

@@ -118,6 +118,10 @@ def test_html_simple1(self):
     def test_hocr_simple1(self):
         run("simple1.pdf", "-t hocr")
 
+    def test_contrib_issue_1004_mediabox(self):
+        """Verify that we do not crash with MediaBox is an object reference"""
+        run("contrib/issue-1004-indirect-mediabox.pdf")
+
 
 class TestDumpImages:
     @staticmethod