Handle parse error stream data (fixes #1123)

A parse error in stream data in which stream data contained a nested object would cause a crash because qpdf was not correctly updating its internal state. Rework the QPDF json reactor to not be sensitive to parse errors in this way.
qpdf · Feb 4, 2024 · cb0f390 · cb0f390
1 parent 7caa9dd
commit cb0f390
Show file tree

Hide file tree

Showing 14 changed files with 311 additions and 151 deletions.
diff --git a/libqpdf/QPDF_json.cc b/libqpdf/QPDF_json.cc
diff --git a/qpdf/qpdf.testcov b/qpdf/qpdf.testcov
@@ -668,7 +668,6 @@ QPDF_json value stream both or neither 0
 QPDFJob need json-stream-prefix for stdout 0
 QPDFJob write json to stdout 0
 QPDFJob write json to file 0
-QPDF_json don't check object after parse error 0
 QPDF_json ignoring unknown top-level key 0
 QPDF_json ignore second-level key 0
 QPDF_json ignore unknown key in object_top 0
@@ -694,3 +693,6 @@ QPDFJob misplaced page range 0
 QPDFJob duplicated range 0
 QPDFJob json over/under no file 0
 QPDF_Array copy 1
+QPDF_json stream data not string 0
+QPDF_json stream datafile not string 0
+QPDF_json stream not a dictionary 0
diff --git a/qpdf/qtest/qpdf-json.test b/qpdf/qtest/qpdf-json.test
@@ -37,6 +37,8 @@ my @badfiles = (
     'obj-key-errors',
     'bad-data',
     'bad-datafile',
+    'bad-data2',
+    'bad-datafile2',
     );
 
 $n_tests += scalar(@badfiles);

diff --git a/qpdf/qtest/qpdf/qjson-bad-data2.json b/qpdf/qtest/qpdf/qjson-bad-data2.json
@@ -0,0 +1,71 @@
+{
+  "qpdf": [
+    {
+      "jsonversion": 2,
+      "pdfversion": "1.3",
+      "maxobjectid": 6
+    },
+    {
+      "obj:1 0 R": {
+        "value": {
+          "/Pages": "2 0 R",
+          "/Type": "/Catalog"
+        }
+      },
+      "obj:2 0 R": {
+        "value": {
+          "/Count": 1,
+          "/Kids": [
+            "3 0 R"
+          ],
+          "/Type": "/Pages"
+        }
+      },
+      "obj:3 0 R": {
+        "value": {
+          "/Contents": ["4 0 R", "7 0 R"],
+          "/MediaBox": [
+            0,
+            0,
+            612,
+            792
+          ],
+          "/Parent": "2 0 R",
+          "/Resources": {
+            "/Font": {
+              "/F1": "6 0 R"
+            },
+            "/ProcSet": "5 0 R"
+          },
+          "/Type": "/Page"
+        }
+      },
+      "obj:4 0 R": {
+        "stream": {
+          "data": [[]],
+          "dict": {}
+        }
+      },
+      "obj:5 0 R": {
+        "value": [
+          "/PDF",
+          "/Text"
+        ]
+      },
+      "obj:6 0 R": {
+        "value": {
+          "/BaseFont": "/Helvetica",
+          "/Encoding": "/WinAnsiEncoding",
+          "/Subtype": "/Type1",
+          "/Type": "/Font"
+        }
+      },
+      "trailer": {
+        "value": {
+          "/Root": "1 0 R",
+          "/Size": 7
+        }
+      }
+    }
+  ]
+}
diff --git a/qpdf/qtest/qpdf/qjson-bad-data2.out b/qpdf/qtest/qpdf/qjson-bad-data2.out
@@ -0,0 +1,2 @@
+WARNING: qjson-bad-data2.json (obj:4 0 R, offset 846): "stream.data" must be a string
+qpdf: qjson-bad-data2.json: errors found in JSON
diff --git a/qpdf/qtest/qpdf/qjson-bad-datafile2.json b/qpdf/qtest/qpdf/qjson-bad-datafile2.json
@@ -0,0 +1,71 @@
+{
+  "qpdf": [
+    {
+      "jsonversion": 2,
+      "pdfversion": "1.3",
+      "maxobjectid": 6
+    },
+    {
+      "obj:1 0 R": {
+        "value": {
+          "/Pages": "2 0 R",
+          "/Type": "/Catalog"
+        }
+      },
+      "obj:2 0 R": {
+        "value": {
+          "/Count": 1,
+          "/Kids": [
+            "3 0 R"
+          ],
+          "/Type": "/Pages"
+        }
+      },
+      "obj:3 0 R": {
+        "value": {
+          "/Contents": ["4 0 R", "7 0 R"],
+          "/MediaBox": [
+            0,
+            0,
+            612,
+            792
+          ],
+          "/Parent": "2 0 R",
+          "/Resources": {
+            "/Font": {
+              "/F1": "6 0 R"
+            },
+            "/ProcSet": "5 0 R"
+          },
+          "/Type": "/Page"
+        }
+      },
+      "obj:4 0 R": {
+        "stream": {
+          "datafile": [[]],
+          "dict": {}
+        }
+      },
+      "obj:5 0 R": {
+        "value": [
+          "/PDF",
+          "/Text"
+        ]
+      },
+      "obj:6 0 R": {
+        "value": {
+          "/BaseFont": "/Helvetica",
+          "/Encoding": "/WinAnsiEncoding",
+          "/Subtype": "/Type1",
+          "/Type": "/Font"
+        }
+      },
+      "trailer": {
+        "value": {
+          "/Root": "1 0 R",
+          "/Size": 7
+        }
+      }
+    }
+  ]
+}
diff --git a/qpdf/qtest/qpdf/qjson-bad-datafile2.out b/qpdf/qtest/qpdf/qjson-bad-datafile2.out
@@ -0,0 +1,2 @@
+WARNING: qjson-bad-datafile2.json (obj:4 0 R, offset 850): "stream.datafile" must be a string containing a file name
+qpdf: qjson-bad-datafile2.json: errors found in JSON
diff --git a/qpdf/qtest/qpdf/qjson-bad-pdf-version1.out b/qpdf/qtest/qpdf/qjson-bad-pdf-version1.out
@@ -1,3 +1,3 @@
-WARNING: qjson-bad-pdf-version1.json (offset 41): invalid JSON version (must be 2)
-WARNING: qjson-bad-pdf-version1.json (offset 70): invalid PDF version (must be x.y)
+WARNING: qjson-bad-pdf-version1.json (offset 41): invalid JSON version (must be numeric value 2)
+WARNING: qjson-bad-pdf-version1.json (offset 70): invalid PDF version (must be "x.y")
 qpdf: qjson-bad-pdf-version1.json: errors found in JSON
diff --git a/qpdf/qtest/qpdf/qjson-bad-pdf-version2.out b/qpdf/qtest/qpdf/qjson-bad-pdf-version2.out
@@ -1,5 +1,5 @@
-WARNING: qjson-bad-pdf-version2.json (offset 41): invalid JSON version (must be 2)
-WARNING: qjson-bad-pdf-version2.json (offset 66): invalid PDF version (must be x.y)
+WARNING: qjson-bad-pdf-version2.json (offset 41): invalid JSON version (must be numeric value 2)
+WARNING: qjson-bad-pdf-version2.json (offset 66): invalid PDF version (must be "x.y")
 WARNING: qjson-bad-pdf-version2.json (offset 97): calledgetallpages must be a boolean
 WARNING: qjson-bad-pdf-version2.json (offset 138): pushedinheritedpageresources must be a boolean
 qpdf: qjson-bad-pdf-version2.json: errors found in JSON
diff --git a/qpdf/qtest/qpdf/qjson-obj-key-errors.out b/qpdf/qtest/qpdf/qjson-obj-key-errors.out
@@ -1,7 +1,7 @@
 WARNING: qjson-obj-key-errors.json (obj:2 0 R, offset 244): object must have exactly one of "value" or "stream"
 WARNING: qjson-obj-key-errors.json (obj:3 0 R, offset 542): object must have exactly one of "value" or "stream"
-WARNING: qjson-obj-key-errors.json (obj:4 0 R, offset 710): "stream" is missing "dict"
-WARNING: qjson-obj-key-errors.json (obj:4 0 R, offset 710): new "stream" must have exactly one of "data" or "datafile"
-WARNING: qjson-obj-key-errors.json (obj:5 0 R, offset 800): new "stream" must have exactly one of "data" or "datafile"
+WARNING: qjson-obj-key-errors.json (obj:4 0 R, offset 690): "stream" is missing "dict"
+WARNING: qjson-obj-key-errors.json (obj:4 0 R, offset 690): new "stream" must have exactly one of "data" or "datafile"
+WARNING: qjson-obj-key-errors.json (obj:5 0 R, offset 780): new "stream" must have exactly one of "data" or "datafile"
 WARNING: qjson-obj-key-errors.json (trailer, offset 1178): "trailer" is missing "value"
 qpdf: qjson-obj-key-errors.json: errors found in JSON
diff --git a/qpdf/qtest/qpdf/qjson-stream-dict-not-dict.out b/qpdf/qtest/qpdf/qjson-stream-dict-not-dict.out
@@ -1,5 +1,4 @@
 WARNING: qjson-stream-dict-not-dict.json (obj:1 0 R, offset 142): "stream.dict" must be a dictionary
-WARNING: qjson-stream-dict-not-dict.json (obj:1 0 R, offset 142): unrecognized string value
-WARNING: qjson-stream-dict-not-dict.json (obj:1 0 R, offset 122): new "stream" must have exactly one of "data" or "datafile"
+WARNING: qjson-stream-dict-not-dict.json (obj:1 0 R, offset 102): new "stream" must have exactly one of "data" or "datafile"
 WARNING: qjson-stream-dict-not-dict.json: "qpdf[1].trailer" was not seen
 qpdf: qjson-stream-dict-not-dict.json: errors found in JSON
diff --git a/qpdf/qtest/qpdf/qjson-stream-not-dict.out b/qpdf/qtest/qpdf/qjson-stream-not-dict.out
@@ -1,3 +1,4 @@
 WARNING: qjson-stream-not-dict.json (obj:1 0 R, offset 122): "stream" must be a dictionary
+WARNING: qjson-stream-not-dict.json (obj:1 0 R, offset 102): "stream" is missing "dict"
 WARNING: qjson-stream-not-dict.json: "qpdf[1].trailer" was not seen
 qpdf: qjson-stream-not-dict.json: errors found in JSON
diff --git a/qpdf/qtest/qpdf/qjson-trailer-stream.out b/qpdf/qtest/qpdf/qjson-trailer-stream.out
@@ -1,2 +1,3 @@
 WARNING: qjson-trailer-stream.json (trailer, offset 1269): the trailer may not be a stream
+WARNING: qjson-trailer-stream.json (trailer, offset 1249): "trailer" is missing "value"
 qpdf: qjson-trailer-stream.json: errors found in JSON
diff --git a/qpdf/qtest/qpdf/update-from-json-errors.out b/qpdf/qtest/qpdf/update-from-json-errors.out
@@ -1,4 +1,4 @@
-WARNING: good13.pdf (obj:4 0 R from qpdf-json-update-errors.json, offset 95): existing "stream" may at most one of "data" or "datafile"
+WARNING: good13.pdf (obj:4 0 R from qpdf-json-update-errors.json, offset 75): existing "stream" may at most one of "data" or "datafile"
 WARNING: good13.pdf (obj:20 0 R from qpdf-json-update-errors.json, offset 335): unrecognized string value
-WARNING: good13.pdf (obj:20 0 R from qpdf-json-update-errors.json, offset 293): new "stream" must have exactly one of "data" or "datafile"
+WARNING: good13.pdf (obj:20 0 R from qpdf-json-update-errors.json, offset 273): new "stream" must have exactly one of "data" or "datafile"
 qpdf: qpdf-json-update-errors.json: errors found in JSON