getsentry · untitaker · Jul 20, 2018 · Jul 20, 2018 · Jul 20, 2018 · Jul 20, 2018
@@ -17,3 +17,4 @@ pip-log.txt
 venv
 .vscode/tags
 .pytest_cache
+.hypothesis
@@ -220,13 +220,7 @@ def skip_internal_frames(frame):
 
 def safe_str(value):
     try:
-        try:
-            rv = text_type(value)
-        except UnicodeError:
-            rv = bytes(value)
-        if isinstance(rv, bytes):
-            rv = rv.decode('utf-8', 'replace')
-        return rv
+        return text_type(value)
     except Exception:
         return safe_repr(value)
 
@@ -236,25 +230,22 @@ def safe_repr(value):
         rv = repr(value)
         if isinstance(rv, bytes):
             rv = rv.decode('utf-8', 'replace')
-        return rv
+        try:
+            return rv.encode('utf-8').decode('unicode-escape')
+        except Exception:
+            return rv
     except Exception:
-        return '<broken repr>'
+        return u'<broken repr>'
 
 
 def object_to_json(obj):
     def _walk(obj, depth):
-        if depth >= 4:
-            return safe_repr(obj)
-        if obj is None or obj is True or obj is False or \
-           isinstance(obj, number_types) or isinstance(obj, text_type):
-            return obj
-        if isinstance(obj, bytes):
-            return obj.decode('utf-8', 'replace')
-        if isinstance(obj, Sequence):
-            return [_walk(x, depth + 1) for x in obj]
-        if isinstance(obj, Mapping):
-            return {safe_repr(k): _walk(v, depth + 1) for k, v in
-                    obj.items()}
+        if depth < 4:
+            if isinstance(obj, Sequence):
+                return [_walk(x, depth + 1) for x in obj]
+            if isinstance(obj, Mapping):
+                return {safe_repr(k): _walk(v, depth + 1) for k, v in
+                        obj.items()}
         return safe_repr(obj)
     return _walk(obj, 0)
 

@@ -0,0 +1,25 @@
+from hypothesis import given, assume, settings
+import hypothesis.strategies as st
+
+from sentry_sdk.utils import safe_repr
+from sentry_sdk._compat import PY2, text_type
+
+any_string = st.one_of(st.binary(), st.text())
+
+@given(x=any_string)
+@settings(max_examples=1000)
+def test_safe_repr_never_broken_for_strings(x):
+    r = safe_repr(x)
+    assert isinstance(r, text_type)
+    assert u'broken repr' not in r
+
+@given(x=any_string)
+@settings(max_examples=1000)
+def test_safe_repr_never_leaves_escapes_in(x):
+    if isinstance(x, bytes):
+        assume(b'\\u' not in x and b'\\x' not in x)
+    else:
+        assume(u'\\u' not in x and u'\\x' not in x)
+    r = safe_repr(x)
+    assert isinstance(r, text_type)
+    assert u'\\u' not in r and u'\\x' not in r
@@ -26,6 +26,7 @@ envlist =
 [testenv]
 deps =
     pytest
+    hypothesis
     django-{16,17,18}: pytest-django<3.0
     django-{19,110,110,111,200,dev}: pytest-django>=3.0,<3.3
     django-{18,19,110}: django-tastypie==0.14