From b70395e831dd7efa023c8d77385adda4b26ef1a5 Mon Sep 17 00:00:00 2001
From: Maciej Urbanski <maciej.rooter.urbanski@reef.pl>
Date: Wed, 24 Jan 2024 13:23:22 +0100
Subject: [PATCH] handle invalid B2 server errors

---
 b2sdk/b2http.py                               | 22 +++++++++++---
 .../+handle_invalid_error_format.fixed.md     |  1 +
 test/unit/b2http/test_b2http.py               | 30 +++++++++++++++++++
 3 files changed, 49 insertions(+), 4 deletions(-)
 create mode 100644 changelog.d/+handle_invalid_error_format.fixed.md

diff --git a/b2sdk/b2http.py b/b2sdk/b2http.py
index c5fce9e8..07f87142 100644
--- a/b2sdk/b2http.py
+++ b/b2sdk/b2http.py
@@ -423,7 +423,9 @@ def _translate_errors(cls, fcn, post_params=None):
                 # Decode the error object returned by the service
                 try:
                     error = json.loads(response.content.decode('utf-8')) if response.content else {}
-                except (json.JSONDecodeError, UnicodeDecodeError):
+                    if not isinstance(error, dict):
+                        raise ValueError('json error value is not a dict')
+                except (json.JSONDecodeError, UnicodeDecodeError, ValueError):
                     logger.error('failed to decode error response: %r', response.content)
                     # When the user points to an S3 endpoint, he won't receive the JSON error
                     # he expects. In that case, we can provide at least a hint of "what happened".
@@ -439,10 +441,22 @@ def _translate_errors(cls, fcn, post_params=None):
                     logger.debug(
                         'received error has extra (unsupported) keys: %s', extra_error_keys
                     )
+
+                try:
+                    status = int(error.get('status', response.status_code))
+                    if status != response.status_code:
+                        raise ValueError('status code is not equal to the one in the response')
+                except (TypeError, ValueError) as exc:
+                    logger.warning(
+                        'Inconsistent status codes returned by the server %r != %r; parsing exception: %r',
+                        error.get('status'), response.status_code, exc
+                    )
+                    status = response.status_code
+
                 raise interpret_b2_error(
-                    int(error.get('status', response.status_code)),
-                    error.get('code'),
-                    error.get('message'),
+                    status,
+                    str(error['code']) if 'code' in error else None,
+                    str(error['message']) if 'message' in error else None,
                     response.headers,
                     post_params,
                 )
diff --git a/changelog.d/+handle_invalid_error_format.fixed.md b/changelog.d/+handle_invalid_error_format.fixed.md
new file mode 100644
index 00000000..48893200
--- /dev/null
+++ b/changelog.d/+handle_invalid_error_format.fixed.md
@@ -0,0 +1 @@
+Handle json encoded, invalid B2 error responses, preventing exceptions such as `invalid literal for int() with base 10: 'service_unavailable'`.
\ No newline at end of file
diff --git a/test/unit/b2http/test_b2http.py b/test/unit/b2http/test_b2http.py
index 936247a8..f9240678 100644
--- a/test/unit/b2http/test_b2http.py
+++ b/test/unit/b2http/test_b2http.py
@@ -145,6 +145,36 @@ def test_b2_error__nginx_html():
     assert response.content.decode('utf-8') in str(exc_info.value)
 
 
+def test_b2_error__invalid_error_format():
+    """
+    Handling of invalid error format.
+
+    If server returns valid JSON, but not matching B2 error schema, we should still raise ServiceError.
+    """
+    response = MagicMock()
+    response.status_code = 503
+    # valid JSON, but not a valid B2 error (it should be a dict, not a list)
+    response.content = b'[]'
+    with pytest.raises(ServiceError) as exc_info:
+        B2Http._translate_errors(lambda: response)
+    assert '503' in str(exc_info.value)
+
+
+def test_b2_error__invalid_error_values():
+    """
+    Handling of invalid error values.
+
+    If server returns valid JSON, but not matching B2 error schema, we should still raise ServiceError.
+    """
+    response = MagicMock()
+    response.status_code = 503
+    # valid JSON, but not a valid B2 error (code and status values (and therefore types!) are swapped)
+    response.content = b'{"code": 503, "message": "Service temporarily unavailable", "status": "service_unavailable"}'
+    with pytest.raises(ServiceError) as exc_info:
+        B2Http._translate_errors(lambda: response)
+    assert '503 Service temporarily unavailable' in str(exc_info.value)
+
+
 class TestTranslateAndRetry(TestBase):
     def setUp(self):
         self.response = MagicMock()