def test_possible_type_mismatch(self):
data, content_type = heuristic_decode(self.json_body, 'application/x-www-form-urlencoded')
assert data == {'key': 'value', 'key2': 'value2'}
assert content_type == 'application/json'
data, content_type = heuristic_decode(self.url_body, 'application/json')
assert data == {'key': ['value'], 'key2': ['value2']}
assert content_type == 'application/x-www-form-urlencoded'
def test_possible_type_mismatch(self):
data, content_type = heuristic_decode(self.json_body, "application/x-www-form-urlencoded")
assert data == {"key": "value", "key2": "value2"}
assert content_type == "application/json"
data, content_type = heuristic_decode(self.url_body, "application/json")
assert data == {"key": ["value"], "key2": ["value2"]}
assert content_type == "application/x-www-form-urlencoded"
def test_possible_type_mismatch(self):
data, content_type = heuristic_decode(
self.json_body, 'application/x-www-form-urlencoded')
assert data == {'key': 'value', 'key2': 'value2'}
assert content_type == 'application/json'
data, content_type = heuristic_decode(self.url_body,
'application/json')
assert data == {'key': ['value'], 'key2': ['value2']}
assert content_type == 'application/x-www-form-urlencoded'
def to_python(cls, data):
is_valid, errors = validate_and_default_interface(data, cls.path)
if not is_valid:
raise InterfaceValidationError("Invalid interface data")
kwargs = {}
if data.get('method'):
method = data['method'].upper()
# Optimize for the common path here, where it's a GET/POST, falling
# back to a regular expresion test
if method not in ('GET',
'POST') and not http_method_re.match(method):
raise InterfaceValidationError("Invalid value for 'method'")
kwargs['method'] = method
else:
kwargs['method'] = None
scheme, netloc, path, query_bit, fragment_bit = urlsplit(data['url'])
query_string = data.get('query_string') or query_bit
if query_string:
# if querystring was a dict, convert it to a string
if isinstance(query_string, dict):
query_string = urlencode([(to_bytes(k), to_bytes(v))
for k, v in query_string.items()])
else:
if query_string[0] == '?':
# remove '?' prefix
query_string = query_string[1:]
kwargs['query_string'] = trim(query_string, 4096)
else:
kwargs['query_string'] = ''
fragment = data.get('fragment') or fragment_bit
cookies = data.get('cookies')
# if cookies were [also] included in headers we
# strip them out
headers = data.get('headers')
if headers:
headers, cookie_header = format_headers(headers)
if not cookies and cookie_header:
cookies = cookie_header
else:
headers = ()
# We prefer the body to be a string, since we can then attempt to parse it
# as JSON OR decode it as a URL encoded query string, without relying on
# the correct content type header being passed.
body = data.get('data')
content_type = next((v for k, v in headers if k == 'Content-Type'),
None)
# Remove content type parameters
if content_type is not None:
content_type = content_type.partition(';')[0].rstrip()
# We process request data once during ingestion and again when
# requesting the http interface over the API. Avoid overwriting
# decoding the body again.
inferred_content_type = data.get('inferred_content_type', content_type)
if 'inferred_content_type' not in data and not isinstance(body, dict):
body, inferred_content_type = heuristic_decode(body, content_type)
if body:
body = trim(body, settings.SENTRY_MAX_HTTP_BODY_SIZE)
env = data.get('env', {})
# TODO (alex) This could also be accomplished with schema (with formats)
if 'REMOTE_ADDR' in env:
try:
validate_ip(env['REMOTE_ADDR'], required=False)
except ValueError:
del env['REMOTE_ADDR']
kwargs['inferred_content_type'] = inferred_content_type
kwargs['cookies'] = trim_pairs(format_cookies(cookies))
kwargs['env'] = trim_dict(env)
kwargs['headers'] = trim_pairs(headers)
kwargs['data'] = fix_broken_encoding(body)
kwargs['url'] = urlunsplit((scheme, netloc, path, '', ''))
kwargs['fragment'] = trim(fragment, 1024)
return cls(**kwargs)
def to_python(cls, data, rust_renormalized=RUST_RENORMALIZED_DEFAULT):
if rust_renormalized:
data.setdefault('query_string', [])
for key in (
"method",
"url",
"fragment",
"cookies",
"headers",
"data",
"env",
"inferred_content_type",
):
data.setdefault(key, None)
return cls(**data)
is_valid, errors = validate_and_default_interface(data, cls.path)
if not is_valid:
raise InterfaceValidationError("Invalid interface data")
kwargs = {}
if data.get('method'):
method = data['method'].upper()
# Optimize for the common path here, where it's a GET/POST, falling
# back to a regular expresion test
if method not in ('GET',
'POST') and not http_method_re.match(method):
raise InterfaceValidationError("Invalid value for 'method'")
kwargs['method'] = method
else:
kwargs['method'] = None
if data.get('url', None):
url = to_unicode(data['url'])
# The JavaScript SDK used to send an ellipsis character for
# truncated URLs. Canonical URLs do not contain UTF-8 characters in
# either the path, query string or fragment, so we replace it with
# three dots (which is the behavior of other SDKs). This effectively
# makes the string two characters longer, but it will be trimmed
# again down below.
if url.endswith(u"\u2026"):
url = url[:-1] + "..."
scheme, netloc, path, query_bit, fragment_bit = urlsplit(url)
else:
scheme = netloc = path = query_bit = fragment_bit = None
query_string = data.get('query_string') or query_bit
if query_string:
if isinstance(query_string, six.string_types):
if query_string[0] == '?':
query_string = query_string[1:]
if query_string.endswith(u"\u2026"):
query_string = query_string[:-1] + "..."
query_string = [
(to_unicode(k), jsonify(v))
for k, v in parse_qsl(query_string, keep_blank_values=True)
]
elif isinstance(query_string, dict):
query_string = [(to_unicode(k), jsonify(v))
for k, v in six.iteritems(query_string)]
elif isinstance(query_string, list):
query_string = [
tuple(tup) for tup in query_string
if isinstance(tup, (tuple, list)) and len(tup) == 2
]
else:
query_string = []
kwargs['query_string'] = trim(query_string, 4096)
else:
kwargs['query_string'] = []
fragment = data.get('fragment') or fragment_bit
cookies = data.get('cookies')
# if cookies were [also] included in headers we
# strip them out
if data.get("headers"):
headers, cookie_header = format_headers(
get_path(data, "headers", filter=True))
if not cookies and cookie_header:
cookies = cookie_header
else:
headers = ()
# We prefer the body to be a string, since we can then attempt to parse it
# as JSON OR decode it as a URL encoded query string, without relying on
# the correct content type header being passed.
body = data.get('data')
content_type = next((v for k, v in headers if k == 'Content-Type'),
None)
# Remove content type parameters
if content_type is not None:
content_type = content_type.partition(';')[0].rstrip()
# We process request data once during ingestion and again when
# requesting the http interface over the API. Avoid overwriting
# decoding the body again.
inferred_content_type = data.get('inferred_content_type', content_type)
if 'inferred_content_type' not in data and not isinstance(body, dict):
body, inferred_content_type = heuristic_decode(body, content_type)
if body:
body = trim(body, settings.SENTRY_MAX_HTTP_BODY_SIZE)
env = data.get('env', {})
# TODO (alex) This could also be accomplished with schema (with formats)
if 'REMOTE_ADDR' in env:
try:
validate_ip(env['REMOTE_ADDR'], required=False)
except ValueError:
del env['REMOTE_ADDR']
kwargs['inferred_content_type'] = inferred_content_type
kwargs['cookies'] = trim_pairs(format_cookies(cookies))
kwargs['env'] = trim_dict(env)
kwargs['headers'] = trim_pairs(headers)
kwargs['data'] = fix_broken_encoding(body)
kwargs['url'] = urlunsplit((scheme, netloc, path, '', ''))
kwargs['fragment'] = trim(fragment, 1024)
return cls(**kwargs)
def test_unable_to_decode(self):
data, content_type = heuristic_decode('string body', 'text/plain')
assert data == 'string body'
assert content_type == 'text/plain'
def test_no_possible_type(self):
data, content_type = heuristic_decode(self.json_body)
assert data == {'key': 'value', 'key2': 'value2'}
assert content_type == 'application/json'
def test_url_encoded(self):
data, content_type = heuristic_decode(
self.url_body, 'application/x-www-form-urlencoded')
assert data == {'key': ['value'], 'key2': ['value2']}
assert content_type == 'application/x-www-form-urlencoded'
def test_unable_to_decode(self):
data, content_type = heuristic_decode("string body", "text/plain")
assert data == "string body"
assert content_type == "text/plain"
def test_no_possible_type(self):
data, content_type = heuristic_decode(self.json_body)
assert data == {"key": "value", "key2": "value2"}
assert content_type == "application/json"
def test_url_encoded(self):
data, content_type = heuristic_decode(
self.url_body, "application/x-www-form-urlencoded")
assert data == {"key": ["value"], "key2": ["value2"]}
assert content_type == "application/x-www-form-urlencoded"
def test_unable_to_decode(self):
data, content_type = heuristic_decode('string body', 'text/plain')
assert data == 'string body'
assert content_type == 'text/plain'
def test_no_possible_type(self):
data, content_type = heuristic_decode(self.json_body)
assert data == {'key': 'value', 'key2': 'value2'}
assert content_type == 'application/json'
def test_url_encoded(self):
data, content_type = heuristic_decode(self.url_body, 'application/x-www-form-urlencoded')
assert data == {'key': ['value'], 'key2': ['value2']}
assert content_type == 'application/x-www-form-urlencoded'
def to_python(cls, data):
if not data.get('url'):
raise InterfaceValidationError("No value for 'url'")
kwargs = {}
if data.get('method'):
method = data['method'].upper()
# Optimize for the common path here, where it's a GET/POST, falling
# back to a regular expresion test
if method not in ('GET', 'POST') and not http_method_re.match(method):
raise InterfaceValidationError("Invalid value for 'method'")
kwargs['method'] = method
else:
kwargs['method'] = None
scheme, netloc, path, query_bit, fragment_bit = urlsplit(data['url'])
query_string = data.get('query_string') or query_bit
if query_string:
# if querystring was a dict, convert it to a string
if isinstance(query_string, dict):
query_string = urlencode(
[(to_bytes(k), to_bytes(v)) for k, v in query_string.items()]
)
else:
query_string = query_string
if query_string[0] == '?':
# remove '?' prefix
query_string = query_string[1:]
kwargs['query_string'] = trim(query_string, 4096)
else:
kwargs['query_string'] = ''
fragment = data.get('fragment') or fragment_bit
cookies = data.get('cookies')
# if cookies were [also] included in headers we
# strip them out
headers = data.get('headers')
if headers:
headers, cookie_header = format_headers(headers)
if not cookies and cookie_header:
cookies = cookie_header
else:
headers = ()
# We prefer the body to be a string, since we can then attempt to parse it
# as JSON OR decode it as a URL encoded query string, without relying on
# the correct content type header being passed.
body = data.get('data')
content_type = next((v for k, v in headers if k == 'Content-Type'), None)
# Remove content type parameters
if content_type is not None:
content_type = content_type.partition(';')[0].rstrip()
# We process request data once during ingestion and again when
# requesting the http interface over the API. Avoid overwriting
# decoding the body again.
inferred_content_type = data.get('inferred_content_type', content_type)
if 'inferred_content_type' not in data and not isinstance(body, dict):
body, inferred_content_type = heuristic_decode(body, content_type)
if body:
body = trim(body, settings.SENTRY_MAX_HTTP_BODY_SIZE)
kwargs['inferred_content_type'] = inferred_content_type
kwargs['cookies'] = trim_pairs(format_cookies(cookies))
kwargs['env'] = trim_dict(data.get('env') or {})
kwargs['headers'] = trim_pairs(headers)
kwargs['data'] = fix_broken_encoding(body)
kwargs['url'] = urlunsplit((scheme, netloc, path, '', ''))
kwargs['fragment'] = trim(fragment, 1024)
return cls(**kwargs)
def to_python(cls, data, rust_renormalized=RUST_RENORMALIZED_DEFAULT):
if rust_renormalized:
data.setdefault('query_string', [])
for key in (
"method",
"url",
"fragment",
"cookies",
"headers",
"data",
"env",
"inferred_content_type",
):
data.setdefault(key, None)
return cls(**data)
is_valid, errors = validate_and_default_interface(data, cls.path)
if not is_valid:
raise InterfaceValidationError("Invalid interface data")
kwargs = {}
if data.get('method'):
method = data['method'].upper()
# Optimize for the common path here, where it's a GET/POST, falling
# back to a regular expresion test
if method not in ('GET', 'POST') and not http_method_re.match(method):
raise InterfaceValidationError("Invalid value for 'method'")
kwargs['method'] = method
else:
kwargs['method'] = None
if data.get('url', None):
url = to_unicode(data['url'])
# The JavaScript SDK used to send an ellipsis character for
# truncated URLs. Canonical URLs do not contain UTF-8 characters in
# either the path, query string or fragment, so we replace it with
# three dots (which is the behavior of other SDKs). This effectively
# makes the string two characters longer, but it will be trimmed
# again down below.
if url.endswith(u"\u2026"):
url = url[:-1] + "..."
scheme, netloc, path, query_bit, fragment_bit = urlsplit(url)
else:
scheme = netloc = path = query_bit = fragment_bit = None
query_string = data.get('query_string') or query_bit
if query_string:
if isinstance(query_string, six.string_types):
if query_string[0] == '?':
query_string = query_string[1:]
if query_string.endswith(u"\u2026"):
query_string = query_string[:-1] + "..."
query_string = [
(to_unicode(k), jsonify(v))
for k, v in parse_qsl(query_string, keep_blank_values=True)
]
elif isinstance(query_string, dict):
query_string = [(to_unicode(k), jsonify(v)) for k, v in six.iteritems(query_string)]
elif isinstance(query_string, list):
query_string = [
tuple(tup) for tup in query_string
if isinstance(tup, (tuple, list)) and len(tup) == 2
]
else:
query_string = []
kwargs['query_string'] = trim(query_string, 4096)
else:
kwargs['query_string'] = []
fragment = data.get('fragment') or fragment_bit
cookies = data.get('cookies')
# if cookies were [also] included in headers we
# strip them out
if data.get("headers"):
headers, cookie_header = format_headers(get_path(data, "headers", filter=True))
if not cookies and cookie_header:
cookies = cookie_header
else:
headers = ()
# We prefer the body to be a string, since we can then attempt to parse it
# as JSON OR decode it as a URL encoded query string, without relying on
# the correct content type header being passed.
body = data.get('data')
content_type = next((v for k, v in headers if k == 'Content-Type'), None)
# Remove content type parameters
if content_type is not None:
content_type = content_type.partition(';')[0].rstrip()
# We process request data once during ingestion and again when
# requesting the http interface over the API. Avoid overwriting
# decoding the body again.
inferred_content_type = data.get('inferred_content_type', content_type)
if 'inferred_content_type' not in data and not isinstance(body, dict):
body, inferred_content_type = heuristic_decode(body, content_type)
if body:
body = trim(body, settings.SENTRY_MAX_HTTP_BODY_SIZE)
env = data.get('env', {})
# TODO (alex) This could also be accomplished with schema (with formats)
if 'REMOTE_ADDR' in env:
try:
validate_ip(env['REMOTE_ADDR'], required=False)
except ValueError:
del env['REMOTE_ADDR']
kwargs['inferred_content_type'] = inferred_content_type
kwargs['cookies'] = trim_pairs(format_cookies(cookies))
kwargs['env'] = trim_dict(env)
kwargs['headers'] = trim_pairs(headers)
kwargs['data'] = fix_broken_encoding(body)
kwargs['url'] = urlunsplit((scheme, netloc, path, '', ''))
kwargs['fragment'] = trim(fragment, 1024)
return cls(**kwargs)
