You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Python's built-in http.server, which Werkzeug's dev server extends, does not drain the request body stream before the next request begins. If the view function didn't read the request body, then it would result in the body appearing prepended to the next request method, like {}GET. Not reading the body is apparently common, perhaps when first developing a view. When we enabled HTTP/1.1 by default, many more users began seeing this issue.
Disable keep-alive connections. The original request for HTTP/1.1 mode was for chunked responses, which will continue to work.
While they technically should report it to Python, they instead reported it repeatedly to Werkzeug and Flask since that's what they were using. If someone wants to report this to Python, and they fix it, please open a PR here to re-enable keep-alive connections.
Remember, the development server should never be used outside development. It is not designed to be particularly secure, stable, or efficient.
The text was updated successfully, but these errors were encountered:
Python's built-in
http.server
, which Werkzeug's dev server extends, does not drain the request body stream before the next request begins. If the view function didn't read the request body, then it would result in the body appearing prepended to the next request method, like{}GET
. Not reading the body is apparently common, perhaps when first developing a view. When we enabled HTTP/1.1 by default, many more users began seeing this issue.Disable keep-alive connections. The original request for HTTP/1.1 mode was for chunked responses, which will continue to work.
While they technically should report it to Python, they instead reported it repeatedly to Werkzeug and Flask since that's what they were using. If someone wants to report this to Python, and they fix it, please open a PR here to re-enable keep-alive connections.
Remember, the development server should never be used outside development. It is not designed to be particularly secure, stable, or efficient.
The text was updated successfully, but these errors were encountered: