http.errors

import http

Everything here is re-exported by http, so import http is enough and the names are called as http.*. Importing http.errors on its own works too and reaches the same definitions.

Every error the HTTP stack raises, under one root.

HttpError is that root, so a caller that only wants to know whether the request failed can catch it and ignore the rest. The subclasses separate the cases worth handling differently: a connection that never opened, one that timed out, a response too large to accept, a redirect loop, and a status the caller asked to be treated as an error.

Classes

HttpError

class http.HttpError < Error

Base class for every error the http module raises.

Catching HttpError catches everything this module can throw on its own; the more specific subclasses below let a caller tell a connection failure apart from a malformed response without matching on message text.

• printable — has a @to_string(), so echo and print() show something useful

HttpError.to_string()

http.HttpError.to_string()

ProtocolError

class http.ProtocolError < HttpError

Raised when a peer sends something that isn’t a well-formed HTTP message: a broken request line or status line, a header field that doesn’t parse, an impossible framing combination, or a chunked body whose chunk sizes don’t add up.

A server answers these with 400 Bad Request and closes the connection, since a stream whose framing is in doubt cannot safely be reused.

• printable — has a @to_string(), so echo and print() show something useful

ProtocolError.to_string()

http.ProtocolError.to_string()

ConnectionError

class http.ConnectionError < HttpError

Raised when a connection could not be established, or when an established connection failed or was closed before the message being read or written was complete.

• printable — has a @to_string(), so echo and print() show something useful

ConnectionError.to_string()

http.ConnectionError.to_string()

TimeoutError

class http.TimeoutError < HttpError

Raised when an operation exceeded its configured deadline: a connect, a read, a write, or the total time budget for a request.

• printable — has a @to_string(), so echo and print() show something useful

TimeoutError.to_string()

http.TimeoutError.to_string()

TooLargeError

class http.TooLargeError < HttpError

Raised when a message exceeds one of the configured size limits - request line, header block, or body. Servers turn this into a 431 or 413 response rather than reading an unbounded amount of attacker-controlled data into memory.

• printable — has a @to_string(), so echo and print() show something useful

TooLargeError.to_string()

http.TooLargeError.to_string()

TooManyRedirectsError

class http.TooManyRedirectsError < HttpError

Raised by the client when a redirect chain exceeds HttpClient.max_redirects, which usually means the chain is a loop.

• printable — has a @to_string(), so echo and print() show something useful

TooManyRedirectsError.to_string()

http.TooManyRedirectsError.to_string()

StatusError

class http.StatusError < HttpError

Raised by HttpResponse.raise_for_status() when the response carries a 4xx or 5xx status. The response itself stays reachable through response, so the body (which for an API is usually where the actual explanation lives) isn’t lost.

• printable — has a @to_string(), so echo and print() show something useful

Fields

Constructor

http.StatusError(message, response)

StatusError.to_string()

http.StatusError.to_string()

UnsupportedProtocolError

class http.UnsupportedProtocolError < HttpError

Raised when a URL names a scheme this module cannot speak, or when a peer insists on a protocol version that was not offered.

• printable — has a @to_string(), so echo and print() show something useful

UnsupportedProtocolError.to_string()

http.UnsupportedProtocolError.to_string()

2026, Richard Ore and Zuri contributors