v0.6.0

Breaking Changes

• Drops support for Python 3.9
• close_timeout no longer raises an exception when the timeout is reached. The error is silently suppressed after issuing a TCP RST which closes the connection.
• A close_timeout arg is now exposed on the top-level request methods.

close_timeout update

# The following call will not wait for the connection to gracefully close.
# A TCP RST will be sent to close the connection.
response = await pywreck.head('www.example.com', '/', close_timeout=0.0)

Updates

• Adds support for Python 3.14

Full Changelog: v0.5.3...v0.6.0

v0.5.3

What's Changed

• Drops support for Python 3.7 and 3.8
• Adds support for Python 3.12 and 3.13
• non-breaking improvements to type annotations

Full Changelog: v0.5.2...v0.5.3

v0.5.2

What's Changed

• Adds support for Python 3.11

Full Changelog: v0.5.1...v0.5.2

v0.5.1

Minor Update

Headers used to require a type of Optional[Dict[str, str]]. This has been changed to Optional[Mapping[str, str]] to support non-dict mapping types.

v0.5.0

Breaking Changes

Timeouts now encompass the entire request/response cycle

The timeout API has been simplified to:

• A timeout argument that tracks the entire request/response cycle
• A connection close timeout (if using the Connection API)

request/response API example

# The following call can take at most 1 second to complete
response = await pywreck.head("www.example.com", "/", timeout=1.0)

Connection API example

# Upon exiting the "async with" block, wait up to 1 second for the connection
# to close before forcing the connection to close
async with await pywreck.Connection.create(
    "www.example.com",
    close_timeout=1.0,
) as conn:
    # Retrieve the response with a 1 second timeout on the request
    response = await conn.request("HEAD", "/", timeout=1.0)

v0.4.1

Bug Fixes

Fixed an issue where chunked request bytes were not fully consumed. This bug caused errors when chunked requests were issued prior to other requests on a reused connection object.

Full Changelog: v0.4.0...v0.4.1

v0.4.0

Bug Fix

Prevent concurrent access to the underlying connection when a request is in progress.

Example:

connection = await pywreck.Connection.create("www.example.com")
coros = (connection.request("GET", "/foo"), connection.request("GET", "/bar"))
await asyncio.gather(*coros)

Note: this request cannot execute concurrently since the underlying connection can only be used for one HTTP/1 request at a time. The exclusive access to the underlying connection is now enforced.

To execute this request concurrently, 2 separate connections must be opened:

coros = (pywreck.request("GET", "www.example.com", "/foo"), pywreck.request("GET", "www.example.com", "/bar"))
await asyncio.gather(*coros)

Full Changelog: v0.3.0...v0.4.0

v0.3.0

Breaking Changes

Mark Response class as frozen. Assignment to attributes of Response after creation is now disallowed.

v0.2.2

Add a Connection class

Multiple requests can now be issued over a single connection. Example:

async with await pywreck.Connection.create("www.example.com") as conn:
    for _ in range(2):
        response = await conn.request("HEAD", "/")
        print(response)

v0.2.1

Support for Python 3.10

Added support and pypi classifiers for python 3.10