kombu中 acquire函数中block参数的解释

一直都想知道acquire中block参数的含义,今天查阅相关文档,如下别有一番洞天

Connection and Producer Pools

Default Pools

Kombu ships with two global pools: one connection pool, and one producer pool.

These are convenient and the fact that they are global may not be an issue as connections should often be limited at the process level, rather than per thread/application and so on, but if you need custom pools per thread see Custom Pool Groups.

The connection pool group

The connection pools are available as kombu.pools.connections. This is a pool group, which means you give it a connection instance, and you get a pool instance back. We have one pool per connection instance to support multiple connections in the same app. All connection instances with the same connection parameters will get the same pool:

>>> from kombu import Connection
>>> from kombu.pools import connections

>>> connections[Connection('redis://localhost:6379')]
<kombu.connection.ConnectionPool object at 0x101805650>
>>> connections[Connection('redis://localhost:6379')]
<kombu.connection.ConnectionPool object at 0x101805650>

Let’s acquire and release a connection:

from kombu import Connection
from kombu.pools import connections

connection = Connection('redis://localhost:6379')

with connections[connection].acquire(block=True) as conn:
    print('Got connection: {0!r}'.format(connection.as_uri()))

Note

The block=True here means that the acquire call will block until a connection is available in the pool. Note that this will block forever in case there is a deadlock in your code where a connection is not released. There is a timeout argument you can use to safeguard against this (see kombu.connection.Resource.acquire()).

If blocking is disabled and there aren’t any connections left in the pool an kombu.exceptions.ConnectionLimitExceeded exception will be raised.