How a server runs its handlers: in a span named
the tag, or <span_prefix>.<tag> (no span with disable_tracing), at most
concurrency at once, a defect ending every request of the client
unless disable_fatal_defects, and waiting for the client's Ack
after each stream chunk unless disable_client_acks.
on_disconnect (read by make and the layers over a protocol) is
called with each client id the protocol reports disconnected, after
the server asked for that client's requests to be interrupted; it does
not wait for them. It is not called for a client that ends its input,
or one still connected when the server stops. Each call runs in a fiber
of its own, interrupted when the server stops; a failure is logged at
UnhandledLogLevel and ignored.
Attributes#
| Attribute | Type | Description |
|---|---|---|
disable_tracing | bool = False | |
disable_span_propagation | bool = False | |
span_prefix | str | None = None | |
span_attributes | Mapping[str, object] = field(default_factory=(dict[str, object])) | |
concurrency | Concurrency = 'unbounded' | |
disable_fatal_defects | bool = False | |
disable_client_acks | bool = False | |
on_disconnect | Callable[[int], Effect[None]] | None = None |
Functions#
__init__#
__init__(disable_tracing=False, disable_span_propagation=False, span_prefix=None, span_attributes=dict[str, object](), concurrency='unbounded', disable_fatal_defects=False, disable_client_acks=False, on_disconnect=None) -> None| Parameter | Type | Description |
|---|---|---|
disable_tracing | bool = False | |
disable_span_propagation | bool = False | |
span_prefix | str | None = None | |
span_attributes | Mapping[str, object] = dict[str, object]() | |
concurrency | Concurrency = 'unbounded' | |
disable_fatal_defects | bool = False | |
disable_client_acks | bool = False | |
on_disconnect | Callable[[int], Effect[None]] | None = None |