package async

You can search for identifiers within the package.

in-package search v0.2.0

package async

async
- Async
  - Command
  - Deferred
    
    Or_error
    
    Expect_test_config
    
    IO
  - Expect_test_config
    
    IO
  - Expect_test_config_or_error
    
    IO
  - Log
  - Quickcheck
async.async_command
- Async_command
  - For_testing
  - Staged
async.async_quickcheck
- Async_quickcheck
  - Configure
    
    Config
async.async_rpc
- Async_rpc
  - Rpc
    
    Connection
    
    For_debugging
    
    Low_latency_transport
    
    Config
    
    Reader
    
    With_internal_reader
    
    With_internal_reader
    
    Writer
    
    Transport
    
    Reader
    
    Tcp
    
    Writer
async.lock_file_async
- Lock_file_async
  - Flock
  - Nfs
  - Symlink
async.log_extended
- Log_extended
  - Command
  - Console
    
    Blocking
  - Syslog
    
    Blocking
async.persistent_connection
- Persistent_connection
  - Make
    
    Conn
    
    Event
  - Make'
    
    Conn
    
    Conn_err
    
    Event
  - Rpc
    
    Event
  - S
    
    Event
  - S'
    
    Event
  - S_rpc
    
    Event
  - Versioned_rpc
    
    Event
async.unpack_sequence
- Unpack_sequence

Legend:
Library
Module
Module type
Parameter
Class
Class type

type t

include Sexplib0.Sexpable.S with type t := t

val t_of_sexp : Sexplib0__.Sexp.t -> t

val sexp_of_t : t -> Sexplib0__.Sexp.t

val create : 
  ?max_message_size:int ->
  ?initial_buffer_size:int ->
  ?max_buffer_size:int ->
  ?write_timeout:Core.Time_ns.Span.t ->
  ?buffering_threshold_in_bytes:int ->
  ?start_batching_after_num_messages:int ->
  unit ->
  t

max_message_size is the maximum message size a reader/writer will accept to receive/send.

max_buffer_size is the maximum size the internal reader/writer's buffer will ever grow.

write_timeout is the maximum time allowed for a write operation to complete before an error is reported.

When the writer's internal buffer is filled at buffering_threshold_in_bytes or more, the writer will try to flush its buffer immediately. This is to get good latency and avoid buffering too much when sending big batches of messages.

If the application hasn't sent any messages in the current Async job and nothing is buffered, the first start_batching_after_num_messages messages will be sent immediately. After that the writer will start buffering. This is to give good latency when the application sends a few messages occasionally but still get good throughput when sending a batch of messages.

Note that start_batching_after_num_messages and buffering_threshold_in_bytes have somewhat opposite meanings: the former determines when to start batching and the latter determines when to write data that has been batched.