Faraday
Serialization primitives built for speed an memory-efficiency.
Faraday is a library for writing fast and memory-efficient serializers. Its core type and related operation gives the user fine-grained control over copying and allocation behavior while serializing user-defined types, and presents the output in a form that makes it possible to use vectorized write operations, such as the writev
system call, or any other platform or application-specific output APIs.
A Faraday serializer manages an internal buffer and a queue of output buffers. The output bufferes may be a sub range of the serializer's internal buffer or one that is user-provided. Buffered writes such as write_string
, write_char
, write_bigstring
, etc., copy the source bytes into the serializer's internal buffer. Unbuffered writes such as schedule_string
, schedule_bigstring
, etc., on the other hand perform no copying. Instead, they enqueue the source bytes into the serializer's write queue directly.
type bigstring =
(char, Bigarray.int8_unsigned_elt, Bigarray.c_layout) Bigarray.Array1.t
val create : int -> t
create len
creates a serializer with a fixed-length internal buffer of length len
. See the Buffered writes section for details about what happens when len
is not large enough to support a write.
of_bigstring buf
creates a serializer, using buf
as its internal buffer. The serializer takes ownership of buf
until the serializer has been closed and flushed of all output.
A serializer manages an internal buffer for coalescing small writes. The size of this buffer is determined when the serializer is created. If the buffer does not contain sufficient space to service a caller's buffered write, the serializer will allocate a new buffer of the sufficient size and use it for the current and subsequent writes. The old buffer will be garbage collected once all of its contents have been flush
ed.
val write_string : t -> ?off:int -> ?len:int -> string -> unit
write_string t ?off ?len str
copies str
into the serializer's internal buffer.
write_bytes t ?off ?len bytes
copies bytes
into the serializer's internal buffer. It is safe to modify bytes
after this call returns.
write_bigstring t ?off ?len bigstring
copies bigstring
into the serializer's internal buffer. It is safe to modify bigstring
after this call returns.
val write_gen :
t ->
length:('a -> int) ->
blit:('a -> src_off:int -> bigstring -> dst_off:int -> len:int -> unit) ->
?off:int ->
?len:int ->
'a ->
unit
write_gen t ~length ~blit ?off ?len x
copies x
into the serializer's internal buffer using the provided length
and blit
operations. See Bigstring.blit
for documentation of the arguments.
val write_char : t -> char -> unit
write_char t char
copies char
into the serializer's internal buffer.
val write_uint8 : t -> int -> unit
write_uint8 t n
copies the lower 8 bits of n
into the serializer's internal buffer.
module BE : sig ... end
Big endian serializers
module LE : sig ... end
Little endian serializers
Unbuffered writes do not involve copying bytes to the serializers internal buffer.
schedule_bigstring t ?off ?len bigstring
schedules bigstring
to be written the next time the serializer surfaces writes to the user. bigstring
is not copied in this process, so bigstring
should only be modified after t
has been flush
ed.
val free_bytes_in_buffer : t -> int
free_bytes_in_buffer t
returns the free space, in bytes, of the serializer's write buffer. If a write_*
call has a length that exceeds this value, the serializer will allocate a new buffer that will replace the serializer's internal buffer for that and subsequent calls.
val has_pending_output : t -> bool
has_pending_output t
is true
if t
's output queue is non-empty. It may be the case that t
's queued output is being serviced by some other thread of control, but has not yet completed.
val pending_bytes : t -> int
pending_bytes t
is the size of the next write, in bytes, that t
will surface to the caller as a `Writev
.
val yield : t -> unit
yield t
causes t
to delay surfacing writes to the user, instead returning a `Yield
. This gives the serializer an opportunity to collect additional writes before sending them to the underlying device, which will increase the write batch size.
As one example, code may want to call this function if it's about to release the OCaml lock and perform a blocking system call, but would like to batch output across that system call. To hint to the thread of control that is performing the writes on behalf of the serializer, the code might call yield t
before releasing the lock.
val flush : t -> (unit -> unit) -> unit
flush t f
registers f
to be called when all prior writes have been successfully completed. If t
has no pending writes, then f
will be called immediately. If yield
was recently called on t
, then the effect of the yield
will be ignored so that client code has an opportunity to write pending output, regardless of how it handles `Yield
operations.
module Flushed_reason : sig ... end
val flush_with_reason : t -> (Flushed_reason.t -> unit) -> unit
flush_with_reason t f
is like flush t f
, but f
is suppplied with the reason that the callback was triggered.
val close : t -> unit
close t
closes t
. All subsequent write calls will raise, and any pending or subsequent yield
calls will be ignored. If the serializer has any pending writes, user code will have an opportunity to service them before it receives the Close
operation. Flush callbacks will continue to be invoked while output is shift
ed out of t
as needed.
val is_closed : t -> bool
is_closed t
is true
if close
has been called on t
and false
otherwise. A closed t
may still have pending output.
val shift : t -> int -> unit
shift t n
removes the first n
bytes in t
's write queue. Any flush callbacks registered with t
within this span of the write queue will be called.
val drain : t -> int
drain t
removes all pending writes from t
, returning the number of bytes that were enqueued to be written and freeing any scheduled buffers in the process.
Low-level operations for runing a serializer. For production use-cases, consider the Async and Lwt support that this library includes before attempting to use this these operations directly.
A view into iovec.buffer
starting at iovec.off
and with length iovec.len
.
The type of operations that the serialier may wish to perform.
`Writev iovecs
: Write the bytes in iovecs
s reporting the actual number of bytes written by calling shift
. You must accurately report the number of bytes written. Failure to do so will result in the same bytes being surfaced in a `Writev
operation multiple times.`Yield
: Yield to other threads of control, waiting for additional output before procedding. The method for achieving this is application-specific, but once complete, the caller can proceed with serialization by simply making another call to operation
or serialize
.`Close
: Serialization is complete. No further output will generated. The action to take as a result, if any, is application-specific.serialize t writev
sufaces the next operation of t
to the caller, handling a `Writev
operation with writev
function and performing an additional bookkeeping on the caller's behalf. In the event that writev
indicates a partial write, serialize
will call yield
on the serializer rather than attempting successive writev
calls.
These functions are included for testing, debugging, and general development. They are not the suggested way of driving a serializer in a production setting.