Struct mbpr::Request

pub struct Request<'a> { /* fields omitted */ }

Memcached Request Packet

The whole thing

Methods

impl<'a> Request<'a>

fn parse(x: &'a [u8]) -> ParseResult<Self>

Parse and borrow a packet from a buffer

fn new(opcode: OpCode,
       vbucket: u16,
       opaque: u32,
       cas: u64,
       extra: Option<&'a [u8]>,
       key: Option<&'a [u8]>,
       body: Option<&'a [u8]>)
       -> Request<'a>

This interface does ABSOLUTELY NO verfication of the packet it is expected if you are calling this method you understand the memcached protocol and you are going to use this to generate a valid packet.

The goal of this interface is to be fast. Encoding a packet with this interface + Encode trait involves very few branches

Memcached only allows Keys that are ASCII and non-white space this interface does not do ANY assertions of this. Please be aware.

fn rebuild(&mut self,
           opcode: OpCode,
           vbucket: u16,
           opaque: u32,
           cas: u64,
           extra: Option<&'a [u8]>,
           key: Option<&'a [u8]>,
           body: Option<&'a [u8]>)

Over write an existing request

This is provided to allow for easier interfacing with SLAB's. The semantics of this method are identical to the above. The primary difference is this doesn't push ~100 bytes to the stack.

This interface does ABSOLUTELY NO verfication of the packet it is expected if you are calling this method you understand the memcached protocol and you are going to use this to generate a valid packet.

The goal of this interface is to be fast. Encoding a packet with this interface + Encode trait involves very few branches

Memcached only allows Keys that are ASCII and non-white space this interface does not do ANY assertions of this. Please be aware.

fn encode_self(&self) -> Encoder

Allocates a new buffer and encodes this packets contents into it. this method works out to a handful of memcp primatives and is fairly quick as their is no bounds checking (buffer length is asserted on construction).

fn encode_into_buffer(&self, x: Vec<u8>) -> Encoder

If you are using a slab to avoid making too many allocations this method will check check the Vec<u8> it is passed only reserving additional capacity if necessary. If the Vec<u8> has enough capacity no action is taken.

fn get_opcode(&self) -> OpCode

fn get_opaque(&self) -> u32

fn get_cas(&self) -> u64

fn get_vbucket_id(&self) -> u16

fn has_extra(&self) -> bool

fn has_key(&self) -> bool

fn get_extra(&'a self) -> Option<&'a [u8]>

fn get_key(&'a self) -> Option<&'a [u8]>

fn get_key_str(&'a self) -> Option<&'a str>

The standard states the key should be an ASCII compatible string so this method preforms that conversion without checking for correctness.

This isnt a problem as the key will be hash/stored as a byte buffer anyways.

This really only opens the door to non standard things.

fn has_body(&self) -> bool

fn get_body(&'a self) -> Option<&'a [u8]>

fn to_owned(self) -> OwnedRequest

Consume this item and take ownership

If a field is Option::None the resulting vector will be a Vec::with_capacity(0) which does not allocate. So this method can be cheap depending on the messages contents.

Trait Implementations

impl<'a> Clone for Request<'a>

fn clone(&self) -> Request<'a>

Returns a copy of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl<'a> Debug for Request<'a>

fn fmt(&self, __arg_0: &mut Formatter) -> Result

Formats the value using the given formatter.

impl<'a> PacketVal for Request<'a>

fn get_keylen(&self) -> usize

Get size of Packet's Key Field

fn get_bodylen(&self) -> usize

Get size of Packet's Body Field (Raw Data)

fn get_extralen(&self) -> usize

Get size of Packet's Extra Field (Flags, Arguments, Etc. command specific)

fn total_len(&self) -> usize

The total length of the packet

impl<'a> Encoding for Request<'a>

fn encode(&self, buffer: &mut Encoder)

Encode a packet