Refactor send.rs

This commit is contained in:
Alex 2022-07-21 17:59:15 +02:00
parent f35fa7d18d
commit 9dffa812c4
Signed by untrusted user: lx
GPG key ID: 0E496D15096376BE
2 changed files with 61 additions and 60 deletions

View file

@ -1,16 +1,13 @@
use std::collections::HashMap;
use std::sync::Arc;
use async_trait::async_trait;
use log::trace;
use futures::channel::mpsc::{unbounded, UnboundedReceiver, UnboundedSender};
use futures::AsyncReadExt;
use async_trait::async_trait;
use crate::error::*;
use crate::send::*;
use crate::util::Packet;

View file

@ -24,6 +24,7 @@ use crate::util::{ByteStream, Packet};
pub(crate) type RequestID = u32;
pub(crate) type ChunkLength = u16;
pub(crate) const MAX_CHUNK_LENGTH: ChunkLength = 0x3FF0;
pub(crate) const ERROR_MARKER: ChunkLength = 0x4000;
pub(crate) const CHUNK_HAS_CONTINUATION: ChunkLength = 0x8000;
@ -62,53 +63,58 @@ enum DataFrame {
data: [u8; MAX_CHUNK_LENGTH as usize],
/// actual lenght of data
len: usize,
/// whethere there may be more data comming from this stream. Can be used for some
/// optimization. It's an error to set it to false if there is more data, but it is correct
/// (albeit sub-optimal) to set it to true if there is nothing coming after
may_have_more: bool,
},
/// An error code automatically signals the end of the stream
Error(u8),
}
struct DataReaderItem {
data: DataFrame,
/// whethere there may be more data comming from this stream. Can be used for some
/// optimization. It's an error to set it to false if there is more data, but it is correct
/// (albeit sub-optimal) to set it to true if there is nothing coming after
may_have_more: bool,
}
impl DataReaderItem {
impl DataFrame {
fn empty_last() -> Self {
DataReaderItem {
data: DataFrame::Data {
data: [0; MAX_CHUNK_LENGTH as usize],
len: 0,
},
DataFrame::Data {
data: [0; MAX_CHUNK_LENGTH as usize],
len: 0,
may_have_more: false,
}
}
fn header(&self) -> [u8; 2] {
let continuation = if self.may_have_more {
CHUNK_HAS_CONTINUATION
} else {
0
let header_u16 = match self {
DataFrame::Data {
len,
may_have_more: false,
..
} => *len as u16,
DataFrame::Data {
len,
may_have_more: true,
..
} => *len as u16 | CHUNK_HAS_CONTINUATION,
DataFrame::Error(e) => *e as u16 | ERROR_MARKER,
};
let len = match self.data {
DataFrame::Data { len, .. } => len as u16,
DataFrame::Error(e) => e as u16 | ERROR_MARKER,
};
ChunkLength::to_be_bytes(len | continuation)
ChunkLength::to_be_bytes(header_u16)
}
fn data(&self) -> &[u8] {
match self.data {
DataFrame::Data { ref data, len } => &data[..len],
match self {
DataFrame::Data { ref data, len, .. } => &data[..*len],
DataFrame::Error(_) => &[],
}
}
fn may_have_more(&self) -> bool {
match self {
DataFrame::Data { may_have_more, .. } => *may_have_more,
DataFrame::Error(_) => false,
}
}
}
impl Stream for DataReader {
type Item = DataReaderItem;
type Item = DataFrame;
fn poll_next(self: Pin<&mut Self>, cx: &mut Context<'_>) -> Poll<Option<Self::Item>> {
let mut this = self.project();
@ -125,10 +131,8 @@ impl Stream for DataReader {
Err(e) => {
let e = *e;
*this.packet = Ok(Vec::new());
return Poll::Ready(Some(DataReaderItem {
data: DataFrame::Error(e),
may_have_more: true,
}));
*this.eos = true;
return Poll::Ready(Some(DataFrame::Error(e)));
}
};
let packet_left = packet.len() - *this.pos;
@ -161,8 +165,9 @@ impl Stream for DataReader {
let len = this.buf.len();
body[..len].copy_from_slice(this.buf);
this.buf.clear();
Poll::Ready(Some(DataReaderItem {
data: DataFrame::Data { data: body, len },
Poll::Ready(Some(DataFrame::Data {
data: body,
len,
may_have_more: !*this.eos,
}))
}
@ -218,38 +223,37 @@ struct SendQueuePollNextReady<'a> {
}
impl<'a> futures::Future for SendQueuePollNextReady<'a> {
type Output = (RequestID, DataReaderItem);
type Output = (RequestID, DataFrame);
fn poll(mut self: Pin<&mut Self>, ctx: &mut Context<'_>) -> Poll<Self::Output> {
for i in 0..self.queue.items.len() {
let (_prio, items_at_prio) = &mut self.queue.items[i];
for _ in 0..items_at_prio.len() {
let mut item = items_at_prio.pop_front().unwrap();
let mut ready_item = None;
for (j, item) in items_at_prio.iter_mut().enumerate() {
match Pin::new(&mut item.data).poll_next(ctx) {
Poll::Pending => items_at_prio.push_back(item),
Poll::Ready(Some(data)) => {
let id = item.id;
if data.may_have_more {
self.queue.push(item);
} else {
if items_at_prio.is_empty() {
// this priority level is empty, remove it
self.queue.items.remove(i);
}
}
return Poll::Ready((id, data));
}
Poll::Ready(None) => {
if items_at_prio.is_empty() {
// this priority level is empty, remove it
self.queue.items.remove(i);
}
return Poll::Ready((item.id, DataReaderItem::empty_last()));
Poll::Pending => (),
Poll::Ready(ready_v) => {
ready_item = Some((j, ready_v));
break;
}
}
}
if let Some((j, ready_v)) = ready_item {
let item = items_at_prio.remove(j).unwrap();
let id = item.id;
if ready_v
.as_ref()
.map(|data| data.may_have_more())
.unwrap_or(false)
{
items_at_prio.push_back(item);
} else if items_at_prio.is_empty() {
self.queue.items.remove(i);
}
return Poll::Ready((id, ready_v.unwrap_or_else(DataFrame::empty_last)));
}
}
// TODO what do we do if self.queue is empty? We won't get scheduled again.
Poll::Pending