clean up some outdated code and implement a few performance
improvements. most notably, the response from `upgradeWebSocket` is now
actually used as a normal response.
a followup pr will rework raw upgrades to a greater extent
Deno.serve `Request` abort signals are aborted by default even when it
is finished successfully. This PR gates this behavior behind the
"legacy_abort" which is the default right now.
Turning the `no_legacy_abort` runtime option on is a **breaking change**
and will only abort request signals when there is a failure, thereby
cannot be used to determine if the request finished. This aligns with
`fetch` API.
Ref https://github.com/denoland/deno/issues/27005
results on my machine (Macbook pro w/ M3 Max)
canary (hello world):
```
❯ oha -c 125 -n 4000000 --no-tui --disable-compression http://localhost:8000
Summary:
Success rate: 100.00%
Total: 31.0160 secs
Slowest: 0.0083 secs
Fastest: 0.0005 secs
Average: 0.0010 secs
Requests/sec: 128965.6454
Total data: 49.59 MiB
Size/request: 13 B
Size/sec: 1.60 MiB
```
this PR (hello world):
```
❯ oha -c 125 -n 4000000 --no-tui --disable-compression http://localhost:8000
Summary:
Success rate: 100.00%
Total: 28.4050 secs
Slowest: 0.0085 secs
Fastest: 0.0001 secs
Average: 0.0009 secs
Requests/sec: 140820.2060
Total data: 49.59 MiB
Size/request: 13 B
Size/sec: 1.75 MiB
```
---
Two changes here:
- use `opt-level` 3 for some of hyper's deps, since profile overrides
are not transitive
- As noted in the [cargo
reference](https://doc.rust-lang.org/cargo/reference/profiles.html#overrides-and-generics)
generic functions _may_ be optimized at the opt-level of the
_instantiating_ crate, rather than the defining crate. So currently it's
possible that some of the functions in `deno_http` are being compiled at
a lower optimization level. This PR ensures the generics are
instantiated in `deno_http`, which theoretically should guarantee they
actually use opt-level 3.
To allow embedders to still provide a custom property extractor, I put
this behind a feature flag.
This commit makes http server parameters configurable on the extension
initialization via two callbacks users can provide.
The main motivation behind this change is to allow `deno_http` users to
tune the HTTP/2 server to suit their needs, although Deno CLI users will
not benefit from it as no JavaScript interface is exposed to set these
parameters currently.
It is up to users whether to provide hook functions. If not provided,
the default configuration from hyper crate will be used.
When the response has been successfully send, we abort the
`Request.signal` property to indicate that all resources associated with
this transaction may be torn down.
Landing part of https://github.com/denoland/deno/pull/21903
This will allow us to more easily refactor `serveHttp` to live on top of
`serve` by splitting the websocket code out. There's probably a lot more
we could do here but this helps.
Main change is that:
- "hyper" has been renamed to "hyper_v014" to signal that it's legacy
- "hyper1" has been renamed to "hyper" and should be the default
Node HTTP/2 was using the default h2 `Bytes` datatype when we can be
making using of `BufView` like we do in `Deno.serve`.
`fetch` and `Deno.serverHttp` can't make use of `BufView` because they
are using `reqwest` which is stuck on hyper 0.x at this time.
Use HttpRecord as response body so requests can be tracked all the way
to response body completion.
This allows Request properties to be accessed while the response body is
streaming.
Graceful shutdown now awaits a future instead of async spinning waiting
for requests to finish.
On the minimal benchmark this refactor improves performance an
additional 2% over pooling alone for a net 3% increase over the previous
deno main branch.
Builds upon https://github.com/denoland/deno/pull/20809 and
https://github.com/denoland/deno/pull/20770.
---------
Co-authored-by: Matt Mastracci <matthew@mastracci.com>
Makes the JavaScript Request use a v8:External opaque pointer to
directly refer to the Rust HttpRecord.
The HttpRecord is now reference counted. To avoid leaks the strong count
is checked at request completion.
Performance seems unchanged on the minimal benchmark. 118614 req/s this
branch vs 118564 req/s on main, but variance between runs on my laptop
is pretty high.
---------
Co-authored-by: Matt Mastracci <matthew@mastracci.com>
This PR implements a graceful shutdown API for Deno.serve, allowing all
current connections to drain from the server before shutting down, while
preventing new connections from being started or new transactions on
existing connections from being created.
We split the cancellation handle into two parts: a listener handle, and
a connection handle. A graceful shutdown cancels the listener only,
while allowing the connections to drain. The connection handle aborts
all futures. If the listener handle is cancelled, we put the connections
into graceful shutdown mode, which disables keep-alive on http/1.1 and
uses http/2 mechanisms for http/2 connections.
In addition, we now guarantee that all connections are complete or
cancelled, and all resources are cleaned up when the server `finished`
promise resolves -- we use a Rust-side server refcount for this.
Performance impact: does not appear to affect basic serving performance
by more than 1% (~126k -> ~125k)
---------
Co-authored-by: Bartek Iwańczuk <biwanczuk@gmail.com>
This tweaks the HTTP response-writer in order to align the two possible
execution flows into using the same gzip default compression level, that
is `1` (otherwise the implicit default level is `6`).
Includes a lightly-modified version of hyper-util's `TokioIo` utility.
Hyper changes:
v1.0.0-rc.4 (2023-07-10)
Bug Fixes
http1:
http1 server graceful shutdown fix (#3261)
([f4b51300](f4b513009d))
send error on Incoming body when connection errors (#3256)
([52f19259](52f192593f),
closes https://github.com/hyperium/hyper/issues/3253)
properly end chunked bodies when it was known to be empty (#3254)
([fec64cf0](fec64cf0ab),
closes https://github.com/hyperium/hyper/issues/3252)
Features
client: Make clients able to use non-Send executor (#3184)
([d977f209](d977f209bc),
closes https://github.com/hyperium/hyper/issues/3017)
rt:
replace IO traits with hyper::rt ones (#3230)
([f9f65b7a](f9f65b7aa6),
closes https://github.com/hyperium/hyper/issues/3110)
add downcast on Sleep trait (#3125)
([d92d3917](d92d3917d9),
closes https://github.com/hyperium/hyper/issues/3027)
service: change Service::call to take &self (#3223)
([d894439e](d894439e00),
closes https://github.com/hyperium/hyper/issues/3040)
Breaking Changes
Any IO transport type provided must not implement hyper::rt::{Read,
Write} instead of tokio::io traits. You can grab a helper type from
hyper-util to wrap Tokio types, or implement the traits yourself, if
it's a custom type.
([f9f65b7a](f9f65b7aa6))
client::conn::http2 types now use another generic for an Executor. Code
that names Connection needs to include the additional generic parameter.
([d977f209](d977f209bc))
The Service::call function no longer takes a mutable reference to self.
The FnMut trait bound on the service::util::service_fn function and the
trait bound on the impl for the ServiceFn struct were changed from FnMut
to Fn.
`ZeroCopyBuf` was convenient to use, but sometimes it did hide details
that some copies were necessary in certain cases. Also it made it way to easy
for the caller to pass around and convert into different values. This commit
splits `ZeroCopyBuf` into `JsBuffer` (an array buffer coming from V8) and
`ToJsBuffer` (a Rust buffer that will be converted into a V8 array buffer).
As a result some magical conversions were removed (they were never used)
limiting the API surface and preparing for changes in #19534.
Under heavy load, we often have requests queued up that don't need an
async call to retrieve. We can use a fast path sync op to drain this set
of ready requests, and then fall back to the async op once we run out of
work.
This is a .5-1% bump in req/s on an M2 mac. About 90% of the handlers go
through this sync phase (based on a simple instrumentation that is not
included in this PR) and skip the async machinery entirely.
Fixes for various `Attemped to access invalid request` bugs (#19058,
#15427, #17213).
We did not wait for both a drop event and a completion event before
removing items from the slab table. This ensures that we do so.
In addition, the slab methods are refactored out into `slab.rs` for
maintainability.
