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>
We can move all promise ID knowledge to deno_core, allowing us to better
experiment with promise implementation in deno_core.
`{un,}refOpPromise(promise)` is equivalent to
`{un,}refOp(promise[promiseIdSymbol])`
I'm not sure what was the purpose of trying to be so clever with the
args were (maybe an optimization?), but it breaks variadic args as
pointed out in #20054.
Signed-off-by: Matt Mastracci <matthew@mastracci.com>
Co-authored-by: Matt Mastracci <matthew@mastracci.com>
Otherwise you can not return `Deno.Server` from async functions.
Co-authored-by: Yoshiya Hinosawa <stibium121@gmail.com>
Co-authored-by: Bartek Iwańczuk <biwanczuk@gmail.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>
Deno.serve's fast streaming implementation was not keeping the request
body resource ID alive. We were taking the `Rc<Resource>` from the
resource table during the response, so a hairpin duplex response that
fed back the request body would work.
However, if any JS code attempted to read from the request body (which
requires the resource ID to be valid), the response would fail with a
difficult-to-diagnose "EOF" error.
This was affecting more complex duplex uses of `Deno.fetch` (though as
far as I can tell was unreported).
Simple test:
```ts
const reader = request.body.getReader();
return new Response(
new ReadableStream({
async pull(controller) {
const { done, value } = await reader.read();
if (done) {
controller.close();
} else {
controller.enqueue(value);
}
},
}),
```
And then attempt to use the stream in duplex mode:
```ts
async function testDuplex(
reader: ReadableStreamDefaultReader<Uint8Array>,
writable: WritableStreamDefaultWriter<Uint8Array>,
) {
await writable.write(new Uint8Array([1]));
const chunk1 = await reader.read();
assert(!chunk1.done);
assertEquals(chunk1.value, new Uint8Array([1]));
await writable.write(new Uint8Array([2]));
const chunk2 = await reader.read();
assert(!chunk2.done);
assertEquals(chunk2.value, new Uint8Array([2]));
await writable.close();
const chunk3 = await reader.read();
assert(chunk3.done);
}
```
In older versions of Deno, this would just lock up. I believe after
23ff0e722e, it started throwing a more
explicit error:
```
httpServerStreamDuplexJavascript => ./cli/tests/unit/serve_test.ts:1339:6
error: TypeError: request or response body error: error reading a body from connection: Connection reset by peer (os error 54)
at async Object.pull (ext:deno_web/06_streams.js:810:27)
```
This PR fixes #19818. The problem was that the new InnerRequest class does not initialize the fields urlList and urlListProcessed that are used during a request clone. The solution aims to be straightforward by simply initializing the missing properties during the clone process. I also implemented a "cache" to the url getter of the new InnerRequest, avoiding the cost of calling op_http_get_request_method_and_url.
This commit stabilizes "Deno.serve()", which becomes the
preferred way to create HTTP servers in Deno.
Documentation was adjusted for each overload of "Deno.serve()"
API and the API always binds to "127.0.0.1:8000" by default.
Fixes #19687 by adding a rejection handler to the write inside the
setTimeout. There is a small window where the promise is actually not
awaited and may reject without a handler.
This commit adds basic support for "node:http2" module. Not
all APIs have been yet implemented, but this change already
allows to use this module for some basic functions.
The "grpc" package is still not working, but it's a good stepping
stone.
---------
Co-authored-by: Bartek Iwańczuk <biwanczuk@gmail.com>
This PR attempts to resolve the first item on the list from
https://github.com/denoland/deno/issues/19330 which is about using a
flat list of interleaved key/value pairs, instead of a nested array of
tuples.
I can tackle some more if you can provide a quick example of using raw
v8 arrays, cc @mmastrac
For the first implementation of node:http2, we'll use the internal
version of `Deno.serve` which allows us to listen on a raw TCP
connection rather than a listener.
This is mostly a refactoring, and hooking up of `op_http_serve_on` that
was never previously exposed (but designed for this purpose).
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.
This commit changes the return type of an unstable `Deno.serve()` API
to instead return a `Deno.Server` object that has a `finished` field.
This change is done in preparation to be able to ref/unref the HTTP
server.
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.
Merges `op_http_upgrade_next` and `op_ws_server_create`, significantly
simplifying websocket construction in ext/http (next), and removing one
JS -> Rust call. Also WS server now doesn't bypass
`HttpPropertyExtractor`.
Migrates some of existing async ops to generated wrappers introduced in
https://github.com/denoland/deno/pull/18887. As a result "core.opAsync2"
was removed.
I will follow up with more PRs that migrate all the async ops to
generated wrappers.
This is a rewrite of the `Deno.serve` API to live on top of hyper
1.0-rc3. The code should be more maintainable long-term, and avoids some
of the slower mpsc patterns that made the older code less efficient than
it could have been.
Missing features:
- `upgradeHttp` and `upgradeHttpRaw` (`upgradeWebSocket` is available,
however).
- Automatic compression is unavailable on responses.
