| Age | Commit message (Collapse) | Author |
|---|
|
Aligns the error messages in the ext/http and a few messages in the
ext/fetch folder to be in-line with the Deno style guide.
This change-set also removes some unnecessary checks in the 00_serve.ts.
These options were recently removed, so it doesn't make sense to check
for them anymore.
https://github.com/denoland/deno/issues/25269
|
|
Towards #22079
---------
Co-authored-by: Bartek Iwańczuk <biwanczuk@gmail.com>
|
|
Max rps without a signal is unchanged, however we can drastically reduce
memory usage by not creating the signal until needed, and we can
optimize the rps in the case where the signal is created.
With a quick memory benchmark, it looks like this helps pretty
drastically with # of GCs when benchmarking w/wrk:
- 1.42.4: 1763
- canary: 1093
- this patch: 874
This branch:
```
Running 10s test @ http://localhost:8080/
2 threads and 10 connections
Thread Stats Avg Stdev Max +/- Stdev
Latency 87.33us 439.95us 20.68ms 99.67%
Req/Sec 66.70k 6.39k 74.11k 83.66%
1340255 requests in 10.10s, 191.73MB read
Requests/sec: 132696.90
Transfer/sec: 18.98MB
cpu: Apple M2 Pro
runtime: deno 1.43.0 (aarch64-apple-darwin)
file:///Users/matt/Documents/scripts/bench_request.js
benchmark time (avg) iter/s (min … max) p75 p99 p995
----------------------------------------------------------------------------------------------- -----------------------------
newRequest 986.5 ns/iter 1,013,682.6 (878.2 ns … 1.18 µs) 1.01 µs 1.18 µs 1.18 µs
newAbortController 18 ns/iter 55,541,104.1 (15.6 ns … 42.62 ns) 17.71 ns 25.05 ns 26.27 ns
newAbortControllerSignal 18.66 ns/iter 53,578,966.7 (16.49 ns … 32.16 ns) 18.71 ns 25.67 ns 26.39 ns
newAbortControllerSignalOnAbort 106.49 ns/iter 9,390,164.9 (97.87 ns … 120.61 ns) 108.6 ns 114.24 ns 115.89 ns
newAbortControllerSignalAddEventListener 86.92 ns/iter 11,504,880.2 (81.88 ns … 103.15 ns) 90 ns 98.28 ns 99.55 ns
newAbortControllerSignalOnAbortNoListener 3.01 µs/iter 331,964.4 (2.97 µs … 3.1 µs) 3.06 µs 3.1 µs 3.1 µs
newAbortControllerSignalOnAbortAbort 3.26 µs/iter 306,662.6 (3.22 µs … 3.36 µs) 3.27 µs 3.36 µs 3.36 µs
```
Latest canary:
```
Running 10s test @ http://localhost:8080/
2 threads and 10 connections
Thread Stats Avg Stdev Max +/- Stdev
Latency 72.86us 71.23us 4.47ms 99.05%
Req/Sec 64.66k 5.54k 72.48k 82.18%
1299015 requests in 10.10s, 185.83MB read
Requests/sec: 128616.02
Transfer/sec: 18.40MB
cpu: Apple M2 Pro
runtime: deno 1.43.0+bc4aa5f (aarch64-apple-darwin)
file:///Users/matt/Documents/scripts/bench_request.js
benchmark time (avg) iter/s (min … max) p75 p99 p995
----------------------------------------------------------------------------------------------- -----------------------------
newRequest 1.25 µs/iter 800,005.2 (1.01 µs … 4.18 µs) 1.16 µs 4.18 µs 4.18 µs
newAbortController 18.56 ns/iter 53,868,204.3 (16.04 ns … 38.73 ns) 18.38 ns 26.1 ns 26.63 ns
newAbortControllerSignal 18.72 ns/iter 53,430,746.1 (16.13 ns … 36.71 ns) 18.71 ns 26.19 ns 26.98 ns
newAbortControllerSignalOnAbort 193.91 ns/iter 5,156,992.4 (184.25 ns … 211.41 ns) 194.96 ns 207.87 ns 209.4 ns
newAbortControllerSignalAddEventListener 171.45 ns/iter 5,832,569.2 (153 ns … 182.03 ns) 176.17 ns 180.75 ns 181.05 ns
newAbortControllerSignalOnAbortNoListener 3.07 µs/iter 326,263.3 (2.98 µs … 3.17 µs) 3.08 µs 3.17 µs 3.17 µs
newAbortControllerSignalOnAbortAbort 3.32 µs/iter 301,344.6 (3.29 µs … 3.4 µs) 3.33 µs 3.4 µs 3.4 µs
```
|
|
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.
|
|
Landing parts of #21903 in preparation for the removal of serveHttp.
|
|
Closes https://github.com/denoland/deno/issues/21900
|
|
First pass of migrating away from `Deno.core.ensureFastOps()`.
A few "tricky" ones have been left for a follow up.
|
|
Closes https://github.com/denoland/deno/issues/21828.
This API is a huge footgun. And given that "Deno.serveHttp" is a
deprecated API that is discouraged to use (use "Deno.serve()"
instead); it makes no sense to keep this API around.
This is a step towards fully migrating to Hyper 1.
|
|
|
|
Deno v1.39 introduces `vm.runInNewContext`. This may cause problems when
using `Object.prototype.isPrototypeOf` to check built-in types.
```js
import vm from "node:vm";
const err = new Error();
const crossErr = vm.runInNewContext(`new Error()`);
console.assert( !(crossErr instanceof Error) );
console.assert( Object.getPrototypeOf(err) !== Object.getPrototypeOf(crossErr) );
```
This PR changes to check using internal slots solves them.
---
current:
```
> import vm from "node:vm";
undefined
> vm.runInNewContext(`new Error("message")`)
Error {}
> vm.runInNewContext(`new Date("2018-12-10T02:26:59.002Z")`)
Date {}
```
this PR:
```
> import vm from "node:vm";
undefined
> vm.runInNewContext(`new Error("message")`)
Error: message
at <anonymous>:1:1
> vm.runInNewContext(`new Date("2018-12-10T02:26:59.002Z")`)
2018-12-10T02:26:59.002Z
```
---------
Co-authored-by: Bartek Iwańczuk <biwanczuk@gmail.com>
|
|
|
|
`opAsync` requires a lookup by name on each async call. This is a
mechanical translation of all opAsync calls to ensureFastOps.
The `opAsync` API on Deno.core will be removed at a later time.
|
|
|
|
This commit refactors how we access "core", "internals" and
"primordials" objects coming from `deno_core`, in our internal JavaScript code.
Instead of capturing them from "globalThis.__bootstrap" namespace, we
import them from recently added "ext:core/mod.js" file.
|
|
This commit implements Symbol.dispose and Symbol.asyncDispose for
the relevant resources.
Closes #20839
---------
Signed-off-by: Bartek Iwańczuk <biwanczuk@gmail.com>
Co-authored-by: Bartek Iwańczuk <biwanczuk@gmail.com>
|
|
|
|
|
|
Startup benchmark shows no changes (within 1ms, identical system/user
times).
|
|
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.
|
|
I would like to get this change into Deno before merging
https://github.com/denoland/deno_lint/pull/1152
|
|
Fix internal "upgradeHttpRaw" API restoring capability to upgrade HTTP
connection in polyfilles "node:http" API.
|
|
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.
|
|
|
|
(#18587)
https://github.com/littledivy/fastwebsockets
```
# This PR
./load_test 100 0.0.0.0 8080 0 0
Running benchmark now...
Msg/sec: 176355.000000
# main
./load_test 100 0.0.0.0 8080 0 0
Running benchmark now...
Msg/sec: 157198.750000
```
|
|
Closes https://github.com/denoland/deno/issues/18582
|
|
Follow up to https://github.com/denoland/deno/pull/18578
We will need to do another pass cleaning up `ext/fetch/23_request.js`
|
|
This commit changes implementation of "Deno.serve()" API to use
"Deno.serveHttp()" under the hood. This change will allow us to
remove the "flash" server implementation, bringing stability to the
"Deno.serve()" API.
"cli/tests/unit/flash_test.ts" was renamed to "serve_test.ts".
Closes https://github.com/denoland/deno/issues/15574
Closes https://github.com/denoland/deno/issues/15504
Closes https://github.com/denoland/deno/issues/15646
Closes https://github.com/denoland/deno/issues/15909
Closes https://github.com/denoland/deno/issues/15911
Closes https://github.com/denoland/deno/issues/16828
Closes https://github.com/denoland/deno/issues/18046
Closes https://github.com/denoland/deno/issues/15869
|
|
This commit adds new "op_http_upgrade_early", that allows to hijack
existing "Deno.HttpConn" acquired from "Deno.serveHttp" API
and performing a Websocket upgrade on this connection.
This is not a public API and is meant to be used internally in the
"ext/node" polyfills for "http" module.
---------
Co-authored-by: Bartek Iwańczuk <biwanczuk@gmail.com>
|
|
This commit adds a new core API `opAsync2` to call an async op with
atmost 2 arguments. Spread argument iterators has a pretty big perf hit
when calling ops.
| name | avg msg/sec/core |
| --- | --- |
| 1.32.1 | `127820.750000` |
| #18506 | `140079.000000` |
| #18506 + #18509 | `150104.250000` |
| #18506 + #18509 + this | `157340.000000` |
|
|
|
|
|
|
for snapshotted modules (#18041)
This commit renames "deno_core::InternalModuleLoader" to
"ExtModuleLoader" and changes the specifiers used by the
modules loaded from this loader to "ext:".
"internal:" scheme was really ambiguous and it's more characters than
"ext:", which should result in slightly smaller snapshot size.
Closes https://github.com/denoland/deno/issues/18020
|
|
(#17727)
|
|
|
|
|
|
This PR refactors all internal js files (except core) to be written as
ES modules.
`__bootstrap`has been mostly replaced with static imports in form in
`internal:[path to file from repo root]`.
To specify if files are ESM, an `esm` method has been added to
`Extension`, similar to the `js` method.
A new ModuleLoader called `InternalModuleLoader` has been added to
enable the loading of internal specifiers, which is used in all
situations except when a snapshot is only loaded, and not a new one is
created from it.
---------
Co-authored-by: Bartek Iwańczuk <biwanczuk@gmail.com>
|
|
Updated third_party dlint to v0.37.0 for GitHub Actions. This PR
includes following changes:
* fix(prefer-primordials): Stop using array pattern assignments
* fix(prefer-primordials): Stop using global intrinsics except for
`SharedArrayBuffer`
* feat(guard-for-in): Apply new guard-for-in rule
|
|
|
|
Yearly tradition of creating extra noise in git.
|
|
Previously, errored streaming response bodies did not cause the HTTP
stream to be aborted. It instead caused the stream to be closed gracefully,
which had the result that the client could not detect the difference
between a successful response and an errored response.
This commit fixes the issue by aborting the stream on error.
|
|
Introduces `SafeSetIterator` and `SafeMapIterator` to primordials
|
|
(#16213)
|
|
We can use Resource::read_return & op_read instead. This allows HTTP
request bodies to participate in FastStream.
To make this work, `readableStreamForRid` required a change to allow non
auto-closing resources to be handled. This required some minor changes
in our FastStream paths in ext/http and ext/flash.
|
|
This commit splits `Deno.upgradeHttp` into two different APIs, because
the same API is currently overloaded with two different functions. Flash
requests upgrade immediately, with no need to return a `Response`
object. Instead you have to manually write the response to the socket.
Hyper requests only upgrade once a `Response` object has been sent.
These two behaviours are now split into `Deno.upgradeHttp` and
`Deno.upgradeHttpRaw`. The latter is flash only. The former only
supports hyper requests at the moment, but can be updated to support
flash in the future.
Additionally this removes `void | Promise<void>` as valid return types
for the handler function. If one wants to use `Deno.upgradeHttpRaw`,
they will have to type cast the handler signature - the signature is
meant for the 99.99%, and should not be complicated for the 0.01% that
use `Deno.upgradeHttpRaw()`.
|
|
Co-authored-by: Bartek Iwańczuk <biwanczuk@gmail.com>
Co-authored-by: Ben Noordhuis <info@bnoordhuis.nl>
Co-authored-by: crowlkats <crowlkats@toaxl.com>
Co-authored-by: Ryan Dahl <ry@tinyclouds.org>
|
|
|
|
Welcome to better optimised op calls! Currently opSync is called with parameters of every type and count. This most definitely makes the call megamorphic. Additionally, it seems that spread params leads to V8 not being able to optimise the calls quite as well (apparently Fast Calls cannot be used with spread params).
Monomorphising op calls should lead to some improved performance. Now that unwrapping of sync ops results is done on Rust side, this is pretty simple:
```
opSync("op_foo", param1, param2);
// -> turns to
ops.op_foo(param1, param2);
```
This means sync op calls are now just directly calling the native binding function. When V8 Fast API Calls are enabled, this will enable those to be called on the optimised path.
Monomorphising async ops likely requires using callbacks and is left as an exercise to the reader.
|
|
RequestEvent (#15100)
Fixes #15099
|
|
|
|
|
