This commit re-implements `ext/fetch` and all dependent crates
using `hyper` and `hyper-util`, instead of `reqwest`.
The reasoning is that we want to have greater control and access
to low level `hyper` APIs when implementing `fetch` API as well
as `node:http` module.
---------
Co-authored-by: Bartek Iwańczuk <biwanczuk@gmail.com>
This commit updates Deno to use `reqwest` at 0.12.4
and `rustls` at 0.22. Other related crates were updated
as well to match versions accepted by `reqwest` and `rustls`.
Note: we are not using the latest available `rustls` yet,
but this upgrade was non-trivial already, so a bump to
0.23 for `rustls` will be done in a separate commit.
Closes #23370
---------
Signed-off-by: Ryan Dahl <ry@tinyclouds.org>
Signed-off-by: Bartek Iwańczuk <biwanczuk@gmail.com>
Co-authored-by: Ryan Dahl <ry@tinyclouds.org>
Co-authored-by: Divy Srivastava <dj.srivastava23@gmail.com>
Also removes permissions being passed in for node resolution. It was
completely useless because we only checked it for reading package.json
files, but Deno reading package.json files for resolution is perfectly
fine.
My guess is this is also a perf improvement because Deno is doing less
work.
Precursor to #23236
This implements the SNI features, but uses private symbols to avoid
exposing the functionality at this time. Note that to properly test this
feature, we need to add a way for `connectTls` to specify a hostname.
This is something that should be pushed into that API at a later time as
well.
```ts
Deno.test(
{ permissions: { net: true, read: true } },
async function listenResolver() {
let sniRequests = [];
const listener = Deno.listenTls({
hostname: "localhost",
port: 0,
[resolverSymbol]: (sni: string) => {
sniRequests.push(sni);
return {
cert,
key,
};
},
});
{
const conn = await Deno.connectTls({
hostname: "localhost",
[serverNameSymbol]: "server-1",
port: listener.addr.port,
});
const [_handshake, serverConn] = await Promise.all([
conn.handshake(),
listener.accept(),
]);
conn.close();
serverConn.close();
}
{
const conn = await Deno.connectTls({
hostname: "localhost",
[serverNameSymbol]: "server-2",
port: listener.addr.port,
});
const [_handshake, serverConn] = await Promise.all([
conn.handshake(),
listener.accept(),
]);
conn.close();
serverConn.close();
}
assertEquals(sniRequests, ["server-1", "server-2"]);
listener.close();
},
);
```
---------
Signed-off-by: Matt Mastracci <matthew@mastracci.com>
Issue https://github.com/denoland/deno/issues/22222
This PR moves `runtime/permissions` code to a upstream crate called
`deno_permissions`. The `deno_permissions::PermissionsContainer` is put
into the OpState and can be used instead of the current trait-based
permissions system.
For this PR, I've migrated `deno_fetch` to the new crate but kept the
rest of the trait-based system as a wrapper of `deno_permissions` crate.
Doing the migration all at once is error prone and hard to review.
Comparing incremental compile times for `ext/fetch` on Mac M1:
| profile | `cargo build --bin deno` | `cargo plonk build --bin deno` |
| --------- | ------------- | ------------------- |
| `debug` | 20 s | 0.8s |
| `release` | 4 mins 12 s | 1.4s |
This change deprecates
`Deno.CreateHttpClientOptions.{certChain,privateKey}` in favour of
`Deno.CreateHttpClientOptions.{cert,key}`.
Closes #22278
Co-authored-by: Matt Mastracci <matthew@mastracci.com>
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.
Switch `ext/fetch` over to `resourceForReadableStream` to simplify and
unify implementation with `ext/serve`. This allows us to work in Rust
with resources only.
Two additional changes made to `resourceForReadableStream` were
required:
- Add an optional length to `resourceForReadableStream` which translates
to `size_hint`
- Fix a bug where writing to a closed stream that was full would panic
Implements `WebSocket` over http/2. This requires a conformant http/2
server supporting the extended connect protocol.
Passes approximately 100 new WPT tests (mostly `?wpt_flags=h2` versions
of existing websockets APIs).
This is implemented as a fallback when http/1.1 fails, so a server that
supports both h1 and h2 WebSockets will still end up on the http/1.1
upgrade path.
The patch also cleas up the websockets handshake to split it up into
http, https+http1 and https+http2, making it a little less intertwined.
This uncovered a likely bug in the WPT test server:
https://github.com/web-platform-tests/wpt/issues/42896
This is a prerequisite for fast streams work -- this particular resource
used a custom `mpsc`-style stream, and this work will allow us to unify
it with the streams in `ext/http` in time.
Instead of using Option as an internal semaphore for "correctly
completed EOF", we allow code to propagate errors into the channel which
can be picked up by downstream sinks like Hyper. EOF is signalled using
a more standard sender drop.
`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.
This commit reimplements most of "node:http" client APIs using
"ext/fetch".
There is some duplicated code and two removed Node compat tests that
will be fixed in follow up PRs.
---------
Co-authored-by: Bartek Iwańczuk <biwanczuk@gmail.com>
Follow-up to #18210:
* we are passing the generated `cfg` object into the state function
rather than passing individual config fields
* reduce cloning dramatically by making the state_fn `FnOnce`
* `take` for `ExtensionBuilder` to avoid more unnecessary copies
* renamed `config` to `options`
This implements two macros to simplify extension registration and centralize a lot of the boilerplate as a base for future improvements:
* `deno_core::ops!` registers a block of `#[op]`s, optionally with type
parameters, useful for places where we share lists of ops
* `deno_core::extension!` is used to register an extension, and creates
two methods that can be used at runtime/snapshot generation time:
`init_ops` and `init_ops_and_esm`.
---------
Co-authored-by: Bartek Iwańczuk <biwanczuk@gmail.com>
This tweaks the signature of `request_builder_hook` in order to support
fallible hooks.
The rationale for this is mostly on two sides:
* it allows a hook to inspect and possibly drop an outgoing request
(e.g. for policying purposes), bubbling up a detailed error message to
the user.
* it wires into newer `reqwest` API which allows to split and then
reassemble a `RequestBuilder`, although only in a fallible way
(https://github.com/seanmonstar/reqwest/pull/1770)
This commit removes "deno_core::RuntimeOptions::extensions_with_js".
Now it's embedders' responsibility to properly register extensions
that will not contains JavaScript sources when running from an existing
snapshot.
Prerequisite for https://github.com/denoland/deno/pull/18080
This commit splits "<ext_name>::init" functions into "init_ops" and
"init_ops_and_esm". That way we don't have to construct list of
ESM sources on each startup if we're running with a snapshot.
In a follow up commit "deno_core" will be changed to not have a split
between "extensions" and "extensions_with_js" - it will be embedders'
responsibility to pass appropriately configured extensions.
Prerequisite for https://github.com/denoland/deno/pull/18080
There's no point for this API to expect result. If something fails it should
result in a panic during build time to signal to embedder that setup is
wrong.
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>
Right now an error in a request body stream causes an uncatchable
global promise rejection. This PR fixes this to instead propagate the
error correctly into the promise returned from `fetch`.
It additionally fixes errored readable stream bodies being treated as
successfully completed bodies by Rust.
https://fetch.spec.whatwg.org/#http-network-or-cache-fetch
> If httpRequest’s header list contains `Range`, then append
(`Accept-Encoding`, `identity`)
> to httpRequest’s header list.
>
> This avoids a failure when handling content codings with a part of an
encoded response.
> Additionally, many servers mistakenly ignore `Range` headers if a
non-identity encoding is accepted.
