This commits adds `DENO_AUTO_SERVE` env var, that when specified
makes `deno run` behave like `deno serve` if the entrypoint satisfies
the `Deno.ServeDefaultExport` interface.
This removes the additional arguments on the function returned by
`Module.wrap`.
Allow libraries that overwrite this behavior (like `v8-code-cache`) to
work correctly.
---------
Co-authored-by: Bartek Iwańczuk <biwanczuk@gmail.com>
This further improves `import.meta.resolve` to not error in many more
scenarios (better alignment with Node).
1. Non-existent files in npm packages
1. Non-existent built-in node modules (ex. `node:non-existent`)
1. Many things that were previously errors with byonm.
1. No longer surfaces some deno_graph resolution errors
Additionally, this defers resolving npm specifiers until loading for
dynamic imports in order to have `prepare_load` properly install them
loading. Before it could potentially error when loading the same npm
specifier on multiple workers (reason for flaky
`specs::npm::worker_shutdown_during_npm_import`).
Fixes#16899.
Fixes https://github.com/denoland/deno/issues/23524.
Fixes https://github.com/denoland/deno/issues/23938.
Fixes https://github.com/denoland/deno/issues/27869.
Unblocks #5501.
This PR adds support for additional stdio pipes to windows, as well as
the detached option in `node:child_process`. I also ported over the
`kill` implementation for windows, which means we now can support
`kill(0)` as well as some other signals.
This means that playwright will now work on windows.
Now that we have a way to support detached processes on all platforms,
we can also easily add a `detached` option to `Deno.Command`, similar to
`child_process.spawn`.
---
The reason for moving away from `std::process::Command` is that the
standard library doesn't expose what we need to control the file
descriptor table of child processes on windows. The implementation here
is based off of parts of `std` and parts of `libuv`, and allows us to
support passing extra pipes in addition to detached processes on
windows.
This commit removes "WorkerGlobalScope" global from the "global
middleware" that we use to provide different set of globals to
user code and npm packages.
This is done, by renaming "WebWorkerType" to "WorkerThreadType"
and introducing a "Node" variant - this variant is used when creating
a worker using "node:worker_threads" module. This worker does
not have a "WorkerGlobalScope" (because it's not a Web Worker)
and the regular Web Worker created using "new Worker" does have
it.
This commit adds support for using
[vsock](https://man7.org/linux/man-pages/man7/vsock.7.html) transport in
fetch API on Linux and macOS.
Similar to #29154, a vsock transport can be specified in the `proxy`
field when calling `Deno.createHttpClient`.
```ts
const client = Deno.createHttpClient({
proxy: {
transport: "vsock",
cid: 2,
port: 80,
},
});
await fetch("http://localhost/ping", { client });
```
todo:
- [ ] cleanup cli, decide what flags we want to commit to
- [x] decide what to do about node addons - (you can mark them external
via `--external`)
- [x] move `esbuild_rs` to the `denoland` org
- [x] figure out the dynamic require issue
- [x] figure out how to test this
- [x] clean up / revert all the random changes
Since `rust 1.87.0` reported `undefined symbol:
ring::pbkdf2::PBKDF2_HMAC_SHA1::*` in CI and it was difficult to debug
locally, use `rust 1.86.0` in CI tests for troubleshoot the errors
This commit adds a signal handler for SIGUSR2 that helps reduce the
memory usage of both main worker and web worker by:
1. Triggering `malloc_trim(0)` to release memory back to the system
2. Invoking V8 isolate's `low_memory_notification` function
This is only available on Linux and enabled when `DENO_USR2_MEMORY_TRIM`
env var is specified.
---------
Co-authored-by: Yusuke Tanaka <yusuktan@maguro.dev>
This commit adds support for using Unix socket proxies in `fetch` API.
This is facilitated by passing an appropriate `Deno.HttpClient` instance
to the `fetch` API:
```
const client = Deno.createHttpClient({
proxy: {
transport: "unix",
path: "/path/to/unix.sock",
},
});
await fetch("http://localhost/ping", { client });
```
Closes https://github.com/denoland/deno/issues/8821
---------
Co-authored-by: Bartek Iwańczuk <biwanczuk@gmail.com>
This commit changes how workers are spawned - instead
of having a "sync" phase that creates a worker and only
then moves to to an "async" block within a Tokio runtime,
now a worker is spawned already on a Tokio runtime.
Closes https://github.com/denoland/deno/issues/28790
Ref https://github.com/denoland/deno/pull/27217
> `op_runtime_cpu_usage` can be optimized by accepting a `&mut [u32]`
instead of returning a serde tuple but it can be done as a follow up.
Accept a `&mut [u32]` instead of serde tuple for `op_runtime_cpu_usage`
and `op_runtime_memory_usage`
---------
Co-authored-by: Divy Srivastava <dj.srivastava23@gmail.com>
implement lazy(?) mode. an unconfigured jsruntime is created if
DENO_UNSTABLE_CONTROL_SOCK is present, and later passed into deno_runtime to be
configured and used.
This commit adds "deno_features" crate that contains definitions of all
unstable features in Deno.
Based on these definitions, both Rust and JS code is generated ensuring
that the two are always in sync.
In addition some of flag handling was rewritten to use the generated
definitions, instead of hand rolling these flag definitions.
---------
Co-authored-by: snek <snek@deno.com>
This is the release commit being forwarded back to main for 2.2.11
---------
Co-authored-by: bartlomieju <bartlomieju@users.noreply.github.com>
Co-authored-by: Bartek Iwańczuk <biwanczuk@gmail.com>
This extracts out the shared libraries and `.node` native modules to a
temp file and opens them from there. **This means that this
implementation will not work in every scenario.** For example, a library
could require other files that only exist in the in-memory file system.
To solve that, we'll introduce
https://github.com/denoland/deno/issues/28918 later or adapt this
solution to solve more issues.
Additionally, this will not work when run on readonly file systems.
This is the release commit being forwarded back to main for 2.2.10
Co-authored-by: nathanwhit <nathanwhit@users.noreply.github.com>
Co-authored-by: Nathan Whitaker <nathan@deno.com>
This is the release commit being forwarded back to main for 2.2.7
Signed-off-by: Divy Srivastava <dj.srivastava23@gmail.com>
Co-authored-by: Divy Srivastava <dj.srivastava23@gmail.com>
