### 📖 What's in there?
As part of Edge function GA, we're allowing `edge` as a runtime value.
This is adding to the allowed list, so we stay backward compatible.
[Linear ticket](https://linear.app/vercel/issue/EC-481/add-edge-to-the-list-of-allowed-runtimes)
I've kept one instance of `experimental-edge` in the test fixtures to ensure we still supports it.
### 📋 Checklist
#### 🧪 Tests
- [x] The code changed/added as part of this PR has been covered with tests
- [x] All tests pass locally with `yarn test-unit`
#### Code Review
- [x] This PR has a concise title and thorough description useful to a reviewer
- [x] Issue from task tracker has a link to this PR
Remove `--forceExit` and wait for the child `vc dev` process to close
before continuing.
Listen to the `'close'` event instead of `'exit'` to ensure stdio has
been flushed and closed before continuing.
Some tests cause `vc dev` to spawn child processes that in turn spawn
nested child processes that do not exit when the parent exits, so a
`nukeProcessTree()` helper was added to ensure all spawned processes are
cleaned up. Furthermore, some of the nested child processes don't
respond well to `SIGTERM` and we need to use `SIGKILL` if `SIGTERM`
doesn't do the job.
I uncovered a larger issue exposed by removing the `--forceExit` flag.
The go builder uses `go run` to build and run the go-based serverless
function. Turns out that `go run` will build the executable, then spawn
it with a parent process id (ppid) of `1` (launchd) on macOS. This means
that the go serverless function executable is not detected as a child
process of `vc dev` and won't be cleaned up, yet has inherited `stdout`
and `stderr` keeping Jest alive. The solution is to explicitly `go
build` the serverless executable, then run it.
### 📋 Checklist
<!--
Please keep your PR as a Draft until the checklist is complete
-->
#### Tests
- [x] The code changed/added as part of this PR has been covered with
tests
- [x] All tests pass locally with `yarn test-unit`
#### Code Review
- [ ] This PR has a concise title and thorough description useful to a
reviewer
- [ ] Issue from task tracker has a link to this PR
Co-authored-by: Sean Massa <EndangeredMassa@gmail.com>
This makes `vc dev` utilize the same Builder installing and importing logic that `vc build` uses. So now the Builders will be installed locally on a per-project basis, rather than being installed to a shared global installation directory.
### Related Issues
Adds configuration override support to `vc build` and `vc dev` commands.
### 📋 Checklist
<!--
Please keep your PR as a Draft until the checklist is complete
-->
#### Tests
- [ ] The code changed/added as part of this PR has been covered with tests
- [ ] All tests pass locally with `yarn test-unit`
#### Code Review
- [ ] This PR has a concise title and thorough description useful to a reviewer
- [ ] Issue from task tracker has a link to this PR
This was broken since https://github.com/vercel/vercel/pull/8242, which accidentally pushed all WebAssembly modules to be nested under `global.wasmBindings` instead of being directly under the global scope.
### Related Issues
- This was broken since https://github.com/vercel/vercel/pull/8242
### 📋 Checklist
<!--
Please keep your PR as a Draft until the checklist is complete
-->
#### Tests
- [ ] The code changed/added as part of this PR has been covered with tests
- [ ] All tests pass locally with `yarn test-unit`
#### Code Review
- [ ] This PR has a concise title and thorough description useful to a reviewer
- [ ] Issue from task tracker has a link to this PR
Middleware server setup wasn't logging errors the same way that dev server setup was. This meant that middleware instantiation errors (like invalid config) would cause requests to 500, but no errors to be logged to the console.
This PR updates the invalid config error, makes sure errors in this area are logged out, and adds a test for this behavior.
**It may be appropriate to fail the deploy (and crash `vc dev`) in this case instead, though. What do others think?**
---
During `vc dev` with middleware that has an invalid `config.matcher` value...
Before: You see a 500 response in the browser and no output in the terminal.
After: You see a 500 response in the browser and this output in the terminal:
```
Error! Middleware's `config.matcher` values must start with "/". Received: not-a-valid-matcher
```
---
Related Issue: https://github.com/vercel/edge-functions/issues/220
* update edge-runtime to allow `instanceof` to work with primitives
* finishin upgrading edge-runtime
* update to latest
* fix merge
* remove dev only from test
When an edge function has no response during `vc dev`, we were seeing an unhelpful error message:
> The event listener did not respond.
Now, we'll see a much more specific error message:
> Unhandled rejection: Edge Function "api/edge-no-response.ts" did not return a response.
> Error! Failed to complete request to /api/edge-no-response: Error: socket hang up
When edge functions error, they were showing a basic text response instead of the error template. They were also sending a 502 status code instead of a 500.
This PR makes the error handling of Edge Functions consistent with Serverless Functions when executed through `vc dev`. If we want to update the error templates themselves, we can do that in a separate PR.
**Note:** Production currently treats Edge Function errors differently from Serverless Function errors, but that's a known issue that will be resolved.
---
*I deleted the original outputs (terminal and browser screenshots) because they are out of date and added a lot of content to this page. See my latest comment for updated examples.*
Add support for Edge Functions when using `vercel dev`. This uses the [`edge-runtime`](https://edge-runtime.vercel.app/) to emulate the proper environment, [`ncc`](https://github.com/vercel/ncc) to parse the entrypoint (supporting TypeScript and importing node modules), and some custom marshaling to pass the request/response values through.
The `DevServer` class as it's currently implement is not really suitable for unit testing, since it calls `process.exit()` when shutting down the server. This causes Jest to exit prematurely, and prevents tests after these one from running. This was in fact hiding some failing tests. So these tests were ported over to be "Dev" suite tests which spawn a child process of `vc dev` rather than running them in-process, so that the `process.exit()` doesn't exit Jest.
* Migrate ava cli dev tests to jest
* remove outdated test
* update tests
* Update utils
* update stdio to fix stalling
* test stdio tests
* add debug logs
* more logs
* use strip ansi
* update
* add exit flag and fix next dev test
* update timeout for mac ci
* update cancel
* update cancel tests.yml -> test.yml
* update other ids -> names
* remove outdated test
* de-dupe scripts
* remove redwood dev test due to size
