Sergey Rubanov
Sergey Rubanov
> whereas Wasm compilation (if asynchronous) can be done in background Yes, but the problem is not main thread blocking, but unnecessary operations (fetch also doesn't block main thread). Also...
> Also I'm wondering what happens when an module (while compiling) gets garbage collected. Good question. I think we can test this case aborting fetch while `instantiateStreaming` / `compileStreaming`.
@lukewagner this won't work for Node.js and other non-browser environments.
Cancellation won't be possible at all until we'll have an API for that in `compile` and `instantiate` methods. (Could we add [JS embedding](https://github.com/WebAssembly/design/issues?q=is%3Aopen+is%3Aissue+label%3A%22JS+embedding%22) label for the issue?)
We now have at least one precedent of using `AbortController` outside of DOM spec — [Web NFC now also uses it](https://github.com/w3c/web-nfc/blob/gh-pages/EXPLAINER.md#integration-with-abortcontroller).
[Web Locks API also uses `AbortController`](https://github.com/WICG/web-locks/blob/master/EXPLAINER.md#signal-option-to-request). Related discussions: - [W3C TAG design principles repo](https://github.com/w3ctag/design-principles/issues/138) - [TC39 Cancellation proposal repo](https://github.com/tc39/proposal-cancellation/issues/22) - [Web NFC repo](https://github.com/w3c/web-nfc/issues/147) - some discussions about `AbortController` in Node.js...
#390 has been merged, so I think we can close this one
@bvisness would you be open to revert changes in gemfile?
VSCode now uses MDN's [compat-data](https://github.com/mdn/browser-compat-data) for testing features support. I hope we'll also be able to use compat-data one day. It's much more machine-readable then compat-table which we use currently...
Safari has implemented it too