Additional tests for E2EE #1283
Conversation
🦋 Changeset detectedLatest commit: f4f9897 The changes in this PR will be included in the next version bump. This PR includes changesets to release 1 package
Not sure what this means? Click here to learn what changesets are. Click here if you're a maintainer who wants to add another changeset to this PR |
src/e2ee/worker/FrameCryptor.ts
Outdated
| @@ -555,7 +555,7 @@ export class FrameCryptor extends BaseFrameCryptor { | |||
|
|
|||
| this.sendCounts.set(synchronizationSource, sendCount + 1); | |||
|
|
|||
| return iv; | |||
| return new Uint8Array(iv); | |||
There was a problem hiding this comment.
any reason why you added this? looks like an unnecessary conversion
There was a problem hiding this comment.
I meant to say something about this....
So, the jsdom WebCrypto implementation did not like iv being an ArrayBuffer.
It would fail with:
Failed to normalize algorithm: 'iv' of 'AesGcmParams' (passed algorithm) is not instance of ArrayBuffer, Buffer, TypedArray, or DataView.
...even though iv is an ArrayBuffer. 🤷
This got me thinking (which I then forgot to ask about): Should we be testing across multiple platforms and running these tests inside real browsers? (the goal would be 100% coverage of .browserslistrc)
There was a problem hiding this comment.
So, having played around with running vitest with real browsers, it feels like we could do this in CI, but would be a pain for local dev.
Another approach: if I set the vitest environment to happy-dom or node (instead of jsdom) then the test passes fine.
Would it be acceptable to set the environment to node or happy-dom? (I'm not personally familiar with the latter)
There was a problem hiding this comment.
we use happy-dom in other repos already successfully, so would be good to switch to that!
|
great, thank you! |
This actually does some encryption and decryption with test vectors.