Improve encrypt/decrypt flow #125
Conversation
Made encrypt/decrypt a "fully functional" flow, so the test dapp can actually be used as an encrypted messenger. Fixes #124
There was a problem hiding this comment.
This design looks good! This will almost certainly break the e2e tests, but we can fix those easily enough. We should publish this as a breaking change to indicate that.
I think the button disabled state logic needs some cleanup. For the two new inputs (public encryption key and encrypted message), we could add listeners to enable the associated button when they are non-empty.
| @@ -258,12 +258,17 @@ <h4 class="card-title"> | |||
| <button | |||
| class="btn btn-primary btn-lg btn-block mb-3" | |||
| id="getEncryptionKeyButton" | |||
| disabled | |||
There was a problem hiding this comment.
Why was this removed? The intention was to have all of the buttons that rely upon an account being connected to start as disabled until we know whether something is connected or not.
There was a problem hiding this comment.
The same goes for the "decrypt" button.
|
|
||
| <button | ||
| class="btn btn-primary btn-lg btn-block mb-3" | ||
| id="encryptButton" | ||
| disabled |
There was a problem hiding this comment.
Similarly, this one was meant to be disabled until we have a public encryption key to encrypt with.
| @@ -6,6 +6,7 @@ import { | |||
| recoverTypedSignatureLegacy, | |||
| recoverTypedSignature, | |||
| recoverTypedSignature_v4 as recoverTypedSignatureV4, | |||
| decrypt, | |||
There was a problem hiding this comment.
This line is causing the linter to fail
| encryptMessageInput.value.length > 0 | ||
| ) { | ||
| if (encryptButton.disabled) { | ||
| encryptButton.disabled = false; |
There was a problem hiding this comment.
Was this deleted to allow encrypting an empty message? I'm not sure why that was disallowed to being with 🤔 Does that work?
Made encrypt/decrypt a "fully functional" flow, so the test dapp can
actually be used as an encrypted messenger.
Changes some element ordering, and the encryption is no longer assumed to the connected address, but requires input, so this could break an e2e test that relies on it.
Fixes #124
Looks like this:
