`derivedSeedHex` and Messages throws `Incorrect HMAC` error

[farsadf]

As the Derived Key implementation is not clearly documented, I thought derivedSeedHex would be interchangeable with seedHex; thus, I copied the implementation provided in this repository (including src/lib/ecies/index.js) and configured it on a React Native project running on top of Expo.

After fiddling around to make the crypto, buffer, and stream implementation load correctly, I tried it with my own seedHex (directly out of my browser's localStorage).

Everything seemed to be working; however, as soon as I replaced the seedHex with the derivedSeedHex (given by doing the derived authorization flow on the identity), the implementation started to throw the following error:

[Unhandled promise rejection: Error: Incorrect MAC]
at src/lib/ecies/index.js:26:1 in kdf
at src/lib/ecies/index.js:170:2 in decrypt
at http://127.0.0.1:19000/node_modules/expo/AppEntry.bundle?platform=ios&dev=true&hot=false&minify=false:192892:18 in decryptShared
at src/pages/Inbox/index.tsx:41:20 in useCallback$argument_0
at [native code]:null in flushedQueue
at [native code]:null in invokeCallbackAndReturnFlushedQueue

The error comes from the following section, and removing it, will cause invalid encryption.

assert(hmacGood.equals(msgMac), "Incorrect MAC");

With that said:

1. Does the Derived Keys support decrypting messages? (or in other words, Am I doing something wrong?)
2. If they don't support decryption: when will this feature be available?
3. If they do support decryption: is it only from the messages encoded using Derived Keys?
4. All and all, how should I proceed to implement such a feature?

[farsadf]

So after talking with Ribal from CloutFeed, he told me that there is an endpoint (/get-shared-secrets) to handle this situation; however, such an implementation does have its limitations, primarily a very unusual UX and character length limitation on Nginx.

@maebeam, Is there any sort of alternative method to achieve the same result without this unusual UX?

[AeonSw4n]

Hey @farsadf sorry for responding to this just now. Also please tag me for anything related to derived keys in the future. Decrypting/encrypting messages with derived keys doesn't work for now, besides the shared secret endpoint you mentioned. It's a non-trivial crypto problem to have derived keys handle message but I already conceptualized how to make it possible (rotating messaging keys working similar to derived keys but with HD wallet hardened-derivation). It's on my priority list, so it will be possible soon.

[farsadf]

@AeonSw4n Fantastic mate! Is there an estimated TAT?
Although I'm not that confident with these sorts of issues, if you think that I might be of help, I'd be more than interested to do some parts.