thank you! one comment inline

after they find a match

accepting with nit inline

Can we use any random string as an AES-256 key? Or are there some additional requirements?

I ended up adding a new setCommServicesAuthMetadata method because we will set user id, device id, and access token all at once initially, but subsequently may need to update just the access token.
also changed getCommServicesAccessToken to getCommServicesAuthMetadata since we will need to retrieve all three at once

address feedback

https://linear.app/comm/issue/ENG-4944/remove-create-inbound-call-from-createpickledolmsession

remove oneTimeKey and correct create_inbound_from args

address feedback

resolve ESLint/Flow/Jest issues

I found it confusing that we were using oneTimeKey instead of oneTimeKeys for arrays of keys so I changed that as well here

address feedback

rebase

address feedback (CommSecureStore changes will come later, but wanted to acknowledge the comment)

looks pretty good, but have some q's/suggestions

introduce new CryptoModule function to get signed prekey the correct way

should be using olm_account_prekey_signature()

LGTM

actual code looks fine to me but i don't really understand the motivation for this work. i've left a question on ENG-4812; we can discuss there

i'll commandeer and fix up grammar later today

In D8975#265478, @michal wrote:

Adding @tomek because of new dependency while @ashoat is unavailable

But also sha2, seems much more popular than sha256. We use in commtest already

LGTM. I wonder if it's possible to simplify the type conversion logic with folly/dynamic

i left a suggestion and nit inline

LGTM. we should consider adding jitter, though

mistakenly put this up without deduping.. fixing rn