@aryak please update to most recent ejabberd. The issue is fixed since a few days (not sure if there's already a release that contains that fix).
Alternatively you could disable TLS1.3.
I thought about re-using the AEAD interface, but that one expects the tag/MAC to be appended and we need it separate, therefore I've created a new one.
Maybe you could add this stuff to the PR in the omemo repo we've been working on and make it depend on this package?
hmm, that'd mean that we don't have a working test of the encrypt/decrypt…
[...] we shouldn't comment out the
.SILENT(you can always do it locally if you need to debug, or we can discuss removing it entirely if you want, but overall I just think it makes the…
Another thought (that we can worry about later, I just want to put it on the table but it's not important right now): I'm slightly worried that
scehas no real adoption [...]
It's a…
Sure, I'll implement it from scratch, but it's not only going to be two lines ;)
OK, I'll open a separate PR with only those changes and will base this PR then on top of that one.
The signatures on these methods will potentially require heap-allocating a new byte slice to return if we can't re-use the input slice and more or less preclude any sort of streaming use.
Mayb…
I renamed the SceEnvelope to EncryptedEnvelope, better?
Yeah, this should be in the omemo module, I've only added it as example here since there's no omemo module yet.
Should I create omemo/omemo2.go and add this there? Or leave it here and…