Passwords will be automatically re-hashed on sign-in. All new password hashes will be argon2 by default. This uses argon2id and is not configurable. In the very unlikely case someone has more specific needs, a fork is recommended. ChangeLog: Added Co-authored-by: Chloe Kudryavtsev <code@toast.bunkerlabs.net>
Breaks Calckey -> Misskey migration, but fixes Foundkey -> Calckey migration
This uses [napi-rs](https://napi.rs/) to allow for automatic generation of node bindings for the native code.
I also changed the `isolatedModules` TS flag to false to allow for `static enum` to be shared across modules. It doesn't seem to be necessary for the build system that CK uses.
Currently this method does not work with ID generators with longer IDs. Likely the best solution is to add another key in the database.
Some benchmarks for 1 million conversions:
```
node, x1_000_000: 2.847s
rust, x1_000_000: 1.265s
```
There are still optimizations that can be made, but I think this is a good starting point and a good way to bring rust into the CK stack.
Co-authored-by: s1idewhist1e <trombonedude05@gmail.com>
Reviewed-on: https://codeberg.org/calckey/calckey/pulls/9786
Co-authored-by: s1idewhist1e <s1idewhist1e@noreply.codeberg.org>
Co-committed-by: s1idewhist1e <s1idewhist1e@noreply.codeberg.org>
Sorry to create PR multiple times. I should have included this in #9778.
Co-authored-by: naskya <m@naskya.net>
Reviewed-on: https://codeberg.org/calckey/calckey/pulls/9783
Co-authored-by: naskya <naskya@noreply.codeberg.org>
Co-committed-by: naskya <naskya@noreply.codeberg.org>
A change sometime ago moved to setting some signature fields in the incoming
object to undefined as opposed to deleting them. The trouble is that downstream
code checks against existence, not undefinedness and rejects the message.
Resolves: #9665
