Damir Jelić

Neon introduced a breaking change in a patch release. The breaking change is this commit https://github.com/neon-bindings/neon/commit/a3503ee8242cc3f1db69af0c8341ae6565a7b0e0#diff-6fa50673a346bbd11536a64e3e825ca5L13. Since neon-serde depends on neon and doesn't pin the version this caused breakage after...

> Apologies from Neon—we started treating the neon_runtime module as internal during the N-API migration because it needed lots of these `Env` changes. I'll try to send a PR this...

The disconnect issue is a Python bug: https://bugs.python.org/issue44050 As for the commands, that's a Weechat limitation, scripts share the scope of the underlying plugin for many things, commands amongst others....

That API seems to be a bit heavy for what the suggested use-case in the screenshot is. Should we really return all the signatures? Even the invalid ones? Even of...

We could, though that would mean we check N signatures even though 2 signature checks might be enough for the functionality. Do we want to check every signature?

Oh this is only for tests? Perhaps a flag that determines if we should check all signatures would be the way forward then? Something like `fn verify_backup(auth_data: Value, check_all: bool)...

I feel like this would end up exploding the API surface. As I previously mentioned, we could have the `short_retry()` request config be override-able or, if this really needs to...

What I meant in chat that the `short_retry()` could be overridden, what you suggest here instead overrides the request config only for the `login()` call. The `short_retry()` config is used...

As mentioned in #1065, no movement as of yet.

A contributor has started to work on this.