feat: move internal to plugin #422
Conversation
The main change is that phase 2 of the wrapping state machine is interactive to accommodate symmetric plugins.
|
Hey! Yeah, the plan is to expose this API, indeed. Seeing it used is extremely helpful, thank you! Looks like besides the current internal plugin package, you'd want a function to parse both native and plugin recipients and identities, and a TUI implementation of the callbacks, is that right? Does the age TUI work as-is for SOPS? |
|
Yes! Surprisingly these two sets of changes allowed me to use sops + age + age-plugin-yubikey. (warning: I did extremely minimal testing.) |
FiloSottile
force-pushed
the
filippo/plugins
branch
4 times, most recently
from
f234662
to
acb1170
Compare
|
Thank you for experimenting with this! I plan to expose the plugin API after some cleanup. Closing this because the filippo/plugins branch is being deleted. |
|
Is it happening? 6976c5f |
|
Hi! Yeah you should find everything but the TUI in https://pkg.go.dev/filippo.io/age/plugin@v1.1.2-0.20230805193447-c89f0b932ecd. Let me know if it looks good, ideally before the next release :) |
|
Awesome! So here's a question, when are we getting a release with this? 😉 |
This is not pretty, but was a very rough attempt to see if it would work to expose the plugin system to sops
It results in a sops binary capable of age1yubikey/AGE-PLUGIN-YUBIKEY plugin actions. So that is promising. Is there continued work to do something along these lines?