Refactor/maccel_core::params::persist -> maccel_core::persist

[sy3c4ll]

This PR includes BREAKING CHANGES. Though I doubt anyone but myself uses maccel_core separately.

As you may be aware, maccel_core::params::persist mainly houses definitions for ParamStore and SysFsStore, a trait for defining behaviour when retrieving and modifying parameters. However the module persist seems to bear no more similarity to params than, say, context, while params.rs is quite sizable and a split may be desirable.

At the same time, while I was using maccel_core in my sy3c4ll/maccel-gui I noticed quirks and inconsistencies with regards to persist::ParamStore, namely that while get and set took in self, get_current_accel_mode and set_current_accel_mode did not while also panicking on error instead of returning an anyhow::Result. This introduced difficulties where I couldn't handle mode switching errors, nor save the mode as a member variable in testing.

This PR addresses these issues.

1. Separate pub mod persist in crates/core/src/params.rs into its own crates/core/src/persist.rs
2. Remove mod validate in crates/core/src/params.rs and extract contents to params.rs while I'm at it, since the module housed only one function which was only necessary for a mysterious Display impl for Fpt in persist.rs...?
3. BREAKING: Change function signatures and implementations of trait ParamStore in persist.rs as described

I've ensured it works for crates cli and tui already, there should be no problems there.

RSVP.

[vercel]

The latest updates on your projects. Learn more about Vercel for Git ↗︎

Name	Status	Preview	Comments	Updated (UTC)
maccel	✅ Ready (Inspect)	Visit Preview	💬 Add feedback	Apr 21, 2025 5:41am

[Gnarus-G]

There are still some compilation errors currently, otherwise LGTM. When you fix those I'll merge.

[sy3c4ll]

Oops, looks like a last-minute addition slipped past my LSP. On it!

To document, my latest commit makes the following additional change:

1. Change argument type of param in ParamStore::get from &Param to Param

[sy3c4ll]

All tests passed! I really skipped everything test-related when I was learning Rust, I shouldn't have done that. TIL.

Oh and uhh...

-    use maccel_core::{fixedptc::Fpt, ContextRef, Parameter};
+    use maccel_core::{ContextRef, Parameter, fixedptc::Fpt};

I may have inadvertently messed up a few import orders... I blame my auto-formatter. Should I revert all of these?

[Gnarus-G]

One more thing, remove that "BREAKING CHANGE" prefix from the commit message on BREAKING CHANGE: Refactor trait ParamStore for consistent error handling and ref self

[Gnarus-G]

All tests passed! I really skipped everything test-related when I was learning Rust, I shouldn't have done that. TIL.

Oh and uhh...

-    use maccel_core::{fixedptc::Fpt, ContextRef, Parameter};
+    use maccel_core::{ContextRef, Parameter, fixedptc::Fpt};

I may have inadvertently messed up a few import orders... I blame my auto-formatter. Should I revert all of these?

Order is not that important to me, what the formatter does is fine. I might add rustfmt configs at some point. But rn you can just format however your rustfmt does it.

[sy3c4ll]

The BREAKING CHANGE: is from Conventional Commits, but whatever you say is king. I'll rebase and force push in a jiffy.

[sy3c4ll]

Done.

[Gnarus-G]

The BREAKING CHANGE: is from Conventional Commits, but whatever you say is king. I'll rebase and force push in a jiffy.

Ain't nobody got time for all these different conventions. But it does look like they recommended putting that in the body/footer of the commit, not the subject line.

I'm generally down with putting whatever in the body.

[sy3c4ll]

Thank you thank you thank you! This happens to be my first successful PR, like, ever! This was such a pleasing experience, and I'm happy to be along in the ride.

[Gnarus-G]

Thank you thank you thank you! This happens to be my first successful PR, like, ever! This was such a pleasing experience, and I'm happy to be along in the ride.

I'm very happy to be the one to merge your first ever PR.