Remove AtatLen

[Rimpampa]

I don't see how it could be used since even when calling AtatCmd::write the buffer size depends on the client's INGRESS_BUF_SIZE and no check is made by default.

I made this PR just to see if it would compile, if others are actually using this feature then fell free to close it.

[Rimpampa]

@MathiasKoch any thoughts on this?

[MathiasKoch]

Thanks for digging into this. You're right that it's effectively unused.

The premise holds: atat's own client never reads MAX_LEN, since it writes into the caller-provided buf with no bounds check. The only thing that actually consumes the computed value is 3 test-only stack-buffer allocations in ublox-short-range-rs (command/edm/mod.rs:216, 259, 395, all inside #[cfg(test)]). Everything else, meaning the MAX_LEN propagation impls and the manual/derived AtatLen impls in both ublox drivers, is just plumbing feeding a value nothing ultimately reads. No production code path uses it.

So I'm happy to merge this as-is. The drivers are rev-pinned to atat, so nothing breaks until someone bumps the rev, and the consumer-side cleanup is small (drop the dead AtatLen impls/derives, resize 3 test buffers). We can do that in a follow-up when we update the pin.

LGTM.

[Rimpampa]

Nice, then I suggest to take a look just at the atat_derive/src/enum_.rs file since I had to change the Default impl for the AtatEnum derive macro, but other than that I think it's ready to be merged.

[MathiasKoch]

Sorry about the delay on this! It has been a busy week.