feat(commands, error_queue): implement 'SYSTem:ERRor:ALL?'

[AdrianDC]

Proposal to add SYSTem:ERRor:ALL? support to list SCPI errors without cleaning the queue.

---

Ugly workaround we had before the patch (unreliable with async multi-clients) :

    // TODO: Migrate to microscpi library
    #[scpi(cmd = "SYSTem:ERRor:ALL?")]
    fn system_error_all_query(&mut self) -> Result<Vec<(i16, String)>, errors::api::Error> {
        let mut errors = Vec::new();
        while let Some(err) = self.error_queue().pop_error() {
            errors.push(err);
        }
        if errors.is_empty() {
            return Ok(vec![(0i16, "".to_string())]);
        }
        let response = errors
            .iter()
            .map(|error| (error.number(), error.to_string()))
            .collect::<Vec<_>>();
        for err in errors {
            self.error_queue().push_error(err);
        }
        Ok(response)
    }

[7h0ma5]

Hi Adrian,

Thank you very much for your contribution!

The change looks good, but it currently fails to compile in no_std environments because Vec is not provided by core. I guess we have three possible approaches:

1. Gate the new functions and command registration behind #[cfg(feature = "std")], making them available only when the standard library is enabled.
2. Add an alloc feature and use alloc::vec::Vec. This would support no_std environments that provide an allocator.
3. Preserve allocation-free no_std support by using a fixed-capacity heapless::Vec. The capacity could be tied to the maximum size of the error queue.

I am leaning to the third option as the error queue size is already bounded.

What do you think?