Sdss-hal

🔧 Fixed

* Remove leftover `break` that would cause the auto-pilot to exist after one iteration.

🔧 Fixed

* Fix an issue with the new auto-pilot loading a new design during an exposure when a fully loaded one was available.
* Correctly set the observed configuration if the a new one has been loaded.

🚀 New

* [19](https://github.com/sdss/HAL/pull/19) Fully rewritten version of the auto-pilot. The behaviour is generally unchanged but this rewrite fixes several bugs and should allow the auto-pilot to be started and stopped at any point. The command has been changed to `hal auto-pilot` although `hal auto` is still being aliased.

✨ Improved

* Fail `expose` and `goto-field` macros if another macro is running.

🔧 Fixed

* `goto-field acquire` was setting the exposure time to 5 seconds for APO but not using the dynamic exposure time flag.

✨ Improved

* [18](https://github.com/sdss/HAL/pull/18) Implement dynamic exposure times during acquisition. `cherno acquire` (and potentially `cherno guide`) can be called with the `--dynamic-exposure-time` and `--max-exposure-time` flags. Initial exposure time during acquisition defaults to 5 seconds.

🏷️ Changed

* Changed the number of `max_iterations` during acquisition to 5 (from 4) at APO and 7 (from 5) at LCO.

🔧 Fixed

* Fix `hal goto-field --guider-time` flag which was previously ignored.
* Potentially fixed and issue in which if the auto pilot mode is turned on while a configuration is being observed and the next configuration is a repeat field, it was treated as a new field and the additional goto-field stages were executed. The old logic required the configuration to be complete (i.e., at least one exposure done). The new logic just requires the goto-field to have been executed in the previous configuration.