bug(capabilities): Core never reads protobuf field 22, so its enabled-channel view can drift from the device and stale the rate cap

[tylerkron]

Summary

Core's idea of which analog channels are enabled is purely what Core itself commanded. It never reads the device's own answer — protobuf field 22, analog_in_port_enabled — so the two can drift, and anything derived from the enabled set is then wrong.

This is scope item 7 of #390, split out because #390 is now closed by #404 and item 7 was deliberately left out of that PR.

Trace (current main, verified after #404 merged)

• No reference to analog_in_port_enabled / AnalogInPortEnabled anywhere outside the generated protobuf class.
• Channels are always constructed disabled — DaqifiDevice.cs:1905 and :1988 both set IsEnabled = false.
• So IsEnabled only ever becomes true because Core set it.

Why it matters

#404 landed the capability document, whose current_max_rate_hz is computed by the device from the device's enabled set. Core caches that value. If Core's enabled view and the device's disagree — a channel enabled by another session, state surviving a reconnect, a partially-applied command — the cached cap is wrong, and Core will either refuse a rate the device would accept or accept one it will reject with SCPI -222.

The drift window is real now in a way it was not before #404: previously the enabled set was only used for Core's own bookkeeping; now a device-computed number depends on it.

Suggested direction

Either:

1. Parse field 22 from the status message and use it as the source of truth for IsEnabled, the way part numbers already drive DeviceCapabilities; or
2. Re-read the capability document at every point the enabled set changes, so the cap is never stale relative to the enabled set.

(1) is the cheaper and more direct answer — the data is already arriving in every status frame and is simply being dropped. (2) costs a SCPI round-trip per change.

Worth confirming on the bench that field 22 is actually populated by firmware ≥ v3.5.0 before building on it.

Refs: #390 (scope item 7), #404, #256, ADR 0001, daqifi-desktop#118.