Task: Regenerate the trade Qt estate

The nine ores.qt Trade files come from the model like the rest of the trade estate, so regenerating trade no longer has to stop at the Qt boundary.

• The nine ores.qt Trade files regenerate, with any bespoke members org-declared through the implements mechanism
• The four files reading execution_timestamp as a string are reconciled with whatever domain type the model settles on
• Regenerating trade is idempotent and leaves the tree clean; the full build and ctest stay green against a live environment

(Implementation strategy. Written when work starts; key decisions are distilled into the parent story's * Decisions at close, but the plan itself stays — it is the historical record of what we did.)

Manual QA scenarios (scaffolded via compass add test_scenario, run through the QA Validation Runner panel) that verify this task. Link new ones here as they're created; the scenario doc itself links back via its "Verifies task" field.

Round 1, PR #2101. The fix itself was confirmed by the reviewer, who traced the template independently rather than taking the body's word for it. Two non-blocking nits, both accepted.

Abandoned before implementation, on the evidence the investigation produced rather than on a change of taste.

Codegen can reach the envelope: the detail-dialog template already supports tabbing by view group, DetailDialogBase is already a QWidget so a generated dialog embeds without a base-class change, the foreign-key graph is available for selectors, and the eleven paste points in the dialog implementation are the same augmentation mechanism the service layer now uses. None of that was the obstacle.

The obstacle is that the effort to make a UI fit codegen and C++ was outrunning what the UI was worth. Two findings made that concrete. The instrument form dispatch is pure data and would generate cleanly, but the .ui template has no paste point, so an extra widget has nowhere declared to sit – a new seam invented for the UI alone. And bond, the freshest model, showed the deeper problem: bond_instrument carries five columns, all foreign keys, while the economics live across bond_issue, bond_leg and bond_leg_rate. A form per entity gives a form over a junction row. The containment that would let codegen follow the family is not modelled – bond_leg names instrument_id as a plain column and declares no foreign key at all – so generating a usable instrument form needs a data-model change before it needs a template change.

The direction is now a TypeScript UI using ordinary code reuse rather than codegen. Trade work ends at the C++ layer.

Status at close: Abandoned. Qt is no longer the target.

One fix is kept and travels separately: the entity struct-coverage guard refused a nullable scoped enum, which stopped trade regenerating altogether. That is not Qt work and it blocks anyone regenerating trade, so it is not abandoned with the rest.