# Two-phase refactor — slice 10 (ResultComp + unbounded sequences)

Branch `ast-stage-c`.  Suite **178/178**; self-hosting fixpoint **OK**
(**3,027,925 bytes**; slice 9 was 3,024,098).

## What landed

### Call-result suffixes (`ResultComp`)

`F()[i]`, `F().f`, `F()^` now wrap the call in an `NkDesignator` with
`NkSelector` children (`SelIndex`/`SelField`/`SelDeref`), exactly like
`Design`.  `Fact` builds the `NkCall` right after `ArgList` into a local
`astRes` and passes it **by reference** to `ResultComp`, which wraps it
on the first suffix and appends each selector.

### Unbounded block / decl sequences

Found while dumping `t_resultsfx.mod`: `NkBlock`/`NkDeclSeq` were capped
at `AST.MaxChild` (8) children, so a body with more than 7 statements
**silently dropped** the rest.  Fixed with a chunked sequence:

- a new module-level `AstAppend(kind, VAR head, tail, item)` appends up
  to `MaxChild-1` items per node and chains the overflow through
  `child[MaxChild-1]` into a fresh node of the same kind;
- `StatSeq`, `DeclSeq`, `ConstBlock`, `VarBlock` now use it.

Sequences are therefore unbounded; a walk reads children `0..MaxChild-2`
then follows `child[MaxChild-1]`.

## Verified (temporary `AST.Dump`, then removed)

- `GetRec()^.a` → `NkDesignator(NkCall(GetRec), SelDeref, SelField a)`.
- A body with 12 assignments produced **12** `NkAssign` nodes across
  **2** `NkBlock` chunks (7 + 5) — nothing dropped.

## Deliberately deferred

- Imports, `CLASS`, nested modules in the unit AST.
- `Lower` + `.ssa` byte-compare.
- The same `MaxChild` cap still applies to `CASE` arms, `WITH`
  designators and call actuals — worth the chunked treatment later.

## Next

Imports / `CLASS` / nested modules, then `Lower`.