ff9e448f8c
Five adversarial reviews of the issue-0160 fix surfaced three more bugs in the touched optional-chain / optional-coercion code; all fixed here: 1. A COLD generic-instance getter through `?.` (`?*Vec(i64)` `.getter`, never called directly first) panicked with "unresolved type reached LLVM emission": a cold instance method is absent from resolveFuncByName, so the getter's return type resolved to .unresolved → a ?unresolved merge type. lowerOptionalChain and getterReturnTypeOnDeref now warm the monomorph (ensureGenericInstanceMethodLowered) before querying its return type. (The 0907 test passed only by luck — List(i64) is warmed by stdlib use; 0907 now also exercises a cold user generic.) 2. A real-field read through a `?*T` chain (`op?.field`, op: ?*T) reinterpreted the pointer bits as the field (silent garbage) — the some-branch real-field path didn't load through the pointer. It now derefs `?*T` before the field access. (Pre-existing — the else-branch predates 0160 — but it's the same function and a silent miscompile, so fixed here.) 3. `?[]T = array` skipped the array→slice promotion (corrupt .len/.ptr): the lowerVarDecl optional arm wrapped the raw array. It now coerces the value to the optional's child type (array→slice) before wrapping. Regression examples 0906/0907 extended to cover all three. Distinct PRE-EXISTING bugs the reviews surfaced in untouched subsystems are filed as issues 0161 (struct-literal vs scalar), 0162 (#run returning an optional aggregate), 0163 (untagged-union payload-binding match).
1.7 KiB
0162 — #run returning an optional aggregate fails the comptime VM reg→value bridge
Symptom
A #run (or comptime const init) whose function returns an OPTIONAL value
(?T, ?i64, any optional) fails comptime evaluation with:
error: comptime init of 'X' failed: reg→value: aggregate shape not bridged yet
A non-optional return of the same type works. This is a pre-existing limitation
in the comptime VM's register→value bridge for optional-typed results; it is
orthogonal to issue 0160 (it reproduces for a value-init optional with no struct
literal anywhere, and for a scalar optional ?i64).
Reproduction
#import "modules/std.sx";
T :: struct { a: i64 = 0; }
mk :: () -> ?T { t : T = .{ a = 7 }; return t; }
mk2 :: () -> ?i64 { return 5; }
X :: #run mk(); // error: reg→value: aggregate shape not bridged yet
Y :: #run mk2(); // same class of failure
main :: () { print("ok\n"); }
Baseline that WORKS: Z :: #run (() -> T { return .{ a = 7 }; })(); (non-optional).
Investigation prompt
src/ir/comptime_vm.zig — the reg→value bridge (search "aggregate shape not
bridged" / regToValue) handles scalars/structs/slices but bails on an
OPTIONAL-typed result. An optional is {payload, has_value} (or a pointer for
?*T / a sentinel for ?Closure); the bridge needs to read the has_value flag
and, when set, bridge the payload as its child type (recursively), producing a
Value optional — and a null optional when clear. Add the .optional arm to
the reg→value bridge (mirror the value→reg direction, which already builds
optionals — see makeStringList/writeField optional handling). Verify with
the repro (expect X/Y to evaluate, main prints ok). Add a
examples/comptime/06xx-... regression.