Files
sx/examples/errors/1063-errors-generic-value-failable-trailing-expr.sx
agra 213cedf0b5 refactor: canonical failable syntax (T, !) — remove the bare -> T ! sugar
The trailing-`!`-after-the-value-type spelling (`-> T !`, `-> Tuple(A,B) !`) was a
redundant second way to write a failable return that the parser folded into the
same AST as the parenthesized `(T, !)` / `(A, B, !)` result list. Remove it so
there is ONE canonical spelling: the error channel always rides as the last slot
of the parenthesized list.

- parser: `parseFnReturnType` no longer folds a trailing `!` after a value type —
  it rejects it with a located diagnostic ("a failable return is written `(T, !)`
  … not `T !`"). This one chokepoint covers fn declarations, lambdas, fn-pointer
  types `(A) -> R`, and closure types `Closure(A) -> R`. The error-ONLY `-> !` /
  `-> !ErrSet` form is unaffected (parsed by parseTypeExpr as an error_type_expr).
- migrated every usage to canonical form across library/ + examples/ + issues/ +
  tests/: `-> T !E` → `-> (T, !E)`; the value-carrying `-> Tuple(A, B) !` (which
  FLATTENED to a multi-value failable) → `-> (A, B, !)`, preserving behavior. A
  genuine single-tuple-value failable stays `-> (Tuple(A,B), !)`.
- parser unit tests: the "bare form folds" tests become "bare form is rejected";
  canonical-form parse tests retained.
- docs: specs.md §12 + scattered refs and readme.md updated to the `(T, !)` form.

Behavior-preserving (the bare form was sugar for the same AST). Adversarial review
confirmed: rejection complete across all positions, every canonical form works on
both success/error paths, error-only `-> !` intact, no crashes. Full suite green
(unit tests + 850 corpus examples).
2026-06-27 18:11:20 +03:00

47 lines
1.7 KiB
Plaintext

// Generic value-carrying failable functions (`($T) -> (T, !E)`) whose body ends in
// a trailing success EXPRESSION (no explicit `return`) must set the success
// error slot to 0 — the caller's `catch` must NOT fire and the value must be
// intact across instantiations (i64 / string / struct), and `or` must yield the
// real value not the fallback.
// Regression (issue 0190): `lowerGenericInstance` hand-rolled a body-return
// (coerceToType+ret) that missed the value-failable success routing, leaving
// the error-tag slot uninitialized → phantom catch on success / value
// corruption for generic instantiations.
#import "modules/std.sx";
E :: error { Bad }
Point :: struct { x: i64; y: i64; }
// Generic trailing-expression success — instantiated at i64 / string / struct.
gen :: ($T: Type, v: T) -> (T, !E) { v }
// Generic that RAISES — the caller's catch must still fire.
gfail :: ($T: Type, v: T, bad: bool) -> (T, !E) {
if bad { raise error.Bad; }
v
}
main :: () -> i32 {
a := gen(i64, 42) catch (e) { print("PHANTOM i64\n"); return 1; };
print("i64: {}\n", a);
s := gen(string, "hello") catch (e) { print("PHANTOM string\n"); return 1; };
print("string: {}\n", s);
p := gen(Point, Point.{ x = 3, y = 4 }) catch (e) { print("PHANTOM struct\n"); return 1; };
print("struct: ({},{})\n", p.x, p.y);
// `or`-form on success must yield the real value, not the fallback.
o := gen(i64, 7) or 999;
print("or: {}\n", o);
// A generic that raises still propagates to the caller's catch.
gfail(i64, 0, true) catch (e) {
print("raise caught\n");
return 0;
};
print("UNEXPECTED no error\n");
return 1;
}