6ed29621ad
A program with no 'main' reached the JIT entry-point call with a garbage address (ORC reports lookup success but leaves main_addr degenerate), then called it -> SIGSEGV. Add a pre-JIT entry-point check in main.zig that emits 'error: no main function found' and exits non-zero before codegen, plus a defensive main_addr==0 guard in target.zig runJITFromObject as a backstop. Regression: examples/1188-diagnostics-run-no-main.sx
40 lines
1.4 KiB
Plaintext
40 lines
1.4 KiB
Plaintext
// issue 0134 — a same-name `error` set collapses into a namespaced import's
|
|
// set (error sets lack per-decl nominal identity).
|
|
//
|
|
// `EventErr` is declared locally as `error { Boom }`, but
|
|
// `#import "modules/std.sx"` also carries `event.EventErr` (an error set with
|
|
// tags Init/Register/Wait). Because error-set DECLARATIONS are not given
|
|
// per-decl nominal identity (unlike struct/enum/union under E6a) —
|
|
// `registerErrorSetDecl` registers via the flat `findByName`-dedup path — the
|
|
// local `EventErr` collapses into the imported one, losing its own `Boom` tag.
|
|
//
|
|
// So `raise error.Boom` / `r == error.Boom` are checked against the IMPORTED
|
|
// set, which has no `Boom`.
|
|
//
|
|
// EXPECT (today): build FAILS —
|
|
// error: error tag 'error.Boom' is not in error set 'EventErr'
|
|
// EXPECT (after fix): prints `own EventErr.Boom`, exit 0.
|
|
//
|
|
// Proof it's the collision: rename `EventErr` -> `MyErr` and it compiles and
|
|
// prints. The reference side (`!EventErr` → resolveNominalLeaf) is already
|
|
// visibility-aware from issue 0132's broader fix, but it is dormant until the
|
|
// local declaration gets its own TypeId. See the .md.
|
|
|
|
#import "modules/std.sx";
|
|
|
|
EventErr :: error { Boom }
|
|
|
|
fail :: () -> !EventErr {
|
|
raise error.Boom;
|
|
}
|
|
|
|
main :: () -> i32 {
|
|
r := fail();
|
|
if r == error.Boom {
|
|
print("own EventErr.Boom\n");
|
|
return 0;
|
|
}
|
|
print("wrong set\n");
|
|
return 1;
|
|
}
|