From 18c22c575b304d9f32897e92fee381af179ff033 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Ralph=20K=C3=BCpper?= Date: Sun, 30 Aug 2026 20:24:28 +0200 Subject: [PATCH] ci: classify string_array_length_9160 so e2e-scoped stops failing every PR MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit `crates/perry-codegen/tests/string_array_length_9160.rs` landed with #9160 without a `_CODEGEN_SUITES` entry. That map is complete-by-construction for the crate — `_assert_map_covers_codegen_suites` requires every `crates/perry-codegen/tests/*.rs` to be either mapped or excluded — so `--self-test` failed, and since `e2e-scoped` runs the self-test before it does anything else, the job (and `pr-gate` behind it) has been red on every core PR since that merge. The gate did exactly what #7708 built it to do: refuse to let a suite be silently invisible to per-PR CI. It goes in the map rather than in `SUITE_EXCLUSIONS` because it passes: `cargo test -p perry-codegen --test string_array_length_9160` is 1/1 green on 84185b5656, and it is the same shape as the other 28 mapped suites — an in-process compile of hand-built HIR, 0.02 s of test time. An exclusion entry requires a FAILING test and an issue number, and there is neither. Verified: `--self-test` passes; a `crates/perry-codegen/src/` change now selects the suite at the mapped 300 s timeout, and a direct edit of the suite file still selects it at the named 1500 s one. Claude-Session: https://claude.ai/code/session_01TE3JXAYXtdnKcLu8TCFWR6 --- scripts/ci_e2e_scope.py | 1 + 1 file changed, 1 insertion(+) diff --git a/scripts/ci_e2e_scope.py b/scripts/ci_e2e_scope.py index a364c98be3..0372b040e4 100755 --- a/scripts/ci_e2e_scope.py +++ b/scripts/ci_e2e_scope.py @@ -143,6 +143,7 @@ "scalar_replaced_slot_roots", "spec_abi_typed_array_local_length", "static_symbol_hygiene", + "string_array_length_9160", "temp_root_operand_temporaries", "typed_array_rmw_8692", "typed_shape_declared_at_allocation",