agent: builder model: opus max_turns: 110 budget_usd: 22 —
Score: bugs lane (docs/priorities.md, bugs-first default; #890 is a checker-passes/rustc-fails bug from rs2nika lane L, plus the narrowing rule the architect decided on 2026-10-27, ADR-352 D1 as amended in this commit).
fn k1(a: String?, b: i64) -> String {
return match (a, b) { ("out", 1) => "o1", (_, _) => "rest" }
}
fn k2(a: String, b: i64) -> String {
return match (a, b) { ("out", 1) => "o1", (_, _) => "rest" }
}
fn k3(a: String?, b: String?) -> String {
return match (a, b) { ("out", s) if s != null => s, (_, _) => "rest" }
}
fn main() { println(k1("out", 1)); println(k2("out", 1)); println(k3("out", "s")) }nikaia check passes all three. In rustc all fail with E0308: a text literal at a String position of a tuple pattern is written as a &str literal pattern over a String, and at a String? position also without Some(..). A text literal at the top level of a match over a String works (text_scrutinees, #704), so tuple positions do not get the same treatment.
Architect’s decision (step 3): a literal at a position counts as “the same part”, as null and a variant do. ADR-352 D1 now reads “…or the same null, variant or literal this arm has there”: after (false, null), (false, why) binds why as the T. Sound: a literal is one fixed value, so the earlier unguarded arm took null at that position for exactly the values the later arm meets. Part I’s table row (10-nikaia-light.md:1198) is general and stays.
Emit: crates/nikaia-std/src/tools/emit_core_match.nika:216 emit_core_pattern_written, :341 emit_core_arm_pattern_written, :834/:853 (text_scrutinees, how a top-level text match is lowered); the fact is filled in the checker (nikaia run scripts/where.nika -- text_scrutinees). Checker: crates/nikaia-std/src/tools/check_core_flow.nika:617 check_core_flow_same_part (only null and paths today), callers :659, :805.
String or String? matches through a view of that position (for example the scrutinee lowered as a tuple of views (a.as_deref(), b) / (a.as_str(), b), so the literal is a &str pattern and Some("out") at a String? position), the same idea as the top-level text match. Bindings at other positions keep their types (k3’s s is still handed back by value as d89aba7 does). Your choice of form; say it in Body:.tests/compiler/tests/a_text_literal_in_a_tuple_pattern.nika with .stdout: k1, k2, k3, each with a matching and a non-matching call, and a nested tuple (("out", n), m) if the checker accepts it.check_core_flow_same_part also says true for two equal literals (text, integer, bool; compare the literal values, not the spelling 0x1/1 if the AST keeps the value). Test tests/compiler/tests/a_literal_is_the_same_part.nika with .stdout: match (flag, why) { (false, null) => "a", (false, w) => w, (true, _) => "c" } over (bool, String?) runs with w used as String (w.len()), and a .stderr test a_different_literal_is_not_the_same_part.nika where (1, null) before (2, w) leaves w a T? (w.len() refused, NK1125 or what the checker says today). Last commit Fixes #890.The new tests; nikaia test tests/compiler --both-settings; the guard set of .claude/agents/builder.md; lower-std unchanged or changed only where a tools file used the form. Changelog: two lines (step 1 the lowering; step 3 “a literal at a tuple position counts as the same part for narrowing (ADR-352 D1)”: a language change, the supervisor raises the version with ADR-352).
An open design choice; an error in another agent’s file.
Read docs/builder-facts.md.
docs/builder-facts.md first. develop in the main checkout; git status at start: other agents’ uncommitted files are not yours, never touch or stage them.git add <files> only; commit after each step; git push origin develop fast-forward, no rebase (rejected: report Pushed: rejected, go on committing).cargo fmt, cargo clippy --all-targets -- -D warnings, scripts/pre-push-check.sh, and the gate of .claude/agents/builder.md (#852).CHANGELOG.md, badges; no new NK code. New tests in Nikaia only; never the full suite. No foreground sleep, no polling; scripts in files.NIKAIA_CACHE_LIMIT=2G; at most two clones with target/ in the machine; delete your own clone’s target/ and target/release/nikaia-scratch at the end; report GB freed./home/user/Nikaia/docs/briefs/b890-text-literal-tuple-pattern.md.report.md)Fields of docs/architect.md, Briefs and reports: Result: done|noop|partial|blocked (the bare word alone on that line: Result: done, never a summary after it, which cuts the batch, b869a 2026-10-27; Owner: is a question for the owner or none, never a role), Files, Commit, Body (optional, one paragraph for the commit message), Fixes, Pushed, Tests, Gate (green|red), Deviation, Found, Needs, Changelog, Owner, Disk freed. none where nothing, and then the bare word: Deviation: none (a note) cuts the batch (batch.nika compares exactly); a note goes into the commit message. Found: is only for a defect outside your change (it cuts the batch); how you solved your own step goes into the commit message. Deviation: is only a step of the brief built otherwise than written or not built (a choice the brief left to you, such as keep what X needs and say so, is no deviation: it goes into the commit message; nor is building two steps in one commit when every step is built, e12-3a 2026-10-27: say so in Body:); how you verified (e.g. a test not run against the unfixed compiler) goes under Tests:. A noop says why under Tests: (what was probed), not under Deviation:. Final answer, one line: done <sha>, report <path> (or BLOCKED: ... / PUSH REJECTED ...).
Attribution: Co-Authored-By: Claude Opus 5.5 noreply@anthropic.com Claude-Session: https://claude.ai/code/session_01GFeMqxRbtUMiGVyDMxsvEq