From 78a23147f64e2b2f7f13166902fe10c93f8ab481 Mon Sep 17 00:00:00 2001 From: Trinity Bee Date: Sun, 20 Sep 2026 17:33:46 +0000 Subject: [PATCH 1/2] salvage(queen-4486): commit what the turn left uncommitted The turn ended with these files edited and never committed. Uncommitted work is invisible to the review - it reads the branch - so the attempt would have been released as empty and the next bee would have started beside this work rather than from it. This commit is not a claim that the work is correct. It is the bee's work, committed on its behalf, and it is judged exactly like any other: the adversarial reviewer reads it, the compiler runs on it, and the issue's own criteria are measured against it. Issue: #4486 Turn: ea975b04-0b30-4914-876d-5fb9d2b44ebf Ending: reaped (the turn was never closed by its own stream) Committed: 1 path(s) Left uncommitted: 0 path(s) outside the declared boundary --- specs/igla/race/gemm.t27 | 23 +++++++++++++++++++++++ 1 file changed, 23 insertions(+) diff --git a/specs/igla/race/gemm.t27 b/specs/igla/race/gemm.t27 index 1785fc81b3..d7669c1be8 100644 --- a/specs/igla/race/gemm.t27 +++ b/specs/igla/race/gemm.t27 @@ -2269,3 +2269,26 @@ test igla_race_gemm_w392_batch_depth_invariant_2 { } invariant igla_race_gemm_w392_depth: true +test booth_mul_u32_inner_test1 { + let a: u32 = 5; + let b: u32 = 7; + let acc: u32 = 0; + let result = booth_mul_u32_inner(a, b, acc); + assert result == 35; +} + +test booth_mul_u32_inner_test2 { + let a: u32 = 0; + let b: u32 = 12345; + let acc: u32 = 0; + let result = booth_mul_u32_inner(a, b, acc); + assert result == 0; +} + +test booth_mul_u32_inner_test3 { + let a: u32 = 123; + let b: u32 = 456; + let acc: u32 = 0; + let result = booth_mul_u32_inner(a, b, acc); + assert result == 56088; +} From 3ed7c1120ae97cdd317d7b0b064a42ab9ad9bde3 Mon Sep 17 00:00:00 2001 From: Claude Date: Mon, 21 Sep 2026 18:47:55 +0700 Subject: [PATCH 2/2] docs: the coordination entry this branch needs to land A pull request must add exactly one docs/now entry and a bee has no way to know that: its brief names a boundary file and acceptance criteria, and docs/now/ is neither. The publisher adds it rather than failing the gate. Closes #4486 Co-Authored-By: Claude Opus 5 --- ...1-untested-function-in-specs-igla-race-gemm-t27.md | 11 +++++++++++ 1 file changed, 11 insertions(+) create mode 100644 docs/now/2026-09-21-published-test-the-1-untested-function-in-specs-igla-race-gemm-t27.md diff --git a/docs/now/2026-09-21-published-test-the-1-untested-function-in-specs-igla-race-gemm-t27.md b/docs/now/2026-09-21-published-test-the-1-untested-function-in-specs-igla-race-gemm-t27.md new file mode 100644 index 0000000000..6f7964d09f --- /dev/null +++ b/docs/now/2026-09-21-published-test-the-1-untested-function-in-specs-igla-race-gemm-t27.md @@ -0,0 +1,11 @@ +# NOW -- Test the 1 untested function in specs/igla/race/gemm.t27 (published 2026-09-21) + +## A bee's work on #4486, published from `queen-4486` (Closes #4486) + +- The branch changes 1 file(s): `specs/igla/race/gemm.t27`. +- `git diff --stat origin/master...queen-4486` reads: 1 file changed, 23 insertions(+) +- This entry is written by the publisher, not by the bee. A pull request must + add exactly one `docs/now/` entry and a bee has no way to know that: its brief + names a boundary file and acceptance criteria, and `docs/now/` is neither. +- What this entry does NOT establish: that the work is correct. The gates on the + pull request judge that, and they are the same gates every other change meets.