Skip to content

t27b: an optional enum compared with a variant is equal only when it holds that tag, as in the reference (Closes #7811) - #7934

Merged
gHashTag merged 2 commits into
masterfrom
t27b-optional-enum
Oct 8, 2026
Merged

t27b: an optional enum compared with a variant is equal only when it holds that tag, as in the reference (Closes #7811)#7934
gHashTag merged 2 commits into
masterfrom
t27b-optional-enum

Conversation

@gHashTag

@gHashTag gHashTag commented Oct 8, 2026 •
edited
Loading

Copy link
Copy Markdown
Owner

Closes #7811. Part of #6063 (t27b coverage), which is item C10 of #6488.

What the reference does

specs/port/trios/crates/trios-store/rings/ST-00/src/lib.t27 passes the reference 6/6. t27b refused it at lines 126-129 with ExprBinary(?T) ('==' on ?Adapter): assert(Adapter_parse("claude") == Adapter.Claude), where Adapter_parse returns ?Adapter. t27b compared an optional with a value of its payload only when the payload was a scalar.

t27c prints the comparison as written. Zig compares an optional with a value of its payload type. Measured on the t27c lab (t27c test-report, zig 0.16.0, x86_64 Debug):

shape reference
?E == v with the optional holding v / another variant / null true / false / false
?E != v, the same three false / true / true
v == ?E, the optional on the right the same
a ?E local against an E local the same
an enum(u16) with tags 1 and 300 the same
?K == L.A, a variant of another enum blocked: "incompatible types: '?spec.K' and 'spec.L'"
!(?E == v) written inline in an assert blocked: gen-zig prints ! before the comparison (Zig: "expected type 'bool', found '?spec.OeKind'"), so the conformance spec binds such a result to a local first

Decision, with the evidence

The decision is a t27 plan, specs/tri/t27b/optional_compare_plan.t27, with 4 tests. t27c gen-rust turns it into gen/rust/tri/t27b/optional_compare_plan.rs, and lower.rs mounts that file with #[path]. The glue in opt_compare reads the payload's kind and asks payload:

opt_compare's other questions are unchanged and asked first: two optionals, an ordering operator, and an operand with effects.

Conformance spec first

specs/tri/t27b/conformance/optional_enum_compare.t27 has 6 tests:

  • ST-00's Adapter_parse shape: a fn returning a variant or null, with its calls counted;
  • not another variant;
  • null is no variant;
  • the optional on the right;
  • two locals;
  • an enum(u16) with a wide tag.
result
reference, t27c test-report 6/6, 0 vacuous, 20 runtime asserts; sealed (seal --save, then --verify: all hashes MATCH)
master's t27b refused: ExprBinary(?T)
this branch's t27b (aarch64 under qemu, t27b test --check) 6/6, 20 runtime asserts
  • Mutants. 20 mutants of the conformance spec flip one assert each. Every one fails exactly one test, in the reference and in t27b alike.
  • The plan. t27c test-report 4/4, 0 vacuous; sealed. All 5 plan mutants are caught by its tests.
  • tail.rs. A null result fails the test, and ?K == L.A is refused as type mismatch at its line.

Corpus

t27b corpus specs --blockers, master against this branch, on the t27c lab:

  • Unblocked: ST-00/src/lib.t27 and the conformance spec.
  • Still blocked: trios-ternary/rings/TR-00/src/lib.t27 and tri/utils/logging.t27 lose ExprBinary(?T) but keep other blockers. Neither is in the ledger, because the reference does not pass them.
  • No file gains a blocker.

Lines (git diff --numstat --no-renames origin/master...HEAD)

file kind added deleted
cli/t27b/src/lower.rs hand-written glue 6 3
cli/t27b/tests/tail.rs hand-written test 15 0
gen/rust/tri/t27b/optional_compare_plan.rs t27c gen-rust, not hand-edited 32 0
specs/tri/t27b/optional_compare_plan.t27 plan spec, 4 tests 75 0
specs/tri/t27b/conformance/optional_enum_compare.t27 conformance spec, 6 tests 91 0
.trinity/seals/*.json (2) t27c seal --save 40 0
docs/reports/t27b_expectations.json ledger 6 4
AGENTS.md t27b remainder line 4 1

Hand-written foreign code: 21 added lines, 3 deleted (6 and 3 in lower.rs, 15 in tests/tail.rs). That is under the #7371 cap of 40 per file and 80 per PR. Both files are on master's tools/policy/foreign-exceptions.txt. This PR does not change that file and needs no label.

AGENTS.md, re-measured with wc -l: cli/t27b/src/*.rs goes from 15224 to 15227, plus 1202 in src/*/*.rs, and cli/t27b/tests/*.rs from 8149 to 8164, on master 824c82d.

Ledger:

  • ST-00/src/lib.t27: blocked -> pass.
  • Two new rows, and both specs pass.
  • Counts, recomputed from the entries on master 824c82d: pass 928 -> 931, not_pass 34 -> 33, max_not_pass 34 -> 33. Master's header says pass 926, but its 1113 entries hold 928 passes; this PR writes the counts the entries give.

Gates

Both gates ran on the t27c lab, using master 824c82d's gen/c/policy/own_language.c and lefthook.yml's one-line C main.

  • check_budget() over git diff --numstat --no-renames origin/master...HEAD: exit 0.
  • check_all() with origin/master:tools/policy/foreign-exceptions.txt, -- and git diff --name-status origin/master...HEAD: exit 0.

cargo test --release -p t27b on the t27c lab (x86_64, interpreter only), on master 824c82d plus this branch: 129 passed, 0 failed. That includes the new an_optional_enum_compares_by_tag. The two seals verify (all hashes MATCH).

t27b lab: signed receipts (#7686)

Head bd0f84c against its parent, master 7c88448. Head bd0f84c is this branch's first commit, on master 7c88448. Both are signed runs of the deployed t27b lab: the base is the lab's own master run, and the head was requested with a fresh 32-byte challenge.

/work/t27c-master corpus-receipt compare BASE HEAD --challenge-head <mine>, run on the t27b lab from /work/t27:

base "7c884488c8983812295d29d8304090e3693495dc" AUTH_MISSING_NONE AUTHOR leaves-bound true
head "bd0f84c38c988f3ddd4a65ea216f25784cc01935" AUTH_MISSING_NONE FRESH leaves-bound true
  lane improved specs/port/trios/crates/trios-store/rings/ST-00/src/lib.t27
  lane improved specs/tri/t27b/conformance/optional_enum_compare.t27
  lane improved specs/tri/t27b/optional_compare_plan.t27
totals false inputs false verdicts false outputs false: IMPROVED_ONLY

Exit 3, IMPROVED_ONLY. (The input, verdict and output lines name the same three files and are left out here.)

base 7c88448 head bd0f84c
files 1662 1664 (+2: the new specs)
t27b pass / pass_vacuous 989 / 153 992 / 153
t27b fail 19 19
jit_interp_mismatch / crash 0 / 0 0 / 0
reference pass 1177 1179
reference disagree (files / tests) 0 / 0 0 / 0
cargo test --release -p t27b (aarch64, qemu) 144 passed, 0 failed 145 passed, 0 failed

ST-00/src/lib.t27 passes with 19 runtime asserts, the conformance spec with 20 and the plan with 11.

The final head is bd0f84c merged with master 824c82d. That brings in #7744, #7863, #7870 and #7853. The merge resolved only the AGENTS.md line and the ledger counts; lower.rs and tests/tail.rs merged cleanly. Its tree was checked on the t27c lab with the results under Gates.

gen-rust gaps

None. The plan is decisions over u8.

🤖 Generated with Claude Code

claude and others added 2 commits October 9, 2026 01:05
@claude
t27b: an optional enum compared with a variant is equal only when it …
bd0f84c 
…holds that tag, as in the reference (Closes #7811)

t27c prints `?E == v` as written, and Zig compares an optional with a value of
its payload type: equal only when the optional holds a value and that value is
v, so null is `!=` every variant. t27b took that comparison only for a scalar
payload and refused an enum payload as ExprBinary(?T), which blocked
specs/port/trios/crates/trios-store/rings/ST-00/src/lib.t27
(`Adapter_parse("claude") == Adapter.Claude`).

The decision is in specs/tri/t27b/optional_compare_plan.t27 (4 tests),
generated with `t27c gen-rust` to gen/rust/tri/t27b/optional_compare_plan.rs
and mounted in cli/t27b/src/lower.rs. A scalar payload compares by value as
before; an enum payload compares by tag through enum_compare, so a variant of
another enum is refused as `type mismatch`, as Zig refuses it; any other
payload keeps ExprBinary(?T).

The conformance spec is specs/tri/t27b/conformance/optional_enum_compare.t27:
`t27c test-report` 6/6, 0 vacuous. Glue: lower.rs +6 -3, tests/tail.rs +15.
Ledger: ST-00/src/lib.t27 now passes; the two new specs are new rows.

Part of #6063.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@claude
Merge master (#7853, #7870) into t27b-optional-enum (Closes #7811)
8469a48 
lower.rs and tests/tail.rs merge cleanly. AGENTS.md takes master's and adds
this branch's step: 15227 plus 1202 and 8164, on master 824c82d's 15224
plus 1202 and 8149. The ledger takes master's and re-applies this branch's
rows: pass 931, not_pass 33.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@gHashTag
gHashTag enabled auto-merge (squash) October 8, 2026 20:58
This was referenced Oct 8, 2026
Merged
Merged
Merged
Merged
Merged
Merged
Open
Open
Merged
@github-actions

github-actions Bot commented Oct 8, 2026

Copy link
Copy Markdown
Contributor

PR Dashboard

Generated at: 2026-10-08 21:28:08 UTC

Summary

Status Count
Total Open PRs 50
PRs with Failing Checks 45
PRs with All Checks Green 5
READY 3
FAILING 45
PENDING 0
NO CHECKS YET 0

These columns do not partition: 3 + 45 + 0 + 0 = 48, and there are 50 open PRs. A PR is being counted twice or not at all.

Seal Status

  • ⚠️ STALE -- sha256(compiler.rs)=3c0ade9e73e4 != manifest seal=87e5cbd3ad94.
    The committed NMSE numbers were certified against an older compiler.rs.
    Run scripts/reseal-check.sh locally for the two-step reseal command (advisory; not a merge gate).

This was referenced Oct 8, 2026
Merged
Merged
Merged
@gHashTag
gHashTag merged commit 408d740 into master Oct 8, 2026
39 of 40 checks passed
gHashTag pushed a commit that referenced this pull request Oct 8, 2026
@claude
Merge master 408d740 into t27b-ptr-eq-7742 (Closes #7742)
1247620 
From master's side (#7934's optional enum compare; lower.rs conflicted only on the doc comment of opt_compare, kept as master wrote it, with this branch's ptr_compare and ptr_word above it; master's optional_enum_compare.t27, const_optional.t27, postfix_optional.t27 and optional_compare_plan.t27 pass with this merge's t27b under --check).
docs/reports/t27b_expectations.json: master's ledger plus this branch's three
rows, recomposed as before (gen_hazard_pointers.t27 -> pass, two new pass
rows): pass 931 -> 934, max_not_pass 33 -> 32. AGENTS.md: master's remainder plus this branch's clause,
re-measured with `wc -l` over master 408d740: src/*.rs 15227 -> 15266,
src/*/*.rs unchanged, tests/*.rs 8164 -> 8172.

Merged-tree checks on the t27c lab: cargo test --release -p t27b 130
passed, 0 failed; under `t27b test --check` gen_hazard_pointers.t27 (4/4),
pointer_equality.t27, ptr_eq_plan.t27 and opaque_pointer.t27 pass.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
gHashTag pushed a commit that referenced this pull request Oct 9, 2026
@claude
Merge master (#7744, #7863, #7870, #7853, #7934, #7875, #7862) into t…
240ddc6 
…27b-const-div (Closes #7812)

lower.rs and tests/tail.rs keep both sides. AGENTS.md takes master's and adds
this branch's step: 15235 plus 1233 and 8200, on master 4270760's 15232
plus 1233 and 8183. The ledger takes master's (no stored counts since #7862)
and re-applies this branch's rows; max_not_pass 31 -> 30.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@gHashTag gHashTag mentioned this pull request Oct 9, 2026
Open
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

No reviews

Assignees

No one assigned

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

t27b: an optional enum compared with a variant (ST-00 lib)

2 participants

@gHashTag @claude