Repository navigation
NFR: route 8 atom-or-not wildcards through node_atom_identity_optional - #12669
gunbai-bot[bot] wants to merge 3 commits into
Conversation
The typed non-fold-residue census (#12610) reported ten sites whose wildcard scrutinee has no resolved inferred type. Each is `match x.kind` with x bound by `Present { value: x }` from list_at_optional; the arm tests for an Atom and the wildcard answers everything else. v2.std.grammar node_atom_identity_optional is exactly that projection, so eight of them (plus pattern_head_identity_optional's outer match, the same shape) become total matches over Optional: the wildcard and the undetermined scrutinee both disappear, with no arm cloned and no behavior change (Present <-> the Atom arm, Absent <-> the wildcard). The two remaining (target_model arrow_domain_binder_names and produced_decl_function_type_domain_edges) test for Conj, which has no projection; they keep their wildcards with their own rows in the joint landing on #12610. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
…optional; drop pattern_head_identity_optional (owned by #12669) Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
…uling) arrow_domain_binder_names: every non-Conj NodeKind arm answers Empty (one token). produced_decl_function_type_domain_edges: every non-Conj arm delegates to produced_decl_single_domain_edge, which holds the one positional-edge body. Exhaustive, no copied body; all ten undetermined census sites are now gone. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
|
On the two points the review could not check: |
briansrls
left a comment
briansrls
left a comment
There was a problem hiding this comment.
APPROVE at 39349db. The Optional projection replacements preserve the Atom/non-Atom split, and the two NodeKind/Connective sites exhaust the closed variants while delegating the former wildcard body once. Current checks, including the strict floor, are green.
…-nfr-undetermined # Conflicts: # src/v2/std/compilers/target_model.dag
|
Consolidated into integration PR #12951 at the operator's request; this branch is merged there unchanged in its own merge commit. Closing here; the branch is kept. — sent from deep-ferret-305 |
1 participant
Burn-down on main for the typed non-fold-residue census behind #12610 (plan ruled by neat-boar-16 via silent-koi-310).
The census reported 10 undetermined sites. Each has a wildcard whose scrutinee has no resolved inferred type:
match x.kind, withxbound byPresent { value: x }out oflist_at_optional. The v1 checker doesn't carry the generic payload type into that pattern binder. Each tests for an Atom, and its wildcard answers everything else.v2.std.grammarnode_atom_identity_optionalis exactly that projection. So 8 of them, pluspattern_head_identity_optional's outer match (same shape), become total matches overOptional.Present { value: id }takes the Atom arm andAbsenttakes the wildcard's body verbatim. No arm is cloned and behavior is unchanged.v2.extdeps.runtimes.v2_evaluatorv2_eval_transform_is_int_addv2.lens.undecidable_verdict_collapsepattern_head_identity_optional(inner + outer)v2.lens.vacuityvacuity_transform_operator_symbol_optionalv2.workflow.bash_command_fold_serialize:bash_fold_built_stmt_production_binding,_assign_node,_exit_node,_with_redir_node,_word_nodeThe two
target_modelsites test forConj. They are enumerated per the delegation ruling. Inarrow_domain_binder_namesevery non-Conj arm answersEmpty. Inproduced_decl_function_type_domain_edgesevery non-Conj arm delegates to the newproduced_decl_single_domain_edge. All 10 undetermined sites are gone; none keeps a row.Green under both scans: none of these sites is on the old parameter-keyed roster, and the wildcards are gone. The floor strictly prepares every touched module, so this PR's floor run type-checks all four.
🤖 Generated with Claude Code