[sergo] Sergo Report: New-Linter Audit (typeassertionnil) - 2026-10-02

[github-actions[bot]]

Overview

Run §36961902897 - strategy: new-linter audit (typeassertionnil, 75th registered analyzer) + reconcile refresh. Registry grew 74->75 since yesterday own run, giving a clean 50/50 split: cached-proven reconcile methodology on one side, fresh exploration of the newly-added linter on the other.

Key Findings

1 issue filed (sg84a1): typeassertionnil (pkg/linters/typeassertionnil/typeassertionnil.go:62-75) gates pointer-type detection on AST shape (typeAssert.Type.(*ast.StarExpr)) before ever consulting pass.TypesInfo. Two valid pointer-assertion shapes escape as a result:

• Type aliases to pointer types (type P = *T; v.(P)) - go/types resolves these transparently to *types.Pointer, but the syntactic gate returns before TypesInfo is reached.
• Parenthesized pointer-type syntax (v.((*T))) - legal per the assertion grammar, but typeAssert.Type becomes an *ast.ParenExpr and the same early-return skips it.

This is the first time the established paren_unwrap_gap pattern class has shown up on the asserted-Type position rather than a value expression - it has now recurred independently in consecutive new-linter audits (closeerrorunchecked at R82, typeassertionnil here). Confirmed clean: the sibling linter typeassertionokdiscarded already covers the x, _ := v.(T) discarded-ok case regardless of pointer-ness, so that is correctly out of scope here, not a gap.

Reconcile

Ran a fresh gh api query against state=open&labels=sergo directly (per last run lesson, rather than trusting a predicted list) - 16 opens, all matching memory expectations exactly. Both of yesterday refiles (sg83a1/sg83a2, #64680/#64681) confirmed landed-as-filed and still open. Zero new reverse-phantom flips this run - first clean zero-flip reconcile since R73.

Metrics

Metric	Value
Serena tools	24 (stable)
Linter registry	75 (+1 from 74)
Issues filed	1
Tasks generated	1
Open sergo issues (post-run)	17
Success score	8/10
Lifetime runs	84
Lifetime avg success	8.74

Next-Run Focus

• Verify sg84a1 landed.
• If the registry grows again (76th), audit fresh and apply a standing day-one paren_unwrap_gap check (bare *ast.BasicLit/*ast.Ident/*ast.CallExpr/*ast.StarExpr assertions without UnwrapParenExpr) to any new linter before anything else, since it keeps recurring independently.
• Check whether closeerrorunchecked (74th) or typeassertionnil (75th) land in cgo.yml LINTER_FLAGS.
• All other established veins (paren_unwrap_gap, branch_state_merge, presnapshot_position_blind, nested_selector_key_collapse, node_filter_too_narrow, suppression_enforce_gap, comment-overlap-guard-parity) remain exhaustively mapped as of R83.

Generated by 🤖 Sergo - Serena Go Expert · claude · agent · 213.3 AIC · ⌖ 7.94 AIC · ⊞ 5.3K · ◷

• expires on Oct 2, 2026, 8:00 PM UTC-08:00

[github-actions[bot]]

This discussion has been marked as outdated by Sergo - Serena Go Expert.

A newer discussion is available at Discussion #65228.