Refactor 7 functions over cyclo 15; enable gocyclo lint#21
Merged
Conversation
Pulls schema/orphan-purge, module-iteration, git-commit capture, and the two JSON-output paths out of runIndex into named helpers (openIndexDB, purgeOrphanedModules, indexAllModules, captureGitCommit, emitBenchJSON, emitIndexJSON). Behavior unchanged; the per-module count tuple becomes an indexCounts struct so the helpers can return it cleanly. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Extracts collectCallees, printCallees, printCalleesUnresolved, and emitCalleesJSON from execCallees. Behavior unchanged. The collect step now always populates a non-nil slice so the JSON envelope keeps emitting "callees":[] / "unresolved":[] for empty results. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Pulls the short-name resolution switch into resolveTypeRef and the two output paths into emitReferencesJSON / printReferences. Hoists the local refRow type to package scope so the helpers can take it directly. Behavior unchanged. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Extracts clampCallersDepth, buildCallersExplain, collectUnresolvedCallers, printCallersUnresolved, and printCallers from execCallers. The original streamed unresolved lines inline during the fetch loop; the refactor buffers them into a slice, then prints in the same position so the emitted text and ordering for --count, --explain, and plain modes stay identical. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
…kagesStore Both fileHashFallback (sql.DB) and fileHashFallbackStore (store.ReadStore) shared the same shape: load file list, check for new prod-only files on disk, compare per-file hashes, then fall back to the package-level files_hash. Each was over cyclo 15 and used a goto for the fallback hop. Replaces the goto with a (stale, complete) tuple from a comparePerFileHashes helper and pulls the package-hash branch into its own function. Shared "new file detection" logic (hasNewProductionFiles) is reused between both paths via a small basename-set helper. The git-fast-path block in StalePackagesStore moves into loadGitFastPathState + pkgStaleViaFastPath so the per-package loop is flat. No behavior change in the per-file path beyond preserving the original short-circuit on the first empty hash. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The package-scope var-initializer loop in indexCallGraph (decl → spec → value → ast.Inspect for FuncLit) was inline and pushed the function over cyclo 15. Extracts scanCallGraphVarInits and scanFuncLitsInVarInit, plus the two insert-and-count closures into makeCallEdgeRecorder / makeUnresolvedEdgeRecorder methods. The package-iteration body moves into indexPackageCallGraph so the outer function is just the gate + per-package dispatch. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
DefSymbol does a four-step search (exact fqname → normalized fqname → exact name → normalized name) with a pkg filter on each step; inline, that exceeded cyclo 15. Extracts a symbolScanner factory and two narrow helpers (scanByFqname, scanByName) so the function reads as four sequential lookups with early returns. Also drops a stray placeholder mismatch the old normalized-name branch built (an extra "20" appended to args while the LIMIT was already inlined into the SQL string) — the new helper builds the query consistently. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Adds gocyclo to the linter set with min-complexity=16 so any function with cyclomatic complexity > 15 fails the lint pass. This is the same threshold goreportcard.com uses, so the badge can no longer drop from A+ via a CI-clean PR. Excluded for *_test.go where long table-driven tests routinely exceed 15 for benign reasons (the badge counts those, but they don't push the overall score below A+). The existing lefthook pre-commit + CI lint job both invoke tools/golangci-lint.sh, so this single config change wires the gate at both layers — no new scripts or hook plumbing needed. Tooling versions (go, golangci-lint, lefthook) remain pinned in mise.toml. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
runIndex,execCallees,execReferences,execCallers,fileHashFallback/fileHashFallbackStore/StalePackagesStore,(*indexRun).indexCallGraph,(*SQLiteStore).DefSymbol). Behavior preserved; output orderings, JSON envelopes, and error wrappings all unchanged.gocycloin.golangci.ymlatmin-complexity: 16, matching goreportcard's> 15threshold. Excluded for*_test.go(long table-driven tests can stay chunky without dragging the badge below A+).tools/golangci-lint.sh, so this single config change wires the gate at both layers — no new scripts, no new hook plumbing.Context
The badge is currently A+ (97.7% — gocyclo at 80.5%, all other checks at 100%). After this lands, gocyclo bumps too: production code is 100%, the two big test functions remain (which is fine — they don't move the badge off A+).
The two test functions (
TestIndexModuleSymbolsAndCallsCC=28,TestPurgeModuleCleansAllTablesCC=20) are intentionally left alone; goreportcard sees them but the overall weighted average stays well above the A+ threshold (≥ 0.95).Test plan
go test ./...passes (cmd, indexer, store/sqlite all green)golangci-lint runreports 0 issues with the new configlefthookpre-commit and bygolangci-lint rungosymdb indexend-to-end on an isolated copy:index_metarow inserted, symbols/calls counts match pre-refactor baseline (497 symbols, 1365 calls, 5 unresolved, 930 type_refs)🤖 Generated with Claude Code