mirror of
https://github.com/Dvorinka/Devour.git
synced 2026-06-03 12:03:06 +00:00
Tomas Dvorak
88 lines
4.9 KiB
Plaintext
88 lines
4.9 KiB
Plaintext
{
|
|
"batch": "Governance & Contracts",
|
|
"batch_index": 3,
|
|
"assessments": {
|
|
"cross_module_architecture": 82.0,
|
|
"test_strategy": 74.0
|
|
},
|
|
"dimension_notes": {
|
|
"cross_module_architecture": {
|
|
"evidence": [
|
|
"`internal/quality/types.go` defines a typed status contract (`type Status string` with constants like `StatusOpen`, `StatusFixed`, `StatusWontfix`).",
|
|
"`internal/quality/types.go` also defines `Scorecard.StatusByType map[string]int`, which bypasses the typed status contract at the module boundary.",
|
|
"`internal/quality/scoring_test.go` asserts raw string keys (`\"open\"`, `\"fixed\"`) instead of using `Status` constants, reinforcing stringly-typed cross-component coupling.",
|
|
"`README.md` claims quality features include tracking resolution states, but the in-code state transport for scorecards is weakly typed."
|
|
],
|
|
"impact_scope": "subsystem",
|
|
"fix_scope": "multi_file_refactor",
|
|
"confidence": "high"
|
|
},
|
|
"test_strategy": {
|
|
"evidence": [
|
|
"`internal/quality/narrative_test.go` validates exact headline/action prose and directly tests internal helper behavior (e.g., `determinePhase`, `generateHeadline`, `classifyDimension`), creating high implementation-coupling.",
|
|
"`internal/quality/scoring_test.go` similarly focuses on exact internal scoring details and string key literals, which makes refactors noisy and discourages safe design changes.",
|
|
"`pkg/rustdocs/parser_test.go` is heavily happy-path: it checks successful parses and minimal field presence but has no malformed-input/error-path cases for parser resilience.",
|
|
"`README.md` marks parts of the CLI as unstable/stubbed, but assigned tests do not provide cross-module contract/integration safety nets for those runtime boundaries."
|
|
],
|
|
"impact_scope": "subsystem",
|
|
"fix_scope": "multi_file_refactor",
|
|
"confidence": "high"
|
|
}
|
|
},
|
|
"findings": [
|
|
{
|
|
"dimension": "cross_module_architecture",
|
|
"identifier": "status_contract_string_map_boundary",
|
|
"summary": "Scorecard state uses string keys instead of shared Status type, weakening module contracts.",
|
|
"related_files": [
|
|
"internal/quality/types.go",
|
|
"internal/quality/scoring_test.go",
|
|
"README.md"
|
|
],
|
|
"evidence": [
|
|
"`internal/quality/types.go` defines `Status` constants but `Scorecard.StatusByType` is `map[string]int`.",
|
|
"`internal/quality/scoring_test.go` asserts `card.StatusByType[\"open\"]` and `card.StatusByType[\"fixed\"]` directly.",
|
|
"README promises resolution-state tracking, but this boundary is not type-safe."
|
|
],
|
|
"suggestion": "Change `Scorecard.StatusByType` to `map[Status]int` (or a dedicated typed struct), update serialization adapters if needed, and update tests to assert using `StatusOpen`/`StatusFixed` constants.",
|
|
"confidence": "high",
|
|
"impact_scope": "subsystem",
|
|
"fix_scope": "multi_file_refactor"
|
|
},
|
|
{
|
|
"dimension": "test_strategy",
|
|
"identifier": "brittle_private_and_copy_assertions_in_quality_tests",
|
|
"summary": "Quality tests are tightly coupled to private helpers and exact copy text, reducing refactor safety.",
|
|
"related_files": [
|
|
"internal/quality/narrative_test.go",
|
|
"internal/quality/scoring_test.go"
|
|
],
|
|
"evidence": [
|
|
"`narrative_test.go` directly asserts exact strings for generated headlines/actions and tests helper internals rather than stable external behavior.",
|
|
"`scoring_test.go` anchors on specific internal weighting outputs and literal status strings, which can fail on benign internal redesigns."
|
|
],
|
|
"suggestion": "Shift to contract-level tests against exported APIs with invariant assertions (phase category, presence of required fields, monotonic score behavior), and keep only a small set of snapshot/copy tests for user-facing text.",
|
|
"confidence": "high",
|
|
"impact_scope": "module",
|
|
"fix_scope": "multi_file_refactor"
|
|
},
|
|
{
|
|
"dimension": "test_strategy",
|
|
"identifier": "rust_parser_missing_negative_and_boundary_cases",
|
|
"summary": "Rust parser tests miss malformed-input and degradation-path coverage.",
|
|
"related_files": [
|
|
"pkg/rustdocs/parser_test.go",
|
|
"README.md"
|
|
],
|
|
"evidence": [
|
|
"`parser_test.go` cases are successful parses with valid fixture HTML and only basic assertions.",
|
|
"No tests verify behavior for malformed HTML, missing selectors, empty documents, or unsupported result rows.",
|
|
"README positions docs ingestion as core functionality, so parser failure behavior is a critical path."
|
|
],
|
|
"suggestion": "Add table-driven negative tests for malformed/partial HTML, empty search blocks, and missing headings; assert stable fallback behavior (explicit error or safe zero-value output) for each parser entrypoint.",
|
|
"confidence": "high",
|
|
"impact_scope": "module",
|
|
"fix_scope": "single_edit"
|
|
}
|
|
]
|
|
} |