You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Fixes the AI reviewer findings from PR #229 that were verified as valid (CodeRabbit + cubic).
Changes
utility_signature_parser.rs — brace_pair now guards against reversed ranges (close < open) so the &line[open + 1..close] slice can never panic on arbitrary source text.
utility_signature_parser.rs — collect_python_return_primitives now checks token boundaries: -> IntervalVO no longer reports int, -> StrategyVO no longer reports str, etc.
taxonomy_filesystem_vo.rs — boundary_suffix is now deterministic: picks the longest (most specific) matching key instead of whichever HashMap iteration finds first.
.repowiseIgnore — narrowed barrel exclusions: **/lib.rs re-includes external-lint/src/lib.rs (has convert_executor_error), dropped **/build.rs (shared/build.rs is a real build script). Only verified-pure barrels stay excluded from health scoring.
cargo nextest run -p shared-lint-arwaky — 302/302 pass ✅
lint-arwaky-cli scan . — 0 violations ✅
Reviewer findings assessed as NOT valid (not changed)
boundary_ends_with off-by-one claim — analysis shows the logic is correct (before is the index of the character preceding the suffix because of ends_with).
LineNumber/Score Eq issues — those types no longer exist (removed in prior refactor).
Summary by cubic
Fixes verified reviewer findings from PR #229 to prevent panics, reduce Python parsing false positives, and make path matching deterministic with normalized keys. Also updates .repowiseIgnore so files with real logic are scored.
Bug Fixes
brace_pair guards reversed {} so slicing can’t panic.
Python parsing: params use whole-token matching (: int fires; : internal_id doesn’t); returns use token boundaries (-> IntervalVO doesn’t match int); list [T]/dict [K, V] (with or without a space) are treated as parameterized in both params and returns.
Path matching: boundary_suffix normalizes paths (strips leading .///), picks the longest normalized key and tie-breaks lexicographically, so /src/a.rs and src/a.rs rank the same; boundary_ends_with normalizes both sides and checks the byte before the suffix to fix the off-by-one.
.repowiseIgnore narrows exclusions: re-includes crates/external-lint/src/lib.rs; drops **/build.rs so real build scripts are scored; only verified-pure barrels stay excluded.
Written for commit 20d41bd. Summary will update on new commits.
Addresses issues raised by CodeRabbit and cubic that were verified valid:
1. utility_signature_parser.rs: guard `brace_pair` against reversed ranges
(`close < open`) so the `&line[open + 1..close]` slice can never panic.
2. utility_signature_parser.rs: token-boundary check in
`collect_python_return_primitives` so `-> IntervalVO` no longer matches
`int`, `-> StrategyVO` no longer matches `str`, etc.
3. taxonomy_filesystem_vo.rs: make `boundary_suffix` deterministic by
picking the longest (most specific) matching key instead of relying on
arbitrary HashMap iteration order.
4. .repowiseIgnore: narrow `**/lib.rs` / drop `**/build.rs` so files with
real logic (`external-lint/src/lib.rs`, `shared/build.rs`) stay scored;
only verified-pure barrels remain excluded.
Verification: clippy clean, 302 shared tests pass, self-lint 0 violations.
Co-authored-by: CommandCodeBot <noreply@commandcode.ai>
This issue fixes valid AI reviewer findings from PR #229, focusing on improving the signature parser's handling of reversed ranges and token boundaries. It also refines barrel exclusions for health scoring.
kapybroBot
changed the title
fix(shared): address valid AI reviewer findings from PR #229
Fix AI Review Findings in PR #229
Aug 11, 2026
⚠️ Health of changed files: 6.6 → 5.7 (-0.8) ⚠️ Change risk: moderate, riskier than 61% of this repo's commits.
📋 At a glance
2 files changed health · 2 hotspots touched · 2 below the 7.0 health floor · 2 new findings introduced · 1 co-change pair left out · 2 files with recent fix history · 10 dead-code findings · no tests touched. Scoped to crates.
⚠️ Health gate (advisory) would block if enforced: crates/shared/src/filesystem/taxonomy_filesystem_vo.rs is at 4.9, below the 7.0 floor (+1 more)
📌 Before you merge
No test file imports the changed files and none were touched: consider adding coverage
.../src/agent_filesystem_orchestrator.rs changed together with .../filesystem/taxonomy_filesystem_vo.rs in 10 past commits and isn't in this PR
🔎 More signals (4)
🗺️ Change map
flowchart LR
subgraph PR ["Changed in this PR (2 with dependents)"]
f_crates_shared_src_common_utility_signature_parser_rs[".../common/utility_signature_parser.rs 🔥"]:::changed
f_crates_shared_src_filesystem_taxonomy_filesystem_vo_rs[".../filesystem/taxonomy_filesystem_vo.rs 🔥"]:::changed
end
f_crates_import_rules_src_capabilities_cycle_import_analyzer_rs[".../src/capabilities_cycle_import_analyzer.rs"]
f_crates_shared_src_common_utility_signature_parser_rs --> f_crates_import_rules_src_capabilities_cycle_import_analyzer_rs
f_crates_mcp_server_benches_bench_mcp_server_rs[".../benches/bench_mcp_server.rs"]
f_crates_shared_src_common_utility_signature_parser_rs --> f_crates_mcp_server_benches_bench_mcp_server_rs
f_crates_naming_rules_src_capabilities_naming_convention_checker_rs[".../src/capabilities_naming_convention_checker.rs"]
f_crates_shared_src_common_utility_signature_parser_rs --> f_crates_naming_rules_src_capabilities_naming_convention_checker_rs
f_crates_role_rules_src_agent_role_orchestrator_rs[".../src/agent_role_orchestrator.rs"]
f_crates_shared_src_common_utility_signature_parser_rs --> f_crates_role_rules_src_agent_role_orchestrator_rs
f_crates_filesystem_src_agent_filesystem_orchestrator_rs[".../src/agent_filesystem_orchestrator.rs"]
f_crates_shared_src_filesystem_taxonomy_filesystem_vo_rs --> f_crates_filesystem_src_agent_filesystem_orchestrator_rs
f_crates_filesystem_src_capabilities_ast_parser_rs[".../src/capabilities_ast_parser.rs"]
f_crates_shared_src_filesystem_taxonomy_filesystem_vo_rs --> f_crates_filesystem_src_capabilities_ast_parser_rs
f_crates_filesystem_src_lib_rs[".../src/lib.rs"]
f_crates_shared_src_filesystem_taxonomy_filesystem_vo_rs --> f_crates_filesystem_src_lib_rs
f_crates_orphan_rules_src_capabilities_orphan_contract_analyzer_rs[".../src/capabilities_orphan_contract_analyzer.rs"]
f_crates_shared_src_filesystem_taxonomy_filesystem_vo_rs --> f_crates_orphan_rules_src_capabilities_orphan_contract_analyzer_rs
more(["+4 more dependents"])
PR --> more
w_crates_filesystem_src_agent_filesystem_orchestrator_rs(["⚠️ .../src/agent_filesystem_orchestrator.rs changed together 10×, not in PR"]):::warn
f_crates_shared_src_filesystem_taxonomy_filesystem_vo_rs -.- w_crates_filesystem_src_agent_filesystem_orchestrator_rs
classDef changed fill:#dbeafe,stroke:#1d4ed8,color:#1e3a5f
classDef warn fill:#fef3c7,stroke:#b45309,color:#78350f
classDef guard fill:#dcfce7,stroke:#15803d,color:#14532d
Loading
Solid arrows: code that imports the changed files (12 direct dependents, from the last indexed snapshot). Dashed: history/tests.
📊 See the full report for this PR Your repo map with this PR's blast radius lit up, every caller of the contracts it changes, and health before and after. No sign-in. · ⭐ Star Repowise · 📥 Install bot · Silence on a single PR with [skip repowise] in the title · Per-repo toggle on repowise.dev/settings?tab=bot · Updated 2026-08-12 04:23 UTC
boundary_suffix normalizes paths, evaluates all boundary-aligned suffixes, and selects the longest match. Equal-length matches use lexicographic order. Tests cover the most specific suffix.
I hop through rules where build files hide,
And guard each token’s outer side.
Braces now point the proper way,
Suffixes choose the longest ray.
Thump, thump—clean paths greet the day!
The reason will be displayed to describe this comment to others. Learn more.
Pull Request Overview
The PR aims to resolve technical debt and bugs identified in previous reviews, specifically around string parsing and filesystem mapping. Although Codacy indicates the PR is up to standards, manual inspection revealed a high-severity off-by-one error in boundary_ends_with that will cause path boundary checks to fail.
Furthermore, while the intent was to fix Python primitive type detection, the fix was only partially applied to return signatures, leaving parameter signatures inconsistent. The determinism logic in boundary_suffix also requires a tie-breaker for equal-length matches to be truly stable. The absence of new test cases for the panic guards and mapping logic is a significant gap given the nature of the fixes.
About this PR
This PR addresses several logic bugs and potential panics, but does not include new test cases to verify these fixes or prevent regressions. It is recommended to add unit tests for the reversed brace scenario in brace_pair and the longest-match logic in boundary_suffix.
The improvement to Python primitive detection was applied to collect_python_return_primitives but missed in collect_python_param_primitives (lines 186-203). The latter still uses .contains(': int'), which will continue to produce false positives (e.g., matching : internal_id as int). Please apply the whole-token matching logic consistently across both functions.
1 comment outside of the diffcrates/shared/src/filesystem/taxonomy_filesystem_vo.rs
line 633 🔴 HIGH RISK
The path boundary check incorrectly inspects the first character of the suffix instead of the character immediately preceding it. This will cause suffix matching to fail unless the separator is explicitly included in the match string.
.get(before - 1)
Test suggestions
Test brace_pair with reversed braces (e.g., '} {') to ensure it returns None instead of a range that would panic.
Verify collect_python_return_primitives correctly ignores types that merely start with a primitive name (e.g., 'IntervalVO', 'strategy').
Verify collect_python_return_primitives correctly identifies bare primitives ('int', 'str', 'float').
Test boundary_suffix with multiple overlapping matches to ensure the longest (most specific) match is returned.
Verify is_token_end helper correctly identifies non-identifier characters and EOF as valid token boundaries.
Prompt proposal for missing tests
Consider implementing these tests if applicable:
1. Test `brace_pair` with reversed braces (e.g., '} {') to ensure it returns None instead of a range that would panic.
2. Verify `collect_python_return_primitives` correctly ignores types that merely start with a primitive name (e.g., 'IntervalVO', 'strategy').
3. Verify `collect_python_return_primitives` correctly identifies bare primitives ('int', 'str', 'float').
4. Test `boundary_suffix` with multiple overlapping matches to ensure the longest (most specific) match is returned.
5. Verify `is_token_end` helper correctly identifies non-identifier characters and EOF as valid token boundaries.
The reason will be displayed to describe this comment to others. Learn more.
🟡 MEDIUM RISK
The implementation remains non-deterministic when multiple keys have the same matched length because HashMap iteration order is unstable. Add a lexicographical tie-breaker to ensure consistent results.
Try updating the boundary_suffix loop to track the matching key and use it as a tie-breaker when matched_len == best_len.
The reason will be displayed to describe this comment to others. Learn more.
⚪ LOW RISK
Suggestion: Use Unicode-aware logic to identify token boundaries, as Python identifiers may contain non-ASCII characters. is_ascii_alphanumeric may incorrectly flag the end of a token if followed by a Unicode identifier character.
The reason will be displayed to describe this comment to others. Learn more.
⚪ LOW RISK
Nitpick: This inclusion is redundant because the general **/build.rs exclusion was removed. Consider reverting the removal of **/build.rs if you want to keep other build scripts ignored.
The reason will be displayed to describe this comment to others. Learn more.
Actionable comments posted: 2
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@crates/shared/src/common/utility_signature_parser.rs`:
- Around line 221-226: Update the bare-type checks in the return-type parsing
logic to trim whitespace from the suffix after the list/dict token before
testing for a generic bracket, so spaced annotations such as list [ResultVO] and
dict [KeyVO, ValueVO] are not flagged. Add regression coverage for both
annotations while preserving detection of genuinely unparameterized list and
dict types.
In `@crates/shared/src/filesystem/taxonomy_filesystem_vo.rs`:
- Around line 606-615: Update the reverse suffix-match branch in the candidate
scoring logic to record k_clean.len() rather than clean.len(), so longer
matching candidate keys receive higher scores. Add a regression test covering
multiple matching suffixes, such as /src/a.rs and /a.rs, and assert that the
more-specific key is selected.
🪄 Autofix
Fix all unresolved CodeRabbit comments on this PR:
##[group]Run set -euo pipefail
�[36;1mset -euo pipefail�[0m
�[36;1mecho "RUNNER_TEMP=$RUNNER_TEMP"�[0m
�[36;1mfind "$RUNNER_TEMP" -maxdepth 1 -type f -name 'git-credentials-*.config' -print -delete�[0m
�[36;1m�[0m
�[36;1m# Generate a unique token and stop processing workflow commands to prevent the runtime from injecting commands�[0m
�[36;1mSTOP_***REDACTED_SECRET_ASSIGNMENT*** /proc/sys/kernel/random/uuid)�[0m
�[36;1m�[0m
�[36;1m# Use a trap to ensure we always resume command processing and check for�[0m
�[36;1m# fallback error annotations, even if the runtime exits with a non-zero code�[0m
�[36;1m# (which would otherwise cause set -e to abort the shell before we get here).�[0m
�[36;1m# The trap preserves the original exit code.�[0m
�[36;1mcopilot_cleanup() {�[0m
�[36;1m �[0m
�[36;1m if [ -n "${GIT_PROXY_PID:-}" ] && kill -0 "$GIT_PROXY_PID" 2>/dev/null; then�[0m
�[36;1m echo "Stopping git-proxy (pid=$GIT_PROXY_PID)..."�[0m
�[36;1m kill "$GIT_PROXY_PID" 2>/dev/null || true�[0m
�[36;1m for _ in {1..25}; do�[0m
�[36;1m if ! kill -0 "$GIT_PROXY_PID" 2>/dev/null; then break; fi�[0m
�[36;1m sleep 0.2�[0m
�[36;1m done�[0m
�[36;1m if kill -0 "$GIT_PROXY_PID" 2>/dev/null; then�[0m
�[36;1m echo "git-proxy did not stop gracefully; forcing termination."�[0m
�[36;1m kill -KILL "$GIT_PROXY_PID" 2>/dev/null || true�[0m
�[36;1m fi�[0m
�[36;1m wait "$GIT_PROXY_PID" 2>/dev/null || true�[0m
�[36;1m fi�[0m
�[36;1m �[0m
�[36;1m echo "::$STOP_***REDACTED_SECRET_ASSIGNMENT***
�[36;1m FALLBACK_FILE="${RUNNER_TEMP}/copilot-fallback-error.txt"�[0m
�[36;1m if [ -f "$FALLBACK_FILE" ]; then�[0m
�[36;1m FALLBACK_MSG=$(head -c 500 "$FALLBACK_FILE" | tr -d '\n\r')�[0m
�[36;1m echo "::error title=Copilot Error::${FALLBACK_MSG}"�[0m
🔇 Additional comments (3)
.repowiseIgnore (1)
28-31: 🎯 Functional Correctness
Keep the global barrel patterns.
The global **/mod.rs and **/lib.rs matches contain only module declarations and re-exports. Logic-bearing files are covered by separate test or workspace rules, and crates/external-lint/src/lib.rs is re-included.
1. boundary_ends_with: normalize leading `./`/`/` on both sides before the
boundary check — a suffix like `/b_vo.rs` is treated the same as
`b_vo.rs`, and the separator check applies at the byte before the
matched suffix. Fixes the off-by-one reported by Codacy without
breaking the existing boundary tests.
2. boundary_suffix: score reverse-direction candidates by k_clean.len()
(key specificity) instead of clean.len(), so the most specific key
always wins; tie-break equal lengths lexicographically.
3. collect_python_param_primitives: whole-token matching via
contains_token_prefix (`: int` fires, `: internal_id` does not) —
matches the return-side fix, Unicode-aware.
4. collect_python_return_primitives: allow whitespace before generic
brackets (`list [ResultVO]`, `dict [KeyVO, ValueVO]` are parameterized,
not bare).
5. Tests: regression coverage for spaced generic brackets, most-specific
key selection, and bare list/dict still flagged.
Verification: clippy clean, 304 shared tests pass, self-lint 0 violations.
Co-authored-by: CommandCodeBot <noreply@commandcode.ai>
1. boundary_suffix: score candidates by normalized key length. Keys are
normalized via normalize_path (strips leading ./ and /) before scoring,
so /src/a.rs and src/a.rs are treated as equally specific — the leading
slash can no longer tip the selection. best now stores an owned String
for the tie-break key.
2. collect_python_param_primitives: use contains_bare_token for list/dict
so `x: list [ResultVO]` (space before bracket) is parameterized, not a
bare list — consistent with the return-side fix. Previously the guard
only matched the no-space `: list[` form.
Verification: clippy clean, 304 shared tests pass, self-lint 0 violations.
Co-authored-by: CommandCodeBot <noreply@commandcode.ai>
The reason will be displayed to describe this comment to others. Learn more.
Actionable comments posted: 1
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@crates/shared/src/common/utility_signature_parser.rs`:
- Around line 209-219: Update contains_bare_token to trim leading whitespace
from rest before checking whether the next token starts with '['; spaced generic
annotations such as list [T] and dict [K, V] must not be classified as bare
primitives. Add parameter regression cases covering both forms.
🪄 Autofix
Fix all unresolved CodeRabbit comments on this PR:
…st/dict
contains_bare_token now trims whitespace after the token before checking
for `[`, so parameter annotations like `items: list [ResultVO]` and
`mapping: dict [KeyVO, ValueVO]` are treated as parameterized — not bare
primitives. Adds parameter-side regression coverage.
Verification: clippy clean, 304 shared tests pass, self-lint 0 violations.
Co-authored-by: CommandCodeBot <noreply@commandcode.ai>
The reason will be displayed to describe this comment to others. Learn more.
2 issues found across 2 files (changes from recent commits).
Confidence score: 4/5
In crates/shared/src/filesystem/taxonomy_filesystem_vo.rs, boundary-suffix winner selection is still nondeterministic when two mapping keys normalize to the same value, so different runs can pick different winners and produce inconsistent taxonomy resolution — add a deterministic tie-break for equal normalized keys (or preserve stable ordering explicitly).
In crates/shared/src/filesystem/taxonomy_filesystem_vo.rs, normalization does not fully collapse mixed prefixes like /./..., which means equivalent paths can score differently by k_norm.len() and select the wrong “most specific” mapping — expand normalization to handle mixed-prefix forms before suffix scoring.
Prompt for AI agents (unresolved issues)
Check if these issues are valid — if so, understand the root cause of each and fix them. If appropriate, use sub-agents to investigate and fix each issue separately.
<file name="crates/shared/src/filesystem/taxonomy_filesystem_vo.rs">
<violation number="1" location="crates/shared/src/filesystem/taxonomy_filesystem_vo.rs:621">
P2: Boundary-suffix winner selection can still vary across runs when two mapping keys normalize to the same value. The equal-length tie-break only uses strict `<` on `k_norm`, so equal normalized keys keep whichever entry HashMap yields first; adding a secondary stable comparator (for example raw key) would keep this deterministic.</violation>
<violation number="2" location="crates/shared/src/filesystem/taxonomy_filesystem_vo.rs:634">
P2: Path normalization misses mixed prefixes like `/./...`, so equivalent keys do not always normalize identically before suffix scoring. That can skew "most specific" selection because `k_norm.len()` is used as the score.</violation>
</file>
Requires human review: Auto-approval blocked because this review re-detected 1 unresolved issue already reported by Cubic. Tip: Review your code locally with the cubic CLI to iterate faster.
The reason will be displayed to describe this comment to others. Learn more.
P2: Path normalization misses mixed prefixes like /./..., so equivalent keys do not always normalize identically before suffix scoring. That can skew "most specific" selection because k_norm.len() is used as the score.
Prompt for AI agents
Check if this issue is valid — if so, understand the root cause and fix it. At crates/shared/src/filesystem/taxonomy_filesystem_vo.rs, line 634:
<comment>Path normalization misses mixed prefixes like `/./...`, so equivalent keys do not always normalize identically before suffix scoring. That can skew "most specific" selection because `k_norm.len()` is used as the score.</comment>
<file context>
@@ -595,39 +595,45 @@ impl InboundLinkMap {
+/// Strip leading `./` and `/` so equivalent path forms compare identically.
+fn normalize_path(p: &str) -> &str {
+ p.trim_start_matches("./").trim_start_matches('/')
+}
+
</file context>
The reason will be displayed to describe this comment to others. Learn more.
P2: Boundary-suffix winner selection can still vary across runs when two mapping keys normalize to the same value. The equal-length tie-break only uses strict < on k_norm, so equal normalized keys keep whichever entry HashMap yields first; adding a secondary stable comparator (for example raw key) would keep this deterministic.
Prompt for AI agents
Check if this issue is valid — if so, understand the root cause and fix it. At crates/shared/src/filesystem/taxonomy_filesystem_vo.rs, line 621:
<comment>Boundary-suffix winner selection can still vary across runs when two mapping keys normalize to the same value. The equal-length tie-break only uses strict `<` on `k_norm`, so equal normalized keys keep whichever entry HashMap yields first; adding a secondary stable comparator (for example raw key) would keep this deterministic.</comment>
<file context>
@@ -595,39 +595,45 @@ impl InboundLinkMap {
- Some((_, best_len, best_key)) => {
- len > best_len || (len == best_len && k_clean < best_key)
+ Some((_, best_len, ref best_key)) => {
+ len > best_len || (len == best_len && k_norm < best_key.as_str())
}
};
</file context>
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
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.
kapybro
Bot
changed the title
![codacy-production[bot]](https://avatars.githubusercontent.com/in/56611?s=60&v=4){kind=small}
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
![cubic-dev-ai[bot]](https://avatars.githubusercontent.com/in/1082092?s=60&v=4){kind=small}
Summary
Fixes the AI reviewer findings from PR #229 that were verified as valid (CodeRabbit + cubic).
Changes
utility_signature_parser.rs—brace_pairnow guards against reversed ranges (close < open) so the&line[open + 1..close]slice can never panic on arbitrary source text.utility_signature_parser.rs—collect_python_return_primitivesnow checks token boundaries:-> IntervalVOno longer reportsint,-> StrategyVOno longer reportsstr, etc.taxonomy_filesystem_vo.rs—boundary_suffixis now deterministic: picks the longest (most specific) matching key instead of whichever HashMap iteration finds first..repowiseIgnore— narrowed barrel exclusions:**/lib.rsre-includesexternal-lint/src/lib.rs(hasconvert_executor_error), dropped**/build.rs(shared/build.rs is a real build script). Only verified-pure barrels stay excluded from health scoring.Verification
cargo clippy -p shared-lint-arwaky --all-targets -D warnings✅cargo nextest run -p shared-lint-arwaky— 302/302 pass ✅lint-arwaky-cli scan .— 0 violations ✅Reviewer findings assessed as NOT valid (not changed)
boundary_ends_withoff-by-one claim — analysis shows the logic is correct (beforeis the index of the character preceding the suffix because ofends_with).LineNumber/ScoreEq issues — those types no longer exist (removed in prior refactor).Summary by cubic
Fixes verified reviewer findings from PR #229 to prevent panics, reduce Python parsing false positives, and make path matching deterministic with normalized keys. Also updates
.repowiseIgnoreso files with real logic are scored.brace_pairguards reversed{}so slicing can’t panic.: intfires;: internal_iddoesn’t); returns use token boundaries (-> IntervalVOdoesn’t matchint);list [T]/dict [K, V](with or without a space) are treated as parameterized in both params and returns.boundary_suffixnormalizes paths (strips leading.///), picks the longest normalized key and tie-breaks lexicographically, so/src/a.rsandsrc/a.rsrank the same;boundary_ends_withnormalizes both sides and checks the byte before the suffix to fix the off-by-one..repowiseIgnorenarrows exclusions: re-includescrates/external-lint/src/lib.rs; drops**/build.rsso real build scripts are scored; only verified-pure barrels stay excluded.Written for commit 20d41bd. Summary will update on new commits.