Conversation
…leaks, pagination Fixes from code audit (issues #175–#192): - jobs.py: Add threading.Lock to RateLimiter.can_scan/record_scan (#175) - networking.py: Use .get("hops") to avoid KeyError on uninitialized state (#176) - google_drive.py: Enforce 10-minute expiry on OAuth PKCE auth state token (#177) - settings.py, google_drive.py: Remove filesystem paths from user-facing errors (#178) - google_drive.py: Drop path interpolation from missing-credentials error message (#178) - auto_scan.py: Catch ValueError in _parse_time_of_day with safe fallback (#188) - runtime_db.py: Add limit/offset pagination to list_report_artifacts (#180) - workflows.py: Use authoritative customer ID instead of linear name search (#190) Fix pre-existing test failures: - test_workflow_failures: stubs now return {"success": False} dict, not bool - test_auth_routes: add missing ensure_google_drive_reports_folder/save_google_drive_credentials deps - test_client_state: stubs return {"success": True} dict and accept **kwargs - test_runtime_contract: update assertions to match current build.sh and Swift source - test_health_modules: add max_scan_minutes/reports_save_to_desktop to expected payloads All 293 tests pass. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Add NMAPUI_DISABLE_UPDATE_CHECKS to skip GitHub API and idle auto-update UI - Respect NMAPUI_SKIP_OPEN and NMAPUI_HEADLESS for update download URLs; use webbrowser.open cross-platform - IdleStateManager honors update_checks_enabled from app startup - Fix runtime _select_release_asset typing (str | None); add updates_disabled helpers - Log google_drive_upload success/failure to module logger for grep-friendly ops logs - Document audit checklist and deployment env (DEPLOYMENT.md, RELIABILITY_AUDIT.md) - Bump VERSION; ignore .venv in gitignore - Extend tests for update modules Made-with: Cursor
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: a8f60fe8ee
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| return [target], bool(chunked_requested), None | ||
|
|
||
| if not chunked_requested and network.version == 4 and network.prefixlen < 24: | ||
| chunk_targets = [str(subnet) for subnet in network.subnets(new_prefix=24)] |
There was a problem hiding this comment.
When a user disables chunking, _plan_report_scan_targets now eagerly builds chunk_targets as all /24 subnets for any IPv4 CIDR broader than /24. For sufficiently broad but valid targets (for example /8 and larger), this can materialize huge lists before scanning begins, causing long stalls or high memory usage in the report worker; unlike the normal chunking path, this override bypasses the existing 2048-chunk safety cap entirely. Please apply a hard cap (or lazy iteration) here as well.
Useful? React with 👍 / 👎.
| self, | ||
| *, | ||
| customer_id: str | None = None, | ||
| limit: int = 500, |
There was a problem hiding this comment.
Changing list_report_artifacts to default limit=500 silently truncates all existing callers that do not pass a limit. In the current codebase, paths such as runtime history/report listing and customer history retrieval still call list_report_artifacts(...) without an explicit limit, so larger installations will now get incomplete results (for example, get_scan_history(limit=1000) can never return more than 500 runtime artifacts). This should either preserve the previous unbounded default or require callers to pass explicit pagination/limits.
Useful? React with 👍 / 👎.
1 participant
This PR includes improvements for headless operation and Google Drive integration reliability.