Pinned Reports

You can now pin the reports that are most useful to your team on the DeepSource dashboard. To do this at both team and repository levels, hover over the ellipsis icon, select change report, and pick the report you wish to pin.

New in Analyzers

• 9 new checks:
• Java: JAVA-W1065, JAVA-W1064, JAVA-W1063, JAVA-S1064
• Rust: RS-W1079, RS-W1103, RS-W1105, RS-W1106, RS-E1021

Improvements and Fixes

• API: You can now query for reports associated with an Account or a Repository. Learn more in the docs.
• Some users saw an empty state in the history page when they ignored, suppressed, or marked an issue as false-positive or manually opened a link to a run. We've fixed that.
• We've fixed an issue where a person could be added as a team member multiple times.
• Some users reported that when they reactivated/deactivated their repository to trigger a complete repo sync, it resulted in a double count in the code health trend report. We've fixed that.
• The repository sync broke if a Gitlab user had a subgroup with the same name as their account. We've fixed that.
• Fixed unexpected analysis timeouts on specific large Python repositories.
• We were facing a parsing bug while Python's dependency installation. This was explicitly happening for flask>=20.0,<23.0 versions. We've fixed that.
• We've fixed several false positives in this release:
• PYL-C0412: Ungrouped imports from the same package inside a try/except or if block was triggering this issue. We've fixed that.
• JS-0464: This issue should not be raised if the codebase user uses React version >= 17. We have improved the detection of React version in the user's project and raise this issue only if we detect the React version to be < 17.
• JS-D007 and JS-0093: These issues were triggered incorrectly on some assertion functions in test files. We've fixed that.
• JS-0129: When function and class declarations were used before their definition, it incorrectly triggered this issue. We've fixed that.
• JAVA-C1002: Constants and class object references were wrongly reported when used as the expected value of a JUnit assert call. We've fixed that.
• RS-W1017: We've added support for write! macro for this issue.
• Disabled JS-0056: Based on the feedback from some users, we have decided to no longer raise this as an issue since it is a stylistic preference whether to follow this pattern or not.
• We've disabled JAVA-E1029 temporarily, as users reported many false positives. We will re-enable it once we have fixed the false positives.
• The C and C++ Analyzer was not reporting any found issues due to a bug in the issue reporting module. We've fixed that.