Support for stacked pull-requests
We're excited to announce first-class support for stacked pull requests on DeepSource. If you're creating pull requests off existing pull requests (also known as stacking), DeepSource will now correctly diff the second pull request and show you only those issues that you're introducing in that pull requests. Prior to this, we reverted to showing you all issues present on the branch in this case (which wasn't very nice, really).
New Transformer: Ruff
Ruff seems to have taken over the Python world lately, and we're excited to support it on DeepSource today! If you're switching your code formatting on Python to Ruff, just enable the Transformer in your `.deepsource.toml` and you should be good to go. More details in the docs.
Support for GitLab subgroups
Large teams that use GitLab often use their subgroups feature to organize their projects more meaningfully. Unlike other VCS providers like GitHub or Bitbucket, however, the subgroups are a concrete namespace for the projects and not just a loose group. We're excited to announce first-class support for GitLab subgroups on DeepSource, which will allow our users on GitLab to bring all of their repositories on DeepSource.
All your nested projects, regardless of the level of nesting, should already be synced and ready on the dashboard. If you don't see a project, just sync the repositories again from the repository activation modal.
Fixes and Improvements
- Subrepos of a repository in the monorepo mode are now supported in Auto Onboard
- When creating a ticket on Atlassian Jira from DeepSource, the reporter is now set to the current user (when available)
- On Azure DevOps Sevices teams, repository syncing no longer fails when there are multiple repositories with the same name under different projects
- Committing the DeepSource config file on GitLab when branch protection rules are enabled no longer fails; it reverts to creating a merge-request now
- Transferring repositories on Bitbucket is now reflected on DeepSource automatically without having to run a manual sync
- We fixed an issue where the severity counts in the OWASP® Top 10 report were incorrect in certain cases
- We fixed an issue in our Atlassian Jira integration where the user could select an incorrect ticket type for a project