- Bug fix: Updated Azure PM data is is now combined correctly with cached date. Previously historical transitions were sometimes lost.
- Bug fix: Fixed display of team labels for "unmodified"
- Bug fix: Fixed the issue with couplings parsing, which results in an extremely slow system map initial load.
- Bug fix: Local repo path is used as regex when transforming paths in Code Coverage data.
- Bug fix: Analyses using the SonarQube plugin now work properly again. (but note that plugins are deprecated)
- Improvement: Improved sorting of the files in the Detailed Delta Analysis page inside the tooltips popups that show a list of files: sort by largest negative impact first in this list; Remove files where there is zero impact to code health; By default sort the list by largest impact to least.
- Feature: Indicate Code Coverage items on the system map that are mismatched, ie. “a discrepancy between the test commit and the analyzed commit.”
- Feature: Code Coverage tab at Software Portfolio
The Software Portfolio now gives a high-level overview of all your projects in terms of CodeScene's four factors and Code Coverage. It is designed to make it easy to see trends impacting your entire organization and to make high-level prioritizations.
More information on how the average coverage calculation is done can be found here.
- Added exclusion filters on a project level for Code coverage
Projects will often have only partial code coverage data. This could be for several reasons:
- Multiple programming languages in a project: they do not have coverage data yet for one or more of the languages they use
- Multiple repositories: they are currently generating CC data on some repos, but not all
There might also be parts of a codebase that cannot or will not be tested.
All of these situations would lead to many files being marked as having 0% code coverage, in other words as false positives. This would result in an incorrect representation of the code coverage situation, both in terms of which files appear as red 0% files and for the averages that we calculate.
We now have a way for the user to tell CodeScene when the absence of CC data does not mean the absence of code coverage.
