- The coverage phase is much faster. I didn’t measure exactly
but intuitively I would say 2 times faster than NCover v2. Clearly both
the the assembly instrumentation phase and the execution phase are
faster, but also my guess is that NCover is now able to re-use
instrumented assembly across multiple tests projects.
- The only tack I had with NCover v2 has been fixed: you can now run
NCover from a command line under a non-administrator account, which is central because who
is running its development/build process box under admin privilege
- The NCoverExplorer usability has been enhanced drastically.
NCover v3 comes with several other cool features, but from my point of view much faster + non-admin changes everything!
The ice on the cake is that the integration of this new NCover v3
version with NDepend is seamless. Although the coverage result XML
schema evolved, the NCover team took
care of not breaking the NDepend parsing code (and IMHO that rocks :o).
NDepend is able to import coverage data from both NCover and VSTS in
order to make the most of code coverage data.
For example you can readily list methods added or refactored since your
last release where developers didn’t take the time to have 100%
coverage. One needs to write the following CQL query:
OR CodeWasChanged) AND PercentageCoverage < 100
Even easier, just use the Search by Coverage of Changes and Search by Coverage panels in VisualNDepend: and get in a mouse click everything you want to know about how your code is covered.