Steve Gilham

AltCover deliberately doesn't dump much logging information (so it can't leak any end user secrets), so there's not a lot can directly be done. However, being prompted by your request,...

Release 8.8.173 removes the hot-spot that was taking most of the instrumentation time.

Related to the main issue - having profiled the AltCover self-tests, there are no simple hot-spots like there were in the branch cover instrumentation, however * default collection - eager...

Quick answer - it spends a lot of time deserialising the coverage data. Long answer - it depends on how you configure coverage collection. The default configuration opts for robustness...

As a data point for the effect that `--single` has, a larger test set (AltCover's self-test)  Beyond that - the total size of your system under test...

On reflection - there is one useful set of logging that could be provided above and beyond the raw times of coverage processing; and that is a performance profiler run...

Thank you for at last providing some specifics - the actual command line - rather than talking generalities and getting generalities back. 1) the `dotnet test` integration uses a data...

Duly noted, and will add that translation to the things to profile. In general, though, use of a native format (NCover 1.x, OpenCover or, now, the coverlet-based JSON) is to...

For the LCov case, profiling shows that the bottleneck is in the brute-force process of finding for each source file, which methods does it contain. I can see ways to...

Getting back to this after other real life issues. The underlying problem is the lossy cobertura format, which doesn't include any file fingerprint information that would give confidence as to...