Incremental analysis question

[Jeff Gentry]

I have a project using pitest and we make use of the incremental analysis feature. It is incredibly handy in reducing the run time.

One thing I've never understood though is at least in our project it'll still run some things even if nothing at all has changed. For our project that space is getting up there in runtime so I wanted to see if I could figure out what about our code/tests was leading to that. I'm not really sure where to start.

My first question would be is this a known situation? If so, what circumstances tend to lead to things not getting cached via the history file? Are there patterns I could look for?

Thanks

Jeff

[Jeff Gentry]

I believe I have figured out the problem. Spent some quality time with a debugger on the plugin. That's the coverage calculation running. This led me down a path that drew my attetion to the SCM option. On paper that'd work great for us, in our case we have one module that's way too large and thus in practice we'll rarely manage to skip that coverage check. That's on us though, provides incentive to split that module apart.

[Henry Coles]

Hi Jeff,

Sorry I missed your initial post. Yes, the coverage calculation has to always be run at the start of incremental analysis. The coverage data is one of its inputs.

For larger codebases SCM integration is the way to go. If you're using git or junit5 you might be interested in the cdg beta programme

https://pitest.groupcdg.com/

This offers faster analysis, much better git support, and integration with Github, Gitlab and Bitbucket.

Henry

--
You received this message because you are subscribed to the Google Groups "PIT Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pitusers+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/pitusers/0af71269-760e-4e70-af61-8599dd739048n%40googlegroups.com.