Add dry run feature

[Narcalen]

Would be nice to have dry run feature similar to cucumber-js (doc)

[badeball]

Hi, @Narcalen, I like the idea. Which use-cases specifically are you looking to see solved with this?

[Narcalen]

Hi @badeball,
I would like to set up PR checks for our automated test framework to make sure that all steps in feature files are implemented, otherwise fail the build. Dry run should be faster than running the tests.

[badeball]

I've introduced a diagnostics utility with v13.0.0, read more about it in docs/diagnostics.md. Please try it out and provide feedback.

NB: This is as experimental as it gets.

[Narcalen]

The feature is working as expected, thanks!
its performance is suboptimal though: for a set of 58 feature files, dry run executes for 80 minutes, and if diagnostics utility is applied for each feature file in loop, it takes 3 minutes to complete the dry run.

[badeball]

Ouf. I tried it on a relatively medium project of mine own, but apparently didn't encounter such terrible performance. I'm guessing I've implemented something with O(n^2) runtime or similar.

[badeball]

Eh, it turns out that cli-table3 is actually the culprit, ref. cli-table/cli-table3#68....

[badeball]

I've released v13.0.3 which contains some improvements, including replacing cli-table3. Please try it out.

[Narcalen]

Ran diagnostics in v13.0.3, performance is good now:
real 0m6.291s
user 0m12.305s
sys 0m1.446s

Thanks!