BREAKING CHANGE: the watcher won't run all tasks but report back on the first failing one
The breaking change is subtle but if one relies on the old behavior of always running all tasks, it's a breaking for them, we'll provide option flags to do so
Reasoning:
I get myself using this form of workflow like TDD red -> green -> refactor:
Change code
Run all checks
Is there a failing check? Then go to 5 (RED), else go to 4
Add to git and commit (GREEN)
Fix what failing the checks (REFACTOR) and go to 2
And with --fail-fast it gives you a nice flow where you "Fix the next failing thing" till all checks are GREEN
BREAKING CHANGE: the watcher won't run all tasks but report back on the first failing one
The breaking change is subtle but if one relies on the old behavior of always running all tasks, it's a breaking for them, we'll provide option flags to do so
Reasoning:
I get myself using this form of workflow like TDD red -> green -> refactor:
And with
--fail-fastit gives you a nice flow where you "Fix the next failing thing" till all checks are GREEN