the workflow would be:
1. git clone the nim compiler repo somewhere
2. to confirm the issue is still present on the latest commit, build the
compiler and use the nimble binary you built to try and reproduce the bug
3. start a bisect in the compiler repo by marking the HEAD as bad and v2.0.0 as
good: `git bisect HEAD v2.0.0 --`
4. it will check out a random commit between HEAD and v2.0.0 - you build the
compiler again, and try to reproduce the bug again
5a. if the bug is still present, you mark the current commit bad: `git bisect
bad`
5b. if the bug is not present, you mark it as good: `git bisect good`
6. repeat steps 4 and 5 until you're only left with one commit (effectively
doing a binary search)
7. the one commit left is what introduced the regression