Identifying New vs Existing issues #1
Description
Musing here on options. Caveats to start:
- It won't be trivial
- It'll likely fail when related cause gets changed
- We'll probably want to allow using simpler methods
Obvious methods to start, worst to best:
- Count the total number of issues for the repo. If it is higher than the baseline, decline the build.
- Count the total number of issues for each file. If one goes up, decline the build.
- Fingerprint by file path and line number
- Fingerprint by logical code location (eg: position in a specific method)
Probably worth ensuring all these options are possible architecturally, then start with the low hanging fruit.