Open tpaulshippy opened 3 months ago
ah yes this is a great idea, would make a great contribution
this happened to me once before, it looped endlessly not realizing the command itself was just broken
Thought of a different (better?) solution to the same root problem. What if, after the first test failure in interactive mode, the user was given options like this?
Probably could combine options 1 and 4.
Could go back to these options after the max is hit also.
Should the tool execute the test command first before generating anything new, to ensure that it is passing? I tried it on a fresh app and the test command failed for structural reasons so the retries were wasteful.