Newly input.json files generated during validate overwrite the previous; and each time the previous version (if any) is saved with a .bak suffix.
So, the first time one runs validate after having edited the input.json (for testing) the situation is merely annoying, because one has to edit/cut/paste values from the .json.bak into the new .json.
But if one runs validate a second time then .json.bak is overwritten with an empty/template version, thus discarding all of one's edits.
This is needlessly fragile and labor-inducing for users. When a new input.json is generated it should retain field values from existing input.json, where applicable, instead of always resetting all fields back to their default values, then the edits will idempotently remain in effect.
dheiman
commented
6 years ago
Newly input.json files generated during validate overwrite the previous; and each time the previous version (if any) is saved with a .bak suffix.
So, the first time one runs validate after having edited the input.json (for testing) the situation is merely annoying, because one has to edit/cut/paste values from the .json.bak into the new .json.
But if one runs validate a second time then .json.bak is overwritten with an empty/template version, thus discarding all of one's edits.
This is needlessly fragile and labor-inducing for users. When a new input.json is generated it should retain field values from existing input.json, where applicable, instead of always resetting all fields back to their default values, then the edits will idempotently remain in effect.