[Defaults] reject incomplete cue.Value when generate defaults #34
Conversation
ying-jeanne
changed the title
|
currently the case would through error instead of because the way we organize tsprintField and tsPrintDefault. the default of a specified list is actually generated in tsprintField, because its' wierd incomplete kind. |
There was a problem hiding this comment.
Seems like it should be good. But we need to expand the matrix of test cases:
- Check that if there's a non-concrete default in a referenced value, the struct that references it
- Check that the above is true in both the case where the referenced value is
@cuetsy-exported, and where it's not - Check that both of the above are true whether the reference is local, or if it crosses an import boundary
i think that set of combinations should work out to four paths to check. They should all be expressible in a single txtar under testdata/imports, similar to the way paths that vary in one dimension are checked in this txtar.
| return ok, err | ||
| } | ||
| for iter.Next() { | ||
| isSubEleConcret, err := isCueObjConcrete(iter.Value()) |
There was a problem hiding this comment.
nit: variable naming.
This is more art than science, but a good way i've heard it described is, variable names should get more detailed as the references to them get further away from their declarations. The only references to this are within the same control block, within a few lines of code; it's preferable to use a shorter, even not-particularly-meaningful name.
This PR is to solve ticket #30, the bottom case is not covered yet, we need to discuss whether we should cover the case.