-
Notifications
You must be signed in to change notification settings - Fork 190
Description
The error spreadsheet (in CSV format here) after closer inspection looks pretty convoluted... I see 2 main issues:
- A lot of errors seem too specific, with no way in sight to integrate them with the schema without significant amounts of explicit reference logic which BST might reasonably cover, but not the schema itself.
- Some entries seem to directly contradict the schema, for example we have
README_FILE_MISSINGwhich is classified as a warning, though the schema explicitly states that this file isrequired.
Going by the example of README_FILE_MISSING I think one possible way to improve error codes in the schema would be to couple them to the requirement level — potentially by deprecating the boolean required and replacing it with requirement (as seen here) which accepts strings corresponding to required, recommended, and optional.
This could in turn be codified as an error of the type {}_FILE_MISSING with error or warning determined by the level of requirement.
Is this something that has been discussed before and I just missed it? If not, would you be interested in discussing this in the next meeting? As it stands I see no way to integrate the present error codes into the schema, as there's nothing to reference them to. We have rules/errors.yaml — but these seem different than the spreadsheet error codes, and the file does not contain any formulation of criteria that can be automatically parsed to assign the error codes... The variable entries other than the code seem to be identical across most cases:
selectors:
- suffix == "events"
- extension == ".tsv"
Metadata
Metadata
Assignees
Labels
Type
Projects
Status