Refactor code to use URL objects instead of file paths wherever possible
💡 Description
Current code flips back and forth between using system file paths (operating system-specific) vs. URL objects. Wherever possible, we should switch to the latter to avoid operating system-specific issues with parse files and paths.
Would this mean an enforcement that URL objects be actually resolvable? @acraugh has concerns about an implicit enforcement of using the canonical schemas rather than locally developed ones
@mdrum no. Sorry. Should have pulled this out of the prioritization.
This is under the hood in the code for how validate is passing around paths and reading files. The user should see no impact besides potentially less bugs related to quirks in file paths.