Split up very large JSON files

[caugner]

What would you like to see added to BCD?

Split up large files.

For example: api/Element.json currently has 11,182 lines. Splitting it up would add 212 files under api/Element/ and reduce api/Element.json itself to its own __compat object:

{
  "api": {
    "Element": {
      "__compat": {
        // ...
      }
  }
}

How impactful do you think this enhancement will be?

Smaller files should make both edits and reviews easier.

Do you have anything more you want to share?

Run git ls-files '*.json' | xargs wc -l | sort -n to identify the largest files in terms of line count.

Existing work:

• feat(scripts): add split script #26491 adds a script to split files.
• Split api/Element.json #26492 splits the largest file.

[ddbeck]

My initial reaction to this is slightly negative. As a contributor, I would have no way of knowing, in advance, whether a given tree was "very large" such that I should expect Element.json to exist or api/Element/<subfeature>.json to exist.

Right now, there are a small enough set of file-naming patterns that I can keep them all in my head. This would introduce some one-off exceptions and I'm not sure that's worth the benefit of an absolute ceiling on the size of a file.