Depend only on Foundation Essentials

[lynchsft]

Addresses #358

N.B. @t089

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 90.94%. Comparing base (9281f8c) to head (5f525b9).

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #453      +/-   ##
==========================================
- Coverage   91.31%   90.94%   -0.37%     
==========================================
  Files          21       21              
  Lines        1797     1790       -7     
==========================================
- Hits         1641     1628      -13     
- Misses        156      162       +6     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[lynchsft]

Hmmm. Seems some cleanup is due around the use of NSNull. It strikes me as an odd choice compared to Optional.none. @t089 do you recall any conversation with jpsim about it?

[lynchsft]

PR supplying NSNull where it is missing: #454

[lynchsft]

@t089 This effort has lead to the necessity of replacing NumberFormatter, which I think is a really terrible decision for all the platforms that do have NumberFormatter.

I've engaged this thought process so far as an abstract exercise – the smallest footprint is the best footprint. But in the face of this challenge, I'll like to ask a more practical question. What kind of platform would actually benefit from diminishing the dependency to merely FoundationEssentials?

Microcontrollers? Kernel modules? What was on your mind when you initially opened the issue? What is on your mind now?

[t089]

That is unfortunate indeed. I think the original observation is valid for any non-Darwin platform:

As an example, a simple "Hello World" Swift binary compiled with swift build -c release --static-swift-stdlib on swift:5.7.1 weighs in at 7.6M. Adding the Yams dependency increases the binary size to 50M.

For a fundamental library such as yams it feels wrong to incur this binary size cost.