fix: apply the encode callback to empty cookie values

[spokodev]

The documented encode option is meant to mirror decode and apply to every value, so that a custom encoder (signing, encrypting, base64, etc.) round-trips correctly with its matching decode.

stringifySetCookie instead used a truthiness test:

const value = rawValue ? enc(rawValue) : "";

For an empty-string value this skips the encoder entirely and emits a raw name=, instead of name=<encode("")>. That silently breaks encode/decode symmetry whenever the value happens to be empty.

The sibling stringifyCookie helper always calls the encoder, and so did cookie@0.7.2, so this is also an inconsistency between the two stringify paths.

The fix swaps the truthiness check for a null check:

const value = rawValue != null ? enc(rawValue) : "";

Empty strings are now routed through the encoder, while the object-form value: undefined still maps to an empty value, preserving existing behavior.

Added a test that fails before the change and passes after:

it("should apply the encoder to an empty value", function () {
  expect(
    cookie.stringifySetCookie("foo", "", { encode: (v) => "[" + v + "]" }),
  ).toEqual("foo=[]");
});

Full suite stays green (230 tests).

[blakeembrey]

Since the documentation states it should be a string, I think this should always just be enc(rawValue). However, I don't think it's worth changing except in a major version. I can't recall why I made it skip encode, it was probably a premature performance optimization.

[spokodev]

Agreed that always enc(rawValue) is the cleaner contract. My change is narrower and non-breaking: it only routes the empty-string value through encode like every other value, so a custom encode sees consistent behavior without a major bump. If you would rather fold this into the broader enc(rawValue) change in the next major, I am fine closing this in favor of that. Your call on timing.

[blakeembrey]

I'd just fold it into a new major, that's fine by me. No one has mentioned it as an issue and explicitly unsetting a cookie is usually done by passing an empty string as the value, so making it potentially longer isn't much of a plus 🤷

[spokodev]

Fair enough - the empty value is usually an unset, so leaving it unencoded is the lighter default. I will close this in favor of folding it into the next major. Thanks for thinking it through.

[blakeembrey]

Would you like to update this PR or make another? A new major will occur this or next week.

[blakeembrey]

😅 It looks like I didn't realize my types do allow undefined, so LGTM, I merged with master and it'll be good to land for the next major.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 100.00%. Comparing base (4898ba2) to head (099ed03).
⚠️ Report is 35 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff            @@
##            master      #277   +/-   ##
=========================================
  Coverage   100.00%   100.00%           
=========================================
  Files            1         1           
  Lines          160       193   +33     
  Branches        69        82   +13     
=========================================
+ Hits           160       193   +33     

☔ View full report in Codecov by Harness.
📢 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.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.