fix: Don't use cached traits to initialize sdk state

[tiagoapolo]

#273

As discussed with @kyle-ssg we're not going to cache traits anymore, but for the sake of simplicity we're still saving in local storage because we're storing the everything from the evaluation context but not using it to initialize sdk inner states

• This pull request fix cached traits overwriting server values by not updating traits inner sdk state from cached traits.
• Traits used in init will overwrite previous server value, preserving existing behavior as documented here

[tiagoapolo]

@kyle-ssg I wonder if we still need to keep these tests:

• flagsmith-js-client/test/cache.test.ts

  Lines 337 to 338 in 95b3dec

  	test('should merge any newly passed traits with cache', async () => {
  	const onChange = jest.fn();

• flagsmith-js-client/test/cache.test.ts

  Line 360 in 95b3dec

  test('should cache transient traits correctly', async () => {

[kyle-ssg]

@kyle-ssg I wonder if we still need to keep these tests:

• flagsmith-js-client/test/cache.test.ts

  Lines 337 to 338 in 95b3dec

  	test('should merge any newly passed traits with cache', async () => {
  	const onChange = jest.fn();

• flagsmith-js-client/test/cache.test.ts

  Line 360 in 95b3dec

  test('should cache transient traits correctly', async () => {

@kyle-ssg I wonder if we still need to keep these tests:

• flagsmith-js-client/test/cache.test.ts

  Lines 337 to 338 in 95b3dec

  	test('should merge any newly passed traits with cache', async () => {
  	const onChange = jest.fn();

• flagsmith-js-client/test/cache.test.ts

  Line 360 in 95b3dec

  test('should cache transient traits correctly', async () => {

Yeah with this approach we don't really care about cached traits at all so we don't need the tests

[matthewelwell]

This looks good to me. Are there any considerations we need to make before releasing this, or are we happy that this is just a fix we can release in a patch?

[tiagoapolo]

This looks good to me. Are there any considerations we need to make before releasing this, or are we happy that this is just a fix we can release in a patch?

From my last chat with Kyle we agreed that no one is using that behavior as part of their business logic, also it would be pretty weird, and perhaps a really tiny bit fraction of users are considering caching traits. In summary, no considerations needed.