fix(hmr-plugin): correct persistence state in runtime

[splincode]

PR Checklist

Please check if your PR fulfills the following requirements:

• The commit message follows our guidelines: https://github.com/ngxs/store/blob/master/CONTRIBUTING.md#commit
• Tests for the changes have been added (for bug fixes / features)
• Docs have been added / updated (for bug fixes / features)

PR Type

What kind of change does this PR introduce?

[x] Bugfix
[ ] Feature
[ ] Code style update (formatting, local variables)
[ ] Refactoring (no functional changes, no api changes)
[ ] Build related changes
[ ] CI related changes
[ ] Documentation content changes
[ ] Other... Please describe:

What is the current behavior?

#917

What is the new behavior?

• Current HMR-plugin worked incorrectly, now this saving works through a global object and and is used before other plugins are initialized and states bootstraped.

• Correct Life-cycle events triggered in root state and others

Does this PR introduce a breaking change?

[ ] Yes
[x] No

[splincode]

@markwhitfeld please take a look

[markwhitfeld]

Which test does this PR fix?
I don't like the HMR stuff in the store internals and I think I can make some small modifications that would fix this but I need to know which test this work aims to fix.

[splincode]

@markwhitfeld https://github.com/ngxs/store/blob/45a997ac022b9bfe11c82c48a2be61d6e25712e0/packages/store/internals/testing/tests/ngxs.setup.spec.ts

all tests

[splincode]

@markwhitfeld

I don't like the HMR stuff in the store internals

I understand, but the correct work of the HMR

• is to write the current state before destroyed module into some variable (in global window)
• and initialize it when initializing the behaviorSubject in stateFactory

[markwhitfeld]

Ok. That helps me understand your reasoning. I think that I can clean this up so that we don't need to have HMR in the internals. Let me try. I'll commit to this branch

[splincode]

@markwhitfeld Thank you!

[splincode]

@markwhitfeld we need remove defaultState parameter, because it turned out to be useless

[markwhitfeld]

@splincode I agree. We need to add the removal of defaultState to the list for v4.

[markwhitfeld]

I'm happy with this now.
@arturovt could you review?
@splincode could you test to make sure that the issue is fixed?

[markwhitfeld]

One step forward two steps back! The nuances of hmr are killing me.
Hopefully all sorted out now.

[splincode]

@markwhitfeld LGTM

[markwhitfeld]

I'm good with this.
@arturovt could you review and approve for merging in.

[arturovt]

LGTM

[splincode]

@markwhitfeld @arturovt I will see later does it correct work

[markwhitfeld]

@splincode great, then you are welcome to merge if it does.

[splincode]

@markwhitfeld everything works perfect