feat(functions): turbo modules implementation

[russellwheatley]

Description

Related issues

Release Summary

Checklist

• I read the Contributor Guide and followed the process outlined there for submitting PRs.
  • Yes
• My change supports the following platforms;
  • Android
  • iOS
  • Other (macOS, web)
• My change includes tests;
  • e2e tests added or updated in packages/\*\*/e2e
  • jest tests added or updated in packages/\*\*/__tests__
• I have updated TypeScript types that are affected by my change.
• This is a breaking change;
  • Yes
  • No

Test Plan

---

Think react-native-firebase is great? Please consider supporting the project with any of the below:

• 👉 Star this repo on GitHub ⭐️
• 👉 Follow React Native Firebase and Invertase on Twitter

[mikehardy]

Hey @russellwheatley - strange one!

@mikehardy - Android e2e tests passing for functions. iOS has hit a snag for this reason: facebook/react-native#49250

Note - from my testing, it is maps with properties that have a null value that are stripped away.

It seems like it is part of v0.81.0-rc.2 release therefore it isn't in stable yet. I did try and bump test app to 0.80.0 but came across build issues for both iOS and android.

This should not be happening. That issue does look like the symptoms you describe, however, the issue you link resulted in a PR that was first released from main in 0.79.0-rc series (if you expand the tag list you'll see it), but more importantly, it was picked into the 0.77 and 0.78 branches, so the fix was actually present in 0.78.0 (since the issue was fixed after branch cut) and then a follow-on fix was required, and it was picked into 0.78.1 - you can see those here

• original 0.78 pick [0.78] Properly handle null values in TM interop layer reactwg/react-native-releases#785
• follow-on fix 0.78.1 pick [0.78.1] Handle null values on iOS in new arch interop without release mode crash reactwg/react-native-releases#850

Our e2e app is on 0.78.2, so we have those fixes. I was able to discard those patches.

So...why is this happening now, again, with TurboModules? Something really subtle must be going on. If we think something is wrong, we'll need to make a reproducer for upstream - they have a template here https://github.com/react-native-community/reproducer-react-native

it is possible to take the reproducer and add native modules to it (of any kind) and run the necessary commands to generate native code or whatever - we'll need this though as it is not possible to efficiently share firebase projects related stuff with react-native upstream with all the config and key handling and such

I've had a further think on not checking in generated code, and I'm not sure I agree with the idea it should be gitignored. Any codegen changes might be worth tracking for any bugs further down the line in my opinion. I haven't fixed the code quality check CI which is failing on generated android code, I will fix it if you agree generated code is worth checking in.

I'll trust you on this one, you've got your sleeves rolled up and will have more taste on it right now definitely. I usually don't get value from seeing diffs in generated code, but you may be right in this case, and either way, it's a reversible decision. There's zero harm in trying so my default stance of "implementor's choice" (read as: 'let the cook cook') wins for sure. Let's see if it has value over time and anyone can re-evaluate if not ?

[russellwheatley]

@mikehardy - I've created an issue with reproducer on react native repo: facebook/react-native#52802

[mikehardy]

Looks like they've got a patch queued for release in 0.81 already. When I'm done traveling (still a week or so) or maybe if I catch some time while traveling I'll try to queue pick requests for older branches otherwise this will tie us to react native 0.81 as a minimum version. We probably need to separate our e2e app so macos (which lags main react native) versions aren't holding us back from testing

[russellwheatley]

@mikehardy - I think we still have a problem with null values, I checked against the latest RN release candidate and it didn't make a difference.

[github-actions]

Hello 👋, this PR has been opened for more than 2 months with no activity on it.

If you think this is a mistake please comment and ping a maintainer to get this merged ASAP! Thanks for contributing!

You have 15 days until this gets closed automatically

[github-actions]

Hello 👋, this PR has been opened for more than 2 months with no activity on it.

If you think this is a mistake please comment and ping a maintainer to get this merged ASAP! Thanks for contributing!

You have 15 days until this gets closed automatically

[mikehardy]

this one's up next - current thinking is to have a serializer/deserializer process for iOS that maintains an array of paths for JSON nulls for serialized objects and send then on the side of the main object, to rehydrate them correctly on the native side.

it's the only way to handle this that doesn't rely on a maybe-they-fix-it/maybe-they-don't process upstream, nor their release cadence even if they do fix it, while avoiding any collision problems with in-place canaries