[HOLD] Improve soft keyboard padding fix with native solution

[Julesssss]

Details

To resolve a fire and ensure we could deploy mobile while in Santiago, we merged a fix for a keyboard padding issue that depends on a Platform conditional.

This PR reverts the original changes and resolves the issue with an improved native fix:

The idea behind this PR is that the underlying cause of the original issue was that both RN and the Android OS were trying to handle the showing/hiding of the soft keyboard. This solution prevents Android from doing this in every single case, ensuring the keyboard control is delegated to RN and the KeyboardAvoidingView.

The native Android windowSoftInputMode attribute is explained here. We were using adjustResize, which would always attempt to resize the Activities View and to make space for the keyboard, whereas adjustUnspecified is the default behavior that uses the context of the underlying view. It's important to note that the Android OS will still reduce the Activities View height to make space for the keyboard when RN descides to show/hide it (this is desired, as the alternative is that the entire view moves up to make room for the keyboard). But now, the OS is not forcing the extra padding.

Fixed Issues

$ #10539

Tests

Please test Android and iOS native apps and mWeb with a physical mobile device

I have tested on:

• Physical Nexus 5, Android 8.1
• Physical Pixel 6a, Android 13
• Emulated Pixel 2, Android 8.1
• Simulated iPhone 13, iOS 16.0
• Physical iPhone 7, 13.3

1) Ensure the original issue is fixed

• Create a chat with a user you have not chatted with before
• Tap the input field
• The keyboard should be displayed correctly
• Close the keyboard
• The keyboard should hide correctly

2) Ensure the keyboard works in other inputs throughout the app

• Test the app features (registration/sign in/settings/workspace/new chats/ect) ad ensure the soft keyboard works as expected
• Applause should run the full regression suite instead

3) Verify that no errors appear in the JS console

Phsyical Device screencapture
Click to view -- not sure why it's not embedded
https://user-images.githubusercontent.com/10736861/193265622-fc8eb721-5db1-450b-a4aa-3287e5c62980.mp4

Emulator Screenshots

[Julesssss]

Next step is to compile a build and attach it here for Applause to test.

[Julesssss]

Reminder to self that this change might actually fix this other issue

[tgolen]

What's the status of this PR? Can we get it merged and deployed?

[Julesssss]

This is low priority at the moment. But I will return to this. Next step is to compile release builds and test a few of the attributes manually, but this takes a lot of time.

[tgolen]

@Julesssss this is actually turning out to be pretty high priority for me since I'm looking into simplifying/improving our KeyboardAvoidingView behaviors. Is there anything I can do to help? I've not ever compiled release builds before. Are these the SOs?

https://stackoverflow.com/c/expensify/questions/7311
https://stackoverflow.com/c/expensify/questions/8290

What do you mean by "test a few of the attributes manually"?

[Julesssss]

I found some time this morning to finish up this PR and I tested it on three Android devices:

• Physical Nexus 5, Android 8.1
• Physical Pixel 6a, Android 13
• Simulator Pixel 2, Android 8.1

Is there anything I can do to help? I've not ever compiled release builds before. Are these the SOs?

It would be great if you could test a native iOS and Android build, and yeah, those are the correct stack overflow posts. Please let me know if you have any further questions here!

What do you mean by "test a few of the attributes manually"?

The idea behind this fix is that we tell the OS not to automatically show/hide the soft-keyboard, but this solution requires that we use one of these attributes on the MainActivity and I wasn't 100% sure that I had made the correct choice. However, after testing I'm happy that adjustUnspecified` is correct.

[Julesssss]

CC-ing a few people who have context about the original changes that this native fix reverts: @roryabraham @marcaaron @Luke9389

[tgolen]

That worked! Thank you

[Julesssss]

@tgolen I should have realized this on Friday, but it looks like you tested the release build on a simulator -- in which case it's expected that the keyboard won't display 😅 Given the context of the PR it was a good thing to raise though.

Anyway, after testing I can confirm the keyboard works as expected on a physical device. Added iPhone 7 to the list of devices I have tested.

I think this is ready for further testing, or to be merged now.

[tgolen]

I think that's some pretty good testing. I'll go ahead and merge this

[tgolen]

Actually, holding off to get a C+ review

[Julesssss]

Actually, holding off to get a C+ review

Requested a C+ review here. Awaiting the review/confirmation.

[thesahindia]

@Julesssss, I was able to repro the blank space issue after applying changes.

Steps:-

1. Search new user
2. Keep the keyboard open
3. Tap on the contact

video_20221004_043659_edit.mp4

[Julesssss]

@Julesssss, I was able to repro the blank space issue after applying changes.

Damn it. Thank you for testing 😞

I recall this initially being inconsistently reproducible, but I'll try to see if I can find the cause. I have a feeling it's related to the type of component.

[tgolen]

Is that reproducible on any Android? I haven't had the time yet, but I'd be really curious to see if my changes in #11586 will fix that or not.

[thesahindia]

Is that reproducible on any Android?

Yeah, it was reproducible for iqoo z5. (haven't tested on other android devices)

but I'd be really curious to see if my changes in #11586 will fix that or not.

I will try testing it again with your changes.

[Julesssss]

I'd be really curious to see if my changes in #11586 will fix that or not.

Oh good point. I'll also test this PR against your changes

[tgolen]

What were the results of the testing? Any luck? What's the next step here for this PR?

[thesahindia]

What were the results of the testing? Any luck? What's the next step here for this PR?

Same results. It's still happening
And as mentioned here, we are seeing one new issue at login page.

[tgolen]

Ah sorry, I got the comments mixed up. Bummer that my PR doesn't fix the empty space issue :(

[Julesssss]

Ah sorry, I got the comments mixed up. Bummer that my PR doesn't fix the empty space issue :(

Yeah... I feel like we're not too far off the underlying issue here, but I'm just not able to spend any further time on this at the moment.

[PauloGasparSv]

I could not reproduce this on iOS but it looks similar to this closed issue #11858 so I'm linking it here!

[tgolen]

Last time I spoke with Jules about this, the solution here sounded pretty hopeless. We decided to go full-steam-ahead with this PR instead: #11586

I'm going to close this PR. We can reopen it later if we want to.