[camera_platform_interface] Add rgba8888 format to ImageFormatGroup

[Mairramer]

Part of #11632

Pre-Review Checklist

• I read the Contributor Guide and followed the process outlined there for submitting PRs.
• I read the AI contribution guidelines and understand my responsibilities, or I am not using AI tools.
• I read the Tree Hygiene page, which explains my responsibilities.
• I read and followed the relevant style guides and ran the auto-formatter.
• I signed the CLA.
• The title of the PR starts with the name of the package surrounded by square brackets, e.g. [shared_preferences]
• I linked to at least one issue that this PR fixes in the description above.
• I followed the version and CHANGELOG instructions, using semantic versioning and the repository CHANGELOG style, or I have commented below to indicate which documented exception this PR falls under¹.
• I updated/added any relevant documentation (doc comments with ///).
• I added new tests to check the change I am making, or I have commented below to indicate which test exemption this PR falls under¹.
• All existing and new tests are passing.

If you need help, consider asking for advice on the #hackers-new channel on Discord.

Note: The Flutter team is currently trialing the use of Gemini Code Assist for GitHub. Comments from the gemini-code-assist bot should not be taken as authoritative feedback from the Flutter team. If you find its comments useful you can update your code accordingly, but if you are unsure or disagree with the feedback, please feel free to wait for a Flutter team member's review for guidance on which automated comments should be addressed.

Footnotes

1. Regular contributors who have demonstrated familiarity with the repository guidelines only need to comment if the PR is not auto-exempted by repo tooling. ↩ ↩²

[gemini-code-assist]

Code Review

This pull request adds the rgba8888 format to the ImageFormatGroup enum and implements platform-specific mappings for Android and iOS. Review feedback recommends adding unit tests to verify the conversion of platform-specific integer values and updating documentation links for consistency with existing code.

[bparrishMines]

LGTM

@stuartmorgan-g or @tarrinneal for secondary review

[stuartmorgan-g]

The addition looks good, but the two switch analyzer errors will need temporary // ignores, which the follow-up PR should clean up.

[Mairramer]

The addition looks good, but the two switch analyzer errors will need temporary // ignores, which the follow-up PR should clean up.

Done.

[stuartmorgan-g]

You added case handling, not ignores. That will fail our safety check against modifying code across dependency layers in a federated plugin.

You need to ignore the warning in this PR, and add the cases in the next PR.

[Mairramer]

You added case handling, not ignores. That will fail our safety check against modifying code across dependency layers in a federated plugin.

You need to ignore the warning in this PR, and add the cases in the next PR.

Yes, I messed up. I'll fix it right away.

[Mairramer]

@stuartmorgan-g I believe it should pass now, hopefully.

[stuartmorgan-g]

I'm very confused by the compile error here; I thought this was only an analyzer warning as long as it wasn't a switch expression (e.g., return switch ...). Did the behavior change?

[stuartmorgan-g]

Wow, apparently it changed several years ago for enums and we hadn't hit it until now.

So there will need to be a prequel PR to change all of those constructions to use default instead of the current return-outside-switch approach. Which unfortunately means that we won't get warnings about places we need to update any more, and it'll just have to be feature-test driven.

[Mairramer]

Wow, apparently it changed several years ago for enums and we hadn't hit it until now.

So there will need to be a prequel PR to change all of those constructions to use default instead of the current return-outside-switch approach. Which unfortunately means that we won't get warnings about places we need to update any more, and it'll just have to be feature-test driven.

How to proceed in this case?

[stuartmorgan-g]

Specifically, all the code in those util files that does:

switch (foo) {
  ...
}
// The enum comes from a different package, which could get a new value at
// any time, so provide a fallback that ensures this won't break when used
// with a version that contains new values. This is deliberately outside
// the switch rather than a `default` so that the linter will flag the
// switch as needing an update.
// ignore: dead_code
return bar;

Will need to become:

return switch(foo) {
 ...
  // The enum comes from a different package, which could get a new value at
  // any time, so provide a fallback that ensures this won't break when used
  // with a version that contains new values. 
  // ignore: no_default_cases, unreachable_switch_default
  default:
    bar;
}

And that will need to land in a separate PR before this one can land.

[stuartmorgan-g]

Oh, but that's not enough, because it would still break the existing versions. So despite our intentions, adding an enum value is breaking for our existing implementation code.

I think we'll need to do a one-time breaking change to the platform interface, as much as I hate to do that. We can add a private enum value to each of the enums that we want to be able to extend in that breaking change, to force clients not to rely on exhaustive matching. @bparrishMines Can you think of any better options?

(Alternately we could ship the implementation change, wait a while, assume that it's very unlikely that anyone would be on an old implementation package version but get a new platform interface version, and then change it without a major version bump, but I don't think we should do that since it violates the intent of using semver.)