Support for Result<MyType, MyError> with exception throwing on Dart

[lattice0]

Close #533

Adds support for Rust's std::Result, where you can return either a value or an error.

Example

pub enum CustomError {
    Error0(String),
    Error1(u32),
}

pub fn return_err_custom_error() -> Result<u32, CustomError> {
    Err(CustomError::Error1(3))
}

Becomes something that can be used like this:

try {
      final r = await api.returnErrCustomError();
      print("received $r");
    } catch (e) {
      print('dart catch e: $e');
      expect(e, isA<CustomError>());
    }
}

Checklist

• New tests are added to ensure new features are working (probably by modifying the frb_example/pure_dart example).
• All existing and new tests are passing.
• If this PR adds/changes features, documentations (in book folder) are modified as well.

[lattice0]

PS: Should fail until we get anyhow support for IntoDart as done here shekohex/allo-isolate#23

[fzyzcjy]

Great job!

Since CI is not functioning before the upstream release, I just comment partially currently.

I am also wondering: Can we make this breaking change as little as possible? I want to cheat a little bit and do not bump the major version (i.e. release 2.0.0), since this feature though great is not as huge as a major version bump.

[fzyzcjy]

Well I mean:

class FrbException implements Exception {} // no fields

abstract class FrbBacktracedException implements FrbException {
  Backtrace get backtrace;
}

class FrbAnyhowException extends FrbBacktraceException {
  final String description; // or something else etc
  final String backtrace;

  FrbAnyhowException(this.description, this.backtrace);
}

@freezed
class CustomNestedError2 with _$CustomNestedError2 {
  // NOTE for data *without* field "backtrace", just implement FrbException
  @Implements<FrbException>()
  const factory CustomNestedError2.customNested2(
    String field0,
  ) = CustomNested2;

  // NOTE **only** implements FrbBacktracedException **if** there is **really** a "backtrace" field
  @Implements<FrbBacktracedException>()
  const factory CustomNestedError2.customNested2Number(
    int field0,
    String backtrace,
  ) = CustomNested2Number;
}

[lattice0]

Why limit to errors with backtrace if we can always put a backtrace? Most people won't put they own backtraces so we should just put it something that is always received

[franklinblanco]

@fzyzcjy any hints on what would it take to finish this, as this is a pretty useful feature!

[fzyzcjy]

@franklinblanco Just fix the CI, and I am going to merge it :)

[franklinblanco]

@franklinblanco Just fix the CI, and I am going to merge it :)

So it would just be the timeout issue and upgrading the project, right?

[fzyzcjy]

@franklinblanco Yes, basically speaking just that :)

Would be great to resolve all unresolved conversations, but that can also be a future work.

[lattice0]

The bug on valgrind is going to take some more days 😔

…

-------- Mensagem Original -------- Em 23 de ago. de 2022 02:54, fzyzcjy escreveu:

Looking forward to merging this PR - It is almost done! — Reply to this email directly, [view it on GitHub](#582 (comment)), or [unsubscribe](https://github.com/notifications/unsubscribe-auth/ABSTHANRE3CFBD5FYYV2CWTV2RRQRANCNFSM535PZQDQ). You are receiving this because you authored the thread.Message ID: ***@***.***>

[fzyzcjy]

@lattice0 The bug on valgrind is going to take some more days 😔

Take your time!

[lattice0]

Something is wrong with the mail system, these messages were supposed to be from months ago. I currently have no time to fix the problems :c

[fzyzcjy]

@lattice0 Something is wrong with the mail system, these messages were supposed to be from months ago. I currently have no time to fix the problems :c

No worries, take your time, and looking forward to your future contribution :)

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

[lattice0]

I wish someone finished this last annoying part, I have no time :(

[fzyzcjy]

@lattice0 Sad to hear that :( Also looking forward to seeing someone implementing it, since this can be quite helpful

[fzyzcjy]

I have added a paragraph b71004c#diff-a10cdc248927ef9888ef0fbb35fdc60697eb7528199fd78c38ef5d1294852a5aR5 pointing to this PR for who is interested :)

3. #582 adds support for error hierarchy, or arbitrary error types. For example, you can create your own CustomError (such as using thiserror), and it will automatically be converted to a new Dart class. The PR is already almost completed (with the hard work done), and the missing piece is fixing the test errors. I have provided some guidance about how to fix it at the PR comments as well.

[lattice0]

Run dart test, without valgrind, once. Does it work? Or does it also timeout? If it timeouts, then we are pretty sure it is our bug.

it does not timeout, only valgrind timeouts. Locally the error runs fine.

I can't continue my app without this feature but I have no time to do complicated fixes, trying to look into the easy paths for solving this

[fzyzcjy]

it does not timeout, only valgrind timeouts. Locally the error runs fine.

Hmm ok. Then we are still unsure whether it is valgrind bug or valgrind helps us to realize a bug in our code (e.g. undefined behavior).

By the way, given the main codebase has changed so much, maybe by merging upstream code something can change.

I can't continue my app without this feature but I have no time to do complicated fixes, trying to look into the easy paths for solving this

:/

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

[lattice0]

noooooo Enviado via Proton Mail para dispositivos móveis

…

-------- Mensagem Original -------- Em 25 de mar. de 2023 01:12, stale[bot] escreveu:

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions. — Reply to this email directly, [view it on GitHub](#582 (comment)), or [unsubscribe](https://github.com/notifications/unsubscribe-auth/ABSTHANXFIMB5RAIAYSAW5TW5ZWBTANCNFSM535PZQDQ). You are receiving this because you were mentioned.Message ID: ***@***.***>

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.