Closed BWhiteApps closed 4 months ago
I see - genq
makes assumptions how the serialization/deserialization functions for your types are named. This is the $SomeFreezedClassFromJson
you see here. Normally genq
will generate these functions for you. When using freezed, private fromJson
and toJson
prefixed with a underscore are generated instead.
I think the least effort would be defining these functions yourself, referencing the function generated by freezed.
@freezed
class SomeFreezedClass with _$SomeFreezedClass {
const factory SomeFreezedClass({
required String value,
}) = _SomeFreezedClass;
factory SomeFreezedClass.fromJson(Map<String, dynamic> json) =>
_$SomeFreezedClassFromJson(json);
}
// Here is the magic:
const $SomeFreezedClassFromJson = _$SomeFreezedClassFromJson;
This way, your genq classes will happily find the missing function, but effectively using the one generated by freezed.
Alternatively, you can specify custom toJson
/fromJson
function with the @JsonKey
annotation. But this basically boils down to providing functions which will need to call _$SomeFreezedClassFromJson
under the hood. So might as well just use the first solution 😉
I work in a very big project and would like to eventually switch most of our data models with a few exceptions from freezed to genq.
However I am running into an issue with json serialization where all properties of genq data models must also be genq models. In a very big project this makes migrating difficult because it will be a big lift to migrate every single data model and even then there are models that need to stay in freezed.
Example:
Gives an error that
The function '$SomeFreezedClassFromJson' isn't defined.
Is there a way around this to make the migration easier?
Thanks