You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
In #5193 functionality was added so that if a user sets the json_name field in the .proto file explicitly, it is used to resolve the path & query params when doing transcoding. This makes perfect sense, if it weren't for a bug in protoc: the proto compiler always populates the json_name field since 3.1.0.
This means that to retain our existing behaviour (using the original field which might contain underscores) we have to add json_name = <field_name> to all our proto definitions.
The Protobuf code base also acknowledges it as a bug and has a workaround to deal with it.
I suppose Armeria needs to use the same workaround as the protobuf codebase: if the json_name that is set is the same as the default calculated one, assume it was not set explicitly by the user.
Thanks!
The text was updated successfully, but these errors were encountered:
I suppose Armeria needs to use the same workaround as the protobuf codebase: if the json_name that is set is the same as the default calculated one, assume it was not set explicitly by the user.
It makes sense. We may provide a new option to ignore json_name. Let us handle it in the next version.
Hi,
In #5193 functionality was added so that if a user sets the
json_name
field in the .proto file explicitly, it is used to resolve the path & query params when doing transcoding. This makes perfect sense, if it weren't for a bug in protoc: the proto compiler always populates the json_name field since 3.1.0.This means that to retain our existing behaviour (using the original field which might contain underscores) we have to add
json_name = <field_name>
to all our proto definitions.The Protobuf code base also acknowledges it as a bug and has a workaround to deal with it.
I suppose Armeria needs to use the same workaround as the protobuf codebase: if the
json_name
that is set is the same as the default calculated one, assume it was not set explicitly by the user.Thanks!
The text was updated successfully, but these errors were encountered: