Kriskras99 commented on PR #342:
URL: https://github.com/apache/avro-rs/pull/342#issuecomment-3611129923

   I unfortunately couldn't get a fuzzer to work, the recursive branchy code 
makes it hard for the fuzzer to select new paths.
   
   I took a look at `Compatibility` again, and I think it should stay. Here's a 
comparison of behaviour:
   
   | Writer | Reader | Old | New |
   | ------ | ------ | --- | --- |
   | `Union<Null, Int>` | `Null` | Ok | Partial |
   | `Union<Null, Int>` | `Int` | Ok | Partial |
   | `Null` | `Union<Null, Int>` | Ok | Partial |
   | `Int` | `Union<Null, Int>` | Ok | Partial |
   | `Union<Null, String>` | `Union<Null, Int>` | Err | Partial |
   | `Int` | `TimeMicros` | Err | Full |
   | `Enum<A, B, C>` | `Enum<A, B, C, D>` | Ok | Full |
   | `Enum<A, B, C>` | `Enum<A, B>` | Err | Partial |
   
   So the new behaviour allows slightly more, but is also more honest about the 
compatibility.
   A user who wants absolutely no schema evolution errors, can check for 
`Ok(Compatibility::Full)`, while a user who can allow that can just check 
`is_ok`.
   
   What do you think based on this table?


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to