jjatria commented on PR #3650:
URL: https://github.com/apache/avro/pull/3650#issuecomment-3915884745

   Hm, this seems a little like an antipattern to me. I sympathise with the 
desire to keep the output deterministic, but I imagine that in this context, 
that loses to efficiency: even though we want it to be predictable, we want it 
more for it to be fast.
   
   Like @martin-g says, the spec does not say anything about key ordering. 
Which means that while it doesn't _disallow_ it, it also does not _guarantee_ 
it. But with this change, the Perl SDK _would_ guarantee it. I worry that 
merging this would make the Perl SDK diverge and from the spec.
   
   If we enforce a given order (with a test, even!) then we are, for better or 
worse, making it part of the documented, user-facing behaviour. We are telling 
our users they can rely on it (which means removing it is now a breaking 
change).
   
   I guess I'd flip this around and ask the following of @tjmac1200: what kind 
of test are you writing that require the order of the keys to be predictable? 
And would making the tests easier to write justify making every other use of 
Avro slower?
   
   


-- 
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