David Phillips
David Phillips
Thanks for the link to #4336. Changing the specification to ignore nulls, which matches the current behavior of the primary implementation, seems like a better option.
I don't think a test is required, as this is extra code which shouldn't exist. It seems strange to have a test for "make sure code doesn't randomly create some...
This should be fixed now.
I tried that and got the same error.
This behavior needs to be specified in https://trino.io/docs/current/develop/client-protocol.html There is a standard `Retry-After` header.
This looks good, but I like to have the Presto changes ready first instead so that we know if anything is missing. I made most of the changes here in...
@nezihyigitbasi `MapredParquetInputFormat` is a simple class. Might be easiest to fork it? Or bypass it entirely? Given we have a native Parquet reader, it might not be needed anymore.
This is great! I'm happy we can deprecate our fork of the RPM plugin
Turns out I was using it incorrectly, but it still doesn't work when used correctly: ``` [17:33 ubuntu@i-64580004:~ prod] galaxy agent terminate 6217 UnexpectedResponseException{request=Request{uri=http://127.0.0.1:64000/v1/admin/agent, method='DELETE', headers={}, bodyGenerator=com.proofpoint.galaxy.cli.HttpCommander$TextBodyGenerator@478e4327}, statusCode=405, statusMessage='Method Not...
Looks like the client sends the agentId in the DELETE body, but the coordinator expects it in the URI.