changefeedccl: Envelope schema name reflects schema config

[HonoreDB]

Fixes a bug where the Avro -envelope schema name, as opposed to subject,
did not honor the schema prefix and full table name options.

Release note (bug fix): Envelope schema in avro registry now honors schema_prefix and full_table_name

[cockroach-teamcity]

This change is 

[HonoreDB]

Ping @stevendanna or @miretskiy, no rush but I think this one got lost. Thanks!

[stevendanna]

It looks like we merged part of this in df7be87, but might be worth getting the test addition in still.

Definitely not a blocker, but part of me wonders if we might find a name other than rawTableName that makes it more clear when it is expected that we use the function. Perhaps that is something we could even encode in the type system a bit so that functions expecting a name that has already been prefixed could take a FooName rather than a string.

Reviewable status: complete! 0 of 0 LGTMs obtained (waiting on @HonoreDB and @miretskiy)

---

pkg/ccl/changefeedccl/encoder.go, line 416 at r1 (raw file):

		opts := avroEnvelopeOpts{afterField: true, beforeField: e.beforeField, updatedField: e.updatedField}
		registered.schema, err = envelopeToAvroSchema(e.rawTableName(row.tableDesc), opts, beforeDataSchema, afterDataSchema)

This isn't a blocker, but I kinda wonder if we could come up with a name other than rawTableName that would make it more clear when we need to call this function.

[HonoreDB]

---

pkg/ccl/changefeedccl/encoder.go, line 416 at r1 (raw file):

Previously, stevendanna (Steven Danna) wrote…

This isn't a blocker, but I kinda wonder if we could come up with a name other than rawTableName that would make it more clear when we need to call this function.

Thanks for the review! I agree, this function name makes even less sense now that it's doing two things. I'll fix in another PR so as to keep the backports aligned.

[HonoreDB]

bors r=[stevendanna]

[craig]

Build succeeded:

• GitHub CI (Cockroach)