fix(go/adbc/drivermgr): adjust ingest helper to set target before BindStream

[arnoldwakim]

Summary

Reorder IngestStream and IngestStreamContext to set OptionKeyIngestTargetTable and OptionKeyIngestMode before calling BindStream, matching drivers that require ingest targets up front (e.g., FlightSQL).

Keep all other ingest option handling and execution flow unchanged.

Evidence (FlightSQL requirement):

go/adbc/driver/flightsql/flightsql_statement.go:630-656:

func (s *statement) BindStream(_ context.Context, stream array.RecordReader) error {
	// For bulk ingest, bind to the statement
	if s.targetTable != "" {
		if s.bound != nil {
			s.bound.Release()
			s.bound = nil
		}
		if s.streamBind != nil {
			s.streamBind.Release()
		}
		s.streamBind = stream
		if s.streamBind != nil {
			s.streamBind.Retain()
		}
		return nil
	}

	if s.prepared == nil {
		return adbc.Error{
			Msg:  "[Flight SQL Statement] must call Prepare or set IngestTargetTable before calling Bind",
			Code: adbc.StatusInvalidState}
	}

	// calls retain
	s.prepared.SetRecordReader(stream)
	return nil
}

The current non-patched code, raises the error:

[Flight SQL Statement] must call Prepare or set IngestTargetTable before calling Bind

[lidavidm]

I suppose this is reasonable but I feel like we should fix the FlightSQL driver as well, @zeroshade

[zeroshade]

I agree, we should fix the FlightSQL driver. But more importantly, can you add a test here?

[arnoldwakim]

I will, add a test, did you have something particular in mind or just asserting that the stream was bound?

Also, are you expecting me to fix the FlightSQL driver or is it something one of you wants to tackle?

Thanks for the swift review!

[zeroshade]

If you are able and willing to fix the flightSQL driver accordingly, go for it! Otherwise one of us will do so as a follow-up to this.

I didn't have anything in particular in mind for a test, just simply a test that would fail on main and succeeds after this fix so that we don't have any regressions in the future.

[arnoldwakim]

If you are able and willing to fix the flightSQL driver accordingly, go for it! Otherwise one of us will do so as a follow-up to this.

I didn't have anything in particular in mind for a test, just simply a test that would fail on main and succeeds after this fix so that we don't have any regressions in the future.

I think 0cd718e addresses both.

[lidavidm]

Can't we combine this clause with the one above?

[arnoldwakim]

Sure let me know if 3c38988 did the job.

[lidavidm]

ditto here

[arnoldwakim]

Oh, the failing CI says it needed an action from me?

[lidavidm]

No, it's just a flake