Ah, you're right. I'm getting myself confused with the config outputs, where the overrides occur (and which even have any effect at all!). In my config output file from recent multibandDriver.py, I have these two entries:

# Mapping of reference columns to source columns

config.forcedPhotCoadd.copyColumns={'id': 'objectId', 'parent': 'parentObjectId', 'deblend_nChild': 'deblend_nChild'}

# Mapping of reference columns to source columns

config.forcedPhotCoadd.measurement.copyColumns={'id': 'objectId', 'parent': 'parentObjectId'}

So your update would override the former, right. Would that even have had an effect given that the copyColumns happens in forcedMeasurement.py?

Yeah, my previous change was incorrect. I've now changed that, and I hope it's all working, but I'm waiting for Jenkins to confirm that. I've added a test (in pipe_tasks) to check that.

Lauren MacArthur, I believe I've fixed everything, and I think the new test demonstrates that. And I hope I answered your questions well enough above. I think this is now ready for another look.

This looks good to me now, and it's great that you added a unittest (and got rid of the temporary skip since I last looked!).

Feel free to merge.

I thought I might have been able to fix that temporary-skip mistake before you noticed, but I guess not. Merged to master.