No pull request so I'm commenting here...
Since you're on it, may you please also correct the outdated comments in line 238-240 in decamMapper.py? Looks like that was the original reason to want the first HDU back then and got outdated with the makeRawVisitInfo work.
Also, a similar pattern of getting the "md0" exist in other places of the file. Is there more that could use a cleanup?
And, I'm not sure how good the DECam processing test coverage is on Jenkins today, so maybe somebody would want to run some manual tests or at least include lsst_ci