Hsin-Fang Chiang do you have time to review this? It is a simple change, though I took the opportunity to also clean up some flake8 warnings (as a separate commit).

This function is not part of afw's public interface (it is in namespace lsst::afw::image::detail)

Since this is imperceptible from python, will pybind11 give us a better way to hide this function?

A comment on the PR: may you modify the comments to reflect what makeWcs does now with an input TPV wcs? Those codes were there to workaround makeWcs.

I'm not sure this part of obs_decam is covered well by the tests and CI. But may you please run validate_drp decam if Jenkins supports it now? (Jenkins didn't last time I heard... not sure how it is now..)

Colin Slater pybind11 will not wrap stripWcsKeywords at all if we can manage it. If we do wrap it, we should put it into lsst.afw.image.detail instead of lsst.afw.image (and that is where it should have been all along, but it's not easy to do that given the way the C++ code is written).

Hsin-Fang Chiang I think the unit tests do a pretty good job of making sure that keywords have been stripped but running validate_drp on DECam is a good idea. I will do that.

I ran the quick DECam demo in validate_drp as far as I could (it died on the validation step) with and without this change. The resulting calexp and icExp were identical in both cases. So I think that proves this change is indeed innocuous (as expected, since calling makeWcs with the second argument true simply calls stripWcsMetadata, thus having the same effect as the current code.

I updated clarified the one comment and merged to master.