Would you be willing to review this small change, Paul Price?

It came about because I wanted to make an image of an entire focal plane to visually inspect the results of processCcd. What I would really like, however, is something similar to visualizeVisit that propagates the calexps' WCS (do you have any code for that, by chance...?)

At any rate, this ticket simply adds a template to the obs_decam mapper for the calexp_camera dataset type which is created by visualizeVisit.

Looks fine.

There's nothing that will propagate the calexp WCS — the WCS for the various CCDs will be slightly discontinuous. No doubt you can smooth over it, but I don't believe anyone has taken the trouble. ds9 -mosaicwcs /path/to/*.fits may help, though.

P.S. The alternative to smoothing over the WCS is to resample the pixels via warping. But our warping code doesn't handle large binning factors well.

OK, thanks, I thought that would be your likely response. For some reason, ds9's mosaic seems to show very wrong astrometry yet the matcher is performing fine (and indeed, visualizeVisit shows sources consistently in the same place on the camera, visit after visit, as expected).