Would this work also include applying astrometric corrections, or is that separate?

We use jointcal for astrometry. This ticket covers both.

This is ready for review. Jenkins going here: https://ci.lsst.codes/job/stack-os-matrix/36645/display/redirect

Note that this ticket addresses the production of PreSources (preSourceTable_visit), Sources (sourceTable_visit), and fgcm/jointcal using preSources as input in all the HSC DRP pipelines and the DECam pipeline, and rc2_subset pipeline.

It does not address that faro in the future will want to use Sources as input during RC2 and production runs. Jeffrey Carlin and I discussed this pair coding last week. Because that requires a reordering of tasks in the subsets, I'm going to address that on DM-34853.

Thanks for volunteering:

• https://github.com/lsst/ci_hsc_gen3/pull/76
• https://github.com/lsst/pipe_tasks/pull/685
• https://github.com/lsst/drp_pipe/pull/20
• https://github.com/lsst-dm/rc2_subset/pull/16

Another note is that in on piece of naming unfortunateness is that the new `writeSourceTable` has skymap added to its dimensions (because jointcal is per tract and it needs to know the tract geometry to choose the best one), which means that I can't reuse the old name `writeSourceTable`. I renamed it to `writeRecalibratedSourceTable` and kept `transformSourceTable` and `consolidateSourceTable` the same, but am open to renaming those too.

Oh and if you're curious what happened with that Jenkins run, between the macos and centos Jenkins runs starting, Dan merged something to pipe_tasks,

This mostly looks fine to me. I just have a few questions on the drp_pipe PR, which may turn out to be off base.