Some notes on git bisect in understanding a regression bug

[njtierney]

After noticing some unexpected errors in test_posterior.R, which we do upon a greta release, doing a git bisect search revealed the following:

• The earliest good commit was: 212f2da66cbb61edb6b8f4da98d2badf1c5c4b4e
• The first "bad" commit was: d37b53d26d65c0c438e6b1857659c7ae837f8f7e

More investigation to happen from here, but just noting this in this issue for reference.