Well, I have good news and I have bad news:
The good news is that I believe that I’ve isolated the commit that introduced this issue. The bad news is that it seems to have been a pretty big one.
Specifically, in the “main” branch it appears to have been introduced in the second commit of November 27 of last year (2017). I think that the relevant link is this.
I found this by using the "git reset --hard " command to hop from one commit to another, searching for the last one that worked and the first that didn’t. I chose the “main” branch, as I recall, because I remembered having previously used one of the builds on the SDK download page (which I presume draw from the “main” branch) without encountering this issue. (Indeed, that memory was an inspiration for this experiment, I believe.)