Well, the offending plot printed today in something like seven hours.
So, I assume it was not a timeout problem.
That pretty much leaves the warped paper, but also the fact that I switched from plotting from Inkscape to using the command line interface.
I’m reasonably sure that the pen up and pen down positions, as well as a few other relevant settings, are not the same when I send the plot from the CLI as they are when sending from Inkscape.
That, combined with the paper issues, left the pen off the paper for part of the plot.