[lmi] One PDF line prints on top of another

[Greg Chicares]

There ought to be a single word for the Subject concept, but I can't
find it. It's as though different sections were printed one at a
time on a physical piece of paper, and the sections had conflicting
ideas about which part of the paper they own, so that both printed
to the same contested territory.

To reproduce: Run the same input file we were just using to test
the "vector<>::at() throws" problem (the string that gcc prints in
that case is a good search term, and to preserve its searchability
I don't want to repeat it here). Examine the PDF. On an "Explanatory
Notes" page, I see a definition of "Policy Year" below the bottom of
the footer, so low that there isn't room for another line to fit
below it on the same page.

Reproducing the anomaly might take a little more work. The steps in
my original post on at() throwing used a customized version of the
'sample' product, but I changed that to a proprietary product before
saving the file locally here, to insulate it against the routine
loss of those 'sample' customizations. Running it with 'sample'
instead, everything fits on the "Explanatory Note" page, and the
anomaly is not seen. But I can reproduce it if, before loading
lmi, I edit 'sample.policy' and insert the first three paragraphs of
the first chapter of this tome:
  https://www.gutenberg.org/files/98/98-h/98-h.htm
in both <IrrCsvFootnote>...

  <IrrCsvFootnote>
    <datum>
It was the best of times,
...
of the Cock-lane brood.
    </datum>

...and <IrrDbFootnote>. Then I see
  after rapping out its messages, as the
on its own line, followed by
  spirits of this very
overwritten on the same line as
  This illustration is not

I was looking for an exotic alphabet to portray the typographical
weirdness of this, but the following example that uses Telugu
script may cause your computer to cras