I stumbled over the fact that sh's read returned 1 on the final part
(call it a line or not) of a file when it didn't end in a newline.
Digging through SUS revealed (in the informative part of read):
Although the standard input is required to be a text file, and
therefore will always end with a <newline> (unless it is an empty file)
as well as other passages that suggested that a "text file" consists of
"lines" and a "line" is a sequence of non-newline characters plus a newline,
but I couldn't find a definition explicitly stating that.
