Update advice for running fossilize-reply
The old advice of using fossils/**/*.foz
doesn't work if there are multiple levels of subdirectory. The old advice also doesn't work (in the distant future) when the number of fossils is very large.
A simpler solution to just the first problem is to use find thusly:
fossilize-replay --enable-pipeline-stats output.csv \
--num-threads 4 $(find fossils -type f -a -iname '*.foz')