Try to find valid downloads for files with a version number in their URL

url-version-finder.sh

curl --head --silent --output /dev/null \
  --write-out "%{http_code}: %{url_effective}\n" \
  https://s3.amazonaws.com/bucket/foo/bar-2.[6-82].[0-9]-[1-9].baz.[0-9].tgz \
  | grep -v ^403

# --head # do a HEAD, not a GET; understand the tradeoffs 
# --silent # hush the usual output progress bars etc.
# --output /dev/null # hush the remaining output (header echoes)
# --write-out "%{http_code}: %{url_effective}\n" # output these specific fields on a line, suitable for grep
# .../bucket/foo/bar-2.[6-82].[0-9]-[1-9].baz.[0-9].tgz # URL pattern, usual bash syntax
# | grep -v ^403 # output ONLY lines that aren't 403 (all the "misses" were 403 in this case;
#                # I wanted to catch both 200s and possible 30x redirects)