use hash for long filenames

[paulmars]

I don't expect you to merge this, there are no tests. But I wanted to propose a solution to a problem.

I was download some URLs that are longer than the filenames my system allows. Which prevents the caching from being saved.

Let me know if I'm not being clear.

[paulmars]

Similar to this

https://github.com/thoughtbot/paperclip/issues/1246#issuecomment-18378975

[the-jackalope]

I've also run into this issue using fetch-vcr!

What might be simpler than modifying the buildHash method is to alter the buildFilenamePrefix to simply truncate overly-long filenames like so:

var maxFilenameLength = 255

function buildFilenamePrefix(url, args, hash, suffixLength) {
  args = args || { }
  url = escape(url).replace(/\//g, '_')
  var method = args.method || 'GET' 
  method = method.toUpperCase()
  var filename =  url + '_' + method + '_' + hash
  var totalLength = filename.length + suffixLength
  return totalLength > maxFilenameLength
    ? filename.slice(0, maxFilenameLength - suffixLength)
    : filename
}

Max filename length is capped by every OS I know at 255. Let me know what you think!

[boxman0617]

https://github.com/philschatz/fetch-vcr/pull/32 I set up this PR to address the issue.