add hashing to dupe check

[azt3k]

No description provided.

[hoho]

1. Files in the Gulp stream do not necessarily exist on the disk. And this hasher reads the disk for every file with no ability to opt-out.
2. The tests fail.
3. I actually do not understand the purpose of this change at all.

[azt3k]

• Files in the Gulp stream do not necessarily exist on the disk. And this hasher reads the disk for every file with no ability to opt-out
  • Fair call - it reads from the buffer now
• The tests fail.
  • Fixed
• I actually do not understand the purpose of this change at all.
  • The current implementation doesn't dedupe on contents
  • We needed a deduper to use gulp-cache properly otherwise duplicate images with diff filenames are seen as the same file due to the hash based index and don't end up in the desired output location - this patch solves our issues.