This release features a performance boost from common logic that got accidentally taken out a couple of releases back.
- Basically only compute a checksum for two different files if their sizes vary which is the normal case and only if the user hasn't explicitly asked to ignore checksum verification -- credit to @tkjacobsen with issue #49
ignore-checksum: The motivations are the are cases for which computing the checksum might be straineous on a system eg with a case credited to @jonno84 with issue: #47 as well as their great suggestion.
With drive running on a system mounted with SSHFS, computing the checksum of a 10GB file before a pull/push incurs the cost of data transfer from the FS host to the local machine, then the cost of reading/buffering that file, computing the checksum and in the best but wasteful case the checksums match, in the worst case the checksums differ and you have to re-upload/download the respective files.
This should produce performance gains.