Need a faster directory walker in data-integration

[gpotts]

The default plugin for kettle get filenames is too slow and seems to scan the entire system before returning the list of files. We need to start as soon as possible and also be able to select what fields to return. No need to calculate file size, perms, ... etc if all we need is the filename.

[gpotts]

Initial plugin is implemented and added to the baseline