Cache list result in P2P network to download recursively

[jim3ma]

Feature request:

Currently, when download recursively, dfget need to list all files in target URL, the result is only in memory.

If too many dfget list all files in same target URL, source has much more pressure.

We need design a mechanism to cache the list result in P2P network.

Use case:

dfget --recursive d7yfs://domain/path/to/dir