Those files as we all know have the naming format of: "username.VeryLongCapitalizedAlphaNumericString.xml.bz2"
and the ones I am talking about had exactly the same filesize with each other, the same user name but different the second part of their names (the long unique string). I assume that this is the hub ID of the user.
Please correct me if I'm wrong, but I would suggest that a future version of DC++ checks (somehow) if this particular user's filelist has already been downloaded (under its hubID different name) before it actually gets it again (i.e. there are filelists of 0.5-1.0 MB or even larger that need a lot of time to be downloaded, specially under the slower 512kbps connections). Unless there is a reason for keeping that feature as it is (which I can't think of right now...) Many thanks