I am searching for documentation about ruper. I found this http://krysalis.org/cgi-bin/krywiki.pl?Ruper
I see that the ruper task depends on common-vfs. If we use Ruper for
antlib,
which can be very good, then we are making ant core dependent upon common-vfs.
I have been working with Ruper recently (doing some reworks on it) and the commons-vfs dependency is not my favourite thing. Ruper uses VFS (IMHO) because of the many "file system/transports" (from WebDAV to others) --- but then proceed to use HTTP 99.9999% of the time. Seems a bit like a sledgehammer for a walnut...
I am game to attempt to make Ruper switchable, i.e. it'll work over commons VFS if provided, or simply the HTTP stack inside a JDK.
of course, the http stack built in to java sucks, so we need to serve stuff in the subset of valid http that it supports.
