Todd Vierling writes:
> However, it is a bug in that our ftp is supposed to `automatically' fall
> back to active mode by default.  However, it is not doing it in the case
> above because the PASV command was successful, yet the actual transfer
> attempt resulted in an error.  This should be fixed.

I don't consider the current behaviour a `bug', per se.

If a site has a firewall that doesn't support `passive', their
ftp server shouldn't respond with a positive response code to
the effect of ``I support passive, and here's the data...''.
Their ftp server should return a ``passive not supported'' code.

Otherwise, how do you determine that you have to fallback to active?
A timeout probably won't suffice, since if there's a timeout chances
are that other factors (other than a borken firewall) may be
responsible, and trying active wouldn't achieve anything anyway.