I am only keen to carry on with the way that I am going as it is not a trivial change to implement the suggested classes, and I only see this problem happening on a Unix client and Windows server scenario.
Cheers for the input. I will consider the suggested way of implementation. On 25 November 2010 13:16, John Hartnup <[email protected]> wrote: > On 25 November 2010 13:04, Aidan Diffey <[email protected]> wrote: > >> I have been doing that with the documentation provided by the Apache >> FTP server. (https://mina.apache.org/ftpserver/ftplet.html). I am >> sending FTP reply codes at every stage. I have looked at the RETR >> command and looks like I am sending the same codes >> >> > I hope I'm not offending anyone when I say that the documentation for those > extending the server is a little light. It's easy to get the impression that > ftplet is the only way to modify the server's behaviour. > > There is a page (https://mina.apache.org/ftpserver/user-manager.html) that > says "You can write your own user manager to integrate it with your existing > applications. Your custom user manager should implement * > org.apache.ftpserver.ftplet.UserManager*interface. In your configuration > file, you will have to use the Spring bean element to configure your custom > user manager. This gives you all the power of Spring, for example > integrating with your other beans. You can also provide a custom XML format > by using the Spring XML extension mechanisms." > > That's a bit vague, but in conjunction with the source code, it tells you > enough. > > There ought to be a similar statement about virtual filesystems. I still > think that if you want to deliver "files" that are really database entities, > writing your own FtpFile and Filesystem classes is the right way to go about > it. But I will shut up now, since you seem very keen to stick with what > you're already doing (even though it's not working). > > -- > "There is no way to peace; peace is the way" >
