Question

I have a problem transferring ascii files from a Windows 7 machine to a Unix Solaris server using FileZilla 3. The problem is that the text files end up on the Unix machine with CR+LF characters, instead of just the LF character, resulting in weird characters at the end of each line.

I understand the problem – the CR character should be stripped out. My questions is rather specific: Is the client or server supposed to strip this character?

This fileZilla page says the deal is that ascii files are always transferred with CR+LF, and it is up to the client to make sure that happens. It is then up to the server to convert it to the native format of the host machine, in this case Unix, so it should strip the CR characters.

However, I did not have this problem using WS_FTP, and I am curious if the FileZilla page is simply incorrect, or if WS_FTP was violating the RFC agreement for FTP and actually transferring the file without the CR character. It’s also possible FileZilla is correct, and the problem is with the FTP server on the Unix machine.

Once again: Is it up to the FTP client or server to convert ascii files to the correct line-break format?

Answer 1

From the RFC:

End-of-Line

     The end-of-line sequence defines the separation of printing
     lines.  The sequence is Carriage Return, followed by Line Feed.

The protocol does not specify what should be stored at the end of an FTP session, on either the client or the server, only what the session should look like.