Hi Daniel, Changes looks good to me, as you said code is copied from one test to another, i found FtpGetContent.java where same FtpServer code is copied. Are you planning to fix FtpGetContent.java as well ?.
Thanks, Vyom On Wed, May 15, 2019 at 2:18 PM Daniel Fuchs <daniel.fu...@oracle.com> wrote: > Hi Arthue, > > On 14/05/2019 23:57, Arthur Eubanks wrote: > > In test/jdk/sun/net/ftp/FtpURL.java, extendedEnabled is always true. > > Same with portEnabled and pasvEnabled. > > Yes - I left them there as they were preexisting. > The bits that handles EPSV is a copy paste from another > test and that minimized the changes. > > > Is the assumption that all > > current servers support these modes? If so, why is there a bool to say > > those modes are not enabled? > > I have no idea. EPSV and EPRT have been introduced do > support IPv6, and are replacements for PASV and PORT > which only support IPv4. I would guess that servers > should all support these modes by now. > > As to your question - my guess is that this test server > got copy & pasted in many tests and some of them might > want to disable some of these modes for testing purposes. > > WRT EPSV/PASV the jdk client tries EPSV ALL first, and > if the server replies with `500 'command not understood'` > it will fallback to PASV. > > Because this particular server in the test didn't support > EPSV before - it was using PASV for IPv4 addresses. > I made sure this would continue to happen - so that > what's happening on the client side doesn't change > when testing with IPv4. > > PASV and PORT will not work on a machine that only has > IPv6. > > best regards, > > -- daniel > > > -- Thanks, Vyom