Ftp-WG: Proposed changes to MLST -15 (will be -16)

Robert Elz <[email protected]> Mon, 22 Jul 2002 05:04:21 -0500
Newsgroups gmane.ietf.ftpext
Message-ID <[email protected]>
    Date:        Sun, 21 Jul 2002 03:33:21 -0500
    From:        Marc Huber <[email protected]>
    Message-ID:  <[email protected]>

  | It should be addressed, but I'm not sure about the best solution.

Nor I am.

  | Non-shortest UTF-8 is supposed to be illegal,

Yes, though I'm not sure of the relevance here.

  | so it might be advisable to adopt parts of the NVT processing
  | for the MLSD data connection: transmit <NUL> after <CR>
  | unless that <CR> starts the regular EOL sequence.

That's certainly an option, but it would be a big change to make at
this stage of the protocol's design.   It would also mean a big change
to implementations, which, at least as far as I know, don't currently
support anything like this, and while implementations supporting new stuff
isn't exactly a major problem (all of MLSx is new after all...) this
one would be kind of a dramatic change.

An option would be to simply say that names that contain the EOL sequence
can't be handled by the protocol at all.

There is already no way to send them in NLST or LIST, and while it would
be nice for MLSx to be able to cope (and because the control connection is
used, MLST can) it isn't as if we'd be giving up all that much by
simply saying that those filenames can't be transmitted via MLSD.

Opinions?

kre

ps: it would be nice to get a resolution on this one real soon now, we're
in a window where there's actually a possibility of getting this doc
finished and out the door within a measurable time frame - I'd hate it
to fall back into the "Oh, that one" state that it has been in for the
past year or so.