RE: Please publish the attached draft-galb-filexfer-extensions-00.txt

["Richard Whalen" <Whalenr%process.com@localhost>]

I think that "copy-data" should either be forbidden for text files, or as follows:

Both files must be opened for text transfer.  The "write-to-offset" is ignored, and the transfer starts at the server's current write position. The "read-from-offset" and "data-length" parameters are ignored.

This avoids problems with differences in the way the file may be stored and problems with the (write?) data length being less than the difference between the write-to-offset and the current allocation o the file being written to.


-----Original Message-----
From: Joseph Galbraith [mailto:galb-list%vandyke.com@localhost]
Sent: Tuesday, January 17, 2006 11:57 AM
To: Richard Whalen
Cc: ietf-ssh%netbsd.org@localhost
Subject: Re: Please publish the attached
draft-galb-filexfer-extensions-00.txt


Richard Whalen wrote:
> Are there any restrictions on "copy-data"?  How are the offsets treated when the file is opened for text transfers?

Excellent questions... I feel a new draft coming on.

I believe the answer is that "copy-data" has no restrictions,
other than that the the user not exceed quota (of course.)

Of course if it is a huge operation, the server is permitted
to complete it out of order or out right refuse to do it.

Clarified in local sources.

I believe, since this is a local-to-local copy operation,
we should forbid the files to be opened for text transfer--
of course, with VMS having multiple different file local
file formats, perhaps you'd prefer something different?
What do you think it should do?

> In section 7 (Temporary files & directories)
> You use "prefered" and "preferreed" that I believe should be "preferred"

Ahh.. I see that my creative speller was at full throttle.
Fixed in local source.

Thanks,

Joseph