openssh bugs - Sep 2012 - [Bug 2021] sftp resume support (using size and offset)

https://bugzilla.mindrot.org/show_bug.cgi?id=2021

--- Comment #3 from Loganaden Velvindron <loganaden at gmail.com> ---
ping :-) ?

-- 
You are receiving this mail because:
You are watching the assignee of the bug.

https://bugzilla.mindrot.org/show_bug.cgi?id=2021

Damien Miller <djm at mindrot.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |djm at mindrot.org

--- Comment #4 from Damien Miller <djm at mindrot.org> ---
We've considered doing this in the past but was worried about the
wording in section 6.1 of
http://www.openssh.com/txt/draft-ietf-secsh-filexfer-02.txt allowing
the server to reorder the responses from the block transfers in such a
way that the client got a wrong idea of the high water mark.

Upon re-reading the spec, the wording is still unclear. The first
paragraph seems to unequivocally ban such a reordering, but then the
second paragraph seems to soften that.

An implementor might read both and conclude "it's okay for me to _send_
the responses in any order, so long as the blocks hit the disk in the
order they were requested".

IMO this would be stretching hard against the spirit of the spec and I
don't know of any implementation that would do this. Perhaps we should
just treat them as broken...

-- 
You are receiving this mail because:
You are watching the assignee of the bug.
You are watching someone on the CC list of the bug.

https://bugzilla.mindrot.org/show_bug.cgi?id=2021

--- Comment #5 from Loganaden Velvindron <loganaden at gmail.com> ---
Very interesting.

Maybe the rfc needs to be updated to clear this point, and prevent
implementors from getting the wrong idea.

In any case, If someone has some doubts, he can still check openssh as
it's the sane implementation.

Maybe this could be added to the man page for the "REGET" section to
document the pitfall.

I find SFTP resume very useful, particularly for big files.

-- 
You are receiving this mail because:
You are watching the assignee of the bug.
You are watching someone on the CC list of the bug.

https://bugzilla.mindrot.org/show_bug.cgi?id=2021

--- Comment #6 from Loganaden Velvindron <loganaden at gmail.com> ---
Very interesting.

Maybe the rfc needs to be updated to clear this point, and prevent
implementors from getting the wrong idea.

In any case, If someone has some doubts, he can still check openssh as
it's the sane implementation.

Maybe this could be added to the man page for the "REGET" section to
document the pitfall.

I find SFTP resume very useful, particularly for big files.

-- 
You are receiving this mail because:
You are watching the assignee of the bug.
You are watching someone on the CC list of the bug.