[OpenAFS-devel] dcerpc.net - freedce

[Sam Hartman]

>>>>> "Luke" == Luke Kenneth Casson Leighton <lkcl@samba-tng.org> writes:

    Luke> On Thu, Aug 16, 2001 at 03:17:45PM -0400, Sam Hartman wrote:
    >> So, it seems that fixing the existing problem in the port of
    >> AFS to NT would be much easier than adding DCE RPC to AFS.

    Luke> [hm, looks like i missed out on some afs-devel messages,
    Luke> will have to look at archives again to see]

    >> How exactly do you hope to save effort adding DCE RPC to AFS?

    Luke> well... ... *thinks*.

    Luke> well, your question is ambiguous.  inasmuch as it's not
    Luke> really _necessary_, it's just that it's been done once
    Luke> already (AFS-4, aka DCE/DFS) and, well, basically i thought
    Luke> it'd be cool to see it done again, is the bottom line.  

I don't think that using a particular RPC is cool or useful unless you
get some benefit from it.

    Luke> and
    Luke> that will take effort.

Right, so the effort must be justified.


    Luke> ... but i'm not sure that's the question you were asking: if
    Luke> you're asking will doing this work now save effort with
    Luke> respect to understanding and possibly merging AFS-4
    Luke> (DCE/DFS) when/if it becomes an Open Source Project, then
    Luke> the answer is definitely yes because the people who actually
    Luke> do the development will by then have a working knowledge of
    Luke> both AFS and dce/rpc, and the differences between AFS-4 from
    Luke> The Open Group and AFS-3 from OpenAFS will stick out much
    Luke> clearer.

But is merging DFS and AFS a goal worth pursuing?  It seems that
grabbing file ACLs (if you can find a way to make them work better
than DFS--I considered its ACL system unacceptable when I used it),
byte range locking and a few other things would be easier than doing a
merge of the code.