[chord] new auth type patch

[Emil Sit]

On Wed, 23 July 2003 at 17:56 (-0400), Michael Walfish wrote:
> Replying to Benjie's mail, cutting out chord at pdos and adding Emil . . .

I don't understand why you want to cut chord at pdos out of the loop.  The
only effect that will have is to remove the record of design discussions
and leave it so that people will continue not to have any idea what is
going on with the source code. 
 
> it doesn't really matter. for several reasons, I'd like to update()
> into the sha1ctx several times, rather than doing a flat hash of the
> thing (currently I go in multiples of 8 bytes), but that's my only
> request. How

Er.  How you produce the salt actually has nothing to do with its
contents right?  So we should just pick a size and go with it.
I guess NFS FHs are 32 bytes so that is a reasonable reason to
go with 32.

> yeah, I thought of that at first but wasn't sure what kind of reception it
> would receive and figured it was easiest to ask about a different block
> type to minimize collateral changes. but if you're in favor of it, I
> support this.

fwiw, I am in favor of code simplicity and readability.  

also, regarding how to proceed, let's wait until benjie finishes
the KEYHASH updates (to be "KEYHASH with salt"), and see how things
look from there.  as I said before, i am inclined to take the code
cleanup and noauth patches if they could be separated.  perhaps
once benjie is done, it will be easy for you to produce these
patches.

-- 
Emil Sit / MIT LCS PDOS / http://pdos.lcs.mit.edu/chord/