David Shaw wrote:
On Sun, Jan 16, 2005 at 08:03:54AM +0000, Ben Laurie wrote:
David Shaw wrote:
On Sat, Jan 15, 2005 at 04:38:32PM -0500, David Shaw wrote:
On Sat, Jan 15, 2005 at 08:54:51PM +0000, Ben Laurie wrote:
3.2 (MPI) doesn't specify what the unused bits should be set to. This
may be deliberate but I think it should either say they MUST be zero
(which I prefer) or that their content is unspecified.
I'm not completely sure what you are referring to here. Do you mean
the difference (given an MPI of value "1") between [00 01 01] and [00
01 02] ? The 0x02 bit of the 3rd octet?
Er, I meant [00 01 03], but the question still holds. Are you
referring to the 0x02 bit of the last octet?
Yes.
I see. I think that the draft does indirectly specify that the unused
bits are 0. In section 3.2 it states "These octets form a big-endian
number; a big-endian number can be made into an MPI by prefixing it
with the appropriate length." [00 01 03] would violate that
statement, since the big-endian number would be 3, rather than 1.
I don't think that would be an unambiguous requirement, since the
implementation might also measure length in bits.
I certainly don't have any objection to making it more explicit than
that, though.
Good.
--
http://www.apache-ssl.org/ben.html http://www.thebunker.net/
"There is no limit to what a man can do or how far he can go if he
doesn't mind who gets the credit." - Robert Woodruff