Finn Arne Gangstad <[EMAIL PROTECTED]> writes:
> Ok, following this procedure from a freshly untarred tree it worked
> without problems:
>
> autoconf
> autoheader
> ./configure
> make
Good. What system are you using? (So I can note that there are no obvious
portability problems there).
> > This is a genuin bug, also reported by Keresztfalvi Gabor Agoston.
> > Change "crypto_rc4_algorithm" to "&crypto_rc4_algorithm" in lsh.c and
> > lshd.c. I really don't understand why the code ever worked for me.
>
> Works now!
Great. I also can't reproduce the other reported bug now. Hmm.
> Is this the right mailing list to discuss lsh?
I think Martin has said that it's ok to use the psst-list for
discussions about lsh.
> What is the preferred way to submit patches - mail the list, mail you, or
> something else?
Mail me. Perhaps you want to mail the list as well, I don't know. I'm
also considering making my cvs-repository available for anonymous
read-only access. But I think I want to integrate any patches
manually. diff -u output is preferred.
> Are there plans to include new ciphers such as DFC, MARS, RC6, Serpent or
> Twofish?
Not immediately. Next cipher to be supported is triple-DES, because
that's the only cipher that is required by the ssh spec, and because
cryptographers agree that it's the most conservative choice.
The lsh snapshot already includes implementations of DES, cast and
idea, what is needed is wrappers to make them fit in lsh (except IDEA
which should perhaps be removed for patent reasons). One thing that
needs consideration is handling of weak keys; the current lsh code
doesn't handle that.
GNU Privacy Guard also contains implementations of several ciphers
which could be borrowed by lsh. When time comes, I think that this
code should be extracted into a stand-alone library which can be used
by both lsh and gpg.
/Niels
