On Sun, 20 Feb 2000, Jordan K. Hubbard wrote:
> > See Jim Bloom's patch of earlier this evening.
>
> Sorry, I'm the release engineer - I only "see" something as fixed when
> it's actually committed to the tree and in my current build. :)
Well, I'm not allowed to commit to that file :) I also haven't sent in the
patch yet because I'm waiting on the outcome of this discussion about
whether we're even going to keep it.
> Yes, it's clear that this is more evil than I thought. :-)
>
> This is just wrong. If I go to build openssh then I expect it to DTRT
> with openssl whether or not openssl depends on RSA, I don't expect to
> go have to install a package manually and then continue with my build.
> That's not how openssh or other RSA-using ports *used* to work and I
> and many others were quite pleased with the previous behavior, we
> would prefer to keep it. I also have to wonder about your assertion
> that we can't just have librsaref.so as either a set of stubs or "the
> real deal" - can you tell me more exactly why couldn't you do this
> with a little help from our friendly weak symbol support?
The problem is that as things are now, when you compile libcrypto with
RSAref support, it has two symbol dependencies on librsaglue, which in
turn has a whole lot of dependencies on librsaref.
I don't know how weak symbols work - someone with more experience in such
matters would have to help me understand it :-)
Kris
> - Jordan
----
"How many roads must a man walk down, before you call him a man?"
"Eight!"
"That was a rhetorical question!"
"Oh..then, seven!" -- Homer Simpson
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-current" in the body of the message
