On Sun, Aug 19, 2001 at 09:40:58PM -0700, Guy Harris wrote:
> On Sun, Aug 19, 2001 at 11:21:28PM -0500, ethereal-dev@xxxxxxxxxxxxxxxxxx wrote:
> > I don't know the specifics but the following is included in the
> > license for fetchmail 5.8.17 (5.9.0 most recent version):
> >
> > Specific permission is granted for this code to be linked to OpenSSL
> > (this is necessary becuse the OpenSSL license is not GPL-compatible).
> >
> > Because ethereal is GPL, I presume it is not legal to redistribute
> > binaries linked against OpenSSL. Any problems adding the above clause
> > to the ethereal license to make it legal?
>
> Do we, in fact, link against any SSL library, as in a library that
> implements SSL?
>
> "AC_ETHEREAL_SSL_CHECK" is a somewhat misnamed macro; in fact, the
> library it deals with is a crypto library, and we link against that
> because some versions of UCD SNMP/NETSNMP apparently require it, not
> because we use it for anything related to SSL.
OpenSSL provides -lcrypto and -lssl. I think you're right about the
requirement for UCD-SNMP and SSL.
The ucd-snmp 4.2.1 configure program has the following option:
--with-openssl=PATH Look for openssl in PATH/lib.
However, UCD-SNMP is under a BSD-like license so I believe it can
cleanly link against OpenSSL. So, if Ethereal links against UCD-SNMP
which links against OpenSSL, what happens?
It seems though that because Ethereal doesn't *use* OpenSSL for
anything, there's not a problem. I should have dug further. Sorry.
> Is any version of that crypto library licensed under a
> non-GPL-compatible license?
As -lcrypto is provided by OpenSSL, I guess this applies to OpenSSL.
> > According to packet-ssl.c, the copyright owner is:
> > /* packet-ssl.c
> > * Routines for ssl dissection
> > * Copyright (c) 2000-2001, Scott Renfro <scott@xxxxxxxxxx>
> >
> > so he would need to give his permission for the license change. I don't
> > know if it is sufficient for only packet-ssl.c to have the license
> > change mentioned above.
>
> As per the above, we don't, as far as I know, use any SSL library, so
> the fact that "packet-ssl.c" happens to be an SSL dissector doesn't, as
> far as I know, matter - Scott's permission is no more or less relevant
> than anybody else's permission. If anybody needs to give permission for
> a license change, *everybody* who's contributed code to Ethereal needs
> to do so, as far as I know.
Thanks for the clarification. You are, of course, correct with the
last sentence.
--
albert chin (china@xxxxxxxxxxxxxxxxxx)