Martijn van Oosterhout wrote:
Please read the OpenSSL-GPL FAQ. They themselves acknowledge it's a
problem, but claim they fall under the "operating system exception",
which is fine for everyone except the distributor of the operating
Thanks for the link. Unfortunately that FAQ does not say anything about
the advertising clause or how they want it to be interpreted.
They recommend that if you want to use OpenSSL, use a licence other
than the GPL.
..which could be seen as a sign that they take their advertising clause
serious. I wonder how much of their code is still copyrighted by authors
refusing to remove that clause...
Wikipedia also has more information about this.
I also found this to be a good description:
[ OT: Generally, I feel that the exceptions which are made to the GPL
are very messy and confusing. And again, the exception implicitly states
that the OpesSSL Projects wants you to adhere to the advertising clause. ]
The original authors have been asked and apparently can't be found or
don't care. I strongly suggest you read the openssl archives before
opening this can of worms. Note the authors involved are no longer part
of OpenSSL, they have another SSL library, so they're probably not
inclined to be nice.
Sure, I've heard about that and won't open that can of worms ;-)
Following that 'better-safe-than-sorry' philosophy, one could ask if
PostgreSQL shouldn't better include the acknowledgements of OpenSSL (and
MIT Kerberos) in all of their advertising materials...
AIUI all compiled distributions of postgresql using openssl do actually
include such. For example the Windows Installer.
The OpenSSL license speaks of "all advertising materials mentioning ..
use of this software". IMO, the PostgreSQL website matches that
criterion very well, doesn't it?
AFAICT, that's why so many people avoid advertising clauses like the
plague. (And why it's called 'advertising clause' and not 'compiled
Probably PostgreSQL should ask for an exception... ;-)
We're in the bizarre situation were both Debian and the OpenSSL groups
beleive it is a problem, and postgresql does not. Quite odd.
I somehow don't understand how this could *not* be a problem. My
reasoning is that one must not not respect authors wishes (licenses)
very much if one feels this is not a problem.
---------------------------(end of broadcast)---------------------------
TIP 9: In versions below 8.0, the planner will ignore your desire to
choose an index scan if your joining column's datatypes do not