On 11/06/2015 10:33 AM, Olaf Buddenhagen wrote: > This sounds rather complex; and I'm not even sure it really helps with > the licensing issue, if both backends are built...
*Probably* true, but that wasn't why I pointed this out. > I hope you do not expect us to implement such a solution :-) We > definitely don't need more than a compile time option for CoreOS. Well, inside the code you will have two versions, one for GnuTLS and one for OpenSSL. The patch will be cleaner and simpler if it doesn't end up being an #ifdef nightmare, but instead uses a clean API towards SSL backends.
