Should we merge glib-openssl and glib-schannel?

[nacho]

This is pretty interesting. At some point I thought of creating this myself as well, but the openssl one was working well enough. Should we create a glib-tls including both backends?

[sdroege]

For being able to share the base classes? That would certainly be useful. Ideally also together with the GNUTLS backend, which IIRC uses the same code yet another time.

Also the tests here are the same as in the openssl backend, except for some changes due to SChannel working differently