[Swan-commit] Changes to ref refs/heads/main

[Andrew Cagney]

New commits:
commit 15065d0dd5292db7646af53a9c102ad0c4b0ff9c
Author: Andrew Cagney <cagney at gnu.org>
Date:   Sat Jan 9 11:26:24 2021 -0500

    revival: set that.host_encap when connection being revived was encapsulated
    
    This way the code re-orienting the local interface knows that
    encapsulation is needed and the local encapsulated port should be
    used.
    
    - but why is new-state re-orienting the connection?
      perhaps to re-initialize fields?
    
    - but why isn't host_port updated once things establish?
      well it is, sometimes :-(
    
    (the whole approach of trying to stuff everything into a single struct
    end seems fragile).
    
    Follow-up a43c1a86ba13111587b6910bf3dc6ee9a4a25d61
    Re-fix ikev2-revive-through-nat.

commit aa45d0c17236602f487d0e712590658ab82d8955
Author: Andrew Cagney <cagney at gnu.org>
Date:   Fri Jan 8 13:29:22 2021 -0500

    revive: when reviving an instance, don't modify remote.raw.host.ikeport
    
    It should contain the unmodified value from the config file
    (now if only it could be made const).
    
    Fixes ikev2-41-rw-replace, for instance (after re-connect was
    displaying a port when it shouldn't as the code though the connection
    had to be encapsulated).
    
    Breaks ikev2-revive-through-nat (which wasn't in TESTLIST).
    
    Follow-up a43c1a86ba13111587b6910bf3dc6ee9a4a25d61

commit 2918600190e54a7a4bf04b048ffe4c2e7da99d17
Author: Andrew Cagney <cagney at gnu.org>
Date:   Sat Jan 9 11:46:32 2021 -0500

    testing: add ikev2-revive-through-nat* tests