Noticed while examining the file.

The only reason this wasn't picked up before was because this header
file was never included more than once in a .cc file.

Not strictly part of the ticket, but added because the first tests I
did with the existing code returned NSAS_INVALID_RESPONSE messages. This
was a common error to three conditions: the change separates these
out and logs a different message for each one.  It also changes the
severity of two of those conditions (which are caused by the contents
of the returned packet) from ERROR to DEBUG.

(Of course, once the change was made, the original problem could not
be reproduced...
 Please enter the commit message for your changes. Lines starting

It is already the 29th in some parts of the world.

See ticket #1430 for more details. (Ticket pending tests before closing.)

Put correct git revision hash.

Remove review comment.

Remove no trac number comment.

Rejustify another entry while here.

       or NXRRSET of wildcard expansion.

this should fix the systest problem; it seems to trigger a race condition in initialization. If the config session isn't up yet, there will certainly not be any queued messages yet, so simply skipping that should be ok.

use two tabs before the committer username.

Line too long

use a tab before the keyword type.

based on some feedback from vorner

(spell checks don't find misspellings that create other correct words :)

consistency.

This reverts commit 9150be5e.

Pushed to master without review by accident. Be patient, it'll get there
eventually...

As it needs it in many libraries.