330. [func] New function isc_log_wouldlog().
(note: #330 is out of place in CHANGES, because it's new in rc4, but if
it was in the place that 330 is supposed to go, no one would see it. So,
leave it at the top)
399. [bug] When reloading the server with a config file
containing a syntax error, it could catch an
assertion failure trying to perform zone
maintenance on, or sending notifies from,
tentatively created zones whose views were
never fully configured and lacked an address
database and request manager.
In lwres_getaddrinfo() when checking for numeric addresses, the ntmp char array
was only initialized and used when LWRES_HAVE_SIN6_SCOPE_ID was defined,
but was declared and used outside a protective #ifdef. I moved the
declaration and one use of it into #ifdef protection.
Sending a TSIG signed query with a TSIG algorithm that's not
HMAC-MD5.SIG-ALG.REG.INT shouldn't crash the server. Also, the description
of dns_tsigkey_create was wrong in at least one place, and the function
incorrectly returned an undocumented result.
Having a non-noerror value inn the tsig error field should always return
DNS_R_TSIGERRORSET in dns_tsig_verify. Change the result text to reflect
this.
dns_dnssec_verifymessage should make sure that it's verifying with the right
key (which should be the case anyway, but it never hurts to check). Also
fix a memory leak and change a few _tostruct() calls to not allocate memory.
Removed the code added in revision 1.57:
"Store CNAMEs and their SIGs (if any) in their own name structure on
the name list. I.e. do not mix CNAMEs with non-CNAME data. Do the
same thing for DNAMEs."
because it caused DNSSEC validation of responses to ANY queries at
nodes with CNAME records to fail, and no one was able to figure out
why it was added in the first place. As a beneficial side effect,
findname() now finds names even if they have attribute bits set
differently from the one being searched for. This supersedes
the fix of revision 1.133.
If a message is built for one host, which isn't powered on, the tsig object
wasn't removed until the xfer object was deleted, so when the xfer tried the
second master, it asserted.
Brian Wellington