Makes me wonder how this ever worked. Anyways:
Update ipv6calc to 0.50 (in the wrong hope that it was fixed, but I had to
do the work myself, *sigh*).
Changes:
* nothing important as far as I could spot from the large & useless ChangeLog
file
* No longer requite "root" (env-var) to be passed to the pkg
* hack to make autoconf stuff going, dumping on author to fix properly
Sorry for being in rant mode...
Changes since last packages version (1.1.2):
Wed Nov 2 10:03:00 CET 2005
* Fixed typos in sipcalc man page, thanks to A Costa <agcosta@gis.net>
for patch.
* Added -w ipv4 option to display some inverse mask information.
- Removed the dependency on the IO::Socket::INET and IO::Socket::INET6
modules. The IO::Socket module is now used for all socket objects.
- The port information can now be included as part of the transport
address specified with the session() Transport Domain arguments.
- Added support for specifying the scope zone index for IPv6 addresses
as described in RFC 4007 - "IPv6 Scoped Address Architecture".
- The default value for the agent-addr in SNMPv1 Trap-PDUs is now the
IP address associated with the interface on which the trap will be
transmitted.
- Support of the AES privacy protocol was updated to be compliant with
RFC 3826 - "The Advanced Encryption Standard (AES) Cipher Algorithm
in the SNMP User-based Security Model".
- Corrected an issue where any non-blocking SNMPv3 message queued prior
to calling snmp_dispatcher() was sent with an empty contextEngineID.
- The first SNMPv3 discovery message is again being sent with a zero-
length msgUserName as suggested by RFC 3414.
- All sockets are now flagged as non-blocking to prevent a possible
deadlock due to an interaction between recv() and select().
- The sending of messages is now bounded by the receive processing rate
to avoid receive buffer overflows.
- The return value of select() is now checked for both "undef" and -1.
- The "usm.t" tests are now skipped if any of the non-core modules
required by the Net::SNMP::Security::USM module are not present.
I got few private comments that one should not use buildlink3.mk but the
standard DEPENDS statement because we do not link against libraries
in this case.
${VARBASE}/db/nsd.db on all platforms and use user/group nsd for the
daemon to run as. Install sample configuration without .sample
extension. Take maintainership. Bump revision.
