[3087] Added state handler unit tests for D2::NameAddTransaction
Created unit tests for all of the states implemented by NameAddTransaction
in b10-dhcp-ddns. The class lacks only proper construction of DNS update
requests.
[3087] Initial implementation of d2::NameAddTransaction
This interrim checkin includes the initial implementation of
NameAddTransaction to b10-dhcp-ddns. It provides the preliminary
state model needed to add forward and/or reverse DNS entries for
and FQDN. It does not yet construct the actual DNS update requests.
Some classes were using references to isc::asiolink::IOService, others
where using d2::IOServicePtr. The latter is now used throughout for
consistency as well as support for future, possible, multi-threaded
implementation.
This is to set the location of the css stylesheet.
Previously it pointed to a path in the source directory.
By default, this was "./" so was okay as a location via HTTP.
But when built using a different source directory, then
the full source directory ended up as a reference on the webpage
so then style's didn't work.
I briefly discussed this on jabber yesterday but didn't get an okay.
It is a minor fix, so committing per my discretion.
Kean Johnston [Tue, 29 Oct 2013 16:01:38 +0000 (18:01 +0200)]
[3172] Fix build failure on BSD systems
The comment at the end of adding to AM_CXXFLAGS was making the total
line too long, and the automake tool was splitting the line. This meant
that the macro assignment had an escaped comment, which the BSD make
interprets differently to GNU make (the latter swallows the next line as
part of the comment, whereas the former doesn't).
Kean Johnston [Tue, 29 Oct 2013 05:34:13 +0000 (07:34 +0200)]
[530] Do not redirect b10-msgq STDOUT when running in verbose mode
When bind10 is run in verbose mode, as is indicated to the user by
certain error messages, do not redirect b10-msgq's STDOUT to /dev/null
so that the user can see the actual cause for startup failure.
Kean Johnston [Tue, 29 Oct 2013 05:31:09 +0000 (07:31 +0200)]
[530] Display socket file used in error message
When there is an error connecting to the message queue socket file
because it wasn't cleaned up properly, and we know the name of the
socket file, display its name as part of the error message.
[3207] Added support for default users and delimited user id input
Vendor option default values are now taken from default users if the
client is not registered. Registry input file user ids can now contain
colons as delimiters allowing cut-and-paste from outcome file.
[3207] Replaced selected subnet with lease address in user check output
User check hook library's outcome output is now generated in pkt_send
callout allowing the output of lease address (or prefix) in place
of selected subnet.
Kean Johnston [Thu, 24 Oct 2013 16:26:23 +0000 (18:26 +0200)]
[433] Verify the existing daemon behaves correctly after multiple invocations
As per review, update the tests to ensure that the existing message
queue daemon still works even after we have had a failed attempt to
start a second copy of the daemon.
Kean Johnston [Thu, 24 Oct 2013 15:49:06 +0000 (17:49 +0200)]
[433] Use a local variable rather than one larger in scope
As per review, don't use self._session for the check to see if there is
an existing message queue running. Rather, use a localally scoped
variable that will get cleaned up when the function exits.
[3207] Added callouts for packet receive and send to user check hooks lib
Initial checkin creates callout functions for packet receive and send
hooks for IPv4 and IPv6. This sets thes stage for outputting user specific
options in outbound packet.
[master] Fixed make install failure in user_chk unit tests
Yet another make issue introduced with trac3186. Unit test data files
are now looked for in the appropriate directory and copied as part of dist
check.
Tomek Mrugalski [Wed, 23 Oct 2013 19:23:14 +0000 (21:23 +0200)]
[3194] Changes after review:
- config parsers updated
- new unit-tests for vendor options written
- libdhcp++ cleaned up
- fixes in option_vendor.cc
- comments added and cleaned up
Tomek Mrugalski [Tue, 22 Oct 2013 18:18:30 +0000 (20:18 +0200)]
[3195] Changes after review:
- Guide updated
- Iface::addUnicast() now refeses duplicates
- addUnicast and clearUnicasts are now commented
- new unit-tests
- argument for addActiveIface is const reference again
-
Tomek Mrugalski [Tue, 22 Oct 2013 13:16:17 +0000 (15:16 +0200)]
[3191] Changes after review
- default value is no longer 0.0.0.0, so global value is no longer
overwritten with subnet-specific values
- added 2 new unit-tests to check if the value is really set in
server responses
- guide updated to explain that
- fix in Dhcp4ParserTest.nextServerNegative