JINMEI Tatuya [Thu, 17 Jan 2013 17:09:30 +0000 (09:09 -0800)]
[2438] make sure the branch is fully synchronized with trac2435
there are some changes and files that don't exist the latest trac2435 branch.
they are probably because of the intermediate merge of a separate working
branch, which was based on an earlier version of trac2435.
Don't write to the same file at two different rules. That led to running
the two rules in parallel, writing there at the same time, confusing the
script that loaded it just after that.
Mukund Sivaraman [Thu, 17 Jan 2013 10:56:26 +0000 (16:26 +0530)]
[2499] Fix client list unittests to supply an NS record always
Tests used to check for an NS record as proof that the zone had
reloaded. But the validation now requires the NS record. So we
switch to checking for an A record instead.
Marcin Siodelski [Thu, 17 Jan 2013 09:14:26 +0000 (10:14 +0100)]
[2637] Cast uint8_t to int when logging them.
When directly logging uint8_t values they are rather printed as char,
not the integer value within the range of 0..255. Converting them to
integer guarantees that they are printed as numbers.
JINMEI Tatuya [Wed, 16 Jan 2013 19:37:16 +0000 (11:37 -0800)]
[2573] count added RR within datasrc loader, and deprecate MasterLoader method
in fact, the master loader's method is redundant because its user can count
them the same way. its own method might be of some use in the future,
but until then I'd simpy remove it based on YAGNI.
Also made some unrelated cleanup based on recent clarification of
null tests on pointer/shared pointers.
Jelte Jansen [Wed, 16 Jan 2013 14:28:49 +0000 (15:28 +0100)]
[2595] Better handling of bad files in b10-cmdctl
Still a fatal error if it gets to that point, but
- exact error message
- basic checks (exists, isfile, isreadable) are done when updating configuration as well
Let isc::datasrc::ZoneLoader validate the data before committing by
isc::dns::checkZone. Also, remove the in-house validation from b10-loadzone,
since it's only a subset of the checkZone, which is now performed.
Marcin Siodelski [Wed, 16 Jan 2013 09:09:46 +0000 (10:09 +0100)]
[master] Initialize the list of strings before passing it to boost foreach.
This overcomes the compilation failure on Sunstudio that is caused by the
boost::foreach_detail_::is_rvalue_() overloading ambiguity when
passing the rvalue to boost foreach.
Stephen Morris [Tue, 15 Jan 2013 15:40:42 +0000 (15:40 +0000)]
[2559] Remove the temporary configuration of the lease database
The DHCP6 server had the lease database configuration temporarily
hard-coded. With the introduction of database configuration
parameters, that is no longer needed.
This way we check the logging works to the correct place and that the
configuration is handled. It is better checked by lettuce than unit
tests, since there's a lot of inter-module communication in it.
Once the config manager is ready, connect to itself to get the config.
Not covered by unit tests, since it is part of startup routine and
requires interaction with other modules, but lettuce succeeds, which
means msgq itself works.
Mukund Sivaraman [Tue, 15 Jan 2013 05:53:43 +0000 (11:23 +0530)]
[2435] Make ZoneUpdater::getRRsetCollection() return a RRsetCollectionBase&
... because in case of some updaters, the returned RRsetCollectionBase
is very well a singleton for that updater instance, and it does not make
sense to return multiple RRsetCollections for it.
JINMEI Tatuya [Tue, 15 Jan 2013 03:15:15 +0000 (19:15 -0800)]
[2572] convert OpenError to Unexpected in the stream ver. of pushSource.
with the change for #2572 this version can now have this exception, and
since it's intended to be hidden within the lexer and input source classes,
we need to convert it. doc/tests are also added/updated.
JINMEI Tatuya [Mon, 14 Jan 2013 22:04:12 +0000 (14:04 -0800)]
[master] place internal exception class in a separate namespace.
this is the same workaround with particular versions of clang++ (on a
particular instance of OS X box) as that we did for isc.log wrapper.
committing it at my discretion.
This will be needed in the following work. If we run in the same thread
and tried to connect to self by the cc library, it would deadlock, since
the library blocks waiting for answer.