]>
git.ipfire.org Git - thirdparty/kea.git/log
Jelte Jansen [Fri, 18 May 2012 09:30:00 +0000 (11:30 +0200)]
[1207] improved configuration handling for inmem special case
i.e. jinmei's patch, see http://bind10.isc.org/ticket/1207?replyto=15#comment:15
Jelte Jansen [Fri, 18 May 2012 08:46:18 +0000 (10:46 +0200)]
[1207] Disable some more tests when static linking
note that this is mostly a way to at least get it working, we need to figure out a more permanent solution related to static linking (or stop supporting it in the first place)
JINMEI Tatuya [Fri, 18 May 2012 07:11:26 +0000 (00:11 -0700)]
[1539] fixed a small regression: unsupportedRequest now excludes update.
JINMEI Tatuya [Wed, 18 Jan 2012 23:25:04 +0000 (15:25 -0800)]
[1539] introduced class hierarchy for SocketSessionForwarder.
so that we can define and use a mock class for testing.
also added some initial tests.
JINMEI Tatuya [Tue, 10 Jan 2012 08:27:44 +0000 (00:27 -0800)]
[master] [1539] introduce a base class for the socket session forwarder. it's only
for the convenience of tests.
JINMEI Tatuya [Thu, 17 May 2012 23:02:10 +0000 (16:02 -0700)]
[1207] some trivial editorial changes
Michal 'vorner' Vaner [Thu, 17 May 2012 19:20:12 +0000 (21:20 +0200)]
[1914] Update tests for the addMessage extension
As the internal data structure contains another element, we need to put
it into each test that looks into the internal data structure.
Michal 'vorner' Vaner [Thu, 17 May 2012 18:51:00 +0000 (20:51 +0200)]
[1914] Implemented the cancelAsyncRecv method
Isn't it a little bit waste, to have 35 lines of tests for 1-line
wrapper function? O:-)
Michal 'vorner' Vaner [Thu, 17 May 2012 18:48:30 +0000 (20:48 +0200)]
[1914] Tests for the cancelAsyncRecv call
JINMEI Tatuya [Thu, 17 May 2012 18:46:33 +0000 (11:46 -0700)]
[1207] minor style fix: adjusted a brace position.
Michal 'vorner' Vaner [Thu, 17 May 2012 18:15:56 +0000 (20:15 +0200)]
[1914] Check more complicated scenarios
We check that the groupRecvMsgAsync can handle different situations,
like when there are multiple messages and multiple callbacks registered
and when the messages come later than the callbacks.
Michal 'vorner' Vaner [Thu, 17 May 2012 17:56:35 +0000 (19:56 +0200)]
[1914] Fix reply matching tests
The fake session needs to include the reply element in the response
envelope. And the is_reply flag was set to the wrong value for replies
(it was false as well).
Michal 'vorner' Vaner [Thu, 17 May 2012 17:44:07 +0000 (19:44 +0200)]
[1914] More tests for matching
Check that we can wildcard match commands and replies.
Check that we don't match what we shouldn't.
Michal 'vorner' Vaner [Thu, 17 May 2012 15:51:19 +0000 (17:51 +0200)]
[1914] More test fixes
The queued messages still didn't work well.
Removed a problem with logging error triggered in older parts of code.
Michal 'vorner' Vaner [Thu, 17 May 2012 15:26:59 +0000 (17:26 +0200)]
[1914] Fix tests
The fake session didn't want to support hasQueuedMessages properly.
Also, we need to be subscribed to the correct group to actually get the
messages through the session.
Michal 'vorner' Vaner [Thu, 17 May 2012 14:49:47 +0000 (16:49 +0200)]
[1914] Cleanup: constify
The values inside the request never need to be changed after they are
stored.
Michal 'vorner' Vaner [Thu, 17 May 2012 14:46:27 +0000 (16:46 +0200)]
[1914] Cleanup: remove shared_ptr
It seems a list can be declared for a class that is not fully defined
yet. So there's no need for pointers, so even less for shared ones.
Michal 'vorner' Vaner [Thu, 17 May 2012 14:43:21 +0000 (16:43 +0200)]
[1914] Check for async requests on a new message
We now check if the message looks like something a callback wants in the
checkCommand() call. For some reason, it does not seem to work with the
tests, it needs to be checked why.
Jeremy C. Reed [Thu, 17 May 2012 14:07:22 +0000 (09:07 -0500)]
[master] also link with libasiolink
The -no-undefined I added previously caused failures on Mac.
On my Linux and NetBSD tests it didn't notice. But the Macmini
detected it. The top of error output is clear:
Undefined symbols:
"isc::asiolink::IOAddress::operator unsigned int() const", referenced
from:
isc::dhcp::IfaceMgr::send(boost::shared_ptr<isc::dhcp::Pkt4>
const&)in libdhcp___la-iface_mgr.o
Okayed by tomek via jabber.
Michal 'vorner' Vaner [Thu, 17 May 2012 13:32:36 +0000 (15:32 +0200)]
[1914] Store the request into a queue
The requests submitted by groupReadMsgAsync are stored into the list.
The tests still fail, since nothing is interested in the content of the
list yet.
Michal 'vorner' Vaner [Thu, 17 May 2012 13:04:54 +0000 (15:04 +0200)]
[1914] Basic tests for groupRecvMsgAsync
* Needed to extend the FakeSession::addMessage to hold the sequence
number for representing answers.
* Clarified the recipient parameter - it is the group, instance is
ignored, since it is not used anywhere in the code anyway.
* The tests simply register a callback and check it is called. More
complex tests to come later.
Jeremy C. Reed [Thu, 17 May 2012 11:33:48 +0000 (06:33 -0500)]
[master]Merge branch 'master' of ssh://git.bind10.isc.org/var/bind10/git/bind10
Jeremy C. Reed [Thu, 17 May 2012 11:31:01 +0000 (06:31 -0500)]
[master] set the library version for libdhcp++
The libdhcp++ API was changed as part of the refactoring.
When linking new tests to old library it fails with:
/libexec/ld-elf.so.1:
/usr/home/jreed/builder/work/BIND10-cppcheck/
20120516153001 -FreeBSD8-amd64-GCC/build/src/bin/dhcp4/tests/.libs/dhcp4_unittests:
Undefined symbol "_ZN3isc4dhcp6OptionC1ENS1_8UniverseEtRKSt6vectorIhSaIhEE"
This is:
isc::dhcp::Option::Option(isc::dhcp::Option::Universe, unsigned short,
std::vector<unsigned char, std::allocator<unsigned char> > const&)
See http://bind10.isc.org/wiki/LibraryVersioning for details.
Tomek Mrugalski [Thu, 17 May 2012 11:30:52 +0000 (13:30 +0200)]
[1967] DHCPv6 tests now use boost::scoped_ptr<>
Michal 'vorner' Vaner [Thu, 17 May 2012 11:07:46 +0000 (13:07 +0200)]
[1914] Unrelated cleanup: removed extra anonymous namespace
There was an anonymous namespace inside another anonymous namespace.
This brings no further advantage and is just confusing, so it got
removed.
Mukund Sivaraman [Wed, 16 May 2012 05:23:30 +0000 (10:53 +0530)]
[1628] Update path where ddns.spec is read when running from installed copy
Mukund Sivaraman [Wed, 16 May 2012 05:12:25 +0000 (10:42 +0530)]
[master] Update .gitignore
Jeremy C. Reed [Thu, 17 May 2012 03:59:26 +0000 (22:59 -0500)]
[master] fix developer user name
per email on dev list
JINMEI Tatuya [Thu, 17 May 2012 03:47:47 +0000 (20:47 -0700)]
[1512] added other .py's to python_PYTHON.
these are necessary for distcheck.
JINMEI Tatuya [Thu, 17 May 2012 02:32:35 +0000 (19:32 -0700)]
[1512] added ddns to SUBDIRS
JINMEI Tatuya [Thu, 17 May 2012 02:32:14 +0000 (19:32 -0700)]
[1512] more comment (doc) update
JINMEI Tatuya [Thu, 17 May 2012 02:22:31 +0000 (19:22 -0700)]
[1512] documentation update
JINMEI Tatuya [Thu, 17 May 2012 02:14:18 +0000 (19:14 -0700)]
[1512] added detailed log description.
JINMEI Tatuya [Thu, 17 May 2012 02:00:12 +0000 (19:00 -0700)]
[1512] logged forwarding (fail) and notauth events separately.
JINMEI Tatuya [Thu, 17 May 2012 01:41:40 +0000 (18:41 -0700)]
[1512] use NOTIMP instead of REFUSED for denying update forwarding.
this is actually closer to the current implementation status, and is
compatible BIND 9's behavior when the forwarding ACL is completely NULLified.
JINMEI Tatuya [Thu, 17 May 2012 01:35:37 +0000 (18:35 -0700)]
[1512] supported 'notauth' case.
JINMEI Tatuya [Thu, 17 May 2012 01:14:48 +0000 (18:14 -0700)]
[1512] introduce a test data source and search it in ZoneConfig.find_zone().
JINMEI Tatuya [Thu, 17 May 2012 00:37:39 +0000 (17:37 -0700)]
[1512] log events when update request somehow failed.
Jeremy C. Reed [Wed, 16 May 2012 23:02:30 +0000 (18:02 -0500)]
[master] use team instead of specific developer for a changelog entry
as requested via jabber
Jeremy C. Reed [Wed, 16 May 2012 22:36:05 +0000 (17:36 -0500)]
[master] add entry for #1627
note I can't find the git hash for the merge so
picked some individual commit.
JINMEI Tatuya [Wed, 16 May 2012 22:31:50 +0000 (15:31 -0700)]
[1512] let UpdateSession.handle() return zone name and class, too.
the caller (b10-ddns) will use that to notify other processes of updates.
Jeremy C. Reed [Wed, 16 May 2012 22:19:35 +0000 (17:19 -0500)]
[master] add three missing entries
from jelte in jabber and jinmei on dev list
also some other additions of trac numbers
Jeremy C. Reed [Wed, 16 May 2012 22:11:02 +0000 (17:11 -0500)]
[master] add a "released" date stamp to changelog
Jeremy C. Reed [Wed, 16 May 2012 21:59:45 +0000 (16:59 -0500)]
[master] changelog cleanup
Lines too long
Use two tabs before the committer username.
Jeremy C. Reed [Wed, 16 May 2012 21:29:14 +0000 (16:29 -0500)]
[master] regenerate bind10-messages.xml and html
catch up reordered entries and also new or changes messages.
Jeremy C. Reed [Wed, 16 May 2012 21:28:25 +0000 (16:28 -0500)]
[master] reorder the .mes entries
used tools/reorder_message_file.py on each mes file.
This should have no content changes.
Jeremy C. Reed [Wed, 16 May 2012 21:27:03 +0000 (16:27 -0500)]
[master] regenerate few manual pages to roff format
catch up for docbook file's changes
Jeremy C. Reed [Wed, 16 May 2012 20:53:33 +0000 (15:53 -0500)]
[master] clarify the zones and filetype configuration
Based on diff from jinmei.
Jeremy C. Reed [Wed, 16 May 2012 20:40:56 +0000 (15:40 -0500)]
[master] document --clear-config and minor cleanup
reviewed by jelte
Jeremy C. Reed [Wed, 16 May 2012 20:40:12 +0000 (15:40 -0500)]
[master] document --clear-config
reviewed by jelte
Michal 'vorner' Vaner [Wed, 16 May 2012 20:36:55 +0000 (22:36 +0200)]
[1914] Documentation for the async receive interface
Jelte Jansen [Wed, 16 May 2012 16:22:32 +0000 (18:22 +0200)]
[1207] revert previous makefile removal
The removal of memory_datasrc.cc in b10_auth_SOURCES was premature; zone loading code is still in auth (but memory_datasrc-specific)
Jeremy C. Reed [Wed, 16 May 2012 14:52:48 +0000 (09:52 -0500)]
[master] version the log library
This is to fix the ongoing problem seen on a builder:
/libexec/ld-elf.so.1: /usr/home/jreed/builder/work/BIND10-cppcheck/
20120516030045 -FreeBSD8-amd64-GCC/build/src/lib/datasrc/tests/.libs/run_unittests: Undefined symbol "_ZN3isc3log23checkExcessPlaceholdersEPSsj"
It was matching the library in the base system instead.
This was discussed on jabber some weeks ago.
Jelte Jansen [Wed, 16 May 2012 14:28:49 +0000 (16:28 +0200)]
[1207] remove makefile memory_datasrc.cc dependency
Which is now no longer necessary
Marcin Siodelski [Wed, 16 May 2012 14:26:45 +0000 (16:26 +0200)]
[1954] Added unit tests to cover remaining options.
Jelte Jansen [Wed, 16 May 2012 14:22:06 +0000 (16:22 +0200)]
[1207] improve load-from-sqlite3 unit test
instead of fighting the API to get empty zones in, actually change the zone in the database and reload
Jelte Jansen [Wed, 16 May 2012 13:40:00 +0000 (15:40 +0200)]
[1207] more review comments addressed
- RAII InMemoryClient's createInstance
- remove dead code
- added more comments
Michal 'vorner' Vaner [Wed, 16 May 2012 13:23:02 +0000 (15:23 +0200)]
[1914] Interface for the async receives
It needs to be tested, documented and of course implemented.
Marcin Siodelski [Wed, 16 May 2012 12:44:49 +0000 (14:44 +0200)]
[1954] Unit tests aded for 8 command line options as well as CommandOptions class members.
Jelte Jansen [Wed, 16 May 2012 11:49:32 +0000 (13:49 +0200)]
[1207] add comments
Jelte Jansen [Wed, 16 May 2012 09:58:11 +0000 (11:58 +0200)]
[1207] initialize FakeClient in FakeContainer constructor
and store it as a scoped ptr
Jelte Jansen [Wed, 16 May 2012 09:43:14 +0000 (11:43 +0200)]
[1207] minor cleanups after review
- fixed a few include statements
- amended a few comments
Jelte Jansen [Wed, 16 May 2012 09:40:43 +0000 (11:40 +0200)]
[1207] FakeInMemoryClient -> FakeClient
it is no longer necessary to derive it from InMemoryClient, but it can be directly derived from DataSourceClient
Jelte Jansen [Wed, 16 May 2012 08:58:58 +0000 (10:58 +0200)]
[1207] Remove special MemoryDatasourceConfig
The special case is still there, but it's been moved into the general DataSourceConfig class; this makes later generalization easier.
It also allowed the removal of the specialized createAuthConfigParser (the overloaded version with the 'internal' argument), and the special case therein.
JINMEI Tatuya [Wed, 16 May 2012 07:49:15 +0000 (00:49 -0700)]
[1512] distinguished successful update and error with a response.
JINMEI Tatuya [Wed, 16 May 2012 07:41:33 +0000 (00:41 -0700)]
[1512] supported the case of update attmpt for a secondary zone.
(sine we're not yet going to support update forwarding, we'll just return
REFUSED, as if it were rejected as some kind of access control)
JINMEI Tatuya [Wed, 16 May 2012 07:38:32 +0000 (00:38 -0700)]
[1512] supported python hash for dns.Name and dns.RRClass.
a bit beyond of this specific task, but it will help the backend implementation
much simpler.
JINMEI Tatuya [Wed, 16 May 2012 07:24:07 +0000 (00:24 -0700)]
[1512] introduced a tentative stub for retrieving zone information.
JINMEI Tatuya [Wed, 16 May 2012 03:58:41 +0000 (20:58 -0700)]
[1512] covered some more invalid zone section cases.
also bit refactored the code to reduce duplicates.
JINMEI Tatuya [Wed, 16 May 2012 03:19:35 +0000 (20:19 -0700)]
[1512] added python wrapper for Message::clearSection().
We'll need this for creating RFC-compliant update responses.
also tightened parameter validation a bit, and added test cases for
bad parameters.
JINMEI Tatuya [Wed, 16 May 2012 02:42:07 +0000 (19:42 -0700)]
[1512] an initial validation: check the # of zone in the zone section.
JINMEI Tatuya [Tue, 15 May 2012 23:34:52 +0000 (16:34 -0700)]
[1512] added mostly-empty build framework for ddns session python module.
JINMEI Tatuya [Tue, 15 May 2012 20:31:56 +0000 (13:31 -0700)]
[master] Merge branch 'trac1893'
Mukund Sivaraman [Tue, 15 May 2012 15:56:53 +0000 (21:26 +0530)]
[1971] Fix broken tree structure that didn't match the one in the figure
Mukund Sivaraman [Mon, 14 May 2012 07:45:20 +0000 (13:15 +0530)]
[1971] Update test rbtree (of rbtrees) to suit our right-child testcase
Mukund Sivaraman [Mon, 14 May 2012 07:22:45 +0000 (12:52 +0530)]
[master] Remove duplicate data
Mukund Sivaraman [Mon, 14 May 2012 06:47:52 +0000 (12:17 +0530)]
[1971] Add another testcase for untested code detected by lcov
When find() returns COMMONANCESTOR and exits to the right of a
last_compared_node which has a valid down pointer, previousNode() has to
return nodes beginning at at the right-most leaf of subtrees in
last_compared_node->down.
Mukund Sivaraman [Mon, 14 May 2012 06:38:20 +0000 (12:08 +0530)]
[1971] Test the node_path.isEmpty() check
If previousNode() has been called enough times that it has popped
all nodes from the node chain when walking up the trees, the
node chain is then empty. If previousNode() is called yet again
on the empty tree, it should return NULL.
Mukund Sivaraman [Mon, 14 May 2012 05:24:01 +0000 (10:54 +0530)]
[master] Update .gitignore
Jelte Jansen [Tue, 15 May 2012 09:53:13 +0000 (11:53 +0200)]
[master] update changelog for merge of #1889
Jelte Jansen [Tue, 15 May 2012 09:51:35 +0000 (11:51 +0200)]
[master] Merge branch 'trac1889'
Mukund Sivaraman [Tue, 15 May 2012 04:53:26 +0000 (10:23 +0530)]
[master] Regen HTML and text versions of the guide
Mukund Sivaraman [Tue, 15 May 2012 04:37:50 +0000 (10:07 +0530)]
[1911] Add paragraph about what the admin needs to do to serve new versions of zones
Mukund Sivaraman [Tue, 15 May 2012 04:33:23 +0000 (10:03 +0530)]
[1911] Make more minor text updates from review
Mukund Sivaraman [Mon, 14 May 2012 02:58:34 +0000 (08:28 +0530)]
[1911] Clarify text about incoming zone transfers
Mukund Sivaraman [Fri, 11 May 2012 03:24:15 +0000 (08:54 +0530)]
[1911] Update guide on how to configure in-mem datasource with sqlite3 backend
JINMEI Tatuya [Tue, 15 May 2012 01:19:46 +0000 (18:19 -0700)]
[1207] editorial/style nits: indentation/brace pos/constify/() for return
JINMEI Tatuya [Mon, 14 May 2012 22:44:05 +0000 (15:44 -0700)]
[master] reverted the change of lettuce test address from IPv6 to IPv4.
The use of IPv6 was intentional and necessary for some environment.
Added a NOTE, hoping it will help remember that and avoid repeating the
mistake.
As a result of this fix it also fixes the error of address+port specification
at the end of xfrin_bind10.feature. It should have been address:port,
but we actually don't even have to specify that because the default combination
should be okay.
JINMEI Tatuya [Mon, 14 May 2012 21:38:16 +0000 (14:38 -0700)]
[1907] added error msg for possible HomeBrew/Lion issue with python.
intending to provide a possible hint to work around it.
Mukund Sivaraman [Mon, 14 May 2012 17:04:01 +0000 (22:34 +0530)]
[1790] Update backend filenames to match the filetype
Mukund Sivaraman [Thu, 10 May 2012 04:00:50 +0000 (09:30 +0530)]
[1790] Add more testcases for missing keys
Strictly, this is not necessary as the spec file schema would check for
it.
Mukund Sivaraman [Thu, 10 May 2012 03:50:08 +0000 (09:20 +0530)]
[1790] Use appropriate boolean assert calls
Mukund Sivaraman [Thu, 10 May 2012 03:48:04 +0000 (09:18 +0530)]
[1790] Update testcase descriptions
Mukund Sivaraman [Tue, 8 May 2012 07:04:12 +0000 (12:34 +0530)]
[1790] Also handle the case where published is not OK
Mukund Sivaraman [Tue, 8 May 2012 04:49:52 +0000 (10:19 +0530)]
[1790] Use a simpler constructor
Mukund Sivaraman [Tue, 8 May 2012 04:35:57 +0000 (10:05 +0530)]
[1790] Add unit tests using mock objects
Mukund Sivaraman [Tue, 8 May 2012 02:53:53 +0000 (08:23 +0530)]
[master] Fix test class name
Mukund Sivaraman [Fri, 4 May 2012 04:31:31 +0000 (10:01 +0530)]
[1790] Don't catch exceptions caused by msgq failures
Mukund Sivaraman [Fri, 4 May 2012 04:25:19 +0000 (09:55 +0530)]
[1790] When throwing RR parameter exceptions, include the bad string as well
Mukund Sivaraman [Thu, 3 May 2012 04:52:28 +0000 (10:22 +0530)]
[1790] Don't use the same log message id twice