]> git.ipfire.org Git - thirdparty/kea.git/log
thirdparty/kea.git
13 years ago[2349] Documentation for the pimpling of two classes
Jelte Jansen [Mon, 15 Oct 2012 13:49:39 +0000 (15:49 +0200)] 
[2349] Documentation for the pimpling of two classes

13 years ago[2223] completely pull out sigs and check entire set
Jelte Jansen [Mon, 15 Oct 2012 12:57:05 +0000 (14:57 +0200)] 
[2223] completely pull out sigs and check entire set

13 years ago[2342] Separate out lease manager creation into separate factory functions
Stephen Morris [Mon, 15 Oct 2012 11:29:17 +0000 (12:29 +0100)] 
[2342] Separate out lease manager creation into separate factory functions

13 years ago[2207] Implementation of the getZoneReloader
Michal 'vorner' Vaner [Mon, 15 Oct 2012 10:14:04 +0000 (12:14 +0200)] 
[2207] Implementation of the getZoneReloader

The concrete version for local table segment and reloader.

13 years ago[2207] Docs and capitalization fix
Michal 'vorner' Vaner [Mon, 15 Oct 2012 10:13:49 +0000 (12:13 +0200)] 
[2207] Docs and capitalization fix

13 years ago[2207] Tests for the getZoneReloader
Michal 'vorner' Vaner [Mon, 15 Oct 2012 10:13:06 +0000 (12:13 +0200)] 
[2207] Tests for the getZoneReloader

13 years ago[2207] interface of the getZoneReloader
Michal 'vorner' Vaner [Mon, 15 Oct 2012 09:28:35 +0000 (11:28 +0200)] 
[2207] interface of the getZoneReloader

13 years ago[2198] Update comment
Mukund Sivaraman [Mon, 15 Oct 2012 04:31:18 +0000 (10:01 +0530)] 
[2198] Update comment

13 years ago[2198] Test the non-blocking variant of Mutex::Locker
Mukund Sivaraman [Mon, 15 Oct 2012 04:29:22 +0000 (09:59 +0530)] 
[2198] Test the non-blocking variant of Mutex::Locker

13 years ago[2198] Check pthread_mutex_trylock()'s return value
Mukund Sivaraman [Mon, 15 Oct 2012 04:08:39 +0000 (09:38 +0530)] 
[2198] Check pthread_mutex_trylock()'s return value

13 years ago[2198] Revert include to local
Mukund Sivaraman [Sun, 14 Oct 2012 18:33:13 +0000 (00:03 +0530)] 
[2198] Revert include to local

13 years ago[2198] Stop using direct lock methods on Mutex inside InterprocessSyncFile
Mukund Sivaraman [Sun, 14 Oct 2012 16:52:35 +0000 (22:22 +0530)] 
[2198] Stop using direct lock methods on Mutex inside InterprocessSyncFile

13 years ago[2198] Make direct methods on Mutex private
Mukund Sivaraman [Sun, 14 Oct 2012 16:49:47 +0000 (22:19 +0530)] 
[2198] Make direct methods on Mutex private

13 years ago[2198] Add a block argument to Mutex::Locker constructor
Mukund Sivaraman [Sun, 14 Oct 2012 16:47:29 +0000 (22:17 +0530)] 
[2198] Add a block argument to Mutex::Locker constructor

13 years ago[2198] Add AlreadyLocked exception to Mutex::Locker
Mukund Sivaraman [Sun, 14 Oct 2012 16:46:41 +0000 (22:16 +0530)] 
[2198] Add AlreadyLocked exception to Mutex::Locker

13 years agoMerge branch 'master' into trac2302
Mukund Sivaraman [Sun, 14 Oct 2012 14:43:42 +0000 (20:13 +0530)] 
Merge branch 'master' into trac2302

13 years ago[2268] Make minor code cleanup
Mukund Sivaraman [Sun, 14 Oct 2012 14:29:55 +0000 (19:59 +0530)] 
[2268] Make minor code cleanup

13 years ago[2168] Untabify
Mukund Sivaraman [Sun, 14 Oct 2012 14:15:37 +0000 (19:45 +0530)] 
[2168] Untabify

13 years ago[2268] Remove per-test iterator classes and use MockVectorIterator instead
Mukund Sivaraman [Sun, 14 Oct 2012 14:14:41 +0000 (19:44 +0530)] 
[2268] Remove per-test iterator classes and use MockVectorIterator instead

13 years ago[2268] Add MockVectorIterator class
Mukund Sivaraman [Sun, 14 Oct 2012 14:03:46 +0000 (19:33 +0530)] 
[2268] Add MockVectorIterator class

13 years ago[2207] Deprecate InstallAction
Michal 'vorner' Vaner [Sat, 13 Oct 2012 11:03:26 +0000 (13:03 +0200)] 
[2207] Deprecate InstallAction

As we want to directly modify the ZoneTable in ZoneReloaderLocal
(because it is memory-type dependant), we don't need the InstallAction
for anything. If we don't need it, we don't pass it as constructor
parameter -- it'll be called from a factory anyway and the factory may
know we don't need it.

Also adding a setTable() method to ZoneTableHeader. The header had no
way to specify the table currently and we need it for the tests. The
method is currently tentative, we'll probably replace it by something.

13 years ago[2207] Check the load action returns something
Michal 'vorner' Vaner [Sat, 13 Oct 2012 09:48:18 +0000 (11:48 +0200)] 
[2207] Check the load action returns something

NULL makes no sense, so throw in that case.

13 years ago[2207] Change signature of LoadAction
Michal 'vorner' Vaner [Sat, 13 Oct 2012 09:38:58 +0000 (11:38 +0200)] 
[2207] Change signature of LoadAction

We allocate the ZoneData in the LoadAction, to match what happens in
ticket #2268 -- there's a function to load zone which allocates it
itself, we want to use it.

13 years ago[2207] Rename updater to reloader
Michal 'vorner' Vaner [Sat, 13 Oct 2012 09:21:44 +0000 (11:21 +0200)] 
[2207] Rename updater to reloader

This might be less confusing.

13 years ago[2205] added DataSrcClientsBuilder class, main part of configurator thread
JINMEI Tatuya [Sat, 13 Oct 2012 02:44:18 +0000 (19:44 -0700)] 
[2205] added DataSrcClientsBuilder class, main part of configurator thread

13 years ago[2205] Merge branch 'trac2332' into trac2205
JINMEI Tatuya [Fri, 12 Oct 2012 21:32:53 +0000 (14:32 -0700)] 
[2205] Merge branch 'trac2332' into trac2205

13 years ago[2332] missing cleanup: removed an unused function.
JINMEI Tatuya [Fri, 12 Oct 2012 21:32:22 +0000 (14:32 -0700)] 
[2332] missing cleanup: removed an unused function.

13 years ago[2205] Merge branch 'trac2332' into trac2205
JINMEI Tatuya [Fri, 12 Oct 2012 21:30:21 +0000 (14:30 -0700)] 
[2205] Merge branch 'trac2332' into trac2205

13 years ago[1858] (unrelated cleanup) removed BIND10_START_AS_NON_ROOT_AUTH log.
JINMEI Tatuya [Fri, 12 Oct 2012 20:48:46 +0000 (13:48 -0700)] 
[1858] (unrelated cleanup) removed BIND10_START_AS_NON_ROOT_AUTH log.

with the introduction of socket creation this log message hasn't been
making sense.  It would be rather confusing and harmful.

13 years ago[2268] minor editorial cleanups
JINMEI Tatuya [Fri, 12 Oct 2012 18:49:07 +0000 (11:49 -0700)] 
[2268] minor editorial cleanups

13 years ago[2349] Same for RunningQuery
Jelte Jansen [Fri, 12 Oct 2012 15:04:15 +0000 (17:04 +0200)] 
[2349] Same for RunningQuery

See 26483ea27726d4c1097485196a0908c565478188

13 years ago[2270] Comment about code overlap added.
Tomek Mrugalski [Fri, 12 Oct 2012 13:06:23 +0000 (15:06 +0200)] 
[2270] Comment about code overlap added.

13 years ago[2270] Documentation for DHCPv4 config. parser written.
Tomek Mrugalski [Fri, 12 Oct 2012 12:55:59 +0000 (14:55 +0200)] 
[2270] Documentation for DHCPv4 config. parser written.

13 years ago[2349] clean up ForwardQuery objects in tests
Jelte Jansen [Fri, 12 Oct 2012 12:19:19 +0000 (14:19 +0200)] 
[2349] clean up ForwardQuery objects in tests

an attempt to clean up memory; ForwardQuery objects are supposed to delete themselves when they are no longer needed, but due to the direct calls in tests this often does not happen.

To keep internals in recursive_query to a minimum, I did add something to the API; ForwardQuery is now a public class, a pointer of which is returned by RecursiveQuery::forward. Tests that mess around can then delete the query themselves, and the way these are handled does not have to be changed.

Ideally, we should however not need unscoped allocated memory in the first place; we either need to use a centrally managed pool of objects to (re)use, or we need to change the entire design that needs these objects in the first place.

However, these are both non-trivial, and probably better suited in the main resolver work that has been planned.

13 years ago[2270] DHCPv6 pool parser fix.
Tomek Mrugalski [Fri, 12 Oct 2012 11:53:10 +0000 (13:53 +0200)] 
[2270] DHCPv6 pool parser fix.

13 years ago[2270] DHCPv4 parser fixes. Tests now pass.
Tomek Mrugalski [Fri, 12 Oct 2012 11:52:36 +0000 (13:52 +0200)] 
[2270] DHCPv4 parser fixes. Tests now pass.

13 years ago[2304] Minor additions in comments.
Marcin Siodelski [Fri, 12 Oct 2012 07:28:40 +0000 (09:28 +0200)] 
[2304] Minor additions in comments.

13 years ago[2155] added some notes to the mapping tables of counter item
Yoshitaka Aharen [Fri, 12 Oct 2012 07:17:20 +0000 (16:17 +0900)] 
[2155] added some notes to the mapping tables of counter item

13 years ago[2304] Fixes for the code review.
Marcin Siodelski [Fri, 12 Oct 2012 07:02:33 +0000 (09:02 +0200)] 
[2304] Fixes for the code review.

Includes adding missing tests for signed integer values

13 years ago[2155] update test case according to the review comment
Yoshitaka Aharen [Fri, 12 Oct 2012 05:21:15 +0000 (14:21 +0900)] 
[2155] update test case according to the review comment

13 years ago[2268] Remove obsolete comment (the bug was fixed)
Mukund Sivaraman [Fri, 12 Oct 2012 05:00:12 +0000 (10:30 +0530)] 
[2268] Remove obsolete comment (the bug was fixed)

13 years ago[2268] Move FileNameDeleter out of InMemoryClient
Mukund Sivaraman [Fri, 12 Oct 2012 04:40:11 +0000 (10:10 +0530)] 
[2268] Move FileNameDeleter out of InMemoryClient

13 years ago[2268] Update API doc
Mukund Sivaraman [Fri, 12 Oct 2012 04:37:13 +0000 (10:07 +0530)] 
[2268] Update API doc

13 years ago[2268] Move ZoneDataLoader into an unnamed namespace
Mukund Sivaraman [Fri, 12 Oct 2012 04:37:04 +0000 (10:07 +0530)] 
[2268] Move ZoneDataLoader into an unnamed namespace

13 years ago[2268] Move ZoneDataLoader class out of its header
Mukund Sivaraman [Fri, 12 Oct 2012 04:24:54 +0000 (09:54 +0530)] 
[2268] Move ZoneDataLoader class out of its header

13 years ago[2268] Use loadZoneData() inside InMemoryClient
Mukund Sivaraman [Fri, 12 Oct 2012 04:21:30 +0000 (09:51 +0530)] 
[2268] Use loadZoneData() inside InMemoryClient

13 years ago[2268] Implement loadZoneData() functions
Mukund Sivaraman [Fri, 12 Oct 2012 04:06:51 +0000 (09:36 +0530)] 
[2268] Implement loadZoneData() functions

13 years ago[2268] Move ZoneDataLoader to a separate file
Mukund Sivaraman [Fri, 12 Oct 2012 03:40:38 +0000 (09:10 +0530)] 
[2268] Move ZoneDataLoader to a separate file

13 years ago[2268] Rename load method to loadInternal()
Mukund Sivaraman [Fri, 12 Oct 2012 03:40:15 +0000 (09:10 +0530)] 
[2268] Rename load method to loadInternal()

13 years ago[2268] Add a comment about InMemoryClient construction after setThrowCount()
Mukund Sivaraman [Fri, 12 Oct 2012 03:18:14 +0000 (08:48 +0530)] 
[2268] Add a comment about InMemoryClient construction after setThrowCount()

13 years ago[2268] Add any attached RRSIGs when loading from an iterator
Mukund Sivaraman [Fri, 12 Oct 2012 02:51:09 +0000 (08:21 +0530)] 
[2268] Add any attached RRSIGs when loading from an iterator

For example, the MemoryIterator returns TreeNodeRRsets which may have
attached RRSIGs.

This also fixes the failing unittest
MemoryClientTest.loadRRSIGsRdataMixedCoveredTypes.

13 years ago[1858] cleanup: reorder log messages
JINMEI Tatuya [Thu, 11 Oct 2012 22:41:36 +0000 (15:41 -0700)] 
[1858] cleanup: reorder log messages

13 years ago[1858] stop sending hopeless signal when it once fails with EPERM.
JINMEI Tatuya [Thu, 11 Oct 2012 22:39:53 +0000 (15:39 -0700)] 
[1858] stop sending hopeless signal when it once fails with EPERM.

13 years ago[1858] added a test case where killing child process results in exception.
JINMEI Tatuya [Thu, 11 Oct 2012 22:05:01 +0000 (15:05 -0700)] 
[1858] added a test case where killing child process results in exception.

13 years ago[1858] rename terminate_children() kill_... for consistency
JINMEI Tatuya [Thu, 11 Oct 2012 21:49:33 +0000 (14:49 -0700)] 
[1858] rename terminate_children() kill_... for consistency

13 years ago[1858] cleanup: removed unnecessary list() conversion from dict values.
JINMEI Tatuya [Thu, 11 Oct 2012 21:42:48 +0000 (14:42 -0700)] 
[1858] cleanup: removed unnecessary list() conversion from dict values.

dict.values() is an iterable, so we can use it directly in the for statement.
(and we don't modify components within the loop).

13 years ago[1858] refactoring: unify the SIGTERM/KILL cases into a single helper method.
JINMEI Tatuya [Thu, 11 Oct 2012 21:41:47 +0000 (14:41 -0700)] 
[1858] refactoring: unify the SIGTERM/KILL cases into a single helper method.

13 years ago[1858] explicitly set close_fds to True in sockcreator's Popen().
JINMEI Tatuya [Thu, 11 Oct 2012 21:03:58 +0000 (14:03 -0700)] 
[1858] explicitly set close_fds to True in sockcreator's Popen().

13 years ago[2207] Add missing test file
Michal 'vorner' Vaner [Thu, 11 Oct 2012 17:52:59 +0000 (19:52 +0200)] 
[2207] Add missing test file

Someone O:-) forgot to add it in the ^^ commit.

13 years ago[2207] Implementation of the zone updater
Michal 'vorner' Vaner [Thu, 11 Oct 2012 17:52:12 +0000 (19:52 +0200)] 
[2207] Implementation of the zone updater

The interface was slightly updated, to suit the needs (added some more
parameters). Some of it might be wrong, but I'm not sure about it.

13 years ago[master] add changelog entry for #2340
JINMEI Tatuya [Thu, 11 Oct 2012 17:42:39 +0000 (10:42 -0700)] 
[master] add changelog entry for #2340

13 years agoMerge branch 'trac2340' of ssh://git.bind10.isc.org/var/bind10/git/bind10 into trac2340
JINMEI Tatuya [Thu, 11 Oct 2012 17:35:09 +0000 (10:35 -0700)] 
Merge branch 'trac2340' of ssh://git.bind10.isc.org/var/bind10/git/bind10 into trac2340

with fixing Conflicts:

13 years agoMerge branch 'master' into trac2340
JINMEI Tatuya [Thu, 11 Oct 2012 17:22:29 +0000 (10:22 -0700)] 
Merge branch 'master' into trac2340
with fixing conflicts in
ChangeLog
examples/README

13 years ago[2207] Tests for the zone updater
Michal 'vorner' Vaner [Thu, 11 Oct 2012 17:08:56 +0000 (19:08 +0200)] 
[2207] Tests for the zone updater

Doesn't compile, there's no (not even empty) implementation of many
methods of the updater.

13 years ago[2270] Configuration parser for DHCPv4
Tomek Mrugalski [Thu, 11 Oct 2012 17:08:23 +0000 (19:08 +0200)] 
[2270] Configuration parser for DHCPv4

13 years ago[2340] removed a bogus 'test' within a test(1) expression.
JINMEI Tatuya [Thu, 11 Oct 2012 16:41:47 +0000 (09:41 -0700)] 
[2340] removed a bogus 'test' within a test(1) expression.

13 years ago[master] changelog for #2339
JINMEI Tatuya [Wed, 10 Oct 2012 20:50:21 +0000 (13:50 -0700)] 
[master] changelog for #2339

13 years ago[2339] changed the search order of python executable: seek python3.x first.
JINMEI Tatuya [Tue, 9 Oct 2012 20:17:59 +0000 (13:17 -0700)] 
[2339] changed the search order of python executable: seek python3.x first.

this works around a bit awkward installation setup where there's a link
from "python3" to "python3.x" but not from "python3-config" to the
corresponding "python3.x-config".  That happens for recent versions of
Homebrew.

13 years agoCompilation fix: usual static_cast<X*>(NULL)
Michal 'vorner' Vaner [Wed, 10 Oct 2012 18:42:46 +0000 (20:42 +0200)] 
Compilation fix: usual static_cast<X*>(NULL)

Sunstudio is unhappy with plain NULL inside EXPECT_EQ and thinks it is
int. This makes it think otherwise.

13 years agoRemove extra comma at the end of enum
Michal 'vorner' Vaner [Wed, 10 Oct 2012 18:28:16 +0000 (20:28 +0200)] 
Remove extra comma at the end of enum

This should fix compilation issue on sunstudio.

13 years ago[master] added changelog for #2244
JINMEI Tatuya [Wed, 10 Oct 2012 16:32:40 +0000 (09:32 -0700)] 
[master] added changelog for #2244

13 years ago[master] a minor grammar fix to changelog entry.
JINMEI Tatuya [Wed, 10 Oct 2012 16:31:30 +0000 (09:31 -0700)] 
[master] a minor grammar fix to changelog entry.

13 years ago[2244] missing update: Component.running() was renamed is_running().
JINMEI Tatuya [Tue, 9 Oct 2012 18:11:51 +0000 (11:11 -0700)] 
[2244] missing update: Component.running() was renamed is_running().

13 years ago[2244] renamed BaseComponent.is_failed() to is_restarting().
JINMEI Tatuya [Mon, 8 Oct 2012 23:14:16 +0000 (16:14 -0700)] 
[2244] renamed BaseComponent.is_failed() to is_restarting().

so it sounds more natural as an English term.

13 years ago[2244] fixed a typo in a log message.
JINMEI Tatuya [Mon, 8 Oct 2012 22:49:24 +0000 (15:49 -0700)] 
[2244] fixed a typo in a log message.

13 years ago[2244] rename Componet.running() is_running() for consistency.
JINMEI Tatuya [Fri, 5 Oct 2012 05:19:27 +0000 (22:19 -0700)] 
[2244] rename Componet.running() is_running() for consistency.

this branch introduced is_failed() (and failed() was already defined for
a different a purpose), so for consistency it would be better to name
the running version is_xxx too.

13 years ago[2244] cleanup: reordered log message file
JINMEI Tatuya [Fri, 5 Oct 2012 05:11:45 +0000 (22:11 -0700)] 
[2244] cleanup: reordered log message file

13 years ago[2244] make sure restarting components only when they are in config.
JINMEI Tatuya [Fri, 5 Oct 2012 05:10:40 +0000 (22:10 -0700)] 
[2244] make sure restarting components only when they are in config.

this should fix the main problem of this ticket.

13 years ago[2244] a minor editorial fix to the previous comment update
JINMEI Tatuya [Fri, 5 Oct 2012 04:08:55 +0000 (21:08 -0700)] 
[2244] a minor editorial fix to the previous comment update

13 years ago[2244] added comment about has_comonent as it may not be very obvious.
JINMEI Tatuya [Fri, 5 Oct 2012 03:53:37 +0000 (20:53 -0700)] 
[2244] added comment about has_comonent as it may not be very obvious.

13 years ago[2244] added a simple has_component() method to configurtor.
JINMEI Tatuya [Fri, 5 Oct 2012 00:55:57 +0000 (17:55 -0700)] 
[2244] added a simple has_component() method to configurtor.

should be pretty trivial, so added a couple of test cases as part of
other tests.

13 years ago[2244] make sure configurtor's build_plan() handles 'failed' component.
JINMEI Tatuya [Fri, 5 Oct 2012 00:38:43 +0000 (17:38 -0700)] 
[2244] make sure configurtor's build_plan() handles 'failed' component.

this fixes the test case introduced at the beginning of the branch.

13 years ago[2244] introduced a new "Failed" state of Component.
JINMEI Tatuya [Fri, 5 Oct 2012 00:37:09 +0000 (17:37 -0700)] 
[2244] introduced a new "Failed" state of Component.

13 years ago[2244] added a test case that reproduces one aspect of the problem.
JINMEI Tatuya [Fri, 5 Oct 2012 00:27:15 +0000 (17:27 -0700)] 
[2244] added a test case that reproduces one aspect of the problem.

the configurator should be able to delete a component to be removed
even if the it's failing (unexpectedly).  The current implementation
doesn't handle that condition correctly.

13 years ago[2204] removed an obsolete comment that only made sense for the previous code.
JINMEI Tatuya [Tue, 9 Oct 2012 20:28:08 +0000 (13:28 -0700)] 
[2204] removed an obsolete comment that only made sense for the previous code.

13 years ago[2204] added notes about future extensions of incremental datasrc config update
JINMEI Tatuya [Tue, 9 Oct 2012 20:26:20 +0000 (13:26 -0700)] 
[2204] added notes about future extensions of incremental datasrc config update

13 years ago[2204] grammar fix in documentation
JINMEI Tatuya [Tue, 9 Oct 2012 00:29:15 +0000 (17:29 -0700)] 
[2204] grammar fix in documentation

13 years ago[2204] updated test comments without using "rollback".
JINMEI Tatuya [Tue, 9 Oct 2012 00:27:23 +0000 (17:27 -0700)] 
[2204] updated test comments without using "rollback".

as the new code actually doesn't do rollback operation internally,
even though the observable effect is the same.

13 years ago[2204] updated comments about releasing old data source client lists
JINMEI Tatuya [Mon, 8 Oct 2012 23:56:10 +0000 (16:56 -0700)] 
[2204] updated comments about releasing old data source client lists

the previous one was not really accurate and was confusing.

13 years ago[2204] cleanup: removed unused func, and renamed "ClientList" with "DataSrc".
JINMEI Tatuya [Sat, 6 Oct 2012 00:04:28 +0000 (17:04 -0700)] 
[2204] cleanup: removed unused func, and renamed "ClientList" with "DataSrc".

getClientListClasses() was removed.  It's not used anymore.
a few methods named with "ClientList" were renamed with adding "DataSrc".
The resulting names are a bit long, but since there's no namespace
qualification, simply saying "client" can be confusing (it could be
interpreted to mean DNS clients, for example).

13 years ago[2204] regression fix: make sure the datasrc config is installed to server.
JINMEI Tatuya [Fri, 5 Oct 2012 23:52:37 +0000 (16:52 -0700)] 
[2204] regression fix: make sure the datasrc config is installed to server.

13 years ago[2204] removed server function/template param from configureDataSourceGeneric.
JINMEI Tatuya [Fri, 5 Oct 2012 23:48:03 +0000 (16:48 -0700)] 
[2204] removed server function/template param from configureDataSourceGeneric.

This is pure cleanup.  This function already doesn't use the server.

13 years ago[2204] extracted swapDataSrcClientLists() from configureDataSource().
JINMEI Tatuya [Fri, 5 Oct 2012 23:40:10 +0000 (16:40 -0700)] 
[2204] extracted swapDataSrcClientLists() from configureDataSource().

now configureDataSource() can take time.

13 years ago[2204] completely replaced setClientList with swapDataSrcClientLists.
JINMEI Tatuya [Fri, 5 Oct 2012 23:02:01 +0000 (16:02 -0700)] 
[2204] completely replaced setClientList with swapDataSrcClientLists.

the test cases using setClientList were updated so they use
swapDataSrcClientLists (some of them work as a test for the "swap" itself).
now we don't need setClientList, so it was removed.

13 years ago[2204] simplify configureDataSource by always creating a new lists and swap.
JINMEI Tatuya [Fri, 5 Oct 2012 22:15:18 +0000 (15:15 -0700)] 
[2204] simplify configureDataSource by always creating a new lists and swap.

so we don't have to worry about what are in the current lists or rollback
operations.
swapDataSrcClientLists() is newly introduced for AuthSrv.  No direc tests
yet (technically bad in terms TDD but the definition is very simple), which
will be provided in the next step.
the lock is now moved inside swapDataSrcClientLists().

note: even though this version builds everything, the amount of work
should be mostly the same because the only save is to create the empty
ClientList when the new and old have the same class of client.  The expensive
part is ClientList::configure().  This version doesn't need any more call
to configure() than the old version.

13 years ago[2204] changed internal representation of auth_srv's "client_lists_" to ptrs.
JINMEI Tatuya [Fri, 5 Oct 2012 21:03:52 +0000 (14:03 -0700)] 
[2204] changed internal representation of auth_srv's "client_lists_" to ptrs.

this will be necessary later in this branch.
also renamed the member variable "datasrc_client_lists_" as the mere "client"
can be ambiguous.

13 years ago[2292] Add a TODO about re-adding tests
Michal 'vorner' Vaner [Wed, 10 Oct 2012 09:58:17 +0000 (11:58 +0200)] 
[2292] Add a TODO about re-adding tests

Planned for merge #2268

13 years ago[2292] Clarify test
Michal 'vorner' Vaner [Tue, 9 Oct 2012 12:11:25 +0000 (14:11 +0200)] 
[2292] Clarify test

And make it slightly stronger (it now checks something that should be
mostly obvious too, but who knows, with software).

13 years ago[2292] Pass the old zone back
Michal 'vorner' Vaner [Tue, 9 Oct 2012 12:07:13 +0000 (14:07 +0200)] 
[2292] Pass the old zone back

Instead of releasing it directly. While the internal release was more
convenient, it didn't allow for swapping things fast under a mutex and
then spending the time releasing it unlocked.

13 years ago[2292] Don't pass SegmentObjectHolder
Michal 'vorner' Vaner [Tue, 9 Oct 2012 11:39:30 +0000 (13:39 +0200)] 
[2292] Don't pass SegmentObjectHolder

It is supposed to be mostly private. Use it internally only.