]>
git.ipfire.org Git - thirdparty/kea.git/log
JINMEI Tatuya [Tue, 14 May 2013 21:04:57 +0000 (14:04 -0700)]
[2916] simplified callback call signatures
Thomas Markwalder [Tue, 14 May 2013 20:53:39 +0000 (16:53 -0400)]
[2954] Created the initial server source directory, configuration,
and make files; and process framework for the DHCP-DDNS daemon,
b10-d2srv.
JINMEI Tatuya [Tue, 14 May 2013 18:52:30 +0000 (11:52 -0700)]
[2836] style fix: adjusted '*' position
Thomas Markwalder [Tue, 14 May 2013 18:38:45 +0000 (14:38 -0400)]
[master] Merge branch 'trac2355'
JINMEI Tatuya [Tue, 14 May 2013 16:40:41 +0000 (09:40 -0700)]
[master] changelog for #2851
JINMEI Tatuya [Tue, 14 May 2013 16:28:35 +0000 (09:28 -0700)]
[master] Merge branch 'trac2851'
Mukund Sivaraman [Tue, 14 May 2013 13:20:29 +0000 (18:50 +0530)]
Merge branch 'trac2850_2'
JINMEI Tatuya [Tue, 14 May 2013 04:33:33 +0000 (21:33 -0700)]
[2850] removed now-unused inclusion and using for boost::scoped_ptr
JINMEI Tatuya [Mon, 13 May 2013 22:40:41 +0000 (15:40 -0700)]
[2916] added tests and implementation of the LocalSocket class
Thomas Markwalder [Mon, 13 May 2013 20:07:02 +0000 (16:07 -0400)]
[2355] Cleaned up some minor issues spotted in review.
Stephen Morris [Mon, 13 May 2013 14:56:58 +0000 (15:56 +0100)]
[2653] Added documentation as to how to create the test DHCP database
Mukund Sivaraman [Mon, 13 May 2013 14:30:57 +0000 (20:00 +0530)]
[2850] Update comment that it doesn't actually fail
Stephen Morris [Mon, 13 May 2013 14:11:06 +0000 (15:11 +0100)]
[master] Merge branch 'trac2909'
Marcin Siodelski [Mon, 13 May 2013 14:05:58 +0000 (16:05 +0200)]
[2786] Minor: todo comment added.
Marcin Siodelski [Mon, 13 May 2013 14:02:21 +0000 (16:02 +0200)]
[2786] Use OptionString class to represent options with a string.
Marcin Siodelski [Mon, 13 May 2013 13:03:15 +0000 (15:03 +0200)]
[2786] Implemented unpack() function of OptionString class.
Michal 'vorner' Vaner [Mon, 13 May 2013 13:00:51 +0000 (15:00 +0200)]
[2836] Attempt a test for whole zone loading
Attempt to write a test that loads a bunch of zones, probably causing a
relocation.
However, it currently crashes and does so before the relocation happens,
which is strange. It does not crash with the local segment, only with
the mapped, which is stranger.
Marcin Siodelski [Mon, 13 May 2013 12:55:31 +0000 (14:55 +0200)]
[2786] Implemented pack function of OptionString class.
Michal 'vorner' Vaner [Mon, 13 May 2013 12:52:42 +0000 (14:52 +0200)]
[2836] Clarify comment of test.
Michal 'vorner' Vaner [Mon, 13 May 2013 12:45:17 +0000 (14:45 +0200)]
[2836] The updater is never NULL in destructor
So remove the condition to check it. Keep an assert there, just in case
there was a reason originally, so we know about it.
Michal 'vorner' Vaner [Mon, 13 May 2013 12:42:03 +0000 (14:42 +0200)]
[2836] Less hacky way to access the zone data in tests
Store it separately as named address and access it every time, instead
of sneaking into the updater internal.
Marcin Siodelski [Mon, 13 May 2013 12:21:19 +0000 (14:21 +0200)]
[2786] Implemented unit tests for OptionString.
Michal 'vorner' Vaner [Mon, 13 May 2013 12:14:20 +0000 (14:14 +0200)]
[2836] Provide virtual destructor
Just to be sure
Michal 'vorner' Vaner [Mon, 13 May 2013 12:10:17 +0000 (14:10 +0200)]
[2836] Disable tests without shared memory
Disable tests using the shared memory utilities (eg. mapped segment)
when there's no shared memory support.
Michal 'vorner' Vaner [Mon, 13 May 2013 11:55:33 +0000 (13:55 +0200)]
[2836] Check for collision on the named address
Michal 'vorner' Vaner [Mon, 13 May 2013 11:19:54 +0000 (13:19 +0200)]
[2836] Store the address for whole lifetime
Store the named address for the whole lifetime of the zone data updater.
This should be better for performance.
Michal 'vorner' Vaner [Mon, 13 May 2013 11:05:19 +0000 (13:05 +0200)]
[2836] Clarify exception thrown from DomainTree::insert
Michal 'vorner' Vaner [Mon, 13 May 2013 11:00:01 +0000 (13:00 +0200)]
[2836] Retry addition to zone table
Currently untested, as there's no reasonable way to create a zone table
with mapped segment :-(.
Michal 'vorner' Vaner [Mon, 13 May 2013 10:56:31 +0000 (12:56 +0200)]
[2836] Note that a method may throw
Mukund Sivaraman [Mon, 13 May 2013 10:42:13 +0000 (16:12 +0530)]
[2850] Remove unnecessary setup of named addresses
... now that we handle this with an offset_ptr inside setNamedAddress()
itself.
Mukund Sivaraman [Mon, 13 May 2013 08:36:58 +0000 (14:06 +0530)]
[2850] Fix MemorySegmentMapped::allMemoryDeallocated() and make it non-const
Mukund Sivaraman [Mon, 13 May 2013 08:17:45 +0000 (13:47 +0530)]
[2850] Looks like flush() finalizes base_sgmt_, so we free before that
Mukund Sivaraman [Mon, 13 May 2013 07:59:05 +0000 (13:29 +0530)]
[2850] Pre-reserve some memory to workaround relocations in setNamedAddress()
This commit introduces a problem that allMemoryDeallocated() doesn't
work anymore, because the reserved memory is freed only when the segment
is destroyed.
One workaround is to temporarily release and re-reserve this memory in
allMemoryDeallocated(), but that would make it a non-const method. I
don't see any other clean way of doing this.
JINMEI Tatuya [Fri, 10 May 2013 23:52:06 +0000 (16:52 -0700)]
[2836-2] use managed_mapped_file to cause remap at different addr.
the original setup was not really portable and quite dangerous, making
a very large file.
JINMEI Tatuya [Fri, 10 May 2013 23:39:32 +0000 (16:39 -0700)]
[2836-2] make sure remapping boost segment even if grow() fails.
otherwise, subsequent cleanup to be exception safe will cause a crash.
and, if the remap fails simply abort(); there's no hope to recover from
this situation reasonably. and clarify in shrinkToFit() why it behaves
differently.
JINMEI Tatuya [Fri, 10 May 2013 23:21:37 +0000 (16:21 -0700)]
[2836-2] make sure flush()ing the base segment before grow/shrink
I was confused at the time I removed it in #2831. these were actually
necessary.
Jeremy C. Reed [Fri, 10 May 2013 17:26:44 +0000 (12:26 -0500)]
[master] update master version datestamp and fix date for 1.0.0beta2 release
Mukund Sivaraman [Fri, 10 May 2013 14:25:40 +0000 (19:55 +0530)]
[2850] Throw std::bad_alloc if ZoneTable::create() throws MemorySegmentGrown
We don't attempt to recover in this case.
Also change some error handling to simply asserts.
Mukund Sivaraman [Fri, 10 May 2013 14:18:07 +0000 (19:48 +0530)]
[2850] Make more updates to ZoneTableSegment API documentation
Mukund Sivaraman [Fri, 10 May 2013 13:58:58 +0000 (19:28 +0530)]
[2850] Make more updates to ZoneTableSegment API documentation
Mukund Sivaraman [Fri, 10 May 2013 13:48:33 +0000 (19:18 +0530)]
[2850] Update Makefile.am for renamed files
Mukund Sivaraman [Fri, 10 May 2013 13:46:27 +0000 (19:16 +0530)]
[2850] Update ZoneTableSegment API documentation
Mukund Sivaraman [Fri, 10 May 2013 12:52:46 +0000 (18:22 +0530)]
[2850] Add API doc comments about plans to make reset(), etc. non-virtual
Mukund Sivaraman [Fri, 10 May 2013 12:48:39 +0000 (18:18 +0530)]
[2850] Add explicit include for zone_table_segment.h
Mukund Sivaraman [Fri, 10 May 2013 12:39:41 +0000 (18:09 +0530)]
[2850] Add default case for switch on MemorySegmentOpenMode
Michal 'vorner' Vaner [Fri, 10 May 2013 11:37:12 +0000 (13:37 +0200)]
[2836] Make segment holder survive relocation
Let the SegmentObjectHolder store the object's address in the segment's
named addresses, so it moves correctly when the segment is relocated.
Mukund Sivaraman [Fri, 10 May 2013 09:52:21 +0000 (15:22 +0530)]
[2850] Add tests to check that resetHeader() is called by ZoneWriter class
Initially I thought it was not possible to test this as
ZoneTableSegment::create() was used to create the zone table segment,
but it seems it's possible to do it this way using a mock class.
Michal 'vorner' Vaner [Fri, 10 May 2013 08:23:06 +0000 (10:23 +0200)]
[2836] Test that segment holder survives relocation
Test that the SegmentObjectHolder survives when the segment relocates
itself and correctly releases all its memory.
Mukund Sivaraman [Fri, 10 May 2013 03:01:04 +0000 (08:31 +0530)]
[2850] Change style of code to fix buildbot failures
JINMEI Tatuya [Thu, 9 May 2013 22:32:31 +0000 (15:32 -0700)]
[2899] renamed namespace for InterprocessSync "interprocess" for consistency.
JINMEI Tatuya [Thu, 9 May 2013 20:20:30 +0000 (13:20 -0700)]
[2899] Merge branch 'trac2899' into trac2899-2
JINMEI Tatuya [Thu, 9 May 2013 20:21:05 +0000 (20:21 +0000)]
[2899] explicitly LDADD libb10-threads from log lock test
seems to be necessary for some OS. also explicitly include the relevant
header file.
JINMEI Tatuya [Thu, 9 May 2013 20:20:15 +0000 (13:20 -0700)]
[2899] Merge branch 'trac2823-regression' into trac2899-2
Tomek Mrugalski [Thu, 9 May 2013 15:12:18 +0000 (17:12 +0200)]
[master] Merge branch 'trac2898' (DHCPv6 relayed traffic)
Conflicts:
src/lib/dhcp/tests/pkt6_unittest.cc
Tomek Mrugalski [Thu, 9 May 2013 13:43:20 +0000 (15:43 +0200)]
Merge branch 'trac2898' of ssh://git.bind10.isc.org/var/bind10/git/bind10 into trac2898
Tomek Mrugalski [Thu, 9 May 2013 13:43:02 +0000 (15:43 +0200)]
[2898] comment updated in one dhcp6/config_parser test
Mukund Sivaraman [Thu, 9 May 2013 13:10:37 +0000 (18:40 +0530)]
[2850] Add isUsable() and getImplType() methods
It would be better if these were static (to save on space and also to
use them inside ZoneTableSegment::create()), but unfortunately we cannot
make it virtual then. Another option would be to pass a static string
from ZoneTableSegment::create() onto the ZoneTableSegment constructor
and make it a non-virtual base class method.
Mukund Sivaraman [Thu, 9 May 2013 12:38:27 +0000 (18:08 +0530)]
[2850] Rename MemorySegmentTest to MemorySegmentMock (to not resemble a fixture)
Mukund Sivaraman [Thu, 9 May 2013 12:28:31 +0000 (17:58 +0530)]
[2850] Rename ZoneTableSegmentTest to ZoneTableSegmentMock (to not resemble a fixture)
Mukund Sivaraman [Thu, 9 May 2013 12:14:49 +0000 (17:44 +0530)]
[2850] Destroy ZoneTableSegment object upon exception (see full log)
Ideally, this should use something like a SegmentObjectHolder, but a
SegmentObjectHolder takes unnecessary arguments. In this limited
usecase, ZoneTableSegment::destroy() just calls its destructor, so
std::auto_ptr should be ok here.
Stephen Morris [Thu, 9 May 2013 09:54:56 +0000 (10:54 +0100)]
[2909] Updated schema_copy.h to more accurately reflect the database schema
When the MySql lease manager tests create the database, they now also
create all the indexes as well.
Mukund Sivaraman [Thu, 9 May 2013 09:28:15 +0000 (14:58 +0530)]
[2850] Add CREATE over corrupted file testcase
Mukund Sivaraman [Thu, 9 May 2013 09:20:40 +0000 (14:50 +0530)]
[2850] Check that the old data is still available after the corruption tests
Mukund Sivaraman [Thu, 9 May 2013 08:52:27 +0000 (14:22 +0530)]
[2850] Remove obsolete comment (isWritable() no longer throws)
Mukund Sivaraman [Thu, 9 May 2013 08:50:51 +0000 (14:20 +0530)]
[2850] Add a test for clear() on an uninitialized segment
Mukund Sivaraman [Thu, 9 May 2013 08:10:44 +0000 (13:40 +0530)]
[2850] Add resetHeader() method
JINMEI Tatuya [Thu, 9 May 2013 06:05:14 +0000 (23:05 -0700)]
[master] exit from SyncUDPServer::handleRead immediately if stopped.
while not really correct, this seems to work as a workaround for the crash
problem reported in #2946.
okayed on jabber.
JINMEI Tatuya [Thu, 9 May 2013 05:08:27 +0000 (22:08 -0700)]
[master] don't initialize test data objects in namescope level.
this caused statici initialization fiasco for some build, e.g.:
http://git.bind10.isc.org/~tester/builder/BIND10/
20130508225127 -MacOSX10.6-x86_64-Clang-Static/logs/unittests.out
string objects moved inside the compare tests did not actually cause harm
in this environment, but should be generally better to be initialized
this way for the same reason.
while it's not a kind of 5-line patch, it should be pretty straightforward,
so I'm committing it at my discretion.
JINMEI Tatuya [Thu, 9 May 2013 03:21:58 +0000 (20:21 -0700)]
[2899] added a README file explaining the intent of the interproc stuff
JINMEI Tatuya [Thu, 9 May 2013 03:15:10 +0000 (20:15 -0700)]
[2899] hide the interprocess stuff from public log API more.
- not include the header file, but use a forward declaration
- added a note that setInterprocessSync shouldn't be used by normal apps
JINMEI Tatuya [Thu, 9 May 2013 00:16:42 +0000 (17:16 -0700)]
[2899] move interprocess-lock related things under lib/log/interprocess
also
- a utility for tests using multi processes was moved under lib/util/unittests
- other Makefile adjustments
Jeremy C. Reed [Wed, 8 May 2013 23:29:23 +0000 (18:29 -0500)]
[trac2945] miscellaneous doxygen warnings worked on
Still have lots to do.
Some of these marked with TODO indicating needing further work
(but better than it was before).
JINMEI Tatuya [Wed, 8 May 2013 23:03:14 +0000 (16:03 -0700)]
[2851] check return value of insertZone of mock data source, just in case.
JINMEI Tatuya [Wed, 8 May 2013 22:47:10 +0000 (15:47 -0700)]
[2851] added missing parameter description.
JINMEI Tatuya [Wed, 8 May 2013 22:45:40 +0000 (15:45 -0700)]
[2851] corrected an error code in documentation (a simple typo)
JINMEI Tatuya [Wed, 8 May 2013 22:41:49 +0000 (15:41 -0700)]
[2851] use "non writable" instead of read-only for a test case on this.
in some sense they mean the same thing, but technically there's a subtle
difference between them; "non writable" includes the case the cache (segment)
isn't usable at atll. calling it "read-only" is not really correct.
JINMEI Tatuya [Wed, 8 May 2013 22:39:19 +0000 (15:39 -0700)]
[2851] updated description of NOT_WRITABLE.
"yet" seemed to be confusing; while we could explain it more, it's probably
not much of interest for the caller anyway, and this is expected to be
a concise description. so I chose it rather simpler.
JINMEI Tatuya [Wed, 8 May 2013 22:37:47 +0000 (15:37 -0700)]
[2851] added a new test case for reloading a zone added to the datasrc later
JINMEI Tatuya [Wed, 8 May 2013 22:25:51 +0000 (15:25 -0700)]
[2851] updated comment wording for reloadZoneGone test.
it now doesn't throw an exception but reject the attempt with an error code.
JINMEI Tatuya [Wed, 8 May 2013 21:58:18 +0000 (14:58 -0700)]
[2836] style fixes: constify, folded a long line.
JINMEI Tatuya [Wed, 8 May 2013 21:26:32 +0000 (14:26 -0700)]
[2836] style fixes: add () for return
JINMEI Tatuya [Wed, 8 May 2013 20:57:53 +0000 (13:57 -0700)]
[2836] editorial cleanup: removed an unnecessary semicolon.
Stephen Morris [Wed, 8 May 2013 17:33:25 +0000 (18:33 +0100)]
[2898] Minor editing changes to DHCPv6 relay text
Paul Selkirk [Wed, 8 May 2013 15:37:31 +0000 (11:37 -0400)]
Merge branch 'trac2521'
Conflicts:
src/lib/dns/gen-rdatacode.py.in
Paul Selkirk [Wed, 8 May 2013 15:29:02 +0000 (11:29 -0400)]
[2522] fix a typo
Paul Selkirk [Wed, 8 May 2013 14:59:40 +0000 (10:59 -0400)]
[2522] initialize impl_ directly in TSIG lexer constructor
Mukund Sivaraman [Wed, 8 May 2013 12:06:45 +0000 (17:36 +0530)]
[2850] Include <cerrno> for errno
Mukund Sivaraman [Wed, 8 May 2013 12:05:50 +0000 (17:35 +0530)]
[2850] Simplify the corruption of checksum
Mukund Sivaraman [Wed, 8 May 2013 12:01:06 +0000 (17:31 +0530)]
[2850] Use pre-created test data so that it's not confusing to check
Mukund Sivaraman [Wed, 8 May 2013 11:38:53 +0000 (17:08 +0530)]
[2850] Remove unnecessary use of scoped_ptr
Mukund Sivaraman [Wed, 8 May 2013 11:32:16 +0000 (17:02 +0530)]
[2850] Handle current_filename_ assignment causing an exception
Mukund Sivaraman [Wed, 8 May 2013 11:24:30 +0000 (16:54 +0530)]
[2850] Make getHeader() return the current address of header
It could have been relocated since its construction.
Mukund Sivaraman [Wed, 8 May 2013 11:13:41 +0000 (16:43 +0530)]
[2850] Remove obsolete comment
Mukund Sivaraman [Wed, 8 May 2013 11:00:17 +0000 (16:30 +0530)]
[2850] Fix address name generation
Stephen Morris [Wed, 8 May 2013 10:54:45 +0000 (11:54 +0100)]
[2467] Merge branch 'master' into trac2467
Conflicts:
src/bin/dhcp4/tests/dhcp4_srv_unittest.cc
src/lib/dhcp/tests/iface_mgr_unittest.cc
Tomek Mrugalski [Wed, 8 May 2013 09:30:50 +0000 (11:30 +0200)]
[2898] Compilation fix after recent changes.
Tomek Mrugalski [Wed, 8 May 2013 09:22:40 +0000 (11:22 +0200)]
[2898] pair<a,b> replaced with make_pair
Tomek Mrugalski [Wed, 8 May 2013 09:21:59 +0000 (11:21 +0200)]
[2898] Second part of changes after review
JINMEI Tatuya [Wed, 8 May 2013 07:20:18 +0000 (00:20 -0700)]
[2823-regression] make test_polling_init indepent from ordering of "getstat"s
with multi-module setup, it can sometimes fail with python3.3 due to its
unpredictable hash values (and the internal ordering of dictionaries).
JINMEI Tatuya [Wed, 8 May 2013 07:18:54 +0000 (00:18 -0700)]
[2823-regression] mock mccs.check_command for MyStatsHttpd.
our tests basically shouldn't call this method, but, as commented, it could
still happen on some buggy OSes. we need to work around it.