JINMEI Tatuya [Wed, 26 Oct 2011 23:07:01 +0000 (16:07 -0700)]
[1028] removed self reference in XfrinConnection, which apparently caused
reference leak. To test this case, introduced a framework for testing
process_xfrin(). This setup will be used for subseuent related fixes.
JINMEI Tatuya [Wed, 26 Oct 2011 21:38:59 +0000 (14:38 -0700)]
[1028] cleanup: use a more appropriate variable name to specify s_ZoneUpdater.
also fixed another possible (but quite unlikely) leak in case tp_alloc() fails.
JINMEI Tatuya [Wed, 26 Oct 2011 05:42:12 +0000 (22:42 -0700)]
[1028] fiexed leak in RRset.get_rdata() as well as other minor bugs
like missing NULL check or exception safety. also as a cleanup got rid
of unnecessary reinterpret_cast.
JINMEI Tatuya [Sat, 22 Oct 2011 05:21:53 +0000 (05:21 +0000)]
[master] use reverse_iterator instead of const_reverse_iterator in a for loop.
this is a workaround for build problem for the solaris buildbot (both g++ and
sunstudio); they don't seem to work well with the const iterator in operator!=.
JINMEI Tatuya [Fri, 21 Oct 2011 23:24:02 +0000 (16:24 -0700)]
[869] install auth_xfrout_conn under package subdirectory not directly under
localstatedir. (e.g.) /usr/local/var/bind10-devel instead of /usr/local/var.
also updated the EOL code of auth/spec_config.h.pre.in from CR+LF (DOS style)
to LF (unix style), mainly for consistency.
Tomek Mrugalski [Fri, 21 Oct 2011 11:47:14 +0000 (13:47 +0200)]
[1224] Pkt4 improvements after review
- dhcp4.h added to Makefile.am
- Pkt4.cc: address fields initialized in more efficient way
- Pkt4.h: many improvements in comments
- Minor improvements in Pkt4 unittests
JINMEI Tatuya [Fri, 21 Oct 2011 06:31:17 +0000 (23:31 -0700)]
[1307] final (unrelated) cleanups:
- renamed private methods to addXXX for consistency
- made them non const member functions (for the same reason why
addNXDOMAINProof can't be so)
- renamed variable 'zone' to 'finder' to adapt to class name change
JINMEI Tatuya [Thu, 20 Oct 2011 23:33:44 +0000 (16:33 -0700)]
[1307] (unrelated) clarification cleanup: process()/putSOA()/addNXDOMAINProof()
cannot be const because they modify response_.
maybe we should pass the response via method parameter rather than at
the construction time, or even make this a simple free function, rather
than a class. but that will be a separate discussion.
JINMEI Tatuya [Tue, 18 Oct 2011 22:45:20 +0000 (15:45 -0700)]
[1307] updated the RRsetMatch logic so that we can differentiate RRSIGs
of different types covered. Not directly related to the main subject
of the branch, but this enhancement will make writing tests easier.
confirmed this change doesn't cause regression for other tests.
JINMEI Tatuya [Mon, 17 Oct 2011 07:44:36 +0000 (00:44 -0700)]
[1305] added new ZoneFinder definitions to the python wrapper and some simple
tests. updated pydoc string (and fixed some unrelated editorial problems
in the original version)
JINMEI Tatuya [Sat, 15 Oct 2011 00:02:11 +0000 (17:02 -0700)]
[1305] clarified the behavior of various variants of NXRRSET.
originally I intended to introduce a new result code, WILDCARD_EMPTY, but
on furhter thought and look it seems we don't need it. so this commit
is to just clarify the doxygen comments more.
Tomek Mrugalski [Wed, 12 Oct 2011 12:43:37 +0000 (14:43 +0200)]
[1186] Part 6 of review changes.
- Added V6ADDRESS_LEN and V4ADDRESS_LEN
- Removed unnecessary parameters to Option constructors
- Many missing Doxygen comments added
- using sizeof(uintXX_t) instead of 2 and 4 in pack()/unpack()
- Many other smaller changes.
Tomek Mrugalski [Wed, 12 Oct 2011 09:35:28 +0000 (11:35 +0200)]
[1186] Part 5 of the review changes
- Option6Lst is now Option6Collection
- Corrected data_len_ mistake in Option::pack4
- Added comments about length calculation
- Added doxygen comments
- Some const and references added
- Many other minor changes.