Provide a "message undeliverable" error by the msgq daemon in case the
sender declares it expects an answer and there's no recipient to send it
to.
Conflicts:
src/bin/bindctl/run_bindctl.sh.in
src/bin/sysinfo/run_sysinfo.sh.in
src/lib/python/isc/cc/Makefile.am
src/lib/python/isc/cc/session.py
Do a hack to make the tests run with separate build and source
directories. The generated file ended up in a different directory than
the rest of the modules and the __init__.py. It therefore didn't work.
We use the same trick as log_messages, a forwarder module and addition
to PYTHON_PATH to find it.
Only the real file is installed.
without [] the output it seems possible that the output is empty for some
environments. also clarified that we can ignore the cases where -L isn't
contained at all.
We don't take any benefit from the parallel harness, since we have only
one test binary in each directory, and automake can't parallelize that.
But it was hiding test output, which was annoying.
Acked on the mailing list.
- renamed 'bind10' to 'b10-init'
- renamed related files as well
- left alone the 'run' script, as well as the directory name
- added new 'bind10' script that execs b10-init
note this commit only handles file names, and internal references have not been updated yet
This is for ticket #2501.
This is trivial and I did without review. I tested on two
systems (one with python3.3 and one without).
I didn't decide to make it open-ended at this time,
just in case we need to fix python3.4 issues first.
I decided not to add a changelog entry since was so trivial.
(Also had a workaround before by using --with-pythonpath.)
this is necessary to keep the previous behavior for FreeBSD where -Werror
is automatically disabled due to other bug of the compiler.
discussed on jabber, not seen a clear go, but I'm committing it at my
discretion to fix the buildbot errors. I've confirmed the behavior on
multiple FreeBSD boxes.
in order to keep the main configure.ac more concise, and so that we can
eventually introduce more fine-grained configuration (e.g., skipping some
checks for DNS/DHCP-only build)
At the moment, this is just a cleanup, and doesn't change the behavior itself.
the previously hardcoded zone data are mostly moved to two new files:
query_testzone_data.txt and query_testzone_data_nsec3.txt.
A new helper python script, gen-query-testdata.py, converts them into
DNS zone files and C++ files defining some C-string variables for the RRs.
The latter will be included from query_unittest.cc (instead of the hardcoded
variables).
Generated zone files are loaded via include-only zone files: example.zone
and example-nsec3.zone.
For the next step corresponding sqlite3 DB file is also generated.
Do the initialization in tests too (the executable script initializes it
only when run separately, to know the correct parameters).
Also, remove deprecated and unused auxiliary script that doesn't and
didn't work.
test parameters were adjusted accordingly.
there are some non trivial adjustments needed for the 'correct' test
cases for the original loadzone:
- completing the origin for some RDATA paramaeters (NS, SOA) does not
work yet until we complete the RDATA support. At the moment
I made them FQDNs with comments
- a few TXT data were actually incorrect in the original tests, which
caused a seeming regression. I fixed the test data.
- there was one real bug in the $INCLUDE + origin support. I'll go
fix it; right now it fails
The 'error' test cases for the original loadzone also fail, but overall
the intended behavior looked preserved. Fixing the tests to make it pass
seems to be quite difficult (because log output are different, and
the new loadzone ng is more verbose), so I plan to simply remove these
tests.
- implemented a base class Counter and a concrete XfroutCounter as an external
module under isc.statistics. Because it is easy to implement another
concrete Counter class for other module in future. The caller module can
statically import it. The new counter class provides a getter method for
each statistics item. It is intended for making loose relationship between
the counter class and the caller module.
- added implementation of unixsocket counter into the existing UnixSockServer
Class.
- added new tests for checking counters implemented in UnixSockServer are
working properly into xfrout_test.py.
- implemented enabling/disabling counting in the new counter class.
