This patch exports a few utility functions and adds testing of shutdown
commands of the guest agent.

Add a basic test framework with two simple tests to test guest agent
interaction.

This patch adds helpers that allow to check for argument values in
commands sent to the monitor.

Use the JSON error messages to report errors back to the caller in
addition to erroring out. The error reported from the event loop from
the mock function of the monitor was later overwritten by the call to
the monitor/agent interaction API. This will also allow testing of error
reporting.

Add helper functions to open guest agent connections and a handler for
replying to the "guest-sync" command.

The normal monitor uses windows line endings, where the agent monitor
uses only newlines. Change this to tolerate both approaches and allow to
use the utilities for guest agent tests.

Refactor the test helpers to allow adding callbacks to verify the
monitor responses instead of simple command name checking and clean up
various parts to prepare for adding guest agent tests.

The instrumentation for the monitor test can be hacked for qemu agent
testing. Split out the monitor specific stuff to allow using the code in
guest agent tests in the future.

The qemumonitorjsontest crashed when one of the initialization steps
done before starting the worker thread failed. This patch fixes this by
trying to pthread_join() the thread only after it was created.

Simplify the code using the existing helpers instead of open coding the
same functionality.

Most APIs in libvirt report errors, thus no need to state that
explicitly.

To allow testing of the cpu updater function, this function needs to be
available separately. Export it from qemu_agent.c where it should
belong.

Although this isn't apparently needed for the guest agent itself, the
test I will be adding later depends on the newline as a separator of
messages to process.

parentIndex is used in virNetworkDefUpdateIPDHCPRange

this patch introduce the console api in libxl driver for both pv and
hvm guest.  and import and update the libxlMakeChrdevStr function
which was deleted in commit dfa1e1dd.
Signed-off-by: NBamvor Jian Zhang <bjzhang@suse.com>

This function is needed for virt-login-shell.  Also modify virGirUserDirectory
to use the new function, to simplify the code.
Signed-off-by: NEric Blake <eblake@redhat.com>

* src/openvz/openvz_driver.c: Use correct spelling.
* src/vmware/vmware_conf.c: Likewise.
* src/vmware/vmware_conf.h: Likewise.
* src/vmware/vmware_driver.c: Likewise.
Signed-off-by: NEric Blake <eblake@redhat.com>

%build section should first define all required macros and then run
commands. Interleaving them makes it harder to spot what commands are
run.

As RHEL provides a stable tool chain, we don't have to worry about
frequent changes in reported compiler warnings (which prevents us from
enabling -Werror unconditionally).

https://bugzilla.redhat.com/show_bug.cgi?id=905513

Libssh2 isn't reliable enough to support the libvirt transport using it.
The problems include mishandling of "known_hosts" files that may confuse
users.

This fixes vircgrouptest when run in a sandbox which already overrides
open() and others.

If libapparmor-devel happens to be installed when building the
RPM, it will failed due to unlisted virt-aa-helper in %files.
Add support for apparmor in the spec, so that we can explicitly
turn it on/off, defaulting to off in all distros. This causes
--without-apparmor to be given to configure, preventing the
build failures if the user happens to have libapparmor-devel
present.
Signed-off-by: NDaniel P. Berrange <berrange@redhat.com>

The python/tests directory contains a number of so called
"tests" for the python API. These are all hardcoded to
look for Xen and cannot be run in any automated fashion,
and no one is ever manually running them. Given that they
don't meaningully contribute to the test coverage, delete
them.

For some reason these tests were also copied into the
filesystem as part of 'make install'. The change to the
RPM in commit 3347a420
caused a build failure, since it removed the code which
deleted these installed tests.
Signed-off-by: NDaniel P. Berrange <berrange@redhat.com>

Signed-off-by: NEric Blake <eblake@redhat.com>

* configure.ac docs/news.html.in: update for the release
* po/*.po*: update localizations and regenerate

Commit 93ec384b was tested on mingw, but broke the build on Linux:

  CCLD     shunloadtest
shunloadtest.o: In function `main':
/home/eblake/libvirt/tests/shunloadtest.c:106: undefined reference to `virFilePrintf'
...
ssh.o: In function `main':
/home/eblake/libvirt/tests/ssh.c:43: undefined reference to `virFilePrintf'
/home/eblake/libvirt/tests/ssh.c:49: undefined reference to `virFilePrintf'

* tests/testutils.h (fprintf): Provide escape hatch.
* tests/shunloadtest.c: Use it.
* tests/ssh.c: Likewise.
Signed-off-by: NEric Blake <eblake@redhat.com>

Mingw *printf is a moving target; newer mingw now provides a version
of asprintf() that fails to understand %lld:

  CC       event_test-event-test.o
../../../../examples/domain-events/events-c/event-test.c: In function 'myDomainEventRTCChangeCallback':
../../../../examples/domain-events/events-c/event-test.c:270:18: error: unknown conversion type character 'l' in format [-Werror=format=]
                  virDomainGetID(dom), offset) < 0)
                  ^

But since our examples already admitted that they were hacking around
a mingw deficiency, it is easier to just use printf() directly, coupled
with <inttypes.h> macros, for a more portable work-around.

* examples/domain-events/events-c/event-test.c
(myDomainEventRTCChangeCallback): Use PRIdMAX instead of asprintf.
Signed-off-by: NEric Blake <eblake@redhat.com>

Commit a2619962 introduced virFilePrintf to work around the fact
that gnulib doesn't (yet) provide guarantees about fprintf() vs.
%z, which in turn causes all sorts of mingw compilation errors:

../../tests/testutils.c: In function 'virtTestResult':
../../tests/testutils.c:101:9: error: unknown conversion type character 'z' in format [-Werror=format=]
         fprintf(stderr, "%3zu) %-60s ", testCounter, name);
         ^

Rather than s/fprintf/virFilePrintf/ (and reformatting loads of
lines) across multiple files, it's easier to just hack the entire
testsuite to take advantage of our helper function.

* tests/testutils.c: s/fprintf/virFilePrintf/ for mingw.
Signed-off-by: NEric Blake <eblake@redhat.com>

A cross-compile to mingw failed:

  CC       virsystemdmock_la-virsystemdmock.lo
../../tests/virsystemdmock.c:29:6: error: 'dbus_connection_set_change_sigpipe' redeclared without dllimport attribute: previous dllimport ignored [-Werror=attributes]
 void dbus_connection_set_change_sigpipe(dbus_bool_t will_modify_sigpipe ATTRIBUTE_UNUSED)
      ^

But when you think about it, systemd is Linux-only, and even our
use of LD_PRELOAD to provide mock syscalls is Linux-only.

* tests/virsystemdmock.c: Avoid compilation outside Linux.
Signed-off-by: NEric Blake <eblake@redhat.com>

The VIR_DOMAIN_PAUSED_GUEST_PANICKED constant is badly named,
leaking the QEMU event name. Elsewhere in the API we use
'CRASHED' rather than 'PANICKED', and the addition of 'GUEST'
is redundant since all events are guest related.

Thus rename it to VIR_DOMAIN_PAUSED_CRASHED, which matches
with VIR_DOMAIN_RUNNING_CRASHED and VIR_DOMAIN_EVENT_CRASHED.

It was added in commit 14e7e0ae
which post-dates v1.1.0, so is safe to rename before 1.1.1
Signed-off-by: NDaniel P. Berrange <berrange@redhat.com>

The VIR_DOMAIN_SHUTDOWN_CRASHED state constant does not appear
to be used in the QEMU code anyway. It also doesn't make much
(any) sense, since the 'shutdown' state is a transient state
between 'running' and 'shutoff' and when a guest crashes, it
does not end up in a 'shutdown' state, only 'shutoff'.

It was added in commit 14e7e0ae
which post-dates v1.1.0, so is safe to remove before 1.1.1
Signed-off-by: NDaniel P. Berrange <berrange@redhat.com>

The way we were casting small (<32bit) integers was broken
on big endian hosts, causing stack smashing. This was detected
in the test suite either by test failures due to incorrect
results, or by libc/gcc abort'ing with its stack canary
triggered.
Signed-off-by: NDaniel P. Berrange <berrange@redhat.com>

Current automake enables parallel test by default, which means test
details are only logged in test-suite.log and not printed to stderr.
This patch makes test failures directly visible in RPM build logs even
when parallel tests are turned on.