Perform a substantial cleanup of the build system with respect to
building and installing testcases.

By default, tests are no longer compiled as part of 'make'.  They will
be built when 'make check' is run.  The old behaviour can be obtained
with --enable-always-build-tests.

--disable-modular-tests is gone (because tests are no longer built by
default).  There is no longer any way to cause 'make check' to be a
no-op, but that's not very useful anyway.

A new glibtests.m4 file is introduced.  Along with a substantially
expanded Makefile.decl, this provides for consistent handling of
--enable-installed-tests and --enable-always-build-tests (mentioned
above).

Port our various test-installing Makefiles to the new framework.  This
is a net reduction of about ~350 lines of Makefile and configure.ac.

This patch substantially improves the situation in the toplevel tests/
directory.  Things are now somewhat under control there.  There were
some tests being built...

After this patch, there is but one remaining use of g_thread_init(),
which is in tests/slice-threadinit.c, a testcase dedicated to testing
the functionality of gslice across a g_thread_init() boundary.

This testcase is pretty meaningless these days... probably we should
delete it.

This should be the last users that need to be ported.

For some of the oldschool non-gtester-ified tests, we call g_test_init()
from main() because it is necessary in order to use
g_test_build_filename().

testgobject.c and timeloop-closure.c are the only two tests in the
toplevel tests/ directory that depend on gobject, so move them to
tests/gobject/ along with the other gobject tests.

Both of these were in noinst_PROGRAMS and not TESTS, so keep them that
way when we move them.

This is the old non-xml schema file format that we were playing around
with at one time.  Nothing is using this file anymore.

Deserialising an emblemed icon would make uninitialised use of a
variable in the error case.  Fix that.

Mark 'test', 'test-report', 'perf-report' and 'full-report' as PHONY in
docs/Makefile.am to prevent recursion of gtester into the documentation
subdirectories.  Stop including Makefile.decl from these directories
since it is no longer necessary.

This will clear up the warnings about EXTRA_DIST being defined once in
gtk-doc.make and again in Makefile.decl.

GObject has previously allowed installing properties after class_init
has finished running.  This means that you could install some of your
own properties on G_TYPE_OBJECT, for example, although they wouldn't
have worked properly.

A previous patch asserted that this was not true and we had to revert it
because it broke the shell.  Instead of reverting, we should have used a
critical, so do that now.

Complaints go to this bug:

https://bugzilla.gnome.org/show_bug.cgi?id=698614

Since this feature is so utterly automake-centric, we may as well be
using the same terminology as automake itself (ie: although it's
BUILT_SOURCES, it's DIST_EXTRA, not DISTED).

Also add some comments to the enum explaining that these terms are
really corresponding directly to the automake terms.

https://bugzilla.gnome.org/show_bug.cgi?id=549783

This is no longer needed with the new test data file finding stuff.

https://bugzilla.gnome.org/show_bug.cgi?id=549783

These ones were slightly non-trivial so they didn't get included in the
previous patches.  Port them now.

https://bugzilla.gnome.org/show_bug.cgi?id=549783

This API was introduced to save a few lines of code here and there, so
let's start by removing a bunch from our own tests.

https://bugzilla.gnome.org/show_bug.cgi?id=549783

Add a pair of functions for returning strings that don't need to be
freed.  This is a bit of a hack but it will turn the 99% case of using
these functions from:

  gchar *tmp;
  tmp = g_test_build_filename (...);
  fd = open (tmp, ...);
  g_free (tmp);

to:

  fd = open (g_test_get_filename (...), ...);

which is a pretty substantial win.

https://bugzilla.gnome.org/show_bug.cgi?id=549783

Port most of the tests to the new g_test_build_filename() API.

https://bugzilla.gnome.org/show_bug.cgi?id=549783

This function allows testcases to find data files in various situations
of srcdir == builddir, srcdir != builddir and for installed tests.

https://bugzilla.gnome.org/show_bug.cgi?id=549783

Lines:          6631    8862    74.8 %
Functions:      747     893     83.7 %

Both g_[file|bytes]_icon_load() leave the `type' out parameter
untouched, while the async methods g_[file|bytes]_icon_load_finish()
always set it to NULL.

For consistency's sake NULLify it in the sync methods too.

https://bugzilla.gnome.org/show_bug.cgi?id=700725

"Install" the newly-introduced gio/gbytesicon.h...

This broke when the tests were converted to be installable.
My apologies.

The target file is a script, so it'll always be in SRCDIR for the uninstalled
case.  Just look there and avoid the libtool trickery.

Update G_LOG_DOMAIN to be "GLib-GObject" so that we are consistent with
the autotools builds, and that tests expecting the log domain to be
"GLib-GObject" would not fail.

Define the G_LOG_DOMAIN of the GLib DLL as "GLib", because:
-This makes it consistent with the autotools builds
-Some tests expect the log domain to be "GLib"

Make the entries of config.h.win32(.in) consistent with the entries
that are generated from the autotools build (config.h.in).

When the interface name is invalid, we don't get an error
back from g_dbus_connection_call_sync.

We actually need the first dependency because it includes the
final executable name.  Rather, fix the original bug by using
the variable $(EXEEXT).