Introduce a GTK_DEBUG_CHECK() macro and use it to check for
GTK_DEBUG flags everywhere. Also guard all such places by

The logic here is that G_ENABLE_DEBUG is for compiling out
debug spew that can be triggered at runtime with the GTK_DEBUG
environment variable, while G_ENABLE_CONSISTENCY_CHECKS is for
consistency checks that are applied unconditionally.

Use G_ENABLE_CONSISTENCY checks for the recursion checks
in gtksizerequest.c

Use a separate G_ENABLE_CONSISTENCY_CHECKS define to guard internal
consistency checks that are applied unconditionally if they are enabled,
such as the widget invariants checking. Interactive debug spew that can
be triggered at runtime with the GTK_DEBUG environment variable is still
guarded by the G_ENABLE_DEBUG define.

The mapping from enable-debug levels to defines is as follows:

yes:     G_ENABLE_DEBUG G_ENABLE_CONSISTENCY_CHECKS
minimum: G_ENABLE_DEBUG G_DISABLE_CAST_CHECKS
no:      G_DISABLE_CAST_CHECKS G_DISABLE_ASSERT G_DISABLE_CHECKS

Rewrite the one check that the G_DISABLE_CHECKS macro was used for
in terms of g_return_if_fail.

It has been ifzeroed for many, many years.

We've already gotten the value type, use it.

g_value_transform already returns whether it was successful
or not, so no need to call g_value_type_transformable beforehand.

which fixes firefox scrollbar not being fittsy when maximixed.

https://bugzilla.gnome.org/show_bug.cgi?id=754551

For the same reasons as g_object_child_notify_by_pspec.

For the same reasons as g_object_child_notify_by_pspec.

Add gtk_container_install_child_properties and
gtk_container_child_notify_by_pspec to mirror te corresponding
GObject APIs.

It turns out that we always have the pspec at hand already.

We can do this without calling strlen() or allocating memory.

Move _gtk_builder_boolean_from_string in the same source file
as the other _from_functions.

Keep the GString in PropertyInfo around, instead of throwing it
away and just keeping the char*.

Add a getter for state flags, and use it where appropriate.

We already have the GParamSpec in hands, lets just use it instead
of looking it up again.

We have an api to create a quark from a static string, lets use it.

That way we disable typechecks done by the compiler and won't notice
when we pass the wrong values to the macro. And that's bad.

Creating the transient node will already copy the widget type. So
copying again will not do anything.

The speed-up in 7da1f8a1 was wrong in
certain conditions, even though it didn't trigger the existing
testsuite.

New testcase /bitmask/invert_range_hardcoded included.