Now these no longer exist, we can't use them and we shouldn't have been
anyway. This updates the test to not use GIO APIs anyway, it was only needed
to remove a file and it's not needed by the GVDB library so we don't want to
add the dep just to fix a unit test. Instead, we now use the unistd unlink()
API.

This API came in during GLib 2.40.

The tracker_info() is used now for the 'tracker' command's info command and it
make sense to remove the old make-shift tracker_info() API which was for
logging and duplicating g_info() anyway.

Used in tracker-stats.

https://bugzilla.gnome.org/show_bug.cgi?id=739149

The "all" target was used because the MAN page generation was not working
previously, but it seems it does. So this line has been removed!

https://bugzilla.gnome.org/show_bug.cgi?id=739260

Part of the solution is this commit, see:
https://bugzilla.gnome.org/show_bug.cgi?id=737663

The other part is to use cgroups

This means:

- all tracker_spawn*() API is removed, now using g_spawn*()
- the mplayer extractor has been removed (no one uses it anyway)
- the setrlimits() API has also been removed (only tracker-extract uses it)
- many includes have been cleaned up (this is a really old header)
- strnlen() definitions (for OS' without this API) are defined locally (only 2
  cases).

The Namespace has been cleaned up too, all APIs now start with:

  TRACKER_PREFIX_

or

  TRACKER_DATASOURCE_

The well known definition for the TrackerMinerFS graph has also been changed
to:

  TRACKER_OWN_GRAPH_URN

because it now applies to more than just the TrackerMinerFS, we're using it
in:

  tracker-writeback
  tracker-miner-apps
  tracker-miner-user-guides
  ...
  libtracker-data

It should probably be internal actually.

We actually do this directly in the miners ourselves now instead, the code is
quite small and it is cleaner too now. We store a file per miner in case one
isn't run for a period of time between locale updates.

General consensus was that we should remove this now, it's not useful and
GNOME at least expects a log out when changing the locale.

Completely unused

Now we've removed the internal key file object, we've had to put in place
another method for supporting the existing TRACKER_USE_CONFIG_FILES
environment variable. Thanks to the GKeyfileSettingsBackend provided by
GLib, we can fallback to old school INI type config files for embedded
solutions or cases where we don't want dconf as a backend. This works rather
well.

IT should be noted, the INI files are *NOT* written out in full if they do not
exist, only options which are saved or different to the default settings are.
This is how it should be too.

Now we build man pages based on GSettings schemas using xsltproc with the
template in docs/manpages/gsettings.xsl. This is a useful aid when trying to
understand what config files can have in them. One thing it does highlight, is
the config documentation could be better :)

Migrated the KeyFile work in most apps to GSettings

Includes:

gchar *  tracker_glong_to_string               (glong         i);
gchar *  tracker_gint_to_string                (gint          i);
gchar *  tracker_guint_to_string               (guint         i);
gchar *  tracker_gint32_to_string              (gint32        i);
gchar *  tracker_guint32_to_string             (guint32       i);
gboolean tracker_string_to_uint                (const gchar  *s,
                                                guint        *ret);
gchar *  tracker_string_list_to_string         (gchar       **strv,
                                                gsize         length,
                                                gchar         sep);
GList *  tracker_glist_copy_with_string_data   (GList        *list);
gchar *  tracker_string_boolean_to_string_gint (const gchar  *value);

It's only used in libtracker-data and libtracker-common isn't a public library

https://bugzilla.gnome.org/show_bug.cgi?id=737768

https://bugzilla.gnome.org/show_bug.cgi?id=738522

Remove some code which tries to set attributes of a file which no
longer exists.

This code was added in the following commit:

    commit 16fe51de
    Author: Philip Van Hoof <philip@codeminded.be>
    Date:   Wed Nov 9 12:56:59 2011 +0100

        tracker-writeback: Preserve file permissions

        Fixes NB#289953.

Seems that the intent was to preserve the permission bits of the
file after tracker-writeback rewrites it. The code already takes
care of this at the time the temporary file is created.

We are running against a clean data directory when run by 'test-runner.sh', so
the miner should not have any initial crawling work to do. Therefore the tests
don't need to wait for it to be idle before starting.

Test cases should use the await_resource_inserted() and
await_resource_deleted() methods as these are (in theory) reliable. Waiting for
the miner to go idle and hoping this means that the desired file was crawled
and extracted is not reliable at all.

This should hopefully make the tests robust enough to be run continuously
without random failures occuring.

The tracker_miner_fs_wait_for_idle() functions are gone forever!

Resources are created when the miner process sees them, but often we
want to wait until they have been processed by the extractor, too. We
can achieve that by waiting for the insertion of a specific triple that
we know is set by tracker-extract.

This required adding a new parameter to the await_resource_inserted()
function.

The 310-fts-indexing and the base writeback test have been fixed to make
use of this, instead of using tracker_miner_fs_wait_for_idle().

The tracker_miner_fs_wait_for_idle() function is not a good way to
detect when something has changed. We should always be listening
to GraphUpdated instead.

Thanks to Martin Kampas who did some of this already in commit
853eb94d.