Re: [Tracker] REVIEW: Xavier Claessens' "priority" branch
- From: Philip Van Hoof <philip codeminded be>
- To: Carlos Garnacho <carlosg gnome org>, tracker-list gnome org
- Cc: Xavier Claessens <xclaesse gmail com>
- Subject: Re: [Tracker] REVIEW: Xavier Claessens' "priority" branch
- Date: Wed, 19 Feb 2014 14:46:02 +0100
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Hi guys,
One general question I have is that if this is configuration per RDF
class (rdf:type), then why aren't we putting this in SPARQL by just
adding a tracker:priority boolean to the ontology of rdfs:Class?
So basically, just like tracker:notify.
And if not in the ontology, then why not in the .rule files?
Kind regards,
Philip
Carlos Garnacho schreef op 19/02/2014 2:00:
Hey,
I've been having a look at xclaesse's branch to get prioritization
of certain rdf:types over others at the time of extracting
metadata:
http://cgit.collabora.com/git/user/xclaesse/tracker.git/log/?h=priority
There, TrackerDecorator gains API to specify those high priority
rdf:types, which will affect how items are sorted/queued.
tracker-extract does expose this setting for its TrackerDecorator
implementation through a new SetRdfTypes DBus method, so running
apps can register these, callers are monitored so the rdf:types
return to regular priority if those disappear.
So from looking at the code, the implementation looks quite
correct, and it looks like a positive addition, I just have minor
improvements:
0. Nice cleanup to tracker-extract's tracker-config.c, long time
overdue :)
1. tracker_decorator_set_priority_rdf_types() is added, but there's
no matching GObject property, would be great to get one.
2. In tracker-decorator.c:607: /* FIXME: Can we know the
class_name_id ? */ element_add (decorator, subject, 0, FALSE);
Probably the right approach there is accumulating those, and
querying those subjects similarly to how
tracker_decorator_started() does.
3. In tracker-extract-controller.c:41: static void
files_miner_status_changed (TrackerExtractController *self, const
gchar *status)
If we just care about status equality with "Idle", I'd make this
function take a boolean, or wrap one taking a boolean, so we don't
need to make up status strings in some calls.
Also, that function does if (..) if (...), when those are mutually
exclusive, would seem clearer with an if (...) else if (...)
4. In tracker-extract-controller.c:190: self->priv->watch_id =
g_bus_watch_name_on_connection (conn,
"org.freedesktop.Tracker1.Miner.Files", ...)
The arguments there aren't indented properly
5. In tracker-extract-decorator.c:578: goto out; } out:
That goto feels a bit pointless :)
And that's all I could spot in that branch, nice work Xavier :)
Cheers, Carlos
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.20 (MingW32)
Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/
iQEcBAEBAgAGBQJTBLWaAAoJEEP2NSGEz4aDbzMH/j4Neb38iqEJsSf16xpyiH9E
egvXMCxYggwkanfHIEmj2uMPxULllJ7PXaCtPhxRIaUUVIEmhXeLmuoE2ft6LMgl
51xdZBDDYez7EFxKjDxfWpzBVHz0SiYtpsQZ3kxV1OURLfGmS6tJylLoE/5pSrNa
lZIZJFBnU/a9z7++jOgafpr+LSI0qFej1fsP/LRhkxTXz2nmNsccIw3qDUSlxPOp
Hm8MM1/haFU7bB/GkYCo+n8H47AAJL88/VRMvG/7fwn9sdRgJ1TBg8B/viWE1Q3Y
vAqbkJdWkqQIdmVKu3wHQerFENzHgbgqZEinAxr11wgaXM8qOKQ8ttMMrufIuG4=
=Xido
-----END PGP SIGNATURE-----
[
Date Prev][
Date Next] [
Thread Prev][
Thread Next]
[
Thread Index]
[
Date Index]
[
Author Index]