Re: [Tracker] REVIEW: Xavier Claessens' "priority" branch

[Philip Van Hoof <philip codeminded be>]

-----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-----