Re: [Tracker] initial indexing doesn't finish

[Martyn Russell <martyn imendio com>]

Tshepang Lekhonkhobe wrote:
> On Wed, Feb 11, 2009 at 12:18 PM, Martyn Russell <martyn imendio com> wrote:
> > Tshepang Lekhonkhobe wrote:
> > > Hi,
> > >
> > > That problem that Biebl once pointed out still occurs, did anyone
> > > notice...
> > WORKSFORME.
> >
> > What does tracker-status say?
> > What tells you that it didn't finish?
>
> After initial indexing run, the indexer and daemon tell me that
> indexing is finito and when I restart I find that more items are
> indexed. I index the whole /usr and /home by the way. Maybe your test
> dataset is far smaller.

OK. Tracker does crawl everything and check all content is up-to-date in 
the database *every time* it is started. If we didn't do this, between 
restarts some file could change and we wouldn't have an up-to-date 
database. Perhaps the terminology and communication could be better 
here. We don't actually "index" everything the consequent times, but 
rather "check" it is there and up to date. The applet could be improved 
here I suppose. Perhaps if we are not initially indexing, we should say 
"Checking" or something. But basically there is little difference 
between indexing and checking. Both require some database interaction. 
One does a database lookup the other inserts and extracts metadata. The 
indexer and daemon don't change states depending on if they are looking 
up a file in the database (for up-to-dateness) or indexing it because it 
is new. Changing state for that would be quite pointless because it 
*could* switch for every other file then.

We know if we are on an initial index. Perhaps we just need to change 
the text on subsequent cases to "Verifying" or "Checking"?

Any suggestions?

> > Can you provide a bit more information please.
>
> I must have been annoying that you use CAPS on me...

No, it is *tongue in cheek* WORKSFORME - it is a Bugzilla resolution :)

--
Regards,
Martyn