[Nepomuk] Bugs, Patches and Scripts. - StrigiService & KInotify

[Arie Peterson]

On Thu, 6 May 2010 16:45:19 +0530, Vishesh Handa <handa.vish at gmail.com>
wrote:
> I think it's kinda intuitive that if a folder is being indexed and one
of
> it's subfolders, which is also being indexed, is moved somewhere. The
> subfolder should still be indexed. Doesn't that seem normal?

If the subfolder was explicitly marked by the user as "to be indexed",
then yes, I agree.

If, on the other hand, it was being indexed only because it was a
subfolder of an indexed folder, then I'm not so sure. The only reason the
subfolder was indexed in the first place, was its location. If its location
changes, why still index it?

Well, to be honest, there are of course situations where this behaviour is
useful. But I don't think it is wanted in all situations.

For example, it means that if you accidentally place a folder 'S' in an
indexed folder, it becomes "tagged"; if you want to place 'S' somewhere
else, so it won't be indexed, you either have to
  1) move all its files to a new dir, which is not "tagged"; if 'S' has
subfolders, then it's not enough to simply move over all its contents,
because these subdirs will also be "tagged" — you'll have to recreate the
whole directory structure; or
  2) use some explicit interface to nepomuk/strigi to convince them to
"untag" 'S' (remove it from the indexed folders list).

I guess the preferred option would be 2).

> Just to be clear, this patch would only add directories to the indexed
> folders list if they were previously being indexed. It does not track
> indexed files.

OK. I guess this asymmetry between files and folders makes it a bit
unintuitive for me.


Regards,

Arie