[Digikam-users] Metadata

Remco Viëtor remco.vietor at wanadoo.fr
Sat Jun 29 09:51:03 BST 2013


On Saturday 29 June 2013 09:59:42 Anders Lund wrote:
> On Lørdag den 29. juni 2013 09:54:15, jdd wrote:
> > Le 29/06/2013 09:43, Anders Lund a écrit :
> > > That is what I sometimes do, for example by tagging a whole album 
with
> > > some
> > > tags.
> > > 
> > > What I thought of was something like an external process and some que
> > > management.

Well, if Digikam would run a large tag update as a background job, you 
might get in trouble if you do anything with one of the files concerned 
while the queue is running.

For instance: all images in an album need a particular tag (e.g. a 
location), and part of them need other tags. You start the queue for a 
location tag, and then for the other tags. It's very easy to get in a 
situation where a file gets read by the 1st job, then read and modified  by 
the second, then modified by the first  => loss of tags added in second 
queue. It's as easy to lose the tags from the first queue. 

Another situation: you start a queue, then decide to edit an image rather 
far ahead, so that the editor has opened the file when the tag queue has to 
modify the tags. How do you want to handle that?

There are ways to handle this kind of things, but they are not easy to 
implement 100% correct, and have an impact on speed.

Personally I prefer a conservative approach, even if it includes waiting a 
bit every now and then, over a method that seems faster, but has way more 
risk of data loss.

Remco



More information about the Digikam-users mailing list