Re: [Tracker] indexing xml files content

[Martyn Russell <martyn lanedo com>]

On 07/01/11 09:00, Tania Farinella wrote:
> Hi all,

Hi there,

> I'm using tracker and it doesn't index xml files content. So when i look
> for a word using tracker-search-tool it doesn't retrieve xml files in
> which the word is present.
>
> can anyone help me?

We have a bug about this:

 https://bugzilla.gnome.org/show_bug.cgi?id=615857

It isn't straight forward to support this. XML is used in a variety of 
ways and knowing how to use the embedded data in tags would be unique to 
each extractor. For specific cases (like HTML) we know what tags to 
expect and how to use them. It shouldn't be too much work to take the 
HTML extractor we already have and extend it for what you want.

Doing an XML extractor generically is really not so useful IMO. I can 
think of a bunch of files which use XML in different ways and 
translating that into the Tracker ontology is quite hard past 
nie:plainTextContent and possibly nie:title. If we did a generic 
extractor, it would just be filled with concatenated tag text all in one 
text block for nie:plainTextContent. Unless anyone has any other ideas 
on how this could be done better?

Any comments are welcome in the bug report.

--
Regards,
Martyn