Re: [Tracker] Tomboy note extractor

[Jamie McCracken <jamiemcc blueyonder co uk>]

Mikkel Kamstrup Erlandsen wrote:
> 2006/11/18, Jamie McCracken <jamiemcc blueyonder co uk 
> <mailto:jamiemcc blueyonder co uk>>:
>
>     Mikkel Kamstrup Erlandsen wrote:
>      > 2006/11/17, Jamie McCracken <jamiemcc blueyonder co uk
>     <mailto:jamiemcc blueyonder co uk>
>      > <mailto:jamiemcc blueyonder co uk
>     <mailto:jamiemcc blueyonder co uk>>>:
>      >
>      >     Mikkel Kamstrup Erlandsen wrote:
>      >      > 2006/11/16, Jamie McCracken <jamiemcc blueyonder co uk
>     <mailto:jamiemcc blueyonder co uk>
>      >     <mailto: jamiemcc blueyonder co uk
>     <mailto:jamiemcc blueyonder co uk>>
>      >      > <mailto:jamiemcc blueyonder co uk
>     <mailto:jamiemcc blueyonder co uk>
>      >     <mailto:jamiemcc blueyonder co uk
>     <mailto:jamiemcc blueyonder co uk>>>>:
>      >      >
>      >      >     Mikkel Kamstrup Erlandsen wrote:
>      >      >      > I created a small tomboy note extractor
>     (attached)... It
>      >     should
>      >      >     run on
>      >      >      > any system with a Python installation - I think the XML
>      >     parser I
>      >      >     use is
>      >      >      > bundled with Python.
>      >      >
>      >      >
>      >      >     hmm would prefer it internal to tracker in C to be honest.
>      >     The external
>      >      >     metadata extractors are driven by mime type so it would be
>      >     tricky/hacky
>      >      >     to do them via that mechanism (this is the same as our
>     email
>      >     stuff -
>      >      >     its
>      >      >     all internal)
>      >      >
>      >      >     It should be real easy to do this in C if anyone's willing
>      >      >
>      >      >
>      >      > I could do that - I digged around the extraction code and
>     it seemed
>      >      > fairly intuitive. Should I use libxml or just plain hardcoded
>      >     grep-like
>      >      > hacks (or libXYZ)?
>      >
>      >     use the glib xml parser (see abiword extractor)?
>      >
>      >
>      > Done. The abiword extractor uses custom "grepping" through the xml
>      > without any real parsing going on. I could probably do that too
>     (and it
>      > might be simpler) but I felt the urge to get into this whole glib
>      > business :-)
>      >
>      > I haven't tested it with tracker though, I didn't know how to get
>      > trackerd to pick up my notes files (they have mime type
>     text/plain and
>      > are in ~/.tomboy which isn't indexed by default).
>      >
>      > There is a small outcommented test main-function so you can
>     compile the
>      > extractor as stand alone.
>      >
>      > Please don't publicly humiliate me - I'm really not a C-hacker...
>     :-D
>      >
>      > Cheers,
>      > Mikkel
>      >
>
>     its mostly okay - just needs some tidying up of the code (spaces in
>     between function name and opening bracket, no underscores for private
>     functions etc)
>
>
> I think I have done this now... There was also some jumbled up space/tab 
> issues making the code look all weird, I hope that is fixed now too. I 
> also added a GPL header.
>
> There are missing GPL headers in many if the other extractors too. If 
> you wish I can add those (on which I can dig out the original author).
>
>     will need to integrate it in trackerd along with code to pick up new or
>     changed tomboy notes - will do this for you after the 0.5.2 release
>     tomorrow (with some new dbus methods to make your life easier!)
>
>
> Rock on!
>
> Another thing; when I run my test prog through valgrind it reports a few 
> leaks. I really can't find any in my code - does this perhaps come from 
> glib messing up with valgrind?

unlikely - you are not freeing up the hashtable in your test code but I 
will sort it out soonish.


--
Mr Jamie McCracken
http://jamiemcc.livejournal.com/