Re: Thoughts about FilterChm

["D Bera" <dbera web gmail com>]

Hi

> Well, I finally  have enough time to  start working on updating the Chm
> filter, but i need some advice from the Beagle hackers. :)
> Html filter is faster than before because it uses an event driver
> architecture rather that the DOM based approach.
> In theory that would allow to index all the Html files inside the Chm file
> (currently it only index index the topics file and the default page due
> processor use overhead). The code as is written inherits the behaviour of
> the Html filter, but rather than being a Html filter, Chm filter should use
> the html filter to index the text. the last time I saw the code of beagle
> this was not possible (I understood  that FSQ did not support Child
> Indexables [1][2]) .
> I still can't use an Filter from another Filter?
> On the other hand due to modifications made to source code of FilterHtml, I
> will have to copy (and modify some parts) of the FilterHtml, (then cutting
> the dependency) but that would be code repeating,  I don't like it :(  .

Do you really need "Child-indexables" ? As you noted, HTML filter is
now event-driven and all the logic is inside the HandleNodeEvent()
method. I think you should be able to derive from FilterHtml, set up
an event handler; on any event, if you need some regular html
processing call HandleNodeEvent() otherwise handle it inside your
custom event handler.
Easier said than done and I am saying this with no clue how the
chmfilter works :).

Let me know if you need some help.
- dBera

--
-----------------------------------------------------
Debajyoti Bera @ http://dbera.blogspot.com
beagle / KDE fan
Mandriva / Inspiron-1100 user