Hi,

the goal of this email is to set the directions for the next release of Kat.
I will try to answer recurring questions.

1. Current status of development

Currently we have version 0.6.x in trunk and 0.7.x (Praveen's branch) in 
branches. I have come to this because Praveen has been unavailable for long 
time and has not been able to communicate anything about his future 
availability.
This means that we will continue developing the 0.6.x series for a while.
I hope he will come back someday, but Kat's development has to continue 
anyway.
Praveen started refactoring the API, documenting every class and method. This 
has to be continued. I will study his code and try to port it to the 0.6.x 
series. Any help with that will be really appreciated.

2. New API and new daemon

Our major efforts have to be directed to the refactoring of the Kat API and 
Kat daemon. The new GUI can wait.
We receive lots of bug reports about kded crashes and memory leaks.
It is important to notice one simple thing: the index is built upon a loop.
Loops, as every software engineer knows, are the most tricky parts of an 
application because even the smallest bug is amplified by the repetition of 
the code through cycles and a tiny memory leak rapidly becomes a huge hole 
that swallows all the memory of the machine.
This means that this part of Kat should be designed with extreme care and 
tested thoroughfully. This also means that the complexity level of that part 
has to be maintained as low as possible.
The KDED crashes will be avoided creating a simple Linux daemon for Kat and a 
KDED proxy for controlling it. This means that a crash of the daemon should 
not propagate to the KDED. Obviously we will try to avoid any crashes at all.

3. CLuceneQt

Ben van Klinken, creator of the CLucene port of Apache Lucene, offered his 
help in introducing Lucene into the Kat project.
He did more than this: he created a Qt port of CLucene. The CLuceneQt library 
will be used by our new indexer.
This should increase the speed of the indexing process and also give Kat the 
possibility to exploit the powerful query capabilities of Lucene. 
We will continue to use the SQLite3 for storing the things that Lucene does 
not handle, like the thumbnails, the metadata and the full text.

4. Problems with the new version of SQLite3 (3.2.6) SQLITE_MISUSE

The new versions of SQLite3 are more strict than the previous ones about the 
way a db handle has to be accessed. 
Versions > 3.2.6 do not allow the same db handle to be used in different 
threads. Since this was the "normal" behavior of our code, we will have to 
change it radically in order to comply with this new policy.

Final considerations:
Kat is an extremely young project and it is still in BETA version.
There are huge expectations about it, but people must understand that we 
cannot make miracles. I also have to convince myself about that :-D
As said before, any help will be really appreciated.

Regards

-- 
Roberto Cappuccio
Independent open source developer
- [EMAIL PROTECTED]
- http://robertocappuccio.blogspot.com

Kat Desktop Search Environment
- http://kat.mandriva.com
- #kat on irc.freenode.org


-------------------------------------------------------
This SF.Net email is sponsored by:
Power Architecture Resource Center: Free content, downloads, discussions,
and more. http://solutions.newsforge.com/ibmarch.tmpl
_______________________________________________
Kat-development mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/kat-development

Reply via email to