RE: New Slide Database Schema Proposal

["Kumar, Ashok" <ak...@metatomix.com>]

Hi Michael

Thanks for looking at the new slide DB design.

>>actual code (I saw NullPointerExceptions 

I had tested it , but did not get any null pointer exceptions. But I
tested the ContentStore and Descriptors Store only for mkcol,
proppatch,propfind, put and get.
Did U test the LockStore, revisionStore etc and other commands ???
Could U give me a trace of what u have run and the domain.xml, u are
using so that I could try it in my test environ:

>>Also, duplicating all of J2EEStore.java just to add two hashtables to
it
>>seems fairly bad design practice - a minor issue, obviously.

My intention is  to have the existing J2EEStore with the hashtables in.
But I intended to do that once we decided upon the new proposal
approval, upto which i thought it better to have a seperate stream of
code. Any way thanks for pointing it out.

Hope to hear from u asap

Ashok
-----Original Message-----
From: Michael Smith [mailto:msmith@xn.com.au]
Sent: Thursday, April 18, 2002 9:04 PM
To: Slide Developers Mailing List
Subject: Re: Ashok's problem: J2EE Version Slide testing



> The better way is to handle it thru indexing the store tables which is
> impossible with the current database schema. Infact I have posted a
> proposal -see my mail with subject "Proposal for a more efficient New
> Slide Database Design" - for a modified schema with proper indexing
> which could avoid this problem as well as handle the db operations
more
> efficiently. I am waiting for a response from the Slide Dev Team.
> 

Certainly an important step forward. However, in my limited testing (~5
minutes, then I had to revert to the previous store because I needed a
working system), it didn't actually work reliably. The database stuff
looks fine, I think there are just some missing cases to handle in the
actual code (I saw NullPointerExceptions all over the place, but didn't
have time to look further. I'll try to find time to do so reasonably
soon, but no guarantees...).

Also, duplicating all of J2EEStore.java just to add two hashtables to it
seems fairly bad design practice - a minor issue, obviously.

Once this works stably, I'd say it should definately go into slide -
performance is the main disadvantage of slide at this point in time, and
significant improvements like this are great.

Michael

--
To unsubscribe, e-mail:
<ma...@jakarta.apache.org>
For additional commands, e-mail:
<ma...@jakarta.apache.org>


--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>