bugzilla && kbugbuster

[Cornelius Schumacher]

On Monday 09 September 2002 00:09, Helio Chissini de Castro wrote:
>
> Now comes kbugbuster..
>
> 1 - The HorribleParseLikeKbugbusterWorksNow:
> It's the simple, but ugly, hard, and longer to move.. I really
> dislike this way of doing things like parse a web page result not
> intended to this specific operation, but if none of others below is
> aproved, patience...

This doesn't make much sense, if KBugBuster is going to become a generic 
Bugzilla frontend which isn't tied to a specific Bugzilla installation 
like the KDE one.

> 2 - Read only open database access:
> Bugzilla operates over MySQL as usual ( KDE Bugzilla too Stephan ? ),
> then kbugbuster could access directly MySQL database. It's a good
> solution from the programmer side, and very possible, but it impose
> two problems, one related to one feature that we uses here, and other
> in the kbugbuster exec requires.
> The feature that we uses in Conectiva, and Stephan doesn't mention if
> we will use then in KDE bugzilla, is about restricted bugs, that is
> visibility just to reporter and maintainer, not open to anyone. With
> worldwide access to database,  we can't block automatically querys in
> this type of records, and need keep in mind that no one desires
> change things in such databases. The second problem is about the need
> of QT database plugin compiled. I know that a large number of
> developers maintain your qt libraries reduced as they need, not
> compiling all things. A kbugbuster working directly with database
> will force people to do compile things that they don't want...

As this method would circumvent the abstraction layer of the Bugzilla 
scripts, I don't think that this would be a desirable solution.

> 3 - A set of server side cgi scripts directly related to kbugbuster (
> My choice and the local team ):
> The holy grail is create server side scripts with querys needed from
> kbugbuster, wrote for this specific operation.
> This will keep all things protected, handled by the server, returning
> all data as we want..
> The major job in administration is encharge of bugzilla maintainer,
> and specific extra tasks could be developed. It's the fast and
> reliable way to port kbugbuster changing a little their gui and way
> of work, and easy to implement.
> This will enable kbugbuster to easily used in bugzillas d over the
> world, not just kde, and moving a specific application to a full
> usage kde application.

Doesn't Bugzilla already have an XML interface for querying bugs which 
could be used for KBugBuster?

-- 
Cornelius Schumacher <schumacher at kde.org>