We apologize for the misunderstanding, we dont't want to frighten developers
that work on maxdb saying that it is unstable.
We aprreciate your effort to support us. Our previous mail was just a
consideration about some problems we are encountering, and trying to solve
with your help.

We'll submit to you our possible further problems with full details.

Thanks

Fabio


----- Original Message ----- 
From: "Zabach, Elke" <[EMAIL PROTECTED]>
To: "'Fabio Pinotti'" <[EMAIL PROTECTED]>; <[EMAIL PROTECTED]>
Sent: Friday, June 11, 2004 3:57 PM
Subject: AW: Maxdb considerations


> Fabio Pinotti wrote:
> >
> > Dear staff
> >     we are porting a complex j2ee/Oracle enterprise application to a
MaxDb
> > database, but we have found several problems.
> > We started using 7.5.00.12 installed on a linux machine, then we have
> > updated to 7.5.00.14.
> > It seems that database is very unstable: it crashes for a lot of
reasons:
> > calling query statements, calling user defined functions inside sql
> > queries, and so on...
> >
> > We have a complex package in Oracle environment, used to check users
> > permissions on documents, activities, etc.: it is made of lot of nested
> > functions.
> > We are trying to translate it into internal maxdb syntax, but we have
been
> > founding several troubles; for example:
> > 1) we call a function with 3 params, the function calls onother one
> > passing two of them; when the second function returns, parameters are
not
> > valid (they are re-initialized)
> > 2) even if functions are succesfully compiled, db crashes at run time
(we
> > have to manually restart db!!) if they contain type mismatch (like: int
to
> > number)
> > 3) recursion seems to be not possible
> >
> > On the other hand, we are trying to use existing sql (that runs on
Oracle
> > db): we see that maxdb has the "oracle mode", but we often encounter
> > problems with types: maxdb seems to be less flexible than oracle. The
main
> > issue is that calling such a query causes the database to crash.
> > Moreover, we often have to use workaround (swapng parameters in decodes,
> > using dummy substr(), etc,...)
> >
> > We don't undertstand what's wrong with our approach, and we would like
to
> > know if the present release of maxdb is stable. Our goal is to develop a
> > fully open source solution, but considering all of these problems, we
> > aren't sure that maxdb is the right choice for our purpose.
>
> Did I miss something?
> Joannes Capitanio and you sent us three topics:
> 1. ADDDATE and value causing an error where noone should be returned (we
are    working on it)
> 2. a question about recursive functions
> 3. the problem with the very important feature DECODE (?, ...) (workaround
given, fixed with the next version)
> No crash, no info about problems with int <--> number, functions called in
functions and so on.
> As we always say if we are asked: functions in 7.5.00.* are on your own
risk. We do not like the idea of crash in that context, but we are not able
to put much effort in this topic in this release.
>
> We do what we can to produce a stable version and those problems we are
told a little bit more precise will be handled as fast as possible,
especially to avoid crashes. But please do not spread the idea of 7.5.00.*
being a very unstable version. That is, as far as I know, not the impression
other users have. Yes, there are bugs in (unfortunately), but only if we are
given a chance to fix them (by describing them as precise as possible,
sending vtrace / data sometimes, and so on) we will be able to increase the
stability of the next versions (much more than only relying on our own
checks, which are finding problems as well).
>
> Elke
> SAP Labs Berlin
>
>
>
>
> > Thanks in advance
> > Best regards
> >
> > Fabio Pinotti
>


-- 
MaxDB Discussion Mailing List
For list archives: http://lists.mysql.com/maxdb
To unsubscribe:    http://lists.mysql.com/[EMAIL PROTECTED]

Reply via email to