V4RB2 bug 648, new system of catch error

[Ruslan Zasukhin]

On 4/21/05 12:55 PM, "Olivier" <vidal_olivier at yahoo.fr> wrote:

Hi Oliver,

> Hi Ruslan and list,
> 
> Did you look at the bug 648? ( V4RB)

I think this is REALbasic issue.
 
> I think that it is an important bug IF all the operations of Valentina
> are VERY slow by the capture of error of Realbasic.

No. 

Exception is throw ONLY when error is raised,
So if you do 100000 adds without errors then you have ZERO overhead

> In my example (see mantis), if I makes an import, for example of 80 000
> records, but that there are 2000 identical records, while the "unique"
> parameter is enabled, the operation will take dozens supplementary
> minutes! (Because of the catch of error)

Well, try to make REALbasic function without Valentina and bench how many
you can raise - catch exceptions in the REALbasic.

If you prove that this is RB issue, then bomb REAL on that.
If you will find that RB itself can do 100 times more exception than
Valentina can do then we have 2 possible cases

1) exceptions from plugins are slow

2) problem in Valentina kernel.

 
> If I remove the new system of catch of error and what I puts back the
> old, everything is fast and there is not problem anymore.

Right, so you have workaround.

> There is a solution? Or it is necessary to use the old system of error
> to increase the performances?

May be, although this is bad of course.

Exception should not add significant overhead.

We need find WHO cause problems


-- 
Best regards,

Ruslan Zasukhin
VP Engineering and New Technology
Paradigma Software, Inc

Valentina - Joining Worlds of Information
http://www.paradigmasoft.com

[I feel the need: the need for speed]