Example stack provided crashes

[Ruslan Zasukhin]

On 4/15/06 8:27 AM, "Robert Sneidar" <bobs at twft.com> wrote:

> Okay, I did the whole example stack thingummy and was pretty proud of
> myself, that is until I put DatabaseCreate in the openStack handler
> of the stack. From then on opening the stack crashed Revolution to
> the desktop. Hmmm... I must have done something wrong, right?
> 
> So I figured out a trick way to edit the script of the stack and
> commented out the call to the function, and all is well, right? Well
> I can open the stack, but if I call DatabaseCreate from the message
> box, I still crash to desktop!!!
> 
> So I started comparing my code with the code from the finished sample
> stack that came with the documentation. Well there are significant
> differences! Dang! That is troubling! You would think it would be
> identical.
> 
> So I get a brainstorm. How about I open the PROVIDED UNTARNISHED
> sample stack and call DatabaseCreate from the message box and see
> what happens? Gee, whattyaknow? IT ALSO crashes to desktop!!
> 
> So I commented out ALL my code, and pasted the example stack's code
> into my stack, and stepped through the code. When I get to the line:
> 
> get VDatabase_Create( mDatabase, dbPAth, "kDscDatBlbInd" )
> 
> it returns an error. That means mDatabase is set to the value of the
> error. The next time I try to do something that references mDatabase,
> it crashes to desktop, obviously because it's referencing a pointer
> to a table that is actually an error code! Ouch!!!

About crash when you send bad object reference. In our docs said:
    to protect from such crash DURING DEBUG, you
    need set Valentina_DebugLevel = 3

In this case Valentina will CHECK all object references to be valid,
before use them. 

This is additional overhead, so it is good idea to use it only in
DEVELOPMENT mode, but in release turn off. You see ?

We have NOT point this tip in tutorial, to keep it as simple as possible.
Well, may be we should inform about it in first turn...
 
> Again my apologies, but I simply do not have time to troubleshoot a
> problem with a stack that was provided as an example of how to
> properly implement Valentina when it crashes to desktop. I know I can
> trap for errors, that isn't the point. I shouldn't be getting errors
> at all with a stack provided by the developer.

Wait a moment please. As I see, you say:

    * I have create new stack using PDF, step by step. It not works
    * stack of Paradigma so work.
    * I see differences between my stack and Paradigma stack.

Have I correctly understand you?

If yes, then this means that *our* tutorial stacks work fine.
Mistakes present in PDF. Right?

We know 100% that our V4REV/Examples 50 examples also work fine,
we test them periodically on all platforms.

Can you confirm that OUR STACKS do work fine?
    tutorial stacks?
    50+ examples ?

----------
P.S. 

I am as old C++ developer have very good habit DO NOT TRUST docs.
I have never see 100% correct docs. NEVER.
Especially for complex developer tools.

E.g. I remember nightmare when I have to write JavaScript scripts for Adobe
InDesign using their 1000-page reference. Yes also there was a lots of
pain...But with hours and days of fight I have become cool expert in that

But of course we will again and gain now try step by step check our
tutorial. Kirill will start do this today. I want underline that Kirill have
NEVER use Revolution, so will be interesting to see how hard for him will be
get into it.


-- 
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]