Peter,

Thanks for taking the time to share your ideas. See comments below.

On 9/1/10 3:09 PM, Peter Jakobsson wrote:
> On 1 Sep 2010, at 21:53, Brad Perkins wrote:
>
>> Any ideas would be appreciated.
>
>
> I spent a great deal of time tracking down 'invalid parameters in
> execute command' when I upgraded to v11 (mainly because I use
> components and they've been revolutionised).
>
> It was one of those things where there's nothing else for it but
> to sink entire days into it systematically to get to the bottom of
> the problem.
I'm well into day 3.
>
> I think what I did was to systematically remove of If(False) whole
> bits of the Active 4D scripts out of the call chain to isolate
> where the problem was. Unfortunately I can't quite remember what
> the various problems were, however, a lot were related to
> component use (e.g. trying to call a component method from an
> Active 4D script which is not allowed).
We don't use any components.
>
> Other stuff which I dealt with but which may not result in the
> symptoms you have were:
>
>  - type mismatches
I've already looked into this. When comparing our old 4D 2003 shell
to the new I noted a lot of  C_STRING to C_TEXT changes. I haven't
found any mismatches though. I've also looked for multiple variable
declarations that might be inconsistent. You would think the
Compiler would catch those though. I miss Insider. I really miss
Sanity Check (sigh).
>  - undefined variables in interpreted mode being passed to a
> compiled component 
We compile all variables are typed.  Wait, you said component. Don't
use them.
>  - forms being called which did not exist (this is a good one -
> works fine in 2004 but throws and error in v11)
Do you mean 4D forms? This should not be an issue in our application.
>  - 4D bugs (mainly 11.6 If your on 11.7 there are less)
We're on 11.7.
>  - custom constants not defined
I think we are OK here. Do you mean in 4D code or Active4D code? If
the latter I have run the conversion routine and am importing
'constants.a4l' from within On Application Start. If the former,
wouldn't the Compiler catch if those were not defined?
>
> Are you using the 'out of the box' shell ? If not I could go on
> because some stuff has changed with regards to A4D Execute 4D
> request.
No, but since we are using NTK that plugin command is not called. I
installed the new shell into our structure. Before I did that I
exported our existing shell methods and the new shell methods and
ran them through a graphical diff tool. I used that as the basis for
merging our changes back in. The truth is we don't really change
default shell behavior that much. I have to do the previously
mentioned CGI branching and I need to initialize a few process
variables that exist in called 4D methods.
>
> Failing any of the above being your problem, I strongly recommend
> the 'get out and dig' option which means take a couple of hours to
> systematically disable your Active4D call chain (i.e. all the code
> that runs inside A4D Execute 4D request  (or in your case A4D
> Execute stream request) until you no longer get the error. 
This is basically where I am now. I have to figure out what 4D code
runs or is called from A4D Execute stream request.
> Then work back the way, re-enabling code at a more granular level
> until you have isolated the buggy code by sheer brute force and
> will. This approach worked for me, but you need some time and a
> lot of coffee. The payback is you'll learn something :)
That isn't where I want to go, but it looks like this is what I will
have to do. This is a huge pain though as we run our web server on
client. For that reason I also develop that way.

Thanks again for all of the suggestions.

-- Brad
_______________________________________________
Active4D-dev mailing list
[email protected]
http://mailman.aparajitaworld.com/mailman/listinfo/active4d-dev
Archives: http://mailman.aparajitaworld.com/archive/active4d-dev/

Reply via email to