Need Startup Script Clarification

[K1200]

In a 2-file separation model, I have an auto-execute startup script in the data file that is intended for execution when I open the data file (only) during development. For runtime, I have an autologin in the GUI file using a user account that is defined in both files, plus the GUI has it's own auto-execute startup script.

The GUI autologin works fine when starting the application. The problem is that the startup script in the data file appears to execute -- at least partially -- up to the point of a custom dialog step. The dialog never appears, and no subsequent steps execute. It's as if a Halt Script step is executed at that point. Yet, a Set Field step immediately before the dialog appears to be executing.

Can someone clarify what behavior should be expected here? Should the startup script in the data file execute at all upon a GUI startup? If so, what are the rules?

Thanks for any help.

[IdealData]

Have you switched on the script debugger to trace through your script.

[IdealData]

The startup script does not execute unless you actually open the file directly or call the script from another file.

Naturally the data file is being opened because it is related to the GUI, however startup script doesn't run in the data file in this circumstance.

[Vaughan]

The startup script runs when the first window of the file is opened.

If the file appears in the "hidden" window menu, selecting it will cause the window to open and the startup script runs.

[Vaughan]

Are you running any script triggering plugins?

[K1200]

Thanks for the responses. But, I'm confused by the conflicting statements:

The startup script does not execute unless you actually open the file directly or call the script from another file.

The startup script runs when the first window of the file is opened.

(BTW, I'm not using a script-triggering plug-in.)

Activating the Script Debugger in development mode and then starting the application has given some odd results. For instance, it doesn't offer step-by-step option from the first of the script, but instead, "becomes active" in the middle of the script in the data file. And then, when executing a script in the data file, the debugger leaves the application on a layout in the data file and does not automatically return to the GUI layout, like the runtime application does. Obviously, this usage is "pushing the envelope" of the script debugger.

But the real environment is, of course, the built runtime, where no script debug is available -- so I was hoping for solid rules by which to navigate this rather tricky ground.

[IdealData]

Vaughan is of course correct. I was recalling how FM6 worked.