DlgSetCtrlEventHandler - argument mismatch

[Les_Neilson]

From trying to help with a posting on c.l.f I found the following :

In the documentation for the function DlgSetCtrlEventHandler it says the (optional) 5th argument to the function is -

iid
(Input; optional) Derived type GUID, which is defined in the IFWINTY module.

Andaccording to IFLOGM.f90 the interface to the function says that the fifth dummy argument is -

type(T_GUID), optional, intent(in) :: iid

Now the definitions of GUID and T_GUID are almost the same except that T_GUID has a preprocessor directive __KEEP_OLD_CHARACTER_ARRAYS to determine whether the 4th item in the type is character(8) data4 or character(1) data4(8)

GUID has character(8) data4

Converting and compiling the DVF samples project FXPLORER thows up errors re argument mismatch etc. So I presume the conflict is between the character array and the character variable?

Also, presumably by editingthe codein the fxplorer project and changing all references of GUID to T_GUID would fix the problem?

Les

[Steven_L_Intel1]

It doesn't matter if the definitions of types GUID and T_GUID are the same or not - what matters is that they are not the same type (as in same name from the same source.)

I got this to build and run by:

Adding in BROWSER.F90:

use ifwinty, only: T_GUID

Replacing GUID by T_GUID.
Disabling all of the !DEC$ ATTRIBUTES REFERENCE directives (which should never have been there in the first place.)

In fxplorer.f90, changing the various .TRUE. to TRUE

I find there's an exception that occurs when the program exits. I'm not sure why that happens.

[Les_Neilson]

Thanks Steve,

The OP (to c.l.f) reported that there was a "problem"on exit using CVF 6.1A up to 6.6C3 I haven't looked at this in any further detail (maybe next week if I get time)I was just interested in the argument mismatch.

Les

[Les_Neilson]

Steve,

Just a thought. Should the documentation for the function be changed to reflectthat the optional argument iid is atually typeT_GUID not GUID ?

Les

[Steven_L_Intel1]

I've put this on my list to look at. I think the REFERENCE issue has been cleared up in the Module Wizard now, but I had not looked at the GUID issue before.