Community
cancel
Showing results for 
Search instead for 
Did you mean: 
martymike
Beginner
37 Views

Debugging host-associated variables XE 2015

We still cannot see the value of host-associated variables. Is there any hope that someday this will be corrected?

Just to be clear, I'm talking about something as simple as the value of x in the subroutine abc, below (not a real program):

subroutine myprog(i)
integer x,i
x=i
contains
subroutine abc
print *,x
end subroutine abc
end subroutine myprog

0 Kudos
3 Replies
Kevin_D_Intel
Employee
37 Views

I am using the PSXE 2016 integrations with Visual Studio but I believe you can do the same with PSXE 2015 integrations into Visual Studio.

To see the value of X, use the stack frame pull-down to change the context to the host. The first screen snap below shows the issue you described, and second shows the value of X after setting the stack frame to the host scope (line 5).

There is an existing open defect against host-associated variables (DPD200283998) not being shown in the debugger that appears related to the underlying behavior here. Hopefully using the stack frame enables you to see host-associated variables of interest.

0 Kudos
martymike
Beginner
37 Views

Yes. we are aware that you can use that method, and we get by with it. But there are times when it becomes extremely tedious. In a loop where many variables are changing, stepping line-by-line, ... step over the next line, select the host program from the call stack, select the current location to see the values in context, step over the next line, etc., ...

I just keep hoping - year after year - that there might be something better coming. At this point I'd settle for a syntax to specify in watch window like host::var. It's the having to use the call stack to switch contexts in order to see variable values that is such an efficiency killer.

I rarely use host variables like this in new code, but we have a lot of old programs that are nothing but.

0 Kudos
Kevin_D_Intel
Employee
37 Views

Ok, thank you for the additional feedback. I'm sorry this has persisted for so long. I will add your case to the earlier cited internal tracking id and do some internal follow-up to see how this issue might be resolved in an upcoming release.

0 Kudos