I have a module with a function similar to this:
MainModule
Sub Test()
On Error Resume Next
Dim O1 As New Class1
O1.DoSomething
On Error GoTo 0
End Sub
and a few classes similar to this:
Class1
Sub DoSomething()
FindStuff
'create similar objects who perform similar operations and raise similar errors
Dim O2 As New Class2
O2.DoSomething
End Sub
Function FindStuff() As Stuff
'scan the WorkBook, the file system, etc. and organize the members of the object
If CorruptedFileSystem Then Err.Raise 514, "File system corrupted"
If CorruptedWorkBook Then Err.Raise 515, "WorkBook corrupted"
If Found Then Set FindStuff = FoundStuff
End Function
There is an error trapping option in VBA menu Tools > Options > General tab:
If I set the error trapping to Break in Class Module then the On Error Resume Next will be ignored and every Err.Raise will stop the execution inside the class.
If I set the error trapping to Break on Unhandled Errors then the Err.Raise will stop the execution at the call on the main module, not inside the class.
So in one case I can't execute the code with handled errors, in the other case I can't debug unhandled errors.
The problem becomes unmanageable when the project grows and the main module creates an object that opens a form (which is another object) that creates more objects. Some of the methods handle their own errors and some are designed to abort and raise the error to be managed by the caller.
Is there a way to handle and debug errors in classes?
EDIT
Apparently my question wasn't clear enough. I changed the title and I will try with a clearer example.
Module1
Sub Test1()
Dim O As New Class1
O.UnhandledCall
End Sub
Sub Test2()
On Error Resume Next
Debug.Print 1 / 0
Dim O As New Class1
O.HandledCall
On Error GoTo 0
End Sub
Class1
Sub UnhandledCall()
Debug.Print 2 / 0
End Sub
Sub HandledCall()
Debug.Print 3 / 0
End Sub
Test1
Set Error Trapping = Break on Unhandled Errors and execute Test1. The debugger will not stop at the unhandled error 2 / 0. Instead it will stop at O.UnhandledCall, making it impossible to know what line caused the error, what were the local variable values, the stack, etc.
Test2
Set Error Trapping = Break in Class Module and execute Test2. The debugger will not stop at 1 / 0, good, because the error is handled. But it will stop at 3 / 0 inside the class even if the error is handled inside the caller function, at the same level as 1 / 0.
Sad summary
So with the first setting I can't see where en error is trhown, with the second setting I cant run a macro that cleanly handles errors.
This is obviously an oversimplified example. The real world case I'm dealing with at this moment is a form that creates dozens of objects, some objects check some text files, other objects open drawings on a CAD via COM, other objects talk to a database, etc. If any of the conditions is inconsistent I want to abort the form opening.
As the objects are created, they execute thousands of lines of code, with hundreds of managed errors. When they find something unmanageable in a file, in a drawing or in a database, they defer the error handling to their caller, climbing the stack up to the form that should fail to open and up to the caller that should detect the error and do something about it.
I would expect the debugger to run smoothly through the managed errors and stop when there is an unmanaged error at the offending line. Instead the debugger works as expected in modules, but in classes it either stops at all the error or it never stops, regardless of whether they are managed or not.
For example if I set Error Trapping = Break in Class Module all the managed errors will break the execution as in Test2, and my debugging session will never end.
While if I set Error Trapping = Break on Unhandled Errors then I will never know what triggered the error, because the debugger will climb through all the classes up to the first object and tell me that that's the line that caused the error as in Test1.

On Error Resume Next?On Error Resume Nextsuppresses errors and lets the code happily continue to run, defeating the purpose of structured error handling. Your top-level caller should have its ownOn Error GoTo CleanFailor something, to handle any error that bubble up to that point in the call stack.CleanFailand0? They are just two ways to do the same thing and don't change the point of my question. I could have added some details on how I manage the failure, but that has no relation with the way the debugger behaves with the handled or the unhandled errors.