[fpc-pascal] fpUnit - AssertEquals gives Access violation

Graeme Geldenhuys graemeg.lists at gmail.com
Fri Oct 6 11:48:02 CEST 2006


Hi,

AssertEquals always gives an Access violation error for some reason.
First some background info is needed....

I first write some of my tests using AssertTrue as follows:
  AssertTrue('Failed on 2', posClean = lNode1.ObjectState);

This just raise the exception when they didn't match. I preferred to
have the style expected <> but got <> message.  So I started change
those tests to the following:

  AssertEquals('Failed on 5', ObjectStateToString(posClean),
    ObjectStateToString(lList.ObjectState));

This worked fine, but after doing this for a while I realized I am
typing my life away with all those ObjectStateToString() calls.  Then
I thought, why don't I create a new AssertEquals method for my
projects that take TPerObjectState as parameters.  So I created this
in my custom test case class.

{ defined under Public }
procedure TM2TestCase.AssertEquals(const pMessage: string; pExpected,
  pActual: TPerObjectState);
begin
  AssertEquals(pMessage,
    ObjectStateToString(pExpected),
    ObjectStateToString(pActual));
end;

BTW: TPerObjectState is define as follows...

  TPerObjectState = (
                      posEmpty,
                      posPK,
                      posCreate,
                      posUpdate,
                      posDelete,
                      posDeleted,
                      posClean
                     ) ;



Now changing my test to the following, which is much easier to type and read:

  AssertEquals('Failed on 1', posEmpty, lLang.ObjectState);


Finally, getting back to the problem.  Using this new AssertEquals
always gives me a Access violation error.  I know there is a
AssertEquals that uses TClass, could that interfere with this?  I even
tried changing my new AssertEquals to use a different name
AssertEqualsObjState(), but that didn't help either.

What could be causing the problem?

Regards,
  - Graeme -


-- 
There's no place like 127.0.0.1



More information about the fpc-pascal mailing list