Re: [wintrust/tests] Add some basic CryptCATOpen tests

[Detlef Riekenberg]

On Mo, 2009-05-18 at 16:07 +0200, Paul Vriens wrote:
> - hcat = pCryptCATOpen(emptyW, 0, 0, 0, 0);
> - todo_wine
> - {
> - ok(hcat != NULL && hcat != INVALID_HANDLE_VALUE, "Expected a
>
> correct handle\n");
> - ok(GetLastError() == ERROR_SUCCESS,
> -
>   "Expected ERROR_SUCCESS, got %08x\n", GetLastError());
> - }

MSDN documents INVALID_HANDLE_VALUE on failure.
When NULL is valid result, then a comment about the MSDN error would be
nice.

Did you have an application, that need ERROR_SUCCESS?
The result of GetLastError() should only be tested, when the function
failed,
or when GetLastError is documented to be valid on success.

[wintrust/tests] Add some basic CryptCATOpen tests

[Paul Vriens]

Detlef Riekenberg wrote:
> On Mo, 2009-05-18 at 16:07 +0200, Paul Vriens wrote:
> > - hcat = pCryptCATOpen(emptyW, 0, 0, 0, 0);
> > - todo_wine
> > - {
> > - ok(hcat != NULL && hcat != INVALID_HANDLE_VALUE, "Expected a
> >
> > correct handle\n");
> > - ok(GetLastError() == ERROR_SUCCESS,
> > -
> >   "Expected ERROR_SUCCESS, got %08x\n", GetLastError());
> > - }
>
> MSDN documents INVALID_HANDLE_VALUE on failure.
> When NULL is valid result, then a comment about the MSDN error would be
> nice.

NULL is not supposed to be valid it was just to show that hcat changed. 
I guess just checking for INVALID_HANDLE_VALUE will do.
> Did you have an application, that need ERROR_SUCCESS?
> The result of GetLastError() should only be tested, when the function
> failed,
> or when GetLastError is documented to be valid on success.

Well, I actually don't know. Our implementation is wrong so this one is 
hard to check. I will remove it from the test.