Re: [ole32/tests 2/2] Fix crash on NT4 and below

[Paul Vriens]

Rob Shearman wrote:
> 2008/6/10 Paul Vriens :
> > This 'fixes' the crash on win9x and NT4.
> >
> > @@ -1130,11 +1130,15 @@ static void test_data_cache(void)
> >      hr = IOleCache_Uncache(pOleCache, 0xdeadbeef);
> >      ok(hr == OLE_E_NOCONNECTION, "IOleCache_Uncache with invalid value should return OLE_E_NOCONNECTION instead of 0x%x\n", hr);
> > - hr = IOleCache_Cache(pOleCache, NULL, 0, &dwConnection);
> > - ok(hr == E_INVALIDARG, "IOleCache_Cache with NULL fmtetc should have returned E_INVALIDARG instead of 0x%08x\n", hr);
> > - /* Both tests crash on NT4 and below */
> > - if (0)
> > - {
> > -
> >    hr = IOleCache_Cache(pOleCache, NULL, 0, &dwConnection);
> > -
> >    ok(hr == E_INVALIDARG, "IOleCache_Cache with NULL fmtetc should have returned E_INVALIDARG instead of 0x%08x\n", hr);
> >
> > - hr = IOleCache_Cache(pOleCache, NULL, 0, NULL);
> > - ok(hr == E_INVALIDARG, "IOleCache_Cache with NULL pdwConnection should have returned E_INVALIDARG instead of 0x%08x\n", hr);
> > -
> >    hr = IOleCache_Cache(pOleCache, NULL, 0, NULL);
> > -
> >    ok(hr == E_INVALIDARG, "IOleCache_Cache with NULL pdwConnection should have returned E_INVALIDARG instead of 0x%08x\n", hr);
> > -
> >
> > }
> >
> > for (fmtetc.cfFormat = CF_TEXT; fmtetc.cfFormat < CF_MAX; fmtetc.cfFormat++)
> > {
>
> This test was added in response to a real application doing exactly
> this, so I think the test does have value when run on Win 2000+.
> Therefore, I think a version check would be better than "if (0)".

Let me check (later) whether it's possible to have some logic in that decides 
whether we are on W2K and up.

If not, all remains is of course a real version check.