user32: Fix a test that now passes in Windows

[James Hawkins]

On 10/25/07, Reece Dunn <msclrhd at googlemail.com> wrote:
> On 25/10/2007, James Hawkins wrote:
> > Hi,
> >
> > Changelog:
> > * Fix a test that now passes in Windows.
>
> Looking at the code, it looks as if they are now passing in Wine :).
>

I don't think you are looking at the code, nor does it seem you've run
the tests.  The tests that now pass in Wine are moved out of the
todo_wine block, and the one remaining failing test is still in the
todo_wine block.

> >  dlls/user32/tests/dde.c |   13 ++++++++++---
> >  1 files changed, 10 insertions(+), 3 deletions(-)
>
> > +    str = (LPSTR)DdeAccessData(hdata, &size);
> > +    ok(!lstrcmpA(str, "requested data\r\n"), "Expected 'requested data\\r\\n', got %s\n", str);
>
> Shouldn't there be quote marks around the %s (i.e. '%s', to match the
> expected value)?
>

No.  The quotes around 'requested data\\r\\n' are for ease of reading the code.

> > +    ok(size == 19, "Expected 19, got %d\n", size);
> > +
> > +    ret = DdeUnaccessData(hdata);
> > +    ok(ret == TRUE, "Expected TRUE, got %d\n", ret);
>
> These are new test cases. Since these are not fixes to existing test
> cases, they should really be in a separate patch.
>

Huh?  You're really coming out of right field.  I don't think it needs
explanation, but you don't seem to get it.  It's not a 'new' test
case, it's a part of the test block 'XTYP_REQUEST, fAckReq = TRUE'.
The call used to fail before in Windows, but was fixed by a timing
patch committed by Julliard.  There's no point in checking the
requested data if the call fails.  Now the call doesn't fail, so we
check the data.

Please read the code thoroughly before making comments like these.

-- 
James Hawkins