[PATCH 5/5] user32/tests: Fix tests on win9x.

[Paul Vriens]

Huw Davies wrote:
> On Tue, Apr 21, 2009 at 11:34:04AM +0200, Paul Vriens wrote:
>> Huw Davies wrote:
>>> On Tue, Apr 21, 2009 at 11:22:29AM +0200, Paul Vriens wrote:
>>>> Huw Davies wrote:
>>>>> ---
>>>>>  dlls/user32/tests/clipboard.c |    7 +++++--
>>>>>  1 files changed, 5 insertions(+), 2 deletions(-)
>>>>>
>>>>>
>>>>> ------------------------------------------------------------------------
>>>>>
>>>>>
>>>> Was just about to sent a patch along those lines.
>>>>
>>>> Isn't it better to mark that win9x (and winME as well btw) one as broken()?
>>> I was debating with myself whether to mark it as broken or not.  My
>>> feeling was that it's not actually broken behaviour rather missing
>>> functionality.  But I don't really care.
>>>
>>> Huw.
>> Maybe it's not broken() behavior but marking it as broken() makes sure  
>> we (Wine) will have CF_UNICODETEXT. Without that broken(),  
>> CF_UNICODETEXT can go missing without us knowing.
> 
> Yeah, fair enough.  I guess the point is it's broken behaviour for
> Wine.  I'll resend, thanks.
> 
> Huw.
Yeah, it's either broken() or some clumsy:

cf = EnumClipboardFormats(cf);
if (cf == CF_METAFILEPICT)
     win_skip("Win9x/WinME don't have CF_UNICODETEXT\n");
else
{
     ok(cf == CF_UNICODETEXT, "cf %08x\n", cf);
     cf = EnumClipboardFormats(cf);
}
ok(cf == CF_METAFILEPICT, "cf %08x\n", cf);

Maybe we should have behavior_we_dont_want_to_replicate() next to 
broken() ;)

-- 
Cheers,

Paul.