Fixing the last remaining failure on W2K (rpcrt4:server)

[Paul Vriens]

Hi,

In an effort to fix that last failure on my box I found the culprit. I'm 
not comfortable enough with just excluding (yet another) that test just 
for the sake of having zero failures though.

Any idea?

-- 
Cheers,

Paul.

diff --git a/dlls/rpcrt4/tests/server.c b/dlls/rpcrt4/tests/server.c
index 18cfa59..2d8c4d5 100644
--- a/dlls/rpcrt4/tests/server.c
+++ b/dlls/rpcrt4/tests/server.c
@@ -616,9 +616,13 @@ s_context_handle_test(void)
     ok(*(ULONG *)buf == 0, "attributes should have been set to 0 instead of 0x%x\n", *(ULONG *)buf);
     ok(!UuidIsNil((UUID *)&buf[4], &status), "uuid should not have been nil\n");
 
+    /* Crashes on W2K */
+    if (0)
+    {
     h = pNDRSContextUnmarshall2(binding, buf, NDR_LOCAL_DATA_REPRESENTATION, NULL, 0);
     ok(h != NULL, "NDRSContextUnmarshall2 returned NULL\n");
     ok(h->userContext == (void *)0xdeadbeef, "userContext of interface didn't unmarshal properly: %p\n", h->userContext);
+    }
 
     /* raises ERROR_INVALID_HANDLE exception on Vista upwards */
     if (0)

[Rob Shearman]

2009/4/15 Paul Vriens <paul.vriens.wine(a)gmail.com>:
> In an effort to fix that last failure on my box I found the culprit. I'm not
> comfortable enough with just excluding (yet another) that test just for the
> sake of having zero failures though.

What exception is being raised?

I think perhaps s_context_handle_test should be removed and replaced
with another that uses widl-generated code for testing context handles
rather than low-level functions.

-- 
Rob Shearman

[Paul Vriens]

Rob Shearman wrote:
> 2009/4/15 Paul Vriens <paul.vriens.wine(a)gmail.com>:
> > In an effort to fix that last failure on my box I found the culprit. I'm not
> > comfortable enough with just excluding (yet another) that test just for the
> > sake of having zero failures though.
>
> What exception is being raised?
>
> I think perhaps s_context_handle_test should be removed and replaced
> with another that uses widl-generated code for testing context handles
> rather than low-level functions.

I just see a RPC_S_CALL_FAILED

-- 
Cheers,

Paul.