Jeff Latimer wrote:
This is the first of some tests for NtCreateMailslotFile. I have included them in kernel32/tests as I thought it better that the Mailslot tests be collocated.
Why do you need to test NtCreateMailslotFile directly? Why can't you use CreateMailslot instead?
Vitaliy.
Vitaliy Margolen wrote:
Jeff Latimer wrote:
This is the first of some tests for NtCreateMailslotFile. I have included them in kernel32/tests as I thought it better that the Mailslot tests be collocated.
Why do you need to test NtCreateMailslotFile directly? Why can't you use CreateMailslot instead?
Vitaliy.
Dmitry Timoshkov wrote: NtCreateMailslotFile is a public API, so probably it should handle the casewhen attr is NULL. I see the same problem at least in NtCreateNamedPipeFile. In general, handling of OBJECT_ATTRIBUTES in ntdll is not consistent, but fixing it properly most likely a test is needed to see what Windows does in that case.
I am following the suggestion from Dmitry. In my testing so far I have noticed that there are a number of edit conditions that are not yet implemented. These won't be tested as easily by CreateMailslot.
Jeff
-
Jeff L -
Vitaliy Margolen