Re: [Gnu-arch-users] hackerlab q + possible bug

[Tom Lord]

    > From: Johannes Berg <address@hidden>

    > Still continuing to port hackerlab to windows, I found that the socket
    > dup function doesn't associate the correct vtable with the new
    > descriptor. Thats not correct behaviour for windows because there you
    > really need to use different functions for sockets (vs. file handles)...

    > I'm really not sure if this is a bug or how to handle it.

    > If I were to call=20
    >     vu_set_fd_handler (result, &url_socket_vtable, closure_argument);
    > (result = dup(...) set before), what would the correct
    > `closure_argument' be?

I'm fairly sure that you should just forward the closure that was
passed to `dup' (it's been a while since I worked on that part of
hackerlab).

That will ultimate lead to a copy being made of the `struct
url_socket_params' that it refers to.

-t