[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Mechanism extensions and the GSSAPI

>>>>> "Christopher" == Christopher  <cmaxwell@themanor.net> writes:

    Christopher> On Mon, 3 May 2004, Sam Hartman wrote:
    >> >>>>> "Christopher" == Christopher <cmaxwell@themanor.net>
    >> writes:
    Christopher> Maybe this is oversimplistic, but why not provide the
    Christopher> shims as library/source to be linked into the

    Christopher> Maybe I misunderstood the earlier conversation, but I
    Christopher> thought the shim was to be the stable API between the
    Christopher> application and the ioctl interface?

It is.  But if you link or ship shims with applications rather than
with mechanisms you need to deal with version skew between the
mechanism and the shim.  That means the ioctl protocol needs to be
extensible and stable.

I actually think in most cases I'd rather have applications fail to
link if extensions are not present than have them have a runtime
error.  In cases where that is not true you can use weak references in
the application or dlsym to find a function pointer.  This style of
programming is quite common on Windows and Mac and works well.