Hi
On Aug 8, 7:46 pm, Albert <
albertpa...@gmail.com> wrote:
> Thanks for the vid link! I learned some other good stuff from there.
> However, I don't think my question was clearly answered.
>
> I hope I get a clear response from the AppEngine Team on this one.
>
> When I start an asynchronous urlfetch, and return my application
> response without calling get_result() or wait(), I'm assuming that
> either of four things can happen.
>
> 1. The url will be successfully called.
> 2. The url will not be successfully called.
> 3. The url will be called, and then "cancelled" (I don't know if this
> is actually a possibility)
> 4. No guarantees.
>
> So which is it?
>
That will depend on how far the fetch from the target url got
processed. If all the of the data has been retrieved from the other
end
then data in a buffer somewhere will be discarded. If the remote
server hasn't responded yet and the urlfetch is cancelled then you
don't
know how much has been executed on the far end. If it where a zope
server for instance whatever was requested would be processed in full
but no end point would exist to write the data back to (from the
servers point of view)
So you really can't tell if you don't check the async call with a
get_result or wait. You really have an indeterminate state.
Why not do some tests. The fact the multiple async urlfetchs can be
performed means that the requests can be made and will reach the
destination
if you don't get hit with a DeadlineExceeded before the connections
are dispatched.
Rgds
T