Judson Valeski wrote:
> I'm noticing some funky behavior w/ OnStopRequest callbacks. First a
> question...
>
> what are the current semantics surrounding OnStartR() and OnStopR()
> callbacks? We used to maintain the guarantee that you would never
> receive an OnStopR() w/ out a preceeding OnStartR(). Is that rule still
> maintained? I *think* it is, but want to make sure I'm clear up to
> speed.
Yes, OnStopRequest will ALWAYS follow an OnStartRequest.
> Back to the funkiness. Should hitting the reload or the stop button fire
> an OnStopR() up the chain of consumers?
Yes.
> If so, I'm not seeing this
> happen in the multipart/x-mixed-replace stream converter case. If I'm
> receiving a bugzilla query for example, and I hit reload or stop, the
> mixed converter's nsMultiMixedConv::OnStopReqest() never gets called,
> and subsequently bad things start happening (corrupt content, crash,
> other randomness). Is this the multi mixed converter's problem somehow?
> Have semantics changed in this area? Is HTTP receiving OnStopR() in
> response to a reload or stop button being pressed?
HTTP is still recieving an onStop notification when I click the stop button
while reloading a bugzilla query, but your totally right, the
nsMultiMixedConv::OnStopReqest is never called. File a bug against me and I
will start looking.
