Denys Duchier wrote:
> Hi Guys,
>
> First, on the release branch in libdp/flowControl.hh I see:
>
> Bool wakeUpCheck(LongTime *t){
> return first != NULL && t > &time;}
>
> shouldn't that be:
>
> Bool wakeUpCheck(LongTime *t){
> return first != NULL && *t > time;}
>
This seems to be a bug, please correct it.
>
> Second, the problem with DP/gcc 3.2 is even stranger than I thought.
> I tried to combine Dragan's "freelist avoidance patches" with my new
> implementation, but that did not help. I managed to get some DP
> tracing info (which I include below) between two processes, both on my
> laptop (off the net): Process 1 offers atom foo and Process 2 takes
> the corresponding ticket.
>
> The short story is that, while Process 2 works a lot, Process 1 never
> sees anything. I include the traces below and hope someone can
> suggest a way to proceed with debugging: the trace information just
> does not speak to me.
>
>From what I can see in the logs it looks like process-2 tries to connect, but
never reaches process-1. To get more info, you could switch on the printouts in the
connection code. Define DEBUG in <SRC>/share/lib/dp/[ConnectionFunctor.oz, AcceptFunctor.oz,
ConnectAcceptFunctor.oz]. These functors implement the actuall connection sequence. From what
I can see in your log it looks like the problem is located in that area. A possible problem would
be that something happends with the Ip number when passed from the DP-lib to the connection code.
/Erik
-
Please send submissions to hackers@mozart-oz.org
and administriva mail to hackers-request@mozart-oz.org.
The Mozart Oz web site is at http://www.mozart-oz.org/.