[rc5] Mac client reporting

Stephen Langasek vorlon at dodds.net
Wed Jul 16 13:36:17 EDT 1997


On Wed, 16 Jul 1997, Nate Begeman wrote:

> 3) if no network errors, it sends blocks back -
> 4) if there is a network error, it grabs more to work on until it can send.

Now perhaps all the fiddling with registers and jumps at the assembly
level has impaired the developers' ability to write higher-level code ;),
I dunno; but am I the only one who sees a serious flaw in the logic here?
If the client gets a network error, then the next time it has access to
the network it should do *exactly* what it was trying to do before when it
failed, not move on to something different.  (Also, if there's a network
error that prevents it from sending, how in the world is it going to be
grabbing blocks from anywhere?)

<points and holds up a sign that says: "bug">

Other symptoms I've seen suggest to me that this may be tied to the
client's behavior when trying to generate random blocks (i.e., generating
one and then stopping), and given the degree of attention that problem has
received here on the list, it's likely that someone is already attending
to it.

                             -Steve Langasek
-doink-

----
To unsubscribe, send email to majordomo at llamas.net with 'unsubscribe rc5' in the body.



More information about the rc5 mailing list