OpenEuphoria

Re: [EuForum] Definition of insanity...

• Posted by CoJaBo <CoJaBo_EUforum_Address at CJBN.net> Nov 23, 2004
• 582 views

Patrick Barnes wrote:
> 
> Doing the same thing twice and getting different results...
> 
> Working on the contest engine... 
> 
> I have 3 bots set up to run a game. Each is given its own identity at
> startup. (12 letter random string, and an incremental number)
> 
> When the engine calls a procedure in one of the bots, it does its
> business, then calls a notify procedure in the engine to signify that
> it's done.
> 
> The notify procedure receives the number from the bot, and increments
> an appropriate flag element (there's 1 for each bot)
> 
> After a certain amount of time, the engine breaks out of the loop,
> checks the notification flags, and aborts if one of the bots didn't
> respond.
> 
> Got that?
> 
> Somehow, the data is getting mixed up in the IPC library. The correct
> identities are being sent to the correct bots, each bot is sending
> their own identity, but at the engine end:
> (Say with three bots)
> The engine is receiving the first two messages correctly.
> The third message is a duplicate of the second. And I mean, *exact*
> duplicate. So, rather than flagging 1,3,2 (which is success), it might
> flag 1, 3, 3!
> This is not a bot problem, I've checked that they're each sending
> different data...
> By the time it gets to the notify function, the 3rd piece of data has
> been messed up already, it seems to be a problem with the ipc.ew
> library.
> 
> 
> So why the subject line? Well, sometimes it works fine, sometimes it
> does this!! I can't figure out the common factors, let alone the
> problem...
You must never have done this:

for o=1 to rand(1000000)*100 do
	--It doesnt seem to happen
	--unless the program runs for a while
end for
integer g
g=time()*100
?g
include get.e
?wait_key()

Run it a few times, you'll eventually get a fairly unexpected error!
This was the cause of headaches for many weeks before I figured it out...

> 
> If anyone would like to have a crack at solving this, I can email your
> the source code...
> If it can't be fixed, no contest!
> -- 
> MrTrick
> 
>