I really think Turbine should update/review the client as the fact stands it is really the fault of the client and it has been proven time and time again. You don't even need decal to be added into the mix to make problems happen.

If they do, I pray to God that they do away with the need to have "Loose UDP Packets", as more and more routers and firewalls reguard these as "intrusions" and you can't bypass them. AC1 still requires it, but AC2 didn't.

Yeah I brought this up with both Jessica and Srand. It was reviewed as a possible Throne of Destiny feature and rejected due to the amount of work required.


We are talking about very old core code based on a design that is not used in any of the other Turbine products. This is an area that when AC2 came around, Turbine started from scratch.


I think the network code falls in the same category as the physics engine. It is what it is and it will never be changed. One big issue with the network code is that it is a scary area to change. Once you start work, you have product that no longer works until you finish. AC is not very useful if the subservers can not communicate with each other and the client can not talk to the server.

How does one indefinatly prove a memory leak in the software without opening room for concern with hardware/software etc.. as part of the culprit?


I think someone could make a utility to scan all the memory ranges of the software in a closed way in which it first checks hardware for failures in device drivers or miscommunication with that hardware.


Then goto the software. Start with the client, and go to each file individually.


This would do a couple things...


Identify.

Condense.

Prove.

Execute resolution.

One quick way to check on a memory leak is to look at the size of the client task when you first enter a world. Periodically alt-tab back out and see if the task size is increasing steadily.


Memory leaks are a pretty common problem. Microsoft has this issue with their operating systems, bundling components and all of their products. It is a good idea to restart a Windows machine at least once a day to clean up memory leaks and other issues.

"It is a good idea to restart a Windows machine at least once a day to clean up memory leaks and other issues."


Hahahahahahahaha.... not laughing at you or your suggestion... It is actually a good idea as stated.. however.. I have a bit of a problem actually restarting my computer at ANY regular interval... this PC has been up for....


1384:22:03


thats 1384 hours, 22 minutes, and 3 seconds...


that's about 57 days, 16 hours, 22 minutes, and 3 seconds.


*sheepishly grins*


although that's only my System Idle time...


my Network Connection says I've been connected for 63 days, 3 hours, 17 minutes, and 25 seconds.

Outside of Microsoft's server and embedded OSs, I did not think it was possible to keep a Windows machine up this long...


My kids have tried but within two weeks, the machine either gets really slow, the print spooler shuts down or some other other problem develops

I have a bad habit of keeping my work machines in a very functional and non-rebooted status.


I try to minimize applications on it that produce leakes, and I try to keep things patched, and fairly defragged/etc.


I think it's possible to leave windows up a LOT longer than that depending on usage.


I have an old 2K Pro machine which I use as a personal development server which went approximately 270 days without a reboot before I decided it was best to patch with updates from microsoft and reboot.


If it weren't for the windows update needing to reboot things, I'm sure I could have kept it up a lot longer.

Just a FYI.. I disabled Meginjarder, and I haven't been crashed for about 4 hours.

 

-----signature-----

If it was ELTank crashing on exit it would crash EVERY time, not just sometimes.