Scourge of War Boards

This patch post-dates 1.405? If so, excellent news and many thanks.

Martin

We used the new Beta patch tonight for the first time since it came out. Played on the Antietam map wth a modified OOB and a mod that provides the corps commander a button to move the entire corps and double line formation for regiments. Entire ANV AI versus 4 human players with US I, II, III, and VI Corps. No cavalry. Couriers to send messages. Played full three hours with no observable problems. At one point had US II, III, and VI fighting 4 rebel divisions with one heckava brutal fight for almost 30 minutes with no lag. Maybe we got lucky but I figured with the load we gave it something would have given way. We will try another map next Thursday with more people and see what happens. No errors of any kind showing in log file.

With 4022?

I've got fingers crossed for 4022.

We've had freeze after freeze this week with 4020, including players vs AI last night ;-(

A lot of the log files that I receive have errors. I'm not sure if I should just stop the game on these errors or continue. They show machines that just don't have all the latest files and although the game can run with these, they can show up later if the game tries to access those files. Especially in MP if these error rear their heads later in the process you could have issues.

How can a machine join and run in an MP game if it lacks some of the latest files? Could a checker be put in to prevent a machine joining an MP game if it has out of date files (I assume from a patch that was applied and didn't install correctly?)

I am in favour if something stops a player joining an MP game rather than lets him join but the state of the game files he has causes the game to crash an hour later and wastes several gamer-hours.

Norb wrote:

A lot of the log files that I receive have errors. I'm not sure if I should just stop the game on these errors or continue

I suppose it would depend on the error. If it is a missing or corrupt sprite, then yes, because something of that nature will crash the game. However, if it is a mini-map or portrait of the wrong size, then no, since those have no apparent effect.

Thanks, Norb. If the player could be automatically given an error report, with some indication of the error and solution needed, that would be enormously useful. I have NO idea whether this is remotely feasible, btw.

Alternatively, some pointers from the team, or tech-savvy players, as to how to diagnose these things ourselves?

They show machines that just don't have all the latest files

Does this mean being patched to the latest version? [We try to check for this]
Or could a player's game show the latest version number and still not "have all the latest files"?

EDIT: just read 'MP Courier Crashing' thread and see that Saddletank is also unsure exactly what Norb means by: They show machines that just don't have all the latest files

It would be really helpful if someone on the team could go into some more detail on this - a guide for not-so-tech-savvy players. Of course, I appreciate you are busy trying to resolve the issues, but I think this would allow the people having the freezes to be more targeted in our testing.

I agree with the above line of thinking - if an inconsistency is trivial and does not affect MP game stability then let the game run.

If its something more significant which your tech skills suggest to you would or should affect MP game stability then it should halt the game, but as Blaugrana says, some useful debug or error code must be produced that will help online players understand why the game crashed and give them a clue what to fix.

This last week the sowmp_log files have been giving me "map mismatch" errors and that is not at all obvious what that is, though one might think so when you first read it.

Map mismatch means that your map files do not exactly match the servers map files.

The sowmp log file is not really useful for players. It saves the last 10mb of all mp traffic. So what I do is compare them. I always need at least two and I make sure that both game were running the same exact way. I can see by comparing them what happened. So now I usually see waiting for courier message. So if I'm lucky I get the log of the guy that sent the message and I check it. If all looks good all I can deduce is the message just didn't reach anyone. Right now I'm working on the theory that the message is too big. We send our packets UDP. That means that they can be dropped. The advantage is that they are fast and do not hold up the game while sending. The disadvantage is that some providers will drop them if they get too big. So I'm going off the theory that they are too big. I'm working on a way to split them up and reassemble them. Hopefully soon we'll have this fix and keep our fingers crossed that this is the issue.

Map mismatch means that your map files do not exactly match the servers map files.

This is a little unimportant but what are "map files" exactly? My experience yesterday is that it isn't obvious. I thought it meant the 3D battlefield space, when it doesn't, it's something else.

Apart from that I think we are definitely getting somewhere.