help desk requested post here for developer

[fastfrank]

Now running latest TOAW IV baseline from member's download page. Ran France 1944 Cobra scenario without any edits (release version). Ran with PO running game, with saves at start of each turn for each side. Game vanished on Allied turn 16 at about combat round 6. Has now blown up 7 times out of 7. Repeatable. Tried to get save just prior to blowup. Finally got save close enough that when I tried runninhg again, game ran to normal conclusion, Made me think maybe this is a timing issue with my Intel multiprocessor chip??? Ran save file with OPart slow but blew up as before. To recreate, run save file with Play/Computer Player Western Allies. On my PC, dies at about 14 minutes during combat turn 6 or later. Saves of all prior turns available if useful.

[fastfrank]

save file

[larryfulkerson]

latest TOAW IV baseline

does this mean you haven't applied any of the released patches?

[larryfulkerson]

I downloaded your save game and right now the computer is playing the computer and it's running just fine. I have applied both patches to the game engine. Maybe it's crashing because you're running an old version of the game engine?

[fastfrank]

deleted TOAW IV, then downloaded/installed 4.1.0.20 and then 4.1.0.21

[fastfrank]

which patches?

[larryfulkerson]

Here's a link to this web page:
tt.asp?forumid=1514

[fastfrank]

somewhere I saw instructions to delete TOAW IV, then install 4.1.0.20 as a complete new program and then install 4.1.0.21 as a patch....

[Zovs]

I don't think that is correct, you'll need to install the base game (mine so happened to be the installer with 4.1.0.20 IIANM) and then run the .21 patch.

[Lobster]

ORIGINAL: fastfrank

somewhere I saw instructions to delete TOAW IV, then install 4.1.0.20 as a complete new program and then install 4.1.0.21 as a patch....

If you had a previous version of TOAWIV (pre .20) then you needed to do a complete uninstall. If you are making a new install you naturally would not need to uninstall anything.

[sPzAbt653]

Version 4.1.0.20 Available! [12/6/2018]

PLEASE NOTE: we have discovered an issue with this build and are presently working on a fix. You should not install this in the meantime. If you have already done so, please see HEREfor a manual fix while we work on the patch.
tm.asp?m=4564492

So WHY is this version still available for download? Six months after the fact?

[larryfulkerson]

So WHY is this version still available for download? Six months after the fact?

+1

[Zovs]

I did not think there was any issue with .20 but with the .22 patch and the .21 was the only one that worked so the .22 was pulled. Maybe we need a new .23 to clear it up but I heard of a .30 patch, and there may also be a new TOAW V, so it’s confusing and disorienting with the numbers but it is safe with the .21 patch just avoid the editor bug that CTD and your good.

[Lobster]

ORIGINAL: Zovs

I did not think there was any issue with .20 but with the .22 patch and the .21 was the only one that worked so the .22 was pulled. Maybe we need a new .23 to clear it up but I heard of a .30 patch, and there may also be a new TOAW V, so it’s confusing and disorienting with the numbers but it is safe with the .21 patch just avoid the editor bug that CTD and your good.

There is no TOAWV.

.30 is not a patch and shouldn't even be mentioned here by a tester. It just sows more confusion. As far as I know all there are is .20 and .21 and nothing more.

It seems we have become orphans once again. [:D]

[fastfrank]

I would not be surprised if it will run fine on any machine configuration except mine. Since on the 8th run I stopped to save just before the fatal point and it then ran to end of tern makes me suspect the Intel multi-processor might be experiencing a timing conflict and the pause might have permitted a timing conflict to resolve???? Maybe. Any idea why the help desk said post on the forum for "the developer" to see? Old staff officer saying "an action passed is an action completed"?

[Shadrach]

Catch 22 I guess... I've long suspected Matrix Support to be staffed by ex-military types more concerned with following procedure than solving actual user problems... [8|]
And the developer has gone AWOL, although I suspect he's reading these forums - he just realised that actually interacting with users just takes up too much time. Fair enough.

You are a customer, and as long as Matrix sells this product you are entitled to support. I would keep pestering support until they give you a concrete solution or at the very least a confirmation that the developer knows about the issue.

If the application is "vanishing" (i.e. crashing) there should be crash dumps available. For Windows 10 these are usually located in the folder:
c:\Users\<username>\AppData\Local\CrashDumps
These dumps can give support clues on what might be causing the crashes. Any halfway decent support organisation will have the tools needed to analyse these dumps. Attach a couple of them to your support ticket.

[larryfulkerson]

Actually the crashlog.txt file is found in the root directory of the game.

[Shadrach]

Well, the crashlog.txt would definitely be useful for support to have a look at.

But it's not the same as a crash dump, which is a binary file used for analysing application crashes. You need special tools to get data from a dump, usually something like WinDbg. And you need to actually know the application and the code, so this is usually something only the developer (or a skilled support tech) can do.

I'm not sure though, that TOAW4 creates a dump when crashing...

[Shadrach]

I suspect a memory leak. I actually was able to reproduce the crash here running a couple of turns of the provided save, with AI playing both sides. During the run I was monitoring CPU and memory usage for TOAW4. Memory starts out at about 400MB at the start of the Allied turn, then slowly grows to 1.5GB during the turn. There appears to be no attempt by the game to do some housekeeping to clear memory. Memory is only cleared at the start of a new turn.



This is during AI turns. Not sure if memory grows during a human player's turn or if it's cleared after each round.
It crashes during the Allies' turn because there are more units to move. I would also suspect this is the cause of many crashes people are seeing on the bigger scenarios.

The limit for a 32-bit process is about 2GB unless it's Large Address Aware. It might also be a lot less before a crash, for reasons like virtual memory management.

No crash dump is generated in Windows, but attaching crashLog.txt, even though I doubt it will shed any more light on it.

[fastfrank]

Thanks for the most useful input to date. I tested my hypothesis that a multi-processer timing conflict was at the root of this. Ran turn 16 using the move PO button and then the execute all combats button to create pause times. Crashed on combat round 6. Tried reproducing your research, had to use task manager/processes screen, couldn't figure out how to get your graph. Had 4.6 GB used before launching TOAW, 5.5GB after launch, 6.1 before crash out of 15.9GB available. So don't know if TOAW eventually exhausts available memory or trashed what it is using....

I appreciate your help, thanks. Will reply to the help desk and ask if the developer has looked at this.