Counter Location and Logifle Divergence

Last week on VASSAL’s ConsimWorld forum (post 6775), I posted the situation where a visible counter was left in one hex when the Logfile showed that it had moved to another. This happened to me twice with two different modules – 3rd edition TRC and Bull Run.

Chris Dodd (post 6778) provided an answer that a divergence had developed between the game state and the initial state. I was unsure whether it was VASSAL or my computer system that made this divergence between the “game state” and the “initial state.” I did not understand Chris Dodd’s explanation.

Someone else indicated that another module was subject to the same problem and that the bug that caused this situation had not been tracked down as yet.

I provide a screen shot (post 6785), Image BR1, of the Bull Run Logfile showing the movement of the unit and the final location of the unit when the Logfile was saved. This image is from the reopened saved Logfile. Image BR2 is from the beginning of the next Logfile. The game was not saved between Logfiles.

Joel Uckelman (post 6778) wanted a list of steps detailing how the divergence developed so that someone else reproduce the situation.

I reviewed the Logfile. The first time I reviewed it I noticed that a sequence:

  • UNDO: UNDO: * P (EV) Gen Evans (0-0-6) moving from [BR R07] to [BR R10]
  • UNDO: * P (EV) Gen Evans (0-0-6) moving from [BR R07] to [BR R10]

These two lines occurred at the point shown by the blue arrow in image BR3. I wrote down the full sequence as I hope to reproduce it to see if the stated problem occurred again in a different Logfile.

However, when I reopened the BR-005.vlog Logfile again, that sequence was missing. The counter was still located at the BR R07 hex from which it was supposed to have moved but did not.

I tried repeatedly in other Logfiles to repeat the written sequence, but could not. And, the counter always moved as indicated by the Logfile. No problem repeated itself.

I am not crazy although it must seem that way. I don’t believe that Logfiles can change one they are written – but this one seems to have done that.

In conclusion, I could not repeat the error or the sequence which I believed to have caused the error. Further, the Logfile seems to have mysteriously corrected, or at least deleted the error sequence, itself although it failed to correct the location of the counter.

I am including the complete Logfile BR-005.vlog. It is used with the AH’s Bull Run v5 module downloaded from VASSAL’s Module Pages. My computer is a Dell Inspiration running Windows 7 Pro.

Second posting adding complete Logfile BR-005.vlog.

Thus spake Charles McLellan:

I tried repeatedly in other Logfiles to repeat the written sequence, but
could not. And, the counter always moved as indicated by the Logfile. No
problem repeated itself.

If you can find a way to reproduce the problem reliably, please do let
us know.


J.

Happened again, but this time with my opponent’s file. Same module but different Logfile. The units that failed to move share moving back a forth with the problem in the former file. I’ll try again later to have the situation repeat itself. We have had lots of UNDOs and movement back and forth, but when this situation occurs appears to be random.

I made a print copy of the Logfile in which the divergence occurred. I then started with the previous Logfile and copied exactly the same moves that created the divergence. No divergence occurred.

Conclusion, either the divergence is total random or the Logfile is not revealing all that takes place during a move. Of course, I know that the Logfile does not record all actions, just those actions that it has been programmed to record. Many actions are not shown – flips of counters, counter movement to tray from right click menu selection, etc.

So, does the Logfile contain other information than that displayed? If not, is there anyway to capture all the action that is occurring?

Thus spake Charles McLellan:

Conclusion, either the divergence is total random or the Logfile is not
revealing all that takes place during a move. Of course, I know that the
Logfile does not record all actions, just those actions that it has been
programmed to record. Many actions are not shown – flips of counters,
counter movement to tray from right click menu selection, etc.

Flips of pieces definitely should appear in logs. Can you show me an
example in which that happens for pieces which aren’t masked?

So, does the Logfile contain other information than that displayed? If
not, is there anyway to capture all the action that is occurring?

Yes, there can be quite a bit which is in a log file which doesn’t show
up in the chat pane. The only way you’d be able to examine that is by
looking at the contents of the log file directly in a text editor.


J.

I have had players report the same problem with the Memoir '44 module – inconsistencies between pieces moved in multiplayer in the log file and on the display – but have not been able to reproduce the problem.