January 27, 20215 yr Jean I think the bee hive scenario is still tied to the legacy battleset as the DB is not shown next to the timer on the bottom right of the full screen page. I wonder if you could test the same scenario but from the more recent USNI battle set. The battle set is scenario number nine and the region is the GIUK. I believe it will be the top USNI battle set in the selection window. I believe it has a different battle set then the older USNI which is the bottom one on the choices. It should show 170913 as the battleset as it seems that is locked in. Tony
January 27, 20215 yr Author 1 minute ago, Tony said: Jean I think the bee hive scenario is still tied to the legacy battleset as the DB is not shown next to the timer on the bottom right of the full screen page. I wonder if you could test the same scenario but from the more recent USNI battle set. The battle set is scenario number nine and the region is the GIUK. I believe it will be the top USNI battle set in the selection window. I believe it has a different battle set then the older USNI which is the bottom one on the choices. It should show 170913 as the battleset as it seems that is locked in. Tony There is somewhat more nuance than that. A BattleSet is composed of: 1. The BattleSet specific information like the map, description, weather starting locations, sonar conditions, etc.; 2. Database of Platforms except for commondb BattleSets which share a database between multiple BattleSets; 3. Scenarios. So USNI.RES has its own database embedded in that file, a different database from HDSA.RES. Back then there was much more manual creation of the BattleSets so occasionally there is some data out of place. But what to take from it is that both USNI and HDSA are legacy BattleSets filled with gremlins but between them potentially different gremlins. In addition to gremlins is that the current game structures are different than when that original USNI BattleSet was created so there is code to convert from the old to new structures and sometimes that code is wrong and sometimes it unleashes gremlins that were previously caged by one-off workarounds.
January 27, 20215 yr Author As to the rest, yes, the USNI re-makes are very up to date and use the latest game functionality so unless a player has a particularly fond memory of the original USNI scenarios rather than the re-makes, I for one would suggest playing the re-make.
January 29, 20215 yr same result with GIUKUSNI9.0. autosave.0110.hp0 autosave.0109.hp0 autosave.0108.hp0 autosave.0107.hp0 autosave.0106.hp0
January 29, 20215 yr Jean Thanks for giving it a shot though remote it would be that. At least that was eliminated. I ran you save games and had the same result ran for hours with no crashes.
January 30, 20215 yr Author Jean, if you do Retry to debug, does it happen to give you additional information like a call stack?
January 31, 20215 yr I no longer seem to be getting any crashes, but will try a few more of the supplied files
January 31, 20215 yr On 1/29/2021 at 11:45 PM, TonyE said: Jean, if you do Retry to debug, does it happen to give you additional information like a call stack? Hi Tony, here is what I did get this morning.
February 1, 20215 yr Author No crash after more than an hour of play from autosave.0107.hp0 but I did just use a Nimrod one-shot-one-kill to take out some fish. Great fun to play though! Time to call it a night <g>.
February 1, 20215 yr Author So for me, back to better instrumenting the GE to glean some more information.
February 1, 20215 yr Author Since throwing darts at the wall isn't getting us very far, though it was super fun to challenge those Flankers in the re-worked USNI BattleSet, I'm just starting to compile a list of tools and techniques to troubleshoot. Historically instrumenting the game code along with play testing would isolate issues but memory corruption tends to build up over time and can rarely be tracked back to a single event. Over the next few weeks we should be able to start using these strategies to identify memory issues and begin cleaning them up.
February 2, 20215 yr Author While starting to lay out instructions for using windbg and gflags (from Memory problem debugging) I managed to see some heap corruption on a test machine then managed to get an error on my development machine. I've cleaned up an out of control Staff Assistant and no longer seem to get either of those problems. It may well not fix Jean's problem but it would be great if you can re-test with 2021.001 or newer. Thanks! https://harpgamer.com/harpforum/topic/29035-current-files/
February 5, 20215 yr On 2/1/2021 at 11:26 PM, TonyE said: While starting to lay out instructions for using windbg and gflags (from Memory problem debugging) I managed to see some heap corruption on a test machine then managed to get an error on my development machine. I've cleaned up an out of control Staff Assistant and no longer seem to get either of those problems. It may well not fix Jean's problem but it would be great if you can re-test with 2021.001 or newer. Thanks! https://harpgamer.com/harpforum/topic/29035-current-files/ I did play with new beta. It did took longer to crash and the error message were different. autosave.0049.hp0autosave.0047.hp0autosave.0048.hp0
February 5, 20215 yr Author Thanks Jean, I'm disappointed we didn't magically fix everything but am heartened a bit at the change in error messages.
Create an account or sign in to comment