PROJECT: Playest II

[Martin Gühmann]

Originally posted by E
but it doesnt allow for sprites or tiles or other stuff.

What doesn't allow for sprites or tiles or other stuff? ModSwapper only allows one tile file and a set of 200 sprites for each type (or 500 when I increase the number) whereas the scenario setup allows 200 sprites and one tile file for each scenario, BlueO's scenario is an example for different sprites for given sprite numbers and Activision WW2 scenario is an example usage of a different tile file.

-Martin

[Ekmek]

Ack!

Why the have we been using modswapper then?

[Martin Gühmann]

Originally posted by E
Ack!

Why the have we been using modswapper then?

Read the last post of this thread from me you quoted. There I gave the reasons.

The scenario setup of the original game was buggy, you got corupted savegames and that is still true for autosaves and whenever you loaded a scenario savegame you got an extra turn, meaning movement of your units were reset, you got extra science, production, food and gold, just like it would have been if you ended the turn and all the AIs did their turn, but you got it without having the AIs their turn.

-Martin

[mog]

Bugged Savegame

Hi,

if you load this save game and press next turn button, the game crashed.

Maybe this could help to fix another bug

mog

[J Bytheway]

Originally posted by Martin Gühmann
That could be the reason why the city capture option causes a resync when you capture a city, you are not the host and tries to enslave the whole popultaion of the city. The result after the resync is that you still have the city.

That problem is (probably) caused by the fact that code executed by pressing buttons on a messagebox is run only locally, not on the other machines.

[Martin Gühmann]

Originally posted by J Bytheway
That problem is (probably) caused by the fact that code executed by pressing buttons on a messagebox is run only locally, not on the other machines.

So you mean if I use a global variable that is set in the messagebox but is used in the original event handler to determine which of the options should be executed might work.

-Martin

[Fromafar]

Re: Bugged Savegame

Originally posted by mog
if you load this save game and press next turn button, the game crashed.

When I load this save game, I get an error message: "Unable to locate the scenario that created this game. Looking for: Welt."

Is this just the German version of the Activision WorldMaps scenario, or is it a personal modification?
Could you post the contents of your the Scenarios\Welt directory tree?

[Martin Gühmann]

Re: Re: Bugged Savegame

Originally posted by Fromafar

When I load this save game, I get an error message: "Unable to locate the scenario that created this game. Looking for: Welt."

Is this just the German version of the Activision WorldMaps scenario, or is it a personal modification?
Could you post the contents of your the Scenarios\Welt directory tree?

Here is the scebario, actual I justed put the SCEN0000 folder of the WorldMaps scenario folder together so that you have to put this folder into the WorldMaps folder. I also did a test run, looks like I need to relad slic to fix the messages. However I got three asserts during the AI turns. It looks like dereferencing a unit that has been deleted. I also added my logs to the SCEN0000 folder.

-Martin

[Colwyn]

Post the ctp2.exe file again with 5 stack

Hi martin/other.

There is something definetly wrong with the posted 5 stack file as it ctd all the time.

It would be greatly appreciated if you could compile it again with 5-7 max stack size.

The normal play test works ok so i'm guessing its just a corruption.

Eagerly awaiting....

[Martin Gühmann]

Re: Post the ctp2.exe file again with 5 stack

Originally posted by Colwyn
Hi martin/other.

There is something definetly wrong with the posted 5 stack file as it ctd all the time.

It would be greatly appreciated if you could compile it again with 5-7 max stack size.

The normal play test works ok so i'm guessing its just a corruption.

Eagerly awaiting....

Or it is an issure with the five stack stuff, however releasing a new five stacked executable would also make me to release a new playtest version. Unfortunatly afterwards I added all the the lastest updates about the .NET compatibility I am unable to enter the lobby of MP and in addition I am working on something to add some more atmosphere to the game, and I would like to release the next playtest with these enhancements.

-Martin

[Martin Gühmann]

OK, here I have the content of two crash.txt the first one happend yesterday:

  0x005e282a  [?GetName@?$SlicDBConduit@VDiplomacyProposalRecord@@UDiplomacyProposalRecordAccessorInfo@@@@UAEPBDXZ + 0xa]
  0x0066dfdc  [?HandleMouseEvent@aui_Region@@QAE?AW4AUI_ERRCODE@@PAUaui_MouseEvent@@H@Z + 0x37]
  0x0066e07c  [?HandleMouseEvent@aui_Region@@QAE?AW4AUI_ERRCODE@@PAUaui_MouseEvent@@H@Z + 0xd7]
  0x00682774  [?HandleMouseEvents@aui_UI@@QAE?AW4AUI_ERRCODE@@JPAUaui_MouseEvent@@@Z + 0x115]
  0x006830b2  [?Process@aui_UI@@UAE?AW4AUI_ERRCODE@@XZ + 0x23]
  0x0040f3c5  [?ProcessUI@CivApp@@QAEJKAAK@Z + 0x200]
  0x0040f71b  [?Process@CivApp@@QAEJXZ + 0xa7]
  0x00407f3d  [?CivWinMain@@YGHPAUHINSTANCE__@@0PADH@Z + 0x3ca]
  0x004079ee  [WinMain@16 + 0x74]
  0x00860e3c  [WinMainCRTStartup + 0x134]
  0xbff8b560  [(kernel) + 0x0]

It happend IIRC during accessing the science victory screen.

The second one happend during an enemies turn, the coverage map of the science victory already looked a little bit wired.

  0x00826ca4  [?Get_Goal_Type@Goal@@QBECXZ + 0xa]
  0x0081c074  [?Prioritize_Goals@Scheduler@@QAE_NXZ + 0x8b]
  0x0081b7a5  [?Process_Goal_Changes@Scheduler@@QAE?AW4TIME_SLICE_STATE@1@XZ + 0x3c]
  0x0083b574  [?GEVHookCallback@CtpAi_BeginSchedulerEvent@@EAE?AW4GAME_EVENT_HOOK_DISPOSITION@@W4GAME_EVENT@@PAVGameEventArgList@@@Z + 0xa2]
  0x005ad9ad  [?Run@GameEventHook@@QAE?AW4GAME_EVENT_ERR@@PAVWalker@?$PointerList@VGameEventHookNode@@@@PAVGameEventArgList@@AAJ@Z + 0x36]
  0x005ad913  [?Activate@GameEventHook@@QAE?AW4GAME_EVENT_ERR@@PAVGameEventArgList@@AAJ@Z + 0x35]
  0x005ae3e1  [?ActivateHook@GameEventManager@@QAE?AW4GAME_EVENT_ERR@@W4GAME_EVENT@@PAVGameEventArgList@@AAJ@Z + 0x38]
  0x005ac24d  [?Process@GameEvent@@QAE?AW4GAME_EVENT_ERR@@AAH@Z + 0x65]
  0x005ae241  [?ProcessHead@GameEventManager@@QAE?AW4GAME_EVENT_ERR@@XZ + 0x33]
  0x005ae189  [?Process@GameEventManager@@QAE?AW4GAME_EVENT_ERR@@XZ + 0x83]
  0x0040f777  [?Process@CivApp@@QAEJXZ + 0x103]
  0x00407f3d  [?CivWinMain@@YGHPAUHINSTANCE__@@0PADH@Z + 0x3ca]
  0x004079ee  [WinMain@16 + 0x74]
  0x00860e3c  [WinMainCRTStartup + 0x134]
  0xbff8b560  [(kernel) + 0x0]

Unfortunatly I cannot load neither the autosave nor the quicksave both are corupted and unfortunatly I last regular save is some hours old.

From the save game it looks like that there is an additional player in or the data for one player got lost. I don't know whether the save game can be recued, if it can I think huge casulties must be accepted. For anyone who wants to figure something out about the problem I attached the quicksave to this thread. It is played with a playtest version includeing all the changes until my last post in the altered source thread, except the memory leak fix about the status bar, as this "fix" did hold what I saw in it.

-Martin

[Fromafar]

Apparently the recent changes are not stable enough. I too, am getting all kinds of crashes - mostly failing memory allocations (i.e. new fails). I am going to rollback to a more stable version, and reinsert the modifications one by one to find out what the problems are.

One obvious problem is the increase of the number of sprites from 200 to 500. The code in UnitActor::Serialize expects a uint8 sprite index, so anything over 255 may have unexpected results.

[Martin Gühmann]

Originally posted by Fromafar
Apparently the recent changes are not stable enough. I too, am getting all kinds of crashes - mostly failing memory allocations (i.e. new fails). I am going to rollback to a more stable version, and reinsert the modifications one by one to find out what the problems are.

How far do you want to go back? One thing was the status bar issure, that should be now ok. I don't think that calvitex lastest source modifications are a problem either, as I playered the game whose quicksave I attached without these modifications about the AI.

The modification of the dbgen\MemberClass.cpp could be a source of the problem even if it is not very likely, as I just have copied and pasted some code from the main record generator. The Record classes themselves shouldn't be a problem as they are generated automaticly. So only the member class part can here the problem if at all.

UnitActor.cpp might be another source, in addition to the problem below in it, but I added default values to the new fields in AgeCityStyleDB. So this is also less likely. And also the changes in greatlibrary.cpp as they just increase the size of an array shouldn't be the problem, agesscreen.cpp and gameselectwindow.cpp are used when MP games are setup. So that wouldn't be the first place either I would look for. scenarioeditor.cpp also just contains some additional code that exists in another form in the same file. Not an candiate on my list either.

Originally posted by Fromafar
One obvious problem is the increase of the number of sprites from 200 to 500. The code in UnitActor::Serialize expects a uint8 sprite index, so anything over 255 may have unexpected results.

That is something odd, m_spriteID is a long but it is saved as unsigned but is saved as unsigned char. Fixing this and allowing us to use an higher sprite limit then 255 means changing the save game format. If we have to do this we should discuss first what we should also include in a new save game format, we need something so that the game remembers whether the Cheat Editor was used and maybe we need something else.

-Martin

[Fromafar]

Went back to the 2004.08.01 source collection, and found that some of the problems were there already - and probably have been since the start of the project . OTOH, the failing new (out of memory) problem did not appear anymore. So there may be still be a problem with all the array size increases.

I have readded all changes of ctplinuxfan and myself, and Martin's statusbar fix. All without any obvious problems.

So, the remaining files to include are the citymod3, the personality&strategy, and the goal&strategy changes.
So far, the citymod3 gameselectwindow.cpp is OK, and so are citystyle.cdb and scenarioeditor.cpp. Seeing that it would not be too difficult, I just couldn't resist making these backwards compatible again.

More later.

[Fromafar]

Better testing revealed even more errors. Strangely enough, with the debug version there were no problems at all. But when playing with a release version, just starting a game could generate a crash.

The following version has the changes of the previous post, Martin's 2004.09.04.dotnetleftover changes, and my 2004.09.07.newgame changes.

Fixed: crashes when starting a new game
Fixed: crashes with illegal units
Fixed: crashes with missing diplomat pictures
Fixed: always display the newest message
Added: option to display the newest message at the top
Fixed: crashes when loading mod files
Fixed: assignment of civilisations at the start of the game
Added: option to define icons in citystyle.txt
Fixed: user profile settings are used for MP games
Fixed: memory leaks (status bar, starting a new game)

ctp2_program\ctp\ctp2.exe
ctp2_program\ctp\ctp2.map

[Edit: links removed. Newer version available.]