PROJECT: Playtest

[Solver]

So a crash before the you can move a unit or do another thing. In that case we are talking about two different problems.

While the save is still loading.

Yep, but just this one post.

Yeah, but that seems to deal with adding various SLIC functions - unless adding them messes up somehow loading of savegames with old functions. Then again, I could load a game that I created under the previous build, also modded.

I think I just got the idea of what may be causing the crash... in which case it would be a stupid overlook from my part.

[Fromafar]

The load problem occurs when you load a pre-2004.04.17 game (e.g. an original Activision scenario, or a file you have saved yourself) with a post-2004.04.17 playtest build.

Originally posted by Martin Gühmann
2004.04.17
..\ctp2_code\gs\slic\SlicBuiltin.cpp

[...]

Three new slic globals added by Martin Gühmann:
- g.num_of_players: Gets the number of players in the current game
- g.last_player: Gets the last player, the player with the highest index in the game
- g.max_player: Gets the maximum number of players that is in a game possible, currently it is 32 including the Barbarians.

The post-2004.04.17 playtest build tries to read 6 members of the g variable. An old file has only 3, so you end up reading displaced data, and will probably crash shortly afterwards. If you start a new game, and save with the post-2004.04.17 playtest build, it will save the 6 members, and load these without problems.

[J Bytheway]

Originally posted by Solver
I think I just got the idea of what may be causing the crash... in which case it would be a stupid overlook from my part.

Any luck with the idea?

Fromafar: Thanks for the explanation .

[Solver]

Any luck with the idea?

Oh yes. Succesfully loaded a 04-18, GoodMod game in 04-18, and still able to load a 04-12 game from SAP.

[BureauBert]

SLIC issues

I tried to replace the Event:GrantAdvance, which caused argument type mismatch-errors at gamestart, by the function GrantAdvance() as suggested in the post above by Martin Gühmann.
This time I didn't get any error messages with DebugSlic=No but GrantAdvance() in playtest build 2004-04-18 definitely crashes the game to desktop . I commented out all GrantAdvance()-lines in my scripts (actually most of them are not "mine" ) and the game ran pretty well. I arbitarily un-commented one of them: CTD.
Later in the game (when visible wonders should appear beneath my capital ) I noticed, that they couldn't be built because the necessary terraforming did not take place: I am pretty shure that Terraform() does nothing at all, although the SLIC-debugger did not complain about not knowing such a function at gamestart . Maybe it's there but broken.
I am not shure which SLIC-functions are supposed to be re-implemented yet - if someone in charge could eventually publish a list testing would be a bit easier ...

[J Bytheway]

Originally posted by Solver

Any luck with the idea?

Oh yes. Succesfully loaded a 04-18, GoodMod game in 04-18, and still able to load a 04-12 game from SAP.

Originally posted by BureauBert
I tried to replace the Event:GrantAdvance, which caused argument type mismatch-errors at gamestart, by the function GrantAdvance() as suggested in the post above by Martin Gühmann.
This time I didn't get any error messages with DebugSlic=No but GrantAdvance() in playtest build 2004-04-18 definitely crashes the game to desktop . I commented out all GrantAdvance()-lines in my scripts (actually most of them are not "mine" ) and the game ran pretty well. I arbitarily un-commented one of them: CTD.
Later in the game (when visible wonders should appear beneath my capital ) I noticed, that they couldn't be built because the necessary terraforming did not take place: I am pretty shure that Terraform() does nothing at all, although the SLIC-debugger did not complain about not knowing such a function at gamestart . Maybe it's there but broken.
I am not shure which SLIC-functions are supposed to be re-implemented yet - if someone in charge could eventually publish a list testing would be a bit easier ...

In theory all functions except GetMostAtRiskCity (or something like that) which were ever working are now working. I guess what you're experiencing is an error with the reimplimentation.

[Frog_Gamer]

Testing version 2004.04.18

Testing version 2004.04.18

Parameters: Laptop Acer Travelmate 290 Lmi Centrino 1.4Mhz, 512 Mb, 40 Gb, Win Xp Pro

1) I had no problem installing the update over the previous patch, and starting the game. Nice new interface! and great mousewheel support!!

2) When trying to play an old game (started last month) the game crashed with error report for Microsoft which I cancelled.

3) Starting a new game everything was fine, I am up to cavalary, infantry etc and so far no bugs.

[rdorau]

Hi,
I tried the newest version (18.4.) and using the Mod "Apolyton+Good-Mod" (with ModUpdater) I was able to start a game. But afterloading it again the game crashed to desktop.
But perhaps it's a compatibility problem between ModUpdater and the newest testing version of the Apolyton Edition?

[J Bytheway]

Did you get any error message with the crash?

[Martin Gühmann]

Originally posted by Fromafar
The post-2004.04.17 playtest build tries to read 6 members of the g variable. An old file has only 3, so you end up reading displaced data, and will probably crash shortly afterwards. If you start a new game, and save with the post-2004.04.17 playtest build, it will save the 6 members, and load these without problems.

The only problem with this is explaination is that it is not valid for this patch. It is right that I did the globals before John released the latest playtest, but I posted it afterwards, so you can load games saved with previous versions. However I looked into it and so far I have no idea why the global container crashes but the player array with the additional government member does not crash. However I consider this as a serious bug that needs to be removed, even if it means to remove the new globals, alternativly they could be realized as functions, but in my opinion it is more logical to realize the way to get to know how much players the game supports as global then as function, espeacilly if you consider that I just exposed a constant from c++ code to slic.

BureauBert: I did not tough the GrantAdvance function just the HasAdvance function, so nothing should be changed here, and in fact if you can start a new game of GoodMod then you can make the GrantAdvance function work. However there must be another problem if it crashes, do you have the crash.txt?

-Martin

[Martin Gühmann]

Originally posted by rdorau
Hi,
I tried the newest version (18.4.) and using the Mod "Apolyton+Good-Mod" (with ModUpdater) I was able to start a game. But afterloading it again the game crashed to desktop.
But perhaps it's a compatibility problem between ModUpdater and the newest testing version of the Apolyton Edition?

I just started a GoodMod game quit the game and relaoded the autosave, no problems, except that I got a lot of asserts with the debug version and I noticed that the goods weren't improved, but my help tile improvements were on the map.

(So let's see if this posts bumbs the thread.)

-Martin

[BureauBert]

Wonderful world of SLIC

In the meantime I became aware of the fact that GoodMod uses GrantAdvance, too . It is really strange: I replayed GoodMod for a while and it ran with GrantAdvance() and no crash. I tried my hybrid Mod without GrantAdvance(): no crash - until my first attempt to destroy an enemy city: there I had forgotten a GrantAdvance() for my CityRuins-feature: imediate crash. I even compared the lines in GoodMod an in my script:

CityRuins:

GrantAdvance(CIR_PrevOwner, AdvanceDB(ADVANCE_SUBNEURAL_ADS)); 

GoodMod:

GrantAdvance(0, AdvanceDB(ADVANCE_SUBNEURAL_ADS)); 

The only difference is in the first argument, where

CIR_PrevOwner = city[0].owner; 

in my script. But I had also this line crashing the game (in NaturalWonders, immediately at the beginning of old barb's first turn):

GrantAdvance(0, AdvanceDB(ADVANCE_SUBNEURAL_ADS)); 

which is pretty much identical to Martin's. Currently I am clueless on how to track this one down . Obviously GrantAdvance() works in build 2004-04-18, but sometimes it doesn't even where Activision's latest build did .
As for the second function, Terraform(), I haven't yet found a working example to compare, but I will continue research .
Martin: crash.txt - this must be a feature that comes with the debug version of the program, if I had installed the source-code-pack, which is not the case ... but if there is a way to get a crash.txt without running the entire CTP2-AE development environment, please let me know

[J Bytheway]

To get the crash.txt you need to enable logs in your UserProfile.txt, and then look in the ...\ctp2_program\ctp\logs\ directory .

[BureauBert]

Thanks a lot - I will do so and hopefully be able to report something useful. (I opened userprofile yesterday and looked for crash ... crash ... crash, log ... log ... log didn't come to my mind )

[Martin Gühmann]

Originally posted by BureauBert
Thanks a lot - I will do so and hopefully be able to report something useful. (I opened userprofile yesterday and looked for crash ... crash ... crash, log ... log ... log didn't come to my mind )

You have a text editor like notepade with a basic find function. You could use it to look for log.

-Martin