COMPILE: Getting really frustrated with crash

[Locutus]

CoT,
There's plenty of references to Civ3 in the ldl and other text files as well, so ooold news

[The Big Mc]

locutos you did tell that to the files you have to compile first

[J Bytheway]

I suggest then that you locate the definition of DPRINTF and make an analagous version (called RPRINTF, perhaps) which does the same thing for the release and final versions.

[Fromafar]

The following set of files will allow you to use DPRINTF statements in the Release version, provided that you add USE_LOGGING to the preprocessor definitions of the project (tested with the Release version).

USE_LOGGING.zip

Notes:

1. CityData.cpp has been based on the version in the Altered Source thread. If you want to use the original version, just compile without the file in this set. The compiler will report a missing declaration, but that should be easy to solve.
2. If the (amount of) logging makes the crash disappear, you may have to fiddle with the c3debug_SetDebugMask arguments in civ3_main.cpp.

[Locutus]

Thanks, Formafar, those files were most useful.

Edit: nevermind, didn't find the cause of the bug after all.

[Locutus]

D'oh! Stupid me, ignore my last post.

[Fromafar]

Maybe you already had noticed it, but reporting it here just to be sure:

When a program exception occurs, the program will dump its call stack to crash.txt in the logs directory.
For the debug and release versions this will occur automatically. For the final version you have to set EnableLogs to Yes in userprofile.txt.

[MrBaggins]

Decided to uninstall/reinstall VS and all the other components again.

This time the game still crashes , but the debugger doesn't...

The crash is a divide by zero in CPUINF32, called from c3cpu.cpp, something to do with the P4?

I'll give turning off hyperthreading a shot... that doesn't explain why another compile worked for me tho...

*smiles* Well... at least I'm getting somewhere.

[MrBaggins]

Woot! Solved it. (Well.. a workaround, but at least I know where to look)

Just explicitly specified the frequency of the processor, and the thing runs just fine... woot!

Anyone else compiled and running fine with a P4(C) on WinXP?

[kaan]

well im running a Celeron 2GHz on win xp and i have no problems at all.

[Locutus]

Great news MrBaggins

Well, I *did* notice my processor speed being listed as 0MHz in the debug version, so I guess that will work for me as well (I have an Athlon XP 2000+ though).

I'm not at home so I can't test it now, but it will be the first thing I do when I get home. Where did you specify the frequency MrBaggins? In c3cpu.cpp?

[MrBaggins]

in c3cpu_Examine, I changed the g_cpuspeed assignation line to

g_cpuSpeed.norm_freq = 2400;

norm_freq is the only variable in the structure it seems to use later...

Its obviously not a solution, though...

[J Bytheway]

I'm using a P4 1.6GHz, and IIRC it lists my processor speed at something like 1.4GHz...

[Martin Gühmann]

My prosseor is also listed as 0Mhz in the debug version, but if I remember correctly there wasn't such a crash in the last non debug exe unfortunatly this build is not very recently.

-Martin

[Fromafar]

From the cpuinfo/cpuinf32.dll documentation:
While this program works on all current Intel processors,
the code is designed to not analyze processor speed if the
processor is NOT an Intel processor.

For my AMD 1000 MHz, all 4 processor speed values are reported 0. But this does not seem to cause any crashes.