Modernization Kit Beta Thread 2

#181 Okay I will have to look into that. I compiled it fine but nothing changed, except the interior mapping bug. I will have to clean out the DSO's it looks like. Thanks!

#182 A couple of more questions... anyone got it to run on Vista? And do you plan to add HLSL support??

#183 Hi,

Solved the compilation problems in VS2005, but noticed something in the Demo.


The dust made by the orc running is not correct, at least on my system (Vista 64, 8800gts) could this be a driver issue ?

Here is the Torque Demo.vcproj (VC8) file. I included the cppunit in the include path (but removed the MKTesting.cpp)
password is: vc84tgemk

rapidshare.com/files/66662196/Torque_Demo.zip.html

#184 Hi, I've tried the Modernization Kit with Torque indie 1.5.2 and copied Artur's Torque Demo.vcproj (thanks BTW) because I use Visual Studio 2005 Express. I've overwritten all from the Modernization Kit to my current Torque directory. When I tried clean / rebuild Torque Demo, it was successful. But when I tried to run it, it said this:

---------------------------
Fatal: (c:\torque\tge_1_5_2\engine\modernizationkit\mkmaterial.cpp @ 93)
---------------------------
SubShader: requested sub shader (InteriorParallax20) does not exist.

Why does this happen? Thanks.

#185 @Bullit: It looks like some files in your project are missing. It really isn't that difficult in VS 2005. Just open the normal source code project, copy over all MK files to same directories, and just add them in the project. There are some step-by-step instructions somewhere up higher in this thread.

@ Koushik: I recompiled, replaced the .exe in the example folder on a fresh 1.5.2 installation, ran, nothing different. Deleted .DSO's, still no change. Buildings look weird, and no water reflection. Write down your exact steps please, Koushik.

#186 Well, I figured it must be because I didn't use a new & clean installation one. So I tried re-installing Torque 1.5.2 again and implement the kit again. When I tried the stronghold mission from starter.fps, it said this:

Quote:
---------------------------
Fatal: (c:\torque\tge_1_5_2\engine\interior\interiorrender.cc @ 1240)
---------------------------
Aw, crap.4
---------------------------
Retry Cancel
---------------------------

From console.log:

Quote:
...
*** Initial Control Object
Activating DirectInput...
keyboard0 input device acquired.
Mapping string: Kork to index: 13
Allocating 540928 bytes of data for 8452 verts (goal: 2694) in the interior VBO
332 texgen eqs and 1948 lmtexgen eqs
Fatal: (c:\torque\tge_1_5_2\engine\interior\interiorrender.cc @ 1240) Aw, crap.4
...

#187 @Artur: The dust looks that way on my system too... guess we'll have to make changes ourselves...

@Bullitt: Does your system support Shader model 2.0? Could you describe what you did exactly? There must be something wrong... it works perfectly fine on my system...

@Cory: Ok here goes... first install a fresh copy of TGE somewhere... just replace all the files. Compile the TorqueDemo project... The default torqueDemo.exe in your example folder should be overwritten with the new executable you just compiled. Then, delete your DSOs and your Preference files using the batch script.

Now, replace the script files with the ones in the modernization kit, and run...

I'm pretty certain you haven't replaced your script files . Replace the script files in starter.fps and the demo folder with the ones from the modernization kit.

The reason I had those artifacts was that I HAD replaced the .cs scripts, but torque uses the precompiled DSOs each time you run it, and unless the DSOs aren't there, it doesn't recompile the .cs scripts. So, to see the changes you made in your script, you have to delete the DSOs first, which forces a recompile of all the required scripts.

Hope that solves your problem...

#188 @Bullitt: I get the same error when i'm trying to execute torqueDemo_DEBUG.exe, compile the release version and it should run fine. (Still trying to understand what fails on the debug version)

@Koushik: Could it be something like the SM version ? For instance the HDR on the vitral (the scene before the mirror example) is all pixelazed, but if i'm correct when you use SM 3.0 the HDR is well displayed. Or is it some kind of code (C++/torque script) bug ?

Any news on when the build 011 will be out ?

The HDR scene:

nVidia uses OpenEXR as a standard for the HDR on SM 3.0
www.slizone.com/docs/IO/26862/True_HDR.jpg

The Torque MK HDR Vitral scene
img399.imageshack.us/img399/5008/torquehdrsmlx5.jpg

#189 Well this is what I did:

1. Reinstall Torque 1.5.2 to a new directory
2. Copied all the files from RELEASE152 to the Torque directory
3. Copied Artur \"Fox8\" Ferreira's Torque Demo.vcproj to "Torque_152\VS2005"
4. Opened the solution from VS 2005 Express
5. Changed the working directory of the project to "..\example" (without this, it couldn't find the resources and scripts needed)
6. Clean & Rebuild the whole solution
7. Ran the program

BTW, when I tried to compile and run the release version, it was successful without the AssertFatal dialog box (like Artur). So, the issue only happens in debug mode. I was wondering, was it the Torque Demo.vcproj that cause the error? Because the only similarities between me and Artur is the vcproj file used. Then again if I converted from VS7, it needed the Platform SDK while using Artur's project file doesn't.

My system is:
Core Duo 3GHz
1GB of memory
nVidia Ge Force 8400
Windows XP SP2
VS2005 Express

#190 I'm missing something big, I can tell. I've installed the MK before, but that was in 1.5. I tried to integrate it into 1.5.2, downloading the source, replacing files, and using someone's modified vcproj file from the other thread. But I get over 500 errors on a debug build, starting with

19>ModernizationKit.cpp
19>c:\program files\torque\tge\engine\modernizationkit\mkshadermanager.h(312) : error C2146: syntax error : missing ';' before identifier 'mProgram'
19>c:\program files\torque\tge\engine\modernizationkit\mkshadermanager.h(312) : error C4430: missing type specifier - int assumed. Note: C++ does not support default-int
19>c:\program files\torque\tge\engine\modernizationkit\mkshadermanager.h(312) : error C4430: missing type specifier - int assumed. Note: C++ does not support default-int
19>c:\program files\torque\tge\engine\modernizationkit\mkshadermanager.h(313) : error C2146: syntax error : missing ';' before identifier 'mFragmentShader'
19>c:\program files\torque\tge\engine\modernizationkit\mkshadermanager.h(313) : error C4430: missing type specifier - int assumed. Note: C++ does not support default-int
19>c:\program files\torque\tge\engine\modernizationkit\mkshadermanager.h(313) : error C4430: missing type specifier - int assumed. Note: C++ does not support default-int
19>c:\program files\torque\tge\engine\modernizationkit\mkshadermanager.h(314) : error C2146: syntax error : missing ';' before identifier 'mVertexShader'

And so on.

#191 I've solved my "Assertion" problem. It turns out it has been answered here by Simon Duggan (thanks BTW). My bad for not searching more thoroughly. :D

BTW, in Simon's code:

//	AssertFatal(currRenderBufferPoint < 512 && currIndexPoint < 2048, "Aw, crap.4");
			AssertFatal(currIndexPoint < 4096, "Aw, crap.");

What will the consequences be from Simon's "hack"? I mean, the assertion is there to handle an error, right? So if the parameters for the error handler are somewhat different, would that make the engine / game itself unstable or what?

#192 @Bullitt - actualy if you keeped on reading, ArmedGeek admitted he missed that change when he converted. So the consequences is from changing something that didn't work to something that does work. Btw, the 'hack' is not so much a made up change, I looked at the rest of the code and found a simular change. so its build on a foundation of existing code changes.

#193 Well, what do you know. Clean installs solve all your problems. Time to redo all my custom code...

Is anyone else experiencing disgustingly long times for mission lighting? I loaded A Stronghold with a debug_FastRelight build, and lighting 3 objects of 11 took over five minutes (I gave up and loaded Barebones instead).

#194 Stronghold takes a long time to light, but I don't think I've ever had it take that long!

#195 I submitted a bug to the actual MK website forum that I hope someone can look at.

#196 Well I am very close to just canning the use of this unless someone can help me get past some things. Hopefully someone can.

1. Dedicated servers - I read above that MK and Dedicated servers are a bad mix, has someone successfully used a method that allows you to create dedicated servers once you have installed the MK?

Then there are my two bugs I am having

2. Overexposure when I restart my game. If I clear all previous lighting files and start up with the MK it looks great. However, if I shut down the game and open back up the same map I have this problem.


3. My other issue has to do with DTS objects that have separate peices to them. Some of the pieces don't appear when the map has water in it. I tested this by several different methods but the most conclusive was to take a map that had water in it and was causing the problem. Then I editted out the water block and tried opening the map again and viola, no problem anymore.


I hope you can tell me something that will allow me to continue using the MK since I particularly like it since it doesn't kill the material mapping system.

Thank you in advance.

#197 It works great with dedicated servers, you just need to shut it off on the dedicated server. I did this with a compile directive I use for dedicated server only stuff, but the easiest way to shut it off is in starter.fps/main.cs, make this:

exec("./data/init.cs");
      DRL::setGoalIntensity(0.23);
      DRL::setBiasLimits(-0.05, 0);
      DRL::setBloomMultiplier(0.5);

Look like this:

if (!$Server::Dedicated)
   {
      exec("./data/init.cs");
      DRL::setGoalIntensity(0.23);
      DRL::setBiasLimits(-0.05, 0);
      DRL::setBloomMultiplier(0.5);
   }

That seems to do the trick also. As far as the overexposure goes, can't help you there! If you mean restart from the same game session, the easiest thing to do is force a quit when exiting a map session, but I'm sure if you did a little digging you'd find the real culprit.

#198 Daniel, I noticed that the normal maps and auto parallax shaders get rebuilt every time, it would be beneficial to build them once then move the definitions into some other .cs file and exec it. And turn off the auto in the engine. This speeds up load time significantly!

#199 Dave what were the definitions to do that? I thought one was MK_AUTOGEN_NMAP.

#200 Well I got an official response from Alex about my vehicle problem. It seems that a GL is not working properly after setting up the water block and it ends up affecting the DTS objects rendering.

I guess I should ask before hanging up on this if anyone has managed to update any of this on their own or even dealt with the problem I am having?

Thanks again.