Jedi Outcast/Academy source code releases

[leileilol]

http://sourceforge.net/projects/jedioutcast/
http://sourceforge.net/projects/jediacademy/

[Julius]

https://github.com/Razish/OpenJK
https://github.com/mrwonko/Blender-2.6-Ghoul-2-addon

[leileilol]

Where the hell did that Ghoul2 thing come from!?









So....... did anyone get this compiled yet?

Some interesting ideas to do with the source:

• Port tr_flares to Q3
• Attempt to get GAMECUBE compiling working (a bit shocked the console versions' stuff is in this tree)
• Port the environment map calculation code to Q3 (it's a little different)
• Port Jedi Academy's dynamic light code to Q3 (lighting up dark areas is the important part of it)
• A media project that has nothing to do with Star Wars in any way, probably going to be stiff players shooting lasers in box maps at best, only maintaining compatibility with mod codebases and player animations (every saber anim would be a poke... in every stance), disregarding map textures and other stuff, with a lack of sense of direction to shut up warnings.
• Port ICARUS to ioSTVEF to get Elite Force's single player working

[Knightmare]

It also contains what appears to be the source for the random map generator used in SOF2 (under code/rmg). Maybe that could be gotten working- its implementation did use a lot of scripts in SOF2.

More ideas:

• (OpenJK) Implement native widescreen/surround support with properly scaled FOV and anamorphic screen coord scaling (avoid stretched menus/HUD).
• Implement its BSP format in Q3 and/or Q2 (it has the lightstyles from Q1 and Q2 that were removed for Q3).

[leileilol]

Turns out there was a mistake in the source upload. There were no project files, and the JediOutcast code seemed to be JediAcademy code. I was puzzled on why there's HDR bloom in the Jedi Outcast tree


I'm compiling Jedi Outcast now, unfortunately there's an error with the shadercommands typedef in tr_local.h


EDIT: Apparently compiles with Visual C++ 6 SP5 without problems

[scar3crow]

Rich, who worked on these projects, is up to lots of good with it.

http://sourceforge.net/u/richwhitehouse ... ba44/tree/

[Rich]

Here's a tag for Visual Studio 2010 solution+projects with all of the errors/warnings fixed for JAMP:
https://sourceforge.net/u/richwhitehous ... d1e1/tree/

I'd recommend starting from that instead of going back to old versions of MSVC. It also makes more sense for longterm projects to go with JAMP and not JK2 or JASP, JA can be adapted to run JK2 mods and load any content that JK2 can, but backporting JA changes to the JK2 codebase would be a lot more work. JAMP also partially runs JK2 as a coop game already, but people tell me that was broken a bit before it went to retail. Now's a good time to fix that!

I'm going to start working on some crazy threading stuff in my fork and making it run superfast on modern machines, since I feel the highest priority thing here is to have 500 Reborn onscreen at once. Then I'm going to make saber collisions nicer and smoother even at low server framerates (because I was retarded when I was working on that initially), then I'll probably start cleaning stuff up and restructuring the rendering backend to not be horrible. Feel free to let me know if you'd like to work on the fork with me and I'll give you write access.

Edit: Oh, and if you copied the code straight out before last night (as I see in the OpenJK repository) instead of forking, you'll need to remove the bink stuff and feelit libs to avoid legal issues. We had to struggle with Git last night to remove evidence of their existence.

[leileilol]

One thing i've noticed missing is 'fontgen'. I can't find a way to generate new font textures with corresponding fontdat files.

I actually have a stubby placeholder-filled media project in progress, which will have its fate decided whether I can make a working player skeleton and basic animation set with it from scratch (which probably can't guarantee compatibilty with existing _humanoid meshes). To get the engine not to crash i've had to make a few blank strings files and a blank gfx2.shader.

Making efx files is sort of fun. I'm going to make everything bleed, attempt atlasing (putting 4 particle textures on one texture sheet and tcMod scale in their stages to select which), and blending cap fallbacks for ragepro (prefer additive/alphatest) and pcx2/mystique (prefer alpha).

[Rich]

I'll see if James still has fontgen code/binary anywhere. I'd be surprised if a binary wasn't included in the JK2 and/or JA SDK though.

[thelolotov]

It also contains what appears to be the source for the random map generator used in SOF2 (under code/rmg). Maybe that could be gotten working- its implementation did use a lot of scripts in SOF2.

More ideas:

• (OpenJK) Implement native widescreen/surround support with properly scaled FOV and anamorphic screen coord scaling (avoid stretched menus/HUD).
• Implement its BSP format in Q3 and/or Q2 (it has the lightstyles from Q1 and Q2 that were removed for Q3).

Throwing my hat in the ring here, native widescreen/surround/triplehead/eyefinity support would be absolutely incredible. I've submitted a feature request here: https://github.com/Razish/OpenJK/issues/36

So go there and make it known or help, if you can!

[Knightmare]

I've already done native widescreen in my RtCW SP 1.42 unofficial patch, it would probably just be a copy/paste of those changes into the JA source tree. A wide range of resolutions are natively supported, even the new 2560x1080 21:9 ultra-wide aspect. This source can be downloaded here: http://www.markshan.com/knightmare/down ... 42b-src.7z

Here are my vidmodes:

Code: Select all

vidmode_t r_vidModes[] =
{
	{ "Mode  0: 320x240",         320,    240,    1 },
	{ "Mode  1: 400x300",         400,    300,    1 },
	{ "Mode  2: 512x384",         512,    384,    1 },
	{ "Mode  3: 640x480",         640,    480,    1 },
	{ "Mode  4: 800x600",         800,    600,    1 },
	{ "Mode  5: 960x720",         960,    720,    1 },
	{ "Mode  6: 1024x768",        1024,   768,    1 },
	{ "Mode  7: 1152x864",        1152,   864,    1 },
	{ "Mode  8: 1280x960",        1280,   960,    1 },    // Knightmare	added
	{ "Mode  9: 1280x1024",       1280,   1024,   1 },
	{ "Mode 10: 1400x1050",       1400,   1050,   1 },    // Knightmare	added
	{ "Mode 11: 1600x1200",       1600,   1200,   1 },
	{ "Mode 12: 1920x1440",	      1920,   1440,   1 },    // Knightmare	added
	{ "Mode 13: 2048x1536",       2048,   1536,   1 },
	{ "Mode 14: 800x480 (wide)",  800,    480,	  1 },    // Knightmare	added
	{ "Mode 15: 856x480 (wide)",  856,    480,    1 },
	{ "Mode 16: 1024x600 (wide)", 1024,   600,    1 },    // Knightmare	added
	{ "Mode 17: 1280x720 (wide)", 1280,   720,    1 },    // Knightmare	added
	{ "Mode 18: 1280x768 (wide)", 1280,   768,    1 },    // Knightmare	added
	{ "Mode 19: 1280x800 (wide)", 1280,   800,    1 },    // Knightmare	added
	{ "Mode 20: 1360x768 (wide)", 1360,   768,	  1 },    // Knightmare	added
	{ "Mode 21: 1366x768 (wide)", 1366,   768,	  1 },    // Knightmare	added
	{ "Mode 22: 1440x900 (wide)", 1440,   900,	  1 },    // Knightmare	added
	{ "Mode 23: 1600x900 (wide)", 1600,   900,    1 },    // Knightmare	added
	{ "Mode 24: 1600x1024 (wide)", 1600,  1024,   1 },    // Knightmare	added
	{ "Mode 25: 1680x1050 (wide)", 1680,  1050,   1 },    // Knightmare	added
	{ "Mode 26: 1920x1080 (wide)", 1920,  1080,   1 },    // Knightmare	added
	{ "Mode 27: 1920x1200 (wide)", 1920,  1200,   1 },    //----(SA)	added
	{ "Mode 28: 2560x1080 (ultra-wide)", 2560,  1080,   1 },    // Knightmare	added
	{ "Mode 29: 2560x1440 (wide)", 2560,  1440,   1 },    // Knightmare	added
	{ "Mode 30: 2560x1600 (wide)", 2560,  1600,   1 }     // Knightmare	added
};

The largest change is an overhaul of the AdjustFrom640 function (RtCW has 4 redundant versions of this that had to be changed) to take an enumerated alignment type parameter that determines if an object is centered, aligned to the center top, bottom right, stretched, etc. All menu items are centered by default, and an "scrAlign" attribute can be added to menu scripts to set the alignment (good thing the menu files are included with the GPLed source). The alignment parameter had to be added to lots of 2D drawing functions.

Add to tr_types.h:

Code: Select all

// Knightmare- screen item alignment types
typedef enum
{
	ALIGN_STRETCH,
	ALIGN_CENTER,
	ALIGN_LETTERBOX,
	ALIGN_TOP,
	ALIGN_BOTTOM,
	ALIGN_RIGHT,
	ALIGN_LEFT,
	ALIGN_TOPRIGHT,
	ALIGN_TOPLEFT,
	ALIGN_BOTTOMRIGHT,
	ALIGN_BOTTOMLEFT,
	ALIGN_TOP_STRETCH,
	ALIGN_BOTTOM_STRETCH
} scralign_t;

Here's my modified SCR_AdjustFrom640 in cl_scrn.c:

Code: Select all

void SCR_AdjustFrom640( float *x, float *y, float *w, float *h, scralign_t align ) {
	float	xscale, yscale, avgscale, vertscale;	// Knightmare added
	float	tmp_x, tmp_y, tmp_w, tmp_h;				// Knightmare added

	// scale for screen sizes
	xscale = cls.glconfig.vidWidth / 640.0;
	yscale = cls.glconfig.vidHeight / 480.0;
	avgscale = min (xscale, yscale);

	// hack for 4:3 modes
	if ( !(xscale > yscale) && align != ALIGN_LETTERBOX)
		align = ALIGN_STRETCH;

	// Knightmare- added anamorphic code
	switch (align)
	{
	case ALIGN_CENTER:
		if (x) {
		tmp_x = *x;
			*x = (tmp_x - (0.5 * SCREEN_WIDTH)) * avgscale + (0.5 * cls.glconfig.vidWidth);
		}
		if (y) {
			tmp_y = *y;
			*y = (tmp_y - (0.5 * SCREEN_HEIGHT)) * avgscale + (0.5 * cls.glconfig.vidHeight);
		}
		if (w) 
			*w *= avgscale;
		if (h)
			*h *= avgscale;
		break;
	case ALIGN_LETTERBOX:
		// special case: video mode (eyefinity?) is wider than object
		if ( w != NULL && h != NULL && ((float)cls.glconfig.vidWidth / (float)cls.glconfig.vidHeight > *w / *h) ) {
			tmp_h = *h;
			vertscale = cls.glconfig.vidHeight / tmp_h;
			if (x != NULL && w != NULL) {
				tmp_x = *x;
				tmp_w = *w;
				*x = tmp_x * xscale - (0.5 * (tmp_w * vertscale - tmp_w * xscale));
			}
			if (y)
				*y = 0;
			if (w) 
				*w *= vertscale;
			if (h)
				*h *= vertscale;
		}
		else {
			if (x)
				*x *= xscale;
			if (y != NULL && h != NULL)  {
				tmp_y = *y;
				tmp_h = *h;
				*y = tmp_y * yscale - (0.5 * (tmp_h * xscale - tmp_h * yscale));
			}
			if (w) 
				*w *= xscale;
			if (h)
				*h *= xscale;
		}
		break;
	case ALIGN_TOP:
		if (w) 
			*w *= avgscale;
		if (h)
			*h *= avgscale;
		if (x) {
			tmp_x = *x;
			*x = (tmp_x - (0.5 * SCREEN_WIDTH)) * avgscale + (0.5 * cls.glconfig.vidWidth);
		}
		if (y)
			*y *= avgscale;
		break;
	case ALIGN_BOTTOM:
		if (w) 
			*w *= avgscale;
		if (h)
			*h *= avgscale;
		if (x) {
			tmp_x = *x;
			*x = (tmp_x - (0.5 * SCREEN_WIDTH)) * avgscale + (0.5 * cls.glconfig.vidWidth);
		}
		if (y) {
			tmp_y = *y;
			*y = (tmp_y - SCREEN_HEIGHT) * avgscale + cls.glconfig.vidHeight;
		}
		break;
	case ALIGN_RIGHT:
		if (w) 
			*w *= avgscale;
		if (h)
			*h *= avgscale;
		if (x) {
			tmp_x = *x;
			*x = (tmp_x - SCREEN_WIDTH) * avgscale + cls.glconfig.vidWidth;
		}
		if (y) {
			tmp_y = *y;
			*y = (tmp_y - (0.5 * SCREEN_HEIGHT)) * avgscale + (0.5 * cls.glconfig.vidHeight);
		}
		break;
	case ALIGN_LEFT:
		if (w) 
			*w *= avgscale;
		if (h)
			*h *= avgscale;
		if (x)
			*x *= avgscale;
		if (y) {
			tmp_y = *y;
			*y = (tmp_y - (0.5 * SCREEN_HEIGHT)) * avgscale + (0.5 * cls.glconfig.vidHeight);
		}
		break;
	case ALIGN_TOPRIGHT:
		if (w) 
			*w *= avgscale;
		if (h)
			*h *= avgscale;
		if (x) {
			tmp_x = *x;
			*x = (tmp_x - SCREEN_WIDTH) * avgscale + cls.glconfig.vidWidth;
		}
		if (y)
			*y *= avgscale;
		break;
	case ALIGN_TOPLEFT:
		if (w) 
			*w *= avgscale;
		if (h)
			*h *= avgscale;
		if (x)
			*x *= avgscale;
		if (y)
			*y *= avgscale;
		break;
	case ALIGN_BOTTOMRIGHT:
		if (w) 
			*w *= avgscale;
		if (h)
			*h *= avgscale;
		if (x) {
			tmp_x = *x;
			*x = (tmp_x - SCREEN_WIDTH) * avgscale + cls.glconfig.vidWidth;
		}
		if (y) {
			tmp_y = *y;
			*y = (tmp_y - SCREEN_HEIGHT) * avgscale + cls.glconfig.vidHeight;
		}
		break;
	case ALIGN_BOTTOMLEFT:
		if (w) 
			*w *= avgscale;
		if (h)
			*h *= avgscale;
		if (x)
			*x *= avgscale;
		if (y) {
			tmp_y = *y;
			*y = (tmp_y - SCREEN_HEIGHT) * avgscale + cls.glconfig.vidHeight;
		}
		break;
	case ALIGN_TOP_STRETCH:
		if (w) 
			*w *= xscale;
		if (h)
			*h *= avgscale;
		if (x)
			*x *= xscale;
		if (y)
			*y *= avgscale;
		break;
	case ALIGN_BOTTOM_STRETCH:
		if (w) 
			*w *= xscale;
		if (h)
			*h *= avgscale;
		if (x)
			*x *= xscale;
		if (y) {
			tmp_y = *y;
			*y = (tmp_y - SCREEN_HEIGHT) * avgscale + cls.glconfig.vidHeight;
		}
		break;
	case ALIGN_STRETCH:
	default:
		if (x)
			*x *= xscale;
		if (y) 
			*y *= yscale;
		if (w) 
			*w *= xscale;
		if (h)
			*h *= yscale;
		break;
	}
}

FOV also needs to be scaled up for wider aspects.
Added define near the top of cg_view.c:

Code: Select all

#define STANDARD_ASPECT_RATIO ((float)640/(float)480)	// Knightmare added

Here's the addition needed in CG_CalcFOV:

Code: Select all

	// Knightmare- adjust fov_x for wide screen aspect
	if (cg_widescreen_fov.value)
	{
		float aspectRatio = (float)cg.refdef.width/(float)cg.refdef.height;
		if (aspectRatio > STANDARD_ASPECT_RATIO)
			fov_x = RAD2DEG( 2 * atan( (aspectRatio / STANDARD_ASPECT_RATIO) * tan(DEG2RAD(fov_x) * 0.5) ) );
		fov_x = min(fov_x, 160);
	}
	// end Knightmare
	
	x = cg.refdef.width / tan( fov_x / 360 * M_PI );
	fov_y = atan2( cg.refdef.height, x );
	fov_y = fov_y * 360 / M_PI;

Also, SCR_DrawScreenField is bugged in widescreen modes. The code that's supposed to clear the sides of pillarboxed cinematics ends up drawing over them, so just comment this out:

Code: Select all

	// wide aspect ratio screens need to have the sides cleared
	// unless they are displaying game renderings
#ifndef _XBOX
	// Xbox no want this
	if ( cls.state != CA_ACTIVE ) {
		if ( cls.glconfig.vidWidth * 480 > cls.glconfig.vidHeight * 640 ) {
			re.SetColor( g_color_table[0] );
			re.DrawStretchPic( 0, 0, cls.glconfig.vidWidth, cls.glconfig.vidHeight, 0, 0, 0, 0, 0 );
			re.SetColor( NULL );
		}
	}
#endif

Instead, CIN_DrawCinematic in cl_cin.cpp needs to be changed to do pillarboxing:

Code: Select all

void CIN_DrawCinematic( int handle ) {
	float	barheight, barwidth, vw, vh;	// Knightmare added
	.
	.
	.
	x = cinTable[handle].xpos;
	y = cinTable[handle].ypos;
	w = cinTable[handle].width;
	h = cinTable[handle].height;
	
	// Knightmare- use letterbox scaling if specified
	if ( cinTable[handle].letterBox )
		SCR_AdjustFrom640( &x, &y, &w, &h, ALIGN_LETTERBOX );
	else
		SCR_AdjustFrom640( &x, &y, &w, &h, ALIGN_CENTER );

	vw = (float)cls.glconfig.vidWidth;
	vh = (float)cls.glconfig.vidHeight;
	.
	.
	.
	// Knightmare- add pillarbox bars when needed
	barwidth = 0.5 * (vw - w);
	barheight = vh;
	if (barwidth > 0) {
		re.SetColor( &colorBlack[0] );
		re.DrawStretchPic( 0, 0, barwidth, barheight, 0, 0, 0, 0, cls.whiteShader );
		re.DrawStretchPic( vw - barwidth - 1, 0, barwidth + 1, barheight, 0, 0, 0, 0, cls.whiteShader );
	}
	// end Knightmare

	if ( cinTable[handle].dirty && ( cinTable[handle].CIN_WIDTH != cinTable[handle].drawX || cinTable[handle].CIN_HEIGHT != cinTable[handle].drawY ) ) {
		.
		.
		.
		.
	}
	.
	.
}

[leileilol]

(good thing the menu files are included with the GPLed source)

The menu scripts are not included for me.

[Knightmare]

The zip I downloaded has them under base/ui.