[RELEASE] WildShips OXP

[JazHaz]

I just docked at a habitat and also an comms array. Both look great, even without shaders.

[mandoman]

Griff, I'm not trying to rain on your parade, but after I added you WildShips oxp, and went out to encounter the cool stuff I'm seeing pictured in this thread, my system crashed. I had to kill the power, reboot, and then I went to Latest.log, and found this. I'm not sure if it's just my computer, or what. Thought you should know.

Code: Select all

20:39:29.054 [shader.compile.failure]: ***** ERROR: GLSL fragment shader compilation failed for wildShips_enginePlasma.fragment:
>>>>> GLSL log:
0:52(52): error: Could not implicitly convert operands to arithmetic operator
0:52(53): error: Operands to arithmetic operators must be numeric
0:52(54): error: Operands to arithmetic operators must be numeric


20:39:29.055 [shader.load.fullModeFailed]: ----- WARNING: Could not build shader wildShips_enginePlasma.vertex/wildShips_enginePlasma.fragment in full complexity mode, trying simple mode.
  20:39:29.063 [shader.compile.failure]: ***** ERROR: GLSL fragment shader compilation failed for wildShips_enginePlasma.fragment:
>>>>> GLSL log:
0:53(52): error: Could not implicitly convert operands to arithmetic operator
0:53(53): error: Operands to arithmetic operators must be numeric
0:53(54): error: Operands to arithmetic operators must be numeric


20:39:29.065 [shader.load.failed]: ***** ERROR: Could not build shader wildShips_enginePlasma.vertex/wildShips_enginePlasma.fragment.
20:39:32.079 [debugTCP.disconnect]: No connection to debug console: "Connection to debug console failed: 'NSPOSIXErrorDomain 111' (outStream status: 7, inStream status: 7)."

[JazHaz]

Griff, I'm not trying to rain on your parade, but after I added you WildShips oxp.....

Who? This is Thargoid's work I believe....!

[mandoman]

Right you are. My apologies.

[Thargoid]

It is, but I may well need his guru advice to try and see what's going wrong there. It looks like somehow something that should be passed to or calculated by the shader as a numeric isn't (or wasn't in this case).

The code section in question is fairly standard:

Code: Select all

{
   vec4 specular = vec4(0.0);
   vec3 eyeVector = normalize(-v_pos);
   
   // Load texture data
   vec2 texCoord = gl_TexCoord[0].st;
   vec4 colorMap = texture2D(uDiffuse, TexScrollcoords.xy);

   // Add flicker when throwing sparks
	float flicker = rand(uSeed.xz);
	if (uNearly_dead)
	{ colorMap *= flicker; }  

    // Create weighting by energy level
    float energyLevel = 0.5 + (uEnergy / (2*uMaxEnergy));	
	
#ifdef OO_LIGHT_0_FIX
   LIGHT(0);
#endif
   LIGHT(1);
   vec4 color = colorMap * energyLevel;
   gl_FragColor = color + (specular * specIntensity);
}

with the line vec4 color = colorMap * energyLevel; near the bottom being where it's choking (and then subsequent ones as color isn't defined properly).

Can you confirm which shader mode you're running under (full or simple) and the exact sequence of events (what were you doing just before the crash - had you just launched, were approaching a Kiota or had just shot one of the WildShips for example).[/color]

[Gimbal Locke]

I have errors from WildShips, it seems these make Oolite crash after dying (see this post).

Using Xubuntu, full shaders on Intel HD 3000 graphics.

EDIT: I see that these are the same errors as the ones Mandoman has. Thargoid has replied to my error log in the other thread (thanks!), so I remove the log here.

[Thargoid]

I just saw the same thing - we cross-posted in each others threads

As a temporary solution, I'd suggest to open the shipdata.plist and remove the shader lines from the definitions for the ships (not the stations, that's a different shader). Specifically lines 35-48, 88-102, 147-161, 202-216 and 252-266. Then do the normal shift-held restart.

It will remove the plasma effect from the ships, but that's preferable to the game hitting the deck...

[mandoman]

with the line vec4 color = colorMap * energyLevel; near the bottom being where it's choking (and then subsequent ones as color isn't defined properly).

Can you confirm which shader mode you're running under (full or simple) and the exact sequence of events (what were you doing just before the crash - had you just launched, were approaching a Kiota or had just shot one of the WildShips for example).[/color]

All I can find on Shaders is that it is supported. It doesn't specify whether it's full, or what ever. All I was doing was coming back from a sun-skim run to refuel, before docking with the Station. I couldn't tell whether it was one of your stations, or not. I would say not, as it just looked like a dodec. to me. I was in torus drive for several seconds, and then the whole thing just froze. I didn't notice anything unusual, so it is a puzzle to me. The ship I was using wasn't even one of my really fast ships, so the torus drive speed should have been within Oolite Core ship speeds. I think it was in the Inzaquma System, in Galactic Sector 2.

[Gimbal Locke]

I just saw the same thing - we cross-posted in each others threads

As a temporary solution, I'd suggest to open the shipdata.plist and remove the shader lines from the definitions for the ships (not the stations, that's a different shader). Specifically lines 35-48, 88-102, 147-161, 202-216 and 252-266. Then do the normal shift-held restart.

It will remove the plasma effect from the ships, but that's preferable to the game hitting the deck...

My preferred solution is not dying.

I have removed the shaders from WildShips as you instructed, I'll report back what happens.

[Thargoid]

You can leave the station shader in place - it's just the ship one that's giving the problem.

It's not something I see here on my system, so it's either something related to the shader mode, or to your chipset/drivers perhaps. Either way not a nice situation, so I've lit the Griff-signal to ask for expert assistance...

I just tried it on my system and I can't reproduce it, even in simple shader mode and dying by picking a fight with a kiota research station. But I'm on an AMD based system, so my driver set is different to yours.

[Ironfist]

Added wildships and I am not seeing any shader errors. The only errors that are in the log are ones about the ball turrets.

Code: Select all

20:45:55.308 [shipData.load.warning.turret.badWeaponRange]: ----- WARNING: ball turret weapon range of 10000 for subenitity of ship wildShips_kiota4Spur is to high, using 7500.
20:45:55.323 [shipData.load.warning.turret.badWeaponRange]: ----- WARNING: ball turret weapon range of 10000 for subenitity of ship wildShips_kiota8RingVariant is to high, using 7500.

This is with trunk R4827 and OpenGL version: 4.2.11399
So my guess this is going to be one of those OpenGL problems.

Ironfist

[Thargoid]

No, that's a (so far) undocumented limit on turret ranges. Given the size of them (and the stations) I tried to use the new feature of being able to properly set the range larger than the 1.76 limit of 6km.

But seemingly the new improved version isn't that improved after all - rather disappointing

[Ironfist]

Thargoid,

Due to the error message I assume the new limit is 7500 in Truck.

Ironfist

[Thargoid]

Yes, that's what I meant.

The ability to set the range is a new feature in post-1.76 (ie trunk at the moment), but it's rather disappointing that it's still limited (albeit slightly higher).

[Eric Walch]

No, that's a (so far) undocumented limit on turret ranges. Given the size of them (and the stations) I tried to use the new feature of being able to properly set the range larger than the 1.76 limit of 6km.

But seemingly the new improved version isn't that improved after all - rather disappointing

The new feature* was not mend to increase the turret range, just to make it customizable. 7500 meters is the range of a plasma weapon as primary weapon. It seems reasonable that a turreted plasma canon should not surpass that range.

* actually not a new feature, but a bugfix, as it was mend to be present in 1.76