Let's discuss jMonkeyEngine 3.3!

sgold · May 22, 2019, 6:34pm

Wait! @pspeed integrated a PR, and now I’m seeing an amber disc in the GitHub commit history. Fingers crossed…

pspeed · May 22, 2019, 6:34pm

Yeah, for some reason, old messages are now coming through on the slack feed… so I wonder if something was log-jammed somewhere.

sgold · May 22, 2019, 6:50pm

Both tagged commits have now been tested, but deployment was skipped both times:
Travis CI - Test and Deploy Your Code with Confidence
Travis CI - Test and Deploy Your Code with Confidence

I plan to try again after lunch.

sgold · May 22, 2019, 7:28pm

GitHub may still be confused: it’s showing an amber-colored disc for commit 6a71a1c and saying that checks haven’t completed, even though Travis says they have. I’ll check back in an hour or 2.

sgold · May 22, 2019, 10:02pm

The JDK11 build is broken, but this time we got some deployment, I think. @pspeed, what do your elf eyes see?

pspeed · May 22, 2019, 10:26pm

I just published 166 artifacts that were waiting.

sgold · May 22, 2019, 11:13pm

The alpha2 release is available from BinTray:
Version jme3-core/3.3.0-alpha2 - jmonkeyengine

oxplay2 · May 23, 2019, 9:55am

already using it, and seems it work fine

immidietly checked PBR AmbientLight, and it works perfect, nice.

private server repo proxy already cached it too.(better to have, than like jBullet source issues later ;))

thanks!

Ali_RS · May 23, 2019, 10:30am

Thank you, guys.

oxplay2 · May 24, 2019, 11:53am

recently i asked @pspeed how to do normal intensity for normalmaps.

NormalType can be misleading for some people who dont know what it is. i also thought its intensity before, while having NormalIntensity would allow change true intensity and and fix misleading of understanding NormalType param.

Why not just add it into JME Materials, at least for PBR?

code is simple and do not slow down rendering if not used.

Still it was told its not fully correct way, but tools like Blender or other engines i seen thare are “strength” or “intensity” so why not in JME. if need some fixes in below code, then could be done too. (because im not sure why need /= instead of *= because it work opposite.

in fragment shader add:

  #if defined(NORMAL_INTENSITY)
    normal.z /= m_NormalIntensity;
    normal = normalize(normal);
  #endif

in

// ***********************
// Read from textures
// ***********************
#if defined(NORMALMAP)
  vec4 normalHeight = texture2D(m_NormalMap, newTexCoord);
  //Note the -2.0 and -1.0. We invert the green channel of the normal map, 
  //as it's complient with normal maps generated with blender.
  //see http://hub.jmonkeyengine.org/forum/topic/parallax-mapping-fundamental-bug/#post-256898
  //for more explanation.
  vec3 normal = normalize((normalHeight.xyz * vec3(2.0, NORMAL_TYPE * 2.0, 2.0) - vec3(1.0, NORMAL_TYPE * 1.0, 1.0)));
  #if defined(NORMAL_INTENSITY)
    normal.z /= m_NormalIntensity;
    normal = normalize(normal);
  #endif
  normal = normalize(tbnMat * normal);
  //normal = normalize(normal * inverse(tbnMat));
#else
  vec3 normal = norm;
#endif

also add

uniform float m_NormalIntensity;

and in material definition file

    Float NormalIntensity : 1.0

and

    Defines {     
        NORMAL_INTENSITY: NormalIntensity

@Ali_RS, maybe you could prepare some commit/pull request to make it easier? i dont have master JME to do so.

here some screens that it work fine:

normalintensity1 normalintensity2

it even allow easier debug of normals (intensity increase intensity for any face direction, not just front)

sgold · May 31, 2019, 3:45pm

For feature requests, please open an issue at GitHub
Issues · jMonkeyEngine/jmonkeyengine · GitHub
and/or create a new “Development” topic at this Forum
https://hub.jmonkeyengine.org/c/development-discussion-jme3

sgold · September 15, 2019, 10:20pm

Unless there are objections, I plan to cut a v3.3.0-alpha3 release of the Engine sometime in the coming week.

But first, there are issues to fix and PRs to integrate…

Ali_RS · September 16, 2019, 4:20am

Yes, one such PR which I hope it shows up in next release is GLES 3 support which seems almost ready. (thanks to @joliver82 and @RiccardoBlb )

github.com/jMonkeyEngine/jmonkeyengine

Android openGL ES 3 support

jMonkeyEngine:master ← joliver82:android-gl3

opened 12:44PM - 18 Jul 19 UTC

joliver82

+664 -238

Added gles30 and newer functionality for android. All the information at the for…um thread: https://hub.jmonkeyengine.org/t/opengl-es-3-0-and-newer-on-android/41913 Chage list: - Initialize the proper opengl context (2.0, 3.0, 3.1…) - Added Caps for all of the new supported features - Added Instancing as core feature of GLES30 and newer - Properly adding versioning to shaders (#version xxx es) to shaders - Added GLSL320,GLSL310 and/or GLSL300 to most materials depending on the required level for them to work properly - Added depth24 texture usage if hardware supports it - Enabled Geometry shaders and tessellation (3.2 core feature and 3.1 as extension) and fixed compilation errors on the examples - Enabled glBlitFramebuffer - Added GLSL300 to filters not able to work with GLSL100. - Created Glsl300ShaderGenerator to support GLSL300 in shadernodes - Fixed most shader compilation issues (strict type checking, adding OpenGL ES specific extensions, function usage depending on version and some other) - Set default precision to highp solving most filter issues although maybe a little slower. Added specific sampler precision definition for gles300 or newer - Make use of GL_TEXTURE_MAX_LEVEL in GLES300 to avoid black rendering of non complete mipmapped textures - Changed some assets from compressed formats not supported in android to RGB/RGBA depending on the needs so all effects and tests can run also on android - updated precompiled android.jar to have access to newest android API - enabled some other core features of GLES300, GLES310 and GLES320 - Fixed build issues in jme3-android-examples - Added 3D texture and texture array support - Added float texture formats, fixed rgb8 format if having cap rgb8 and fixed glTexImage2D call on AndroidGL - PBRLighting works and any other effect requiring HDR texture formats should work also - Effects and postprocessing - Multisampling Known issues/pending work: - Fixed most issues with shadow rendering in android (GLES 2.0 and GLES 3.0 or better) although some glitches still apear randomly depending on ShadowMode of the nodes and the device - Fixed water filter. Still some rendering glitches in my old redmi3 (gles3.0). The water renders properly but the rest of the scene is overwritten in red. Not reproduced in any other device I've tried

Ali_RS · September 16, 2019, 4:39am

btw, I keep updating the important change logs @nehon created for 3.3 on the Google Doc.

sgold · September 16, 2019, 4:49am

Thank you. I’ve linked that document from the 3.3.0-alpha1 and -alpha2 release descriptions.

sgold · September 19, 2019, 5:27am

3.3.0-alpha3 is rapidly coming together:

It will have a 100% up-to-date jme3-bullet-native-android library, thanks to @darkchaos, who figured out how to build Android native libraries at TravisCI.
It will have the capability to serialize classes with protected no-arg constructors, thanks to timely reviews of PR 1181 by @Ali_RS, @murph9, and @pspeed.
It will fix long-standing issues with terrain picking thanks to @darkchaos, who submitted PR 1049 and patiently waited almost 6 months for it to be integrated.
It will include many other enhancements and fixes, thanks to @pspeed, @RiccardoBlb, @grizeldi, @Ali_RS, @louhy, @Empire_Phoenix, @tlf30, @remy_vd, @richtea, @karanehra, @jayfella, and others.

sgold · September 19, 2019, 5:37am

How soon can this be finished?

sgold · September 21, 2019, 1:24am

Once again, Travis-CI believes a tagged commit “is not a tagged commit.” But I think alpha3 is now up at Bintray. I’ll double-check after dinner.

pspeed · September 21, 2019, 1:29am

I don’t see anything to approve yet.

sgold · September 21, 2019, 1:30am

I now have access to the approval button. I think I beat you it.