Graphic artifacts in 3d scenes.

[kubecz3k]

Operating system or device - Godot version:
Linux, e12b422, gpu: GTX 1060

Issue description:
After opening materials 3d demo artifacts can be observed:

VIDEO: https://www.youtube.com/watch?v=lT_s3C1VNxc

Also platformer 3d looks weird:

Steps to reproduce:

1. Open materials demo
2. Check Platformer 3d demo

Link to minimal example project:
Use Material demo and eventually Platformer 3D demo

[eon-s]

Same on Master, nvidia proprietary drivers, demo with default settings.

I see excessive light too, lower sky energy results in a new set of amazing glitches.

Some tonemaps are totally broken too (on the demo at least).

On the terminal got many

ERROR: _gl_debug_print: GL ERROR: Source: OpenGL Type: Performance ID: 131186 Severity: Medium Message: Buffer performance warning: Buffer object 34 (bound to NONE, usage hint is GL_DYNAMIC_DRAW) is being copied/moved from VIDEO memory to HOST memory.
At: drivers/gles3/rasterizer_gles3.cpp:122.

And

ERROR: _gl_debug_print: GL ERROR: Source: OpenGL Type: Performance ID: 131218 Severity: Medium Message: Program/shader state performance warning: Vertex shader in program 67 is being recompiled based on GL state.
At: drivers/gles3/rasterizer_gles3.cpp:122.

And

ERROR: shadow_atlas_set_size: Condition ' !li ' is true. Continuing..:
At: drivers/gles3/rasterizer_scene_gles3.cpp:145.
ERROR: getornull: Condition ' !id_map.has(p_rid.get_data()) ' is true. returned: __null
At: core/rid.h:164.

And this at the end

[PropertyValueEvaluator]: Error eval! Error code: 1

[JiahuiCai]

Same issues on Window with GTX1070

[eon-s]

Same glitches on AMDGPU (R5).

[HummusSamurai]

I have similar problems with the 3D Platformer, haven't checked the materials demo.
It's been like this since Saturday (or a bit earlier), I thought it was my fault.

Fedora 26, Nouveau drivers, GT930M.

[himaghnam]

yes it looks like .5 is full metal and roughness in latest

[Listwon]

Same glitches on Intel® HD Graphics 530, Windows 10 Pro x64

[reduz]

do you know when this started? It used to work, did it not?

…

On Thu, Aug 17, 2017 at 9:08 AM, Bartłomiej T. Listwon < ***@***.***> wrote: Same glitches on Intel® HD Graphics 530, Windows 10 Pro x64 — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#10149 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AF-Z23ne8YBvJw-zCEIwck2lgfo-NuWLks5sZC0kgaJpZM4Ovxuk> .

[akien-mga]

It's likely a regression from a couple weeks ago, since it was first reported 10 days ago and then there were many duplicates and confirmations.

That's a job for bisect-master @kubecz3k ;)

[reduz]

any bisecting appreciated!

…

On Thu, Aug 17, 2017 at 9:58 AM, Rémi Verschelde ***@***.***> wrote: It's likely a regression from a couple weeks ago, since it was first reported 10 days ago and then there were many duplicates and confirmations. That's a job for bisect-master @kubecz3k <https://github.com/kubecz3k> ;) — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#10149 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AF-Z214d9rx6YqT2S5-OKjfxev7Wl3BDks5sZDjngaJpZM4Ovxuk> .

[kubecz3k]

With help from @Calinou we did a bisection and it seems that 5c36148 is the first bad commit

[Zireael07]

Why would GLTF support lead to those bugs?

[Calinou]

Why would GLTF support lead to those bugs?

Some shader changes were done in that commit, which can lead to 3D rendering regressions like the one in the issue.