And yes, I can see the problem in the file from the report (it is just that I cannot reproduce it from scratch or with the steps provided)

I picked blendfile before with corrupted object. Here I pick actual scene where all started.

Also I can assume that it's happens only with procedural materials without image textures. And used Annotation tool with draw mode above meshes. It's just happens and unfortunately I can not say more. At some moment I just noticed it.

@Philipp Oeser (lichtwerk) I'm looking a t the bug too. Try to find where BKE_material_init_gpencil_settings() is called. If we found why this is called, we will catch the bug.

@Aleksandr (viadvena) While we found the solution to the bug, you can recover your material running these lines in console:

ma = bpy.context.active_object.material_slots[0].material
bpy.data.materials.remove_gpencil_data(ma)

In T71533#810151, @Antonio Vazquez (antoniov) wrote:

@Aleksandr (viadvena) While we found the solution to the bug, you can recover your material running these lines in console:

ma = bpy.context.active_object.material_slots[0].material
bpy.data.materials.remove_gpencil_data(ma)

Thank you! I appreciate this!

I'm going to create an script to fix your file in one step.

This script fix the file:

import bpy

for ob in bpy.data.objects:
    if ob.type == 'MESH':
        for slot in ob.material_slots:
            bpy.data.materials.remove_gpencil_data(slot.material)

In T71533#810159, @Antonio Vazquez (antoniov) wrote:

This script fix the file:

import bpy

for ob in bpy.data.objects:
    if ob.type == 'MESH':
        for slot in ob.material_slots:
            bpy.data.materials.remove_gpencil_data(slot.material)

Thanks a lot! It works) Could you please to help me with another thing: how can I force recompile shaders? Sometimes materials completely won't to updates. I struggled with this few times and in this situations helps install new build of blender. But may be there is much more esthetic way.

We could merge both, but I'm unable to reproduce it, so no idea how fix it.

In T71533#810175, @Antonio Vazquez (antoniov) wrote:

We could merge both, but I'm unable to reproduce it, so no idea how fix it.

Ah, I was thinking you were squashing the bug already.

While we found the solution to the bug

Anyways, since both issues seems to stem from the same roots [Annotations in play...], will merge these reports.
@Aleksandr (viadvena) : lets continue discussion in T63707, and please keep us up to date if you have found a way to reproduce [thanx for providing test_MatBug.blend -- but the corruption has already taken place in that file as well, and I am still not able to reproduce... will try more though...]