Committed in 0e7187f3598d

For scene annotations only can fail if it's done with python, but the user could enable fake user flag, no?

I have the doubt if annotations (using GPencil datablock too) are asigned to Scene or other editor, the US number is correct.

@Bastien Montagne (mont29) Could you review this change? I think this solve the issue, but as I'm not sure 100% about how works US number, I want your opinion before committing.

@Clément Foucault (fclem) I can reproduce it only using Debug mode

Resolved with commit 3fb099bf8152

See D5126 for proposed fix.

@Campbell Barton (campbellbarton) Sorry for that, I didn't see the warning for this field in MSVS2019. I removed two other fields, but this was not in the compiler messages. Anyway, thanks.

Committed in 8bf1977d311f

· Fix texture pointer indirection character & removed by error.

· Remove one mistake in gpencil_primitive.c (don't need to be changed in this patch)

· Remove redundant flags and make more compact the checking of Background texture.
· Try to reuse background drawing and don't disable for primitives always.

Working in the fix in D5115

· Remove the painting session for Primitives. This is not required because this was done to capture fast mouse events, and doing primitives this is not the case.

· Fixed some problems with background texture.

I have tested and in Windows 10 NVIDIA RTX2080TI I'm unable to reproduce it.

@Clément Foucault (fclem) I will take a look of the bug, but not sure now it's the moment to make the other changes so near of 2.80 release date. I thinkis better wait for 2.81.

Example of same stroke with the fixing I'm testing:

@Clément Foucault (fclem) Could you review this simple function before committing?

The FX are working in Screen Space, so any change in the viewport or camera angle will change the final result. Really, FX will work as compositor using the final render/viewport image and apply a filter on it. If you chnage image, you get different result.

IMHO we are missing the initial reason for creating this cursor. In grease pencil, when you are drawing the cursor is very annoying, so you need a very subtle cursor to know where is the cursor, but not disturbing your drawing. If we back to more complex cursor, we will back to the same problem we had before changing the cursor.... maybe, the cursor is too thin for sculpting and other modes, but for sure the cross is not a solution for grease pencil.

Thanks @Germano Cavalcante (mano-wii) for this change. The draw project was a mess and now it's more clear.

I think we can archive this task because is not a bug. If you find a problem, we can reopen it.

The join is always end of first stroke with start of next stroke. You can enable view where the stroke starts and ends (shows a green and red dot).

@Harley Acheson (harley) About update of active layer when click in outliner, the active layer changes but it's missing to call the update of the properties panel, and you need put mouse over to update. If you know what flag is missing in the WM message, just add it.

I really like the idea of collapse duplicated icons in one.

@Dalai Felinto (dfelinto) Could you look at this? it's related with the mail I have sent to you.

Are you joining the same stroke? this is only to join several strokes.

What is not clear for me is why you press V and B keys?

This is solved too.

Already committed with @Brecht Van Lommel (brecht) suggestions.

@Brecht Van Lommel (brecht) Could you review D5071 ?

I think we can archive because the reason in the children particles and the number is very high. Really, this is more a ToDo to improve Child particles creation and multithreading, but not a bug.

If the problem is only in one file and this cannot reproduce in any other file, not sure if we must keep open the bug.... maybe the file was damaged for any reason related to some beta bug and now the problem does not exist.

What you are trying to do is totally wrong. Annotation uses the same API that Grease Pencil, so you need create the Layers and the Frames using the corresponding "new" function.

Why not delete the strokes? If you remove the material is the same if you remove a layer.

This is not a bug, but a To Do task we already know. Close the strokes is not easy as you can think, especially for strokes with alpha. We need first fix a problem with alpha glitches before fixing this, so now we have the best possible solution.
As soon as we solve the alpha problem, we will be able to improve closing.

@Sebastian Parborg (zeddb) I cannot reproduce it... really I don't know what is the bug here (never had this situation).

@Campbell Barton (campbellbarton) I think this can be related to T65108 commit b69ed14ff745