Fix crash when drawing line gizmo with less than 2 vertices

[tim-blackbird]

Objective

Fix #9089

Solution

Don't try to draw lines with less than 2 vertices. These would not be visible either way.

[nicopap]

It should return success no?

[tim-blackbird]

I'm not sure what the semantics of RenderCommandResult are intended to be.
The RenderCommandResult doesn't appear to be used anywhere.

[Elabajaba]

I don't know if the result is correct or not, but it does fix the bug.

[nicopap]

I'll approve for the sake of getting this merged quickly, but I would prefer if it was returning RenderCommandResult::Success

[mockersf]

RenderCommandResult are currently ignored

bevy/crates/bevy_render/src/render_phase/draw.rs

Lines 270 to 271 in eb485b1

	// TODO: handle/log `RenderCommand` failure
	C::render(item, view, entity, param, pass);

But it would make much more sense for me too to return a Success here.

Also, surprisingly, the solution is not what this PR is actually doing. if you add 3 empty lines, the vertex count will be 3 and not hit that condition, still "rendering" empty lines

[mockersf]

adding a gizmo line adds the line position to the list, then unconditionally push an "empty" slot with NAN. could we instead not push those nans when we didn't push actual data to the lines?

[nicopap]

Also I learned recently that we should prefer u32::MAX over NaN. It is the "official" way of telling the GPU that our line list is ending.

[alice-i-cecile]

@devil-ira can you make the RenderCommandResult change requested and then I'll merge this in for you.

[mockersf]

@devil-ira can you make the RenderCommandResult change requested and then I'll merge this in for you.

I would prefer if someone tries to fix the issue at the source rather than this PR. We are pushing garbage in the mesh, and this PR just says "oh if there's only one garbage, don't draw it"

[tim-blackbird]

When we add a retained version of the line gizmo well need this check either way as users could easily create a linestrip of length 0 or 1.

[alice-i-cecile]

@mockersf @superdump; I'll defer to one of you on whether we merge this PR as is.

[alice-i-cecile]

I'm going to merge this following the approval from @DGriffin91, whose expertise I trust here: crashes are very sever bad and we can revert this very simple fix in a PR with a better fix later.

[DGriffin91]

@alice-i-cecile imo we should try to get this into 0.11.1 if we're merging it now. I noticed it was removed from the milestone, not sure if that means it would no longer be cherry picked for the point release.

[alice-i-cecile]

I'll defer to @cart for that decision, but I do agree.

[cart]

Sorry I didn't check my notifications before publishing so I missed this.