Improve isosurface loading for rendering a single cell

[philippotto]

Right now, it's a bit tedious to get a nice rendering of a complete cell. As I wrote here, there are two possible culprits:

• our artificial limit for how much isosurface chunks can be loaded at once
• thin processes might abort the full computation

We should have a look at how to improve this. For example, maybe the limit can be removed altogether as long as the computation is throttled reasonably.

• Also, listing the isosurfaces/meshes explicitly could greatly improve the UX. See isosurfaces for volume tracings #4567 (comment)

[philippotto]

Also, there seems to be a bug when navigating to another area while an isosurfaces is computed. See: https://discuss.webknossos.org/t/experimental-isosurface-rendering/1267/9?u=potto

[philippotto]

I think, this issue got irrelevant with the support for precomputed meshes.

[philippotto]

Users are reporting that ad-hoc meshes are not loaded completely. This is likely happening due to the limit mentioned above.

For example, maybe the limit can be removed altogether as long as the computation is throttled reasonably.

I think, this would be the easiest way to improve the situation right now.