PatrickHaecker

Both ideas sound worth exploring to me. `Pluto.run(store_cell_order=false)` is (hopefully) logically simple, but is effectively a new data storage format which involves some variant handling. Being an option both formats...

> As a tie breaker when creating a topological order, one could use the cell having the alphabetically smallest hash. The problem with this approach is that you get a...

> We already use visual cell order as a tie-breaker. That's probably best when you do stateful things like creating a file, then reading it. Currently, the run-all execution order...

> Not sure though, it might require the "Cell order" string to parse correctly. In order to use this in production, this would need to be a guaranteed property, however,...

> By "cell hash", I refer to the identifier starting with `bb8` in the comment above a cell's content Ah, thanks for the clarification. I think strictly speaking this is...

Ok, sounds like a plan! However, I probably won't find the time to get familiar with the code base and implement it during the next weeks. So if someone finds...

The first step should be to implement the two actions you have listed, @KeithWM. In the second step we need to see what the user interface should be. I could...

It is a good idea, to mention `FieldFlags.jl`. I think documenting the padding issues, which are also present in `C` as I learned, would still be needed.

Thanks for considering this issue. > I would still lean towards defining a default `widen` I think a default implementation of some abstract type, which can be overridden for each...

Thanks, @adienes, for bringing this up again. I think I've found a Julia-generic improvement idea and haven't realized until now that it also applies to this issue. It would solve...