Daniel Buch Hansen

This closes #992 btw.

I think you should leave it as-is. I by your Arc/Rc argument.

> I wanted to see a reproducible example or a test because I cannot reproduce the panic. Seems OK to land but I'd wait a bit if @dbuch replies and...

I plan to put this commit bc9309e46f3dbc9a14c5a854210aa7e6119b5611 in this inplace. Reedline do this even with assert!() (Perhaps, this codebase should do the same? See: [reedline/base.rs](https://github.com/nushell/reedline/blob/265e59966cea8d66672ae178d68f68b55edd178b/src/completion/base.rs#L17)) this will certainly shred light...

> @dbuch do you have repro steps? and can you add a test to this pr for this situation? No I'm currently not using the configuration i used back then....

I've tested this, and i solves the right prompt problem. I vote get it in, and progress from here. There are some issues still that are annotated by todo/fix in...

@fdncred Okay, I'll try. I still think this piece should be merged, do you know what the trade are? This PR is after all - afaik - fairly okay now...

@fdncred you are right! I use kitty, and because (i think) i use HDPI and kitty resizes late i get that prompt in the middle. But only at the beginning....

> That sounds like an interesting way to catch bugs here. The first look at the test failures looks like we have some issues that get caught that way. Yeah...

> Last we talked about this one, I thought we decided it was good to land. We just need to get the ci green. Please correct me if I'm wrong....