Alexander Sosedkin

Not sure about that, since, e.g, `ctrl_i`, `ctrl_0` or `ctrl_щ` also error out with the same message (with 1 char difference, of course).

Oh, wait, sorry, now I get it. So, we'll never have `ctrl_h` or `ctlr_i`, we're limited to that list?

@DavidSagan, sorry, does your comment suggest storing it as a float encoded as a float that happens to be an integer or as an integer encoded as an integer in...

`nixpkgs.overlays = [ (_: super: { nix-serve = super.nix-serve.override { nix = super.nix_2_3; }; } ) ];` fixes that, so I presume it's Nix 2.4.

Yes, that's what I meant, I'm able to use my cache now without internal server errors. Are there more bugs at play here in your case?

@scopatz, your creation is pure wonder and your wisdom is deep. However that's probably not all that is required here. How often is the prompt updated? It seems that it...

I have no idea what I'm doing, but it seems like I just wrote a related xontrib: https://github.com/t184256/xontrib-prompt-vi-mode. I've added its metadata in pull request xonsh/xonsh/pull/1762.

The most problematic thing IMO is redrawing the whole prompt after each and every keypress, complete with requerying VCS info. I currently don't use my own xontrib because of the...

Thanks for a detailed analysis. I think you are overcomplicating stuff with your latter suggestions. I don't remember anyone except the users of my admittedly poorly implemented xontrib complaining that...

Hard question. I'm thinking whether changing the cursor shape (I-beam on insert, block on normal) could come out as a better substitute to displaying it in the prompt, what's your...