Jakub Žádník

> Restarting the whole REPL also seems undesirable, as state would be lost on restart. You'd at least have a working REPL that you can use to investigate. Note that...

More concretely, we agreed that we should try wrapping the entire Nushell REPL into `catch_unwind()` and then recover from the panic by going back to the last known good state...

The stack trace doesn't point at chrono, though, and somehow doesn't show the line numbers. Is it the line 318 in seq_date that's panicking? It would be great if you...

The panic occurs in the `to_string()`. According to [the docs](https://doc.rust-lang.org/std/string/trait.ToString.html#impl-ToString-for-T), `to_string()` panics “if the `Display` implementation returns an error”. Which means that the result of `&next.format(&out_format)` (which I think is...

Yeah, it shouldn't be needed. It is safe to call it there because `merge_env()` is always called in a single thread, but we could also try removing it because it...

Haha, thanks for digging it up! Well, it's my fault. I think I moved it there because previously we were calling it in `cd` command etc. which was wreaking havoc,...

I think that in this case, we don't need a new flag because you can easily use the “not” operator: `not ('spam' | str starts-with 's')`. In Nushell, we try...

IMO, the `str contains --not` should be removed for the same reason I stated above. Otherwise, it's inconsistent, as you said, and we'd have to add a `--not` flag to...

I'll bring it up with other core team members so we get some consensus. Backwards compatibility is not yet a big problem because we haven't reached 1.0 yet. We're still...

OK, so we decided on removing `--not` from `str contains`, you can take it :-). One relatively convenient way to go around the missing `--not` is `... | not $in`...