Stefan Holderbach

We should only support this if everything that sees those spans is tolerant to bad indices. We don't want reedline to panic when you index out of bounds or between...

Also applies to the auto cd behavior in the REPL: ```nushell > cd / > ... > pwd /../..

Our working directory should only contain absolute paths according to my understanding so that part seems to be definitely a bug. There may be valid reasons for it to behave...

Thanks for looking into this, can you add a test that reproduces the original panic? I think it would still be a `fail_test()` kind of test as the example from...

The current behavior is as designed, it follows the same logic as our structural subtyping for records. It is sufficient to have the desired fields but any superfluous fields are...

This is currently hard coded as part of the vi binding parsing stage, so changing is currently not yet possible. Let's move this to reedline as a future improvement.

As the tests now seem to time out in CI, do we need to adjust the setup of any timeouts we need to configure or is there potentially a bug...

Thanks so much for the in depth research here and into the `Terminal.app` behavior. True that 10ms is probably pessimistic for the general events. For pure text events we would...

nushell did support big-ints for a while before version `0.59`. With the refactor of the underlying data model at the time it was removed for performance reasons and to ensure...

I think with the added difficulty of defining clear semantics for the interactions of unsigned and the existing signed types, I think adding a bigint representation internally is the more...