Stefan Holderbach comments

Results 458 comments of


                                            Stefan Holderbach

Behavior of `+` with dataframes seems counterintuitive and wrong

Certainly a bit weird default behavior for the otherwise mathematical add operator. Other libraries implement the broadcasted addition if the column/row indices match. Maybe this should explicitly use the append...

Refactor client.rs of the `http` commands

Worth referencing #13686

Chang http post to permit a pure integer value as the body

Thanks for tackling this. To me looks like we have a pretty bad implementation in `http post` if everything ends up in this one convoluted match/conditional mess to dispatch based...

Chang http post to permit a pure integer value as the body

Yeah sounds good!

Chang http post to permit a pure integer value as the body

Should there be some delay with refactoring that into proper components, I would be fine with landing that in combination with a little test to ensure the behavior works across...

Allow parse-time evaluation of more commands

From a utility value `which` would be nice but to some extent it doesn't satisfy the expectations of the `const` commands as it takes into account environment variables (more concretely...

`SyntaxShape::OneOf` does not support blocks (closures without arguments) for named arguments

Just to clarify `block`s have different semantics to closures, like permitting mutation of the outside scope. Thus in nushell this syntax shape is only used with keyword level commands like...

Singleton lists should be automatically converted to single element when command needs it

Propose a rule under which condition those coercions should happen and when not. Because I can imagine there are a ton of situations where the coercion would be equally surprising....

Unbalanced delimiters on parens when stripping `"` from output of 2nd/last `$"...(... | str trim -c '"')"` before string ends

Seems to be a version of #10327 Somehow the parser state is messed up after the `#` inline.

refactor signature with more verbose easily understandable fn names

I am all for more refined and consistent names. The things in `Signature` should be consistent with the accessors in `Call`/`CallExt` and the respective naming in the parser. Also the...