parse -r capture group naming inconsistency with standard regex.

[CabalCrow]

Describe the bug

When you use the internal command parse -r $regex* with a regex containing capture groups with no names (ex: (\w+)(\w)) it would name the groups capture0, capture1 and so on. This is inconsistent with the regex standard where capture group 0 is the full match, and individual capture group start from 1 onwards.

How to reproduce

1. Run this command 'test' | parse -r '(\w+)(\w)'

Expected behavior

I expect the parse -r command to return capture groups naming starting with index 1 rather than 0.

Screenshots

No response

Configuration

key	value
version	0.92.2
branch	makepkg
commit_hash	2a08a18
build_os	linux-x86_64
build_target	x86_64-unknown-linux-gnu
rust_version	rustc 1.77.1 (7cf61ebde 2024-03-27) (Arch Linux rust 1:1.77.1-2)
cargo_version	cargo 1.77.1
build_time	2024-04-11 08:40:50 +00:00
build_rust_channel	release
allocator	mimalloc
features	dataframe, default, sqlite, trash, which
installed_plugins

Additional context

No response

[fdncred]

I think this is due to the skip and capture{i}, probably should be capture{i+1}, but I haven't tested.

nushell/crates/nu-command/src/strings/parse.rs

Lines 140 to 149 in cc9f41e

	let columns = regex
	.capture_names()
	.skip(1)
	.enumerate()
	.map(|(i, name)| {
	name.map(String::from)
	.unwrap_or_else(|| format!("capture{i}"))
	})
	.collect::<Vec<_>>();