Inconsistency in representation of binary strings

[nkarl]

Describe the bug

• Format 1: 001b
• Format 2: 0b01

Only the second format is functional when used with into bits. In help bits the format used is the first format.

How to reproduce

1. 010b | into bits
2. 0b10 | into bits

Expected behavior

1. 010b | into bits should evaluate to 00000010.
2. Instead, it evaluates to 00001010.

Screenshots

Configuration

key	value
version	0.88.1
branch
commit_hash
build_os	linux-x86_64
build_target	x86_64-unknown-linux-gnu
rust_version	rustc 1.74.1 (a28077b28 2023-12-04)
rust_channel	stable-x86_64-unknown-linux-gnu
cargo_version	cargo 1.74.1 (ecb9851af 2023-10-18)
build_time	2023-12-19 19:09:15 -08:00
build_rust_channel	release
allocator	mimalloc
features	dataframe, default, extra, sqlite, trash, which, zip
installed_plugins

Additional context

No response

[fdncred]

I don't follow. 10 bytes is 0000 1010. We don't have any bits representation in nushell. 10b and 10B are both bytes.

[IanManske]

The command's first example is probably confusing. It says:

convert a binary value

but then uses a filesize, 01b (1 byte), as input to the command. It should be a binary value: 0x[1].

Also, the integer example should maybe use a binary literal 0b10 (= 2 in base 10) to highlight the difference from a filesize.

[fdncred]

those changes, Ian, seem reasonable.

[nkarl]

I don't follow. 10 bytes is 0000 1010. We don't have any bits representation in nushell. 10b and 10B are both bytes.

This was the source of my confusion. I assumed that B is for bytes and b is for bits and that is the general consensus. For example, in Vim the unit for characters is B for bytes.

There is also no documentation on various formats of integer at the moment. Traditionally in C we have:

1. 0b11111111 in binary for 255, b for binary.

• changed to 0x[11111111]
• this is confusing.

2. 0xff in hex for 255, x for hex.

In Rust, the formats also follow the same convention, 0b for bin and 0x for hex.

Because of these assumptions, I thought that Nu allowed an extra format, e.g. 11111111b as binary literal.

[fdncred]

we have 0b, 0x, and 0o as literal inputs but i believe the output is always in bytes for literals. If you want to force binary you can use into binary, which then gives a hexadecimal display since binary isn't represented well in a terminal.

[nkarl]

That makes sense now. Thanks.

[amtoine]

@nkarl
i did not follow the whole discussion, is there still an issue to be addressed? 😇

[nkarl]

I was misunderstanding the behavior. It's all clear to me now. Thanks