nushell/crates/nu-parser/tests/test_lex.rs
Mango The Fourth d0e0701a88
fix: fix lexing of comments, such that a#b becomes a coherent Item (#8151)
# Description

Previously `nix run nixpkgs#hello` was lexed as `Item, Item, Item,
Comment`, however, `#hello` is *not* supposed to be a comment here and
should be parsed as part of the third `Item`.

This change introduces this behavior by not interrupting the parse of
the current token upon seeing a `#`.

Thank you so much for considering this, I think many `nix` users will be
grateful for this change and I think this will lead to more adaptation
in the ecosystem.

- closes #8137 and #6335 

# User-Facing Changes

- code like `somecode# bla` and `somecode#bla` will not be parsed as
`somecode, comment` but as `somecode#bla`, hence this is a breaking
change for all users who didn't put a space before a comment introducing
token (`#`)

# Tests + Formatting

I've added tests that cover this behavior in `test_lex.rs`

- [x] `cargo fmt --all -- --check` to check standard code formatting
(`cargo fmt --all` applies these changes)
- [x] `cargo clippy --workspace -- -D warnings -D clippy::unwrap_used -A
clippy::needless_collect` to check that you're using the standard code
style
- [x] `cargo test --workspace` to check that all tests pass

# After Submitting

> If your PR had any user-facing changes, update [the
documentation](https://github.com/nushell/nushell.github.io) after the
PR is merged, if necessary. This will help us keep the docs up to date.

I think this is expected behavior in most other shells, so the
documentation was lacking for not documenting the unexpected behavior
before and hence now is automatically more complete >D
2023-02-22 12:59:47 +00:00

190 lines
4.1 KiB
Rust

use nu_parser::{lex, ParseError, Token, TokenContents};
use nu_protocol::Span;
#[test]
fn lex_basic() {
let file = b"let x = 4";
let output = lex(file, 0, &[], &[], true);
assert!(output.1.is_none());
}
#[test]
fn lex_newline() {
let file = b"let x = 300\nlet y = 500;";
let output = lex(file, 0, &[], &[], true);
assert!(output.0.contains(&Token {
contents: TokenContents::Eol,
span: Span::new(11, 12)
}));
}
#[test]
fn lex_empty() {
let file = b"";
let output = lex(file, 0, &[], &[], true);
assert!(output.0.is_empty());
assert!(output.1.is_none());
}
#[test]
fn lex_parenthesis() {
// The whole parenthesis is an item for the lexer
let file = b"let x = (300 + (322 * 444));";
let output = lex(file, 0, &[], &[], true);
assert_eq!(
output.0.get(3).unwrap(),
&Token {
contents: TokenContents::Item,
span: Span::new(8, 27)
}
);
}
#[test]
fn lex_comment() {
let file = b"let x = 300 # a comment \n $x + 444";
let output = lex(file, 0, &[], &[], false);
assert_eq!(
output.0.get(4).unwrap(),
&Token {
contents: TokenContents::Comment,
span: Span::new(12, 24)
}
);
}
#[test]
fn lex_is_incomplete() {
let file = b"let x = 300 | ;";
let output = lex(file, 0, &[], &[], true);
let err = output.1.unwrap();
assert!(matches!(err, ParseError::ExtraTokens(_)));
}
#[test]
fn lex_incomplete_paren() {
let file = b"let x = (300 + ( 4 + 1)";
let output = lex(file, 0, &[], &[], true);
let err = output.1.unwrap();
assert!(matches!(err, ParseError::UnexpectedEof(v, _) if v == ")"));
}
#[test]
fn lex_incomplete_quote() {
let file = b"let x = '300 + 4 + 1";
let output = lex(file, 0, &[], &[], true);
let err = output.1.unwrap();
assert!(matches!(err, ParseError::UnexpectedEof(v, _) if v == "'"));
}
#[test]
fn lex_comments_no_space() {
// test for parses that contain tokens that normally introduce comments
// Code:
// let z = 42 #the comment
// let x#y = 69 #hello
// let flk = nixpkgs#hello #hello
let file = b"let z = 42 #the comment \n let x#y = 69 #hello \n let flk = nixpkgs#hello #hello";
let output = lex(file, 0, &[], &[], false);
assert_eq!(
output.0.get(4).unwrap(),
&Token {
contents: TokenContents::Comment,
span: Span::new(11, 24)
}
);
assert_eq!(
output.0.get(7).unwrap(),
&Token {
contents: TokenContents::Item,
span: Span::new(30, 33)
}
);
assert_eq!(
output.0.get(10).unwrap(),
&Token {
contents: TokenContents::Comment,
span: Span::new(39, 46)
}
);
assert_eq!(
output.0.get(15).unwrap(),
&Token {
contents: TokenContents::Item,
span: Span::new(58, 71)
}
);
assert_eq!(
output.0.get(16).unwrap(),
&Token {
contents: TokenContents::Comment,
span: Span::new(72, 78)
}
);
}
#[test]
fn lex_comments() {
// Comments should keep the end of line token
// Code:
// let z = 4
// let x = 4 #comment
// let y = 1 # comment
let file = b"let z = 4 #comment \n let x = 4 # comment\n let y = 1 # comment";
let output = lex(file, 0, &[], &[], false);
assert_eq!(
output.0.get(4).unwrap(),
&Token {
contents: TokenContents::Comment,
span: Span::new(10, 19)
}
);
assert_eq!(
output.0.get(5).unwrap(),
&Token {
contents: TokenContents::Eol,
span: Span::new(19, 20)
}
);
// When there is no space between the comment and the new line the span
// for the command and the EOL overlaps
assert_eq!(
output.0.get(10).unwrap(),
&Token {
contents: TokenContents::Comment,
span: Span::new(31, 40)
}
);
assert_eq!(
output.0.get(11).unwrap(),
&Token {
contents: TokenContents::Eol,
span: Span::new(40, 41)
}
);
}