Snapshots incorrectly treat CRLF and LF the same

[melusc]

Describe the bug

Vitest treats CRLF and LF as the same in snapshots which leads to the wrong snapshot being used. E.g. if a string has CRLF it'll incorrectly be normalised to LF.

import {test, expect} from 'vitest';

test.for([
	'a\nb', // ["a", "\n", "b"], snapshot will have ["a", "\n", "b"]
	'a\r\nb', // ["a", "\r", "\n", "b"], snapshot will have ["a", "\n", "\n", "b"]
])('"%s".split("")', (input) => {
	expect(input.split("")).toMatchSnapshot();
})

On Windows it correctly generates two snapshots for both. But then when it compares the snapshot on a second run this test will fail because the line endings aren't right.

On Ubuntu it writes the first snapshot to the file. Then it incorrectly uses the first snapshot for the second test and fails.

Reproduction

https://github.com/melusc/vitest-snapshot-crlf

You can see how it behaves on Windows or Ubuntu in the Github Actions logs: https://github.com/melusc/vitest-snapshot-crlf/actions/runs/12869505983

System Info

System:
    OS: Windows 11 10.0.26100
    CPU: (8) x64 11th Gen Intel(R) Core(TM) i5-1135G7 @ 2.40GHz
    Memory: 770.81 MB / 7.84 GB
  Binaries:
    Node: 23.6.0 - C:\nvm4w\nodejs\node.EXE
    Yarn: 4.6.0 - C:\nvm4w\nodejs\yarn.CMD
    npm: 10.9.2 - C:\nvm4w\nodejs\npm.CMD
  Browsers:
    Edge: Chromium (124.0.2478.105)
  npmPackages:
    vitest: ^3.0.2 => 3.0.2

Used Package Manager

yarn

Validations

• Follow our Code of Conduct
• Read the Contributing Guidelines.
• Read the docs.
• Check that there isn't already an issue that reports the same bug to avoid creating a duplicate.
• Check that this is a concrete bug. For Q&A open a GitHub Discussion or join our Discord Chat Server.
• The provided reproduction is a minimal reproducible example of the bug.

[melusc]

Workaround would be:

test.for([
	'a\nb', // ["a", "\n", "b"]
	'a\r\nb', // ["a", "\r", "\n", "b"]
-])('"%s".split("")', (input) => {
+])('%j.split("")', (input) => {
-	expect(input.split("")).toMatchSnapshot();
+	expect(JSON.stringify(input.split(""))).toMatchSnapshot();
})

I think an actual solution would be in this direction. Instead of writing actual CRs and LFs in the title and snapshot, write \r and \n so that CRLF cannot be normalised to LF.