Parsing in pathWithoutExt(filePath) does not work for Windows paths

[thewalla07]

Environment is Windows 10 Version 1809

Environment Info:

Node version: v10.15.0
npm version: v6.4.1
Local ESLint version: v6.5.1 (Currently used)
Global ESLint version: Not found

What parser (default, Babel-ESLint, etc.) are you using? default (N/A)

Please show your full configuration: (N/A)

What did you do? Please include the actual source code causing the issue, as well as the command that you used to run ESLint.

We run ESLint in Jenkins CI on a Windows host, and get the following type of output from failures

Currently, this means we do not know the filepath for the files with errors, only for the ones which passed.

This issue arises as a result of PR #11683. See below:

eslint/lib/cli-engine/formatters/junit.js

Lines 28 to 36 in 287ca56

	/**
	* Returns a full file path without extension
	* @param {string} filePath input file path
	* @returns {string} file path without extension
	* @private
	*/
	function pathWithoutExt(filePath) {
	return path.posix.join(path.posix.dirname(filePath), path.basename(filePath, path.extname(filePath)));
	}

The above code assumes that POSIX filePaths are used, meaning that the function does not work as expected in Windows.

npx eslint home_page.js --format junit

What did you expect to happen?
I expect that in the classname tag below I can see the full path to the file, without the file extension for visualization within Jenkins CI on a Windows host. i.e. C:\path\to\file

<?xml version="1.0" encoding="utf-8"?>
<testsuites>
<testsuite package="org.eslint" time="0" tests="1" errors="1" name="C:\git\projectdirectory\home_page.js">
<testcase time="0" name="org.eslint.prefer-const" classname="C:\git\projectdirectory\home_page"><failure message="&apos;msg&apos; is never reassigned. Use &apos;const&apos; instead."><![CDATA[line 7, col 17, Error - &apos;msg&apos; is never reassigned. Use &apos;const&apos; instead. (prefer-const)]]></failure></testcase>
</testsuite>
</testsuites>

What actually happened? Please include the actual, raw output from ESLint.
The function above parses filePath incorrectly on Windows. path.posix.dirname('C:\\path\\to\\file.js') resolves to . on Windows, as posix forces an assumption of POSIX style paths.

This results in the following XML, without the full path to the file in the classname attribute:

<?xml version="1.0" encoding="utf-8"?>
<testsuites>
<testsuite package="org.eslint" time="0" tests="1" errors="1" name="C:\git\projectdirectory\home_page.js">
<testcase time="0" name="org.eslint.prefer-const" classname="home_page"><failure message="&apos;msg&apos; is never reassigned. Use &apos;const&apos; instead."><![CDATA[line 7, col 17, Error - &apos;msg&apos; is never reassigned. Use &apos;const&apos; instead. (prefer-const)]]></failure></testcase>
</testsuite>
</testsuites>

Are you willing to submit a pull request to fix this bug?
Yes.

I would propose a fix where we do not use the posix specific implementation of Node's path library, and we let Node decide which implementation to use based on the underlying OS.

The main question arises in testing. In https://github.com/eslint/eslint/blob/master/tests/lib/cli-engine/formatters/junit.js is it possible to run a test with both POSIX and Windows style paths? My suggestion would be to have the test cases adapt to the current OS, but opinions would be valued here.

[thewalla07]

Working on this, created PR to fix this issue

[eslint-deprecated]

Unfortunately, it looks like there wasn't enough interest from the team
or community to implement this change. While we wish we'd be able to
accommodate everyone's requests, we do need to prioritize. We've found
that issues failing to reach accepted status after 21 days tend to
never be accepted, and as such, we close those issues.
This doesn't mean the idea isn't interesting or useful, just that it's
not something the team can commit to.

Thanks for contributing to ESLint and we appreciate your understanding.

[thewalla07]

@kaicataldo any chance this will be reopened, or should the associated PR be abandoned?

[kaicataldo]

Thanks for the ping. Let me see if I can get some eyes on this PR!