build: added ignore path for eslint

[chrisbudy]

Checklist

• make -j4 test (UNIX), or vcbuild test (Windows) passes
• commit message follows commit guidelines

Affected core subsystem(s)

build

[Trott]

.eslintignore is the default, so this change should be a no-op. I think the problem we saw at Code + Learn probably had to do with a problem in the .eslintignore itself?

[addaleax]

Doesn’t seem to hurt anybody, and if it helps someone, sure

[Trott]

Can someone with a Windows machine (preferably @chrisbudy!) do this to confirm that this fixes anything or else confirm that it has no effect?

1. Create a file called test/tmp/test.js. Make the contents nothing but a garbage string like Fhwqhgads.
2. Run vcbuild lint-js. Does it show errors? If so, we have a bug. If not, something weird was up at Code + Learn.
3. Now make the change in this PR. Run vcbuild lint-js. Does it show errors?

If the second step shows errors and the third step does not, then this PR should land. If any other combination occurs, then this PR should not land.

[refack]

BTW a better command is vcbuild noprojgen nobuild lint-js

[refack]

I can't repro.
Also, trying to force check the canary file gives:

D:\code\node-github-desktop$ node tools/eslint/bin/eslint.js --rulesdir=tools/eslint-rules test\tmp\test.js

D:\code\node-github-desktop\test\tmp\test.js
  0:0  warning  File ignored because of a matching ignore pattern. Use "--no-ignore" to override

✖ 1 problem (0 errors, 1 warning)

I'd also be happy to hear from @chrisbudy that this positively solves the issue.

[refack]

@chrisbudy maybe you have a global .eslintrc with a no-ignore rule?

[chrisbudy]

@Trott I assumed it should default as well, but I was getting build errors on a tmp file (event though excluded in the .eslintignore file) at the code and learn, and adding the line I did stopped the build errors from happening. But I will test it again today.

Thanks for the feedback @refack and @Trott.

[BridgeAR]

@chrisbudy did you get the chance to test this again?

[Trott]

Someone else tested this and unfortunately this doesn't solve the problem. See details here: #16278 (comment)

[Trott]

I'm going to close this one but feel free to re-open or comment if you think this should stay open. The solution is likely a bug-fix in ESLint. Alternatively, we could add a common.refreshTmpDir() to the end of the test that creates the directory with multi-byte characters, but I'd prefer this be run up to ESLint instead, TBH. Will need some Windows folks willing to dig a bit, though. That's not me, unfortunately.