When running eslint from another directory and using --ignore-path, paths are not relative to .eslintignore location

What version of ESLint are you using? 3.1.1

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

Please show your full configuration: Project structure:

.eslintignore
.eslintrc
bar.js
src/foo.js

.eslintrc

{
    "root": true,
    "extends": "eslint:recommended"
}

.eslintignore

src/foo.js

What did you do? Please include the actual source code causing the issue.

$ eslint foo.js --ignore-path ../.eslintignore

C:\dev\eslint-bug\src\foo.js
  1:1  error  'test' is not defined  no-undef

✖ 1 problem (1 error, 0 warnings)

What did you expect to happen? Expected C:\dev\eslint-bug\src\foo.js to be ignored

What actually happened? Please include the actual, raw output from ESLint. Errors were reported for C:\dev\eslint-bug\src\foo.js:

C:\dev\eslint-bug\src\foo.js
  1:1  error  'test' is not defined  no-undef

✖ 1 problem (1 error, 0 warnings)

In Configuring ESLint it says “Paths are relative to .eslintignore location or the current working directory”. As a user, I would assume that if I passed in --ignore-path that it would ignore files as if I had called ESLint from that directory.

I see that there have been a lot of issues opened regarding ESLint ignore and parent directories, but after eslint/eslint#5694, ignoring files in SublimeText with SublimeLinter-eslint has been broken. There is an open PR to fix an existing issue with loading .eslintignore in SublimeLinter-eslint (roadhump/SublimeLinter-eslint#145), but it doesn’t work after ESLint 2.5.0.

There is an option in the CLIEngine to specify a cwd. If my above process is “working as intended” (or perhaps I have misinterpreted the documentation), then it would be great to be able to pass in a --cwdoption for the command line so that the relative paths can be loaded correctly.

[IanVS]

It seems reasonable to me to use a provided --ignore-path directory as the base dir when we convert the filename to a relative path. I’ve tried a possible implementation of this and it seems to work and not break any of our existing tests, would be happy to submit a PR if this is accepted by the TSC.

[IanVS]

Or would it be a bugfix? I can’t think of a single reason why someone would want the behavior of the .eslintignore file to depend on cwd when linting is performed. I doubt that was intended, but maybe I’m wrong. That’s something for the TSC to decide anyway. I’m tossing it on the agenda.