add no-useless-path-segments rule Fixes #471 #912
Conversation
avoid problems with ./ham/ -> ./ham/index.js by simply consuming '..', path segment pairs.
seems to not normalize trailing path seps. ./foo -> foo ./ -> ./ . -> . .. -> .. ../ -> ../
2 similar comments
1 similar comment
|
@benmosher not sure why it's still failing on appveyor |
| import moduleVisitor from 'eslint-module-utils/moduleVisitor' | ||
|
|
||
| function toRel(rel, sep) { | ||
| const rSep = escapeRegExp(sep) |
There was a problem hiding this comment.
is there a way this could be done without having to escape regexes, or use them at all?
There was a problem hiding this comment.
Ah it can be removed, this is only used on the posix side now.
Maybe on the no absolute rule? |
|
If I want to permit absolute paths, but not ones that could instead be relative, this one seems the appropriate place. |
|
Would need to know how far up is too far for absolute paths. '../../../../users/ljharb/Downloads/bower/jquery-ui' As that escapes out of the project |
|
Right, you'd locate the package.json, and consider that the root - anything within that should be relative instead of absolute. When phrased like that, however, it might make sense as an option to |
graingert
force-pushed
the
no-useless-path-segments
branch
from
7768af7
to
f7ef6ea
Compare
2 similar comments
| **/ | ||
| function toRel(rel) { | ||
| const stripped = rel.replace(/\/$/g, '') | ||
| return stripped.match(/^((\.\.)|(\.))($|\/)/) ? |
There was a problem hiding this comment.
since we're not using the match object here, let's use regex.test(str) instead.
| errors: [ 'Useless path segments for "../files/malformed", should be "./malformed"'], | ||
| }), | ||
| test({ | ||
| code: 'import "./test-module/"', |
There was a problem hiding this comment.
We should test all of these cases:
foo/index.somethingexists → path should be "foo"foo.somethingexists → path should be "foo"- both
foo/index.somethingandfoo.somethingexist →foo/andfooare different, and the difference is important, and this rule should not warn on it.
There was a problem hiding this comment.
foo.something exists → path should be "foo"
I think that should be handled by https://github.com/benmosher/eslint-plugin-import/blob/master/docs/rules/extensions.md
There was a problem hiding this comment.
yes, that should handle whether there's an extension or not - i'm saying that "whether this rule can warn on a path or not" actually depends on what's on disk, because the trailing slash is meaningful.
There was a problem hiding this comment.
Can we add a test case for catching multiple slashes, e.g. foo//something? See #915 (comment).
graingert
force-pushed
the
no-useless-path-segments
branch
from
f7ef6ea
to
26fb213
Compare
|
@danny-andrews @ljharb is this ready to go? |
|
Ready on my end! |
|
I'm going to rebase this soon and merge it. |
|
Hype |
ljharb
force-pushed
the
no-useless-path-segments
branch
2 times, most recently
from
61bce61
to
3090e84
Compare
3 similar comments
|
@benmosher do you have the release process documented anywhere? If so, I'll cut a release; if not, would you mind? :-D |
add no-useless-path-segments rule Fixes #471