Add a quick note on how unbound imports and --fix

[mi-na-bot]

Having unbound imports mixed among the bound ones causes unexpected and incorrect seeming results. I spent several hours trying to fix this problem only to find it was well known!

Since this is only a change to one markdown file, I assume no other tests are required. Hopefully this is helpful.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 95.27%. Comparing base (c77c1a6) to head (6564860).

❗ Current head 6564860 differs from pull request most recent head aa17e43. Consider uploading reports for the commit aa17e43 to get more accurate results

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #2640      +/-   ##
==========================================
+ Coverage   95.14%   95.27%   +0.13%     
==========================================
  Files          68       68              
  Lines        2967     2966       -1     
  Branches     1044     1008      -36     
==========================================
+ Hits         2823     2826       +3     
+ Misses        144      140       -4     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[ljharb]

This sounds fine, but in every eslint rule, you can never rely on autofix covering everything - you always have to fix some things manually.

[mi-na-bot]

@ljharb This doc change is not intended to speak negatively about the implementation of --fix, which is IMO totally correct. If it reads that way perhaps there is better wording? I am coming at this from the perspective of a new plugin user, and how the fix settled was somewhat unexpected around some css imports. I kind of thought something was broken, and with a hint from the manual I would have been in business much faster!

[ljharb]

@minervabot right - that's why it's fine :-) but perhaps eslint's own documentation would have been a better place to set up proper expectations about autofixing?

[mi-na-bot]

@minervabot right - that's why it's fine :-) but perhaps eslint's own documentation would have been a better place to set up proper expectations about autofixing?

Like "it doesn't always work"? Wow, not very helpful actually. How/why it doesn't work is 100% on these docs to explain (or not?)

[ljharb]

@minervabot yes, because that's potentially the case in every eslint rule, including the ones in core.