indented arguments weirdness about coffeelint

[clutchski]

Thanks for reporting this. This is definitely a bug. CoffeeLint doesn't really handle invoking methods on different lines from the object. I actually added a failing test case for it last week. Stay tuned for a fix.

https://github.com/clutchski/coffeelint/blob/master/test/test_indent.coffee#L119

[clutchski]

@ddlsmurf: Out of curiosity, do you consider all of these cases valid or just inconsistent?

[ddlsmurf]

That's a good question. I wouldnt mind avoiding some formats to avoid indenting-related surprises in the generated js, but I don't really have any suggestions about that...

[dustinboston]

Bump. Is there a fix planned for this? I noticed it's been quite some time since it was opened....

[ssokolow]

Has any progress been made on this? It's pushing other errors in my Syntastic (Vim) readout out of the first screenful.

[schliffer]

+1

[docwhat]

👍 It's confusing when the line # is wrong for indent warnings (re: #248)

[AsaAyers]

The only indentation bug left that I know of is related to indention of chained calls. Until someone can really close this issue the solution is to not indent the chain:

doSomething()
.then(somethingElse)
.then ->
  inlineCallback()
  youCanEven()
  .then ->
    chaingThingsInHere()
  .withNoProblem()
.then ->
  secondInlineCallback()

It's not that there's no plan to fix it. Multiple people have attempted, but the root cause seems to live in CoffeeScript and no one has found a workaround. If you'd like to work on this I created http://asaayers.github.io/clfiddle/ for examining the token stream.

I think it would help if anyone has time to write a rule (or modification to the rule) to explicitly forbid these indented chains. @docwhat is right, it is definitely confusing. Using indented chains the false positives always seem to show up on the wrong line.

[docwhat]

I have to admit that I am very partial to indenting call chains... Vim indents them automatically. I imagine other indenters do as well.

sigh

[AsaAyers]

Given that CoffeeScript is a whitespace significant language I think you can argue that you shouldn't indent chains. When you indent the chain CoffeeScript doesn't generate an INDENT token because tokens exist to generate Javascript code. If it's not a new block, there's no reason to generate an INDENT token. If you checkout the generated Javascript, these chains all get crammed onto one line.

That said, CoffeeLint is here to help you write better code, not to enforce subjective opinions.

[ibrahima]

Is this supposed to be flagged? It seems like whether or not the below code is flagged depends on whether the dots appear on even column numbers or odd (odd numbers trigger the error):

$('#some_element').attr('disabled', false)
                  .removeClass('disabled')
                  .val(someValue)

We follow this style of indentation for our Ruby code as well (using Rubocop) and I couldn't figure out if there was a linter option for this or not, but it seems like this style of indentation is sometimes marked as valid and sometimes invalid depending on what column the first dot appears at.