Style/IndentationWidth might be conflicting with start_of_line Lint/EndAlignment style

[brandonweiss]

I have:

Lint/EndAlignment:
  EnforcedStyleAlignWith: "start_of_line"

And if I turn on Style/IndentationWidth I get weird suggestions, like this:

def carry_attribution_data_from_previous_customer_invoice
  items = if line_items.retrospective.pre_subtotal.any?
            line_items.retrospective.pre_subtotal
  else
    line_items.bill_ahead.pre_subtotal
  end
end

It's obviously trying to indent two lines relative to the if statement, but that's not what I'd want/expect. What I'd expect is:

def carry_attribution_data_from_previous_customer_invoice
  items = if line_items.retrospective.pre_subtotal.any?
    line_items.retrospective.pre_subtotal
  else
    line_items.bill_ahead.pre_subtotal
  end
end

[stakach]

I have the same issue:

Lint/EndAlignment:
  EnforcedStyleAlignWith: start_of_line

and

msg << if queued.empty?
  'send queue is empty'
else
  "send queue contains: \n#{queued.join("\n")}"
end

outputs

Use 2 (not -7) spaces for indentation.
        'send queue is empty'
        ^^^
Align else with if.
        else
        ^^^^

If I disable Style/ElseAlignment then I only get the Use 2 (not -7) spaces for indentation.
So I also have to disable Style/IndentationWidth which is undesirable

[brandonweiss]

@bbatsov Hey! I've started a PR where I refactored the tests such that they also run Lint/EndAlignment configured with start_of_line in order to create a set of failing tests. I spent several hours today stepping through the on_if block of the cop, introspecting the various nodes and trying to figure out where the bug is, but I don't think I really got anywhere.

My very rudimentary understanding seems to be that when it gets to here, the base_loc#column is wrong. But base_loc is passed through into the on_if hook (via base_node). So… maybe the problem is occurring higher up somewhere? If you could help me by taking a look yourself and/or pointing me in the right direction that would be awesome. Thanks!

[jonas054]

@brandonweiss I think the problem is that in IndentationWidth#check_assignment, we're looking at the Lint/EndAlignment: EnforcedStyleAlignWith configuration when calculating base, but there's only support for the variable alignment style. The start_of_line style is treated the same as keyword, which is a bug.