Fix FC058 false positives

[jaym]

There were multiple issues here. First, the test did not
correctly catch this because the wrong match specification
was being created due to line number not being specified.
Instead of defaulting to nil, it defaulted to 1, making
2 incorrect things look correct together.

Second, FC058 needed to check the length of the array
was greater than so, as [] defaults to truthy.

cc @lamont-granquist @tas50 @sersut

Should fix #419

[lamont-granquist]

I suspect that line number defaulting to 1 was intentional and it should be fixed somewhere else up the stack to use 'nil' more explicitly...

[jaym]

I thought that as well, however it checks for nil in part of it

[jaym]

See https://github.com/jaym/foodcritic/blob/jdm/058-false-positives/features/support/command_helpers.rb#L134

[jaym]

I guess we can push it up the stack based on the order of that merge

[jaym]

I actually disagree with that default being there, this is an example of where badness happened because of it

So formal argument for leaving the default as nil:
expect_warning feels weird to have to specify a line for the not case. If you specify a line other than nil and you mess up the line number, you've screwed yourself. If you don't specify a line number, you've once again screwed yourself unless the warning would have been on line 1.

[jaym]

After discussion with @lamont-granquist, changing expect_warning to default to nil could drop a lot of cases where the default of 1 was part of the test. Instead, I've added on any line to the match to specify that the line does not matter.

[lamont-granquist]

👍 LGTM @tas50

[nathwill]

thanks @jaym! can confirm this solved it for me 👍