Possible false positive for `enforce-slice-style` in an expression context

[connorszczepaniak-wk]

Describe the bug
enforce-slice-style dictates using var x []int rather than x := []int{} for slice definitions when configured with the "nil" argument, but it also reports the usage of []int{} in an expression context where var x []int can't be used

To Reproduce
Steps to reproduce the behavior:

1. I updated revive go install github.com/mgechev/revive@latest
2. I run it with the following flags & configuration file:

revive -config=revive.toml ./...

[rule.enforce-slice-style]
Arguments = ["nil"]

With the following code:

package main

func main() {
    foo([]int{})
}

func foo(x []int) {}

And got this output:

main.go:4:6: use nil slice declaration (e.g. var args []type) instead of []type{}

Expected behavior
I didn't expect this to be a lint finding because this particular rule seems to be all about declarations, not expressions. The error message makes me think the linter is suggesting the following because it mentions var args []type

func main() {
    var x []int
    foo(x)
}

So I wonder if either:

1. Only statements should be checked
2. Or, maybe the error message could say something about use nil instead of []type when findings are in an expression context; using foo(nil) here also suppresses the lint.

[denisvmedia]

Have you tried this?

foo([]int(nil))

[connorszczepaniak-wk]

I'm sure that works (as does foo(nil)) -- if that's the desired behavior, then can the lint message change for findings in an expression context? Maybe instead of

use nil slice declaration (e.g. var args []type) instead of []type{}

It could say something like

use nil slice declaration (e.g. nil) instead of []type{}

The former suggestion doesn't really make sense in an expression context.