Consider returning 0 if we can't parse a file

At the moment, if you run `stylish-haskell` on a file it can't parse, then it returns 1.

However, this makes it a bit harder than necessary for tools that want to run `stylish-haskell` over a bunch of files and establish whether there is anything the user needs to do. For example, a pre-commit hook that runs `stylish-haskell` on changed files.

For such tools, "there's nothing I *can* do with this file" is more-or-less the same as "there is nothing *to* do with this file".

To put it another way, if `stylish-haskell` can't parse a file, there's not really anything the user can do about it. This is quite different to something like "your `stylish-haskell.yaml` file is malformed", where there really is a problem for the user to fix.

So maybe we should return 0 in this case.