AlexHaxe

I took a quick look at the issue on a Windows 10 machine, here is what I found: `neko.exe` crashes in `regexp.ndll` with a stackoverflow I have narrowed it down...

I think it might be more of an "evil" regex problem ( https://www.owasp.org/index.php/Regular_expression_Denial_of_Service_-_ReDoS ) Maybe it works on Linux / OSX, because neko uses larger stacksizes on those systems. Just...

As a workaround we could make lines above 1000 (or 2000, needs testing) spit out a checkstyle message for all line based checks. So that users get some kind of...

I would recommend Haxe 4rc5 for compiling git checkstyle. I haven't done extensive tests, but it seems running a nodejs version of checkstyle might be bit more reliable. Maybe we...

This should be fixed in hxparse, because if I remove the BOM from content before sending it to Lexer, all positions are off by three. And that would likely affect...

How about calling it `countEmpty` (see http://checkstyle.sourceforge.net/config_sizes.html#MethodLength)? We might even go as far as making it an enum, e.g. with `All`, `NoEmptyLines` and `NoEmptyLinesAndComments` (maybe find shorter names).

I'm not sure users will actually use it, but if we change it, then we might as well go for a maximum of flexibility. I mean it would be possible...

No, MethodLength can handle both cases (or all four), you can even define them in the same config file. A check for measuring code to non-code ratios is a whole...

I've come to the conclusion, that `ignoreEmptyLines` is much easier to understand than `countEmpty`. Compatibility with Java checkstyle is not really necessary, since config files are not interchangeable between Haxe...

Regular Haxe files should start with an upper case letter, so `UnusedImportCheck` can just skip performing checks on `import.hx` files, regardless of Haxe version. Since import.hx applies to all files...