I had a generic pointer that was T: ?Sized and could be str in some cases, where it wouldn't work. I got around this using generic trickery, but it's still not nice

Features:

• dbg! macro
• More impl for std types

Right now I'm using git, but I should get arround to releasing these on crates.io. If someone else is using this and wants this, let me know, and I'll be much more likely to do it.

I have my own macros for printing stuff on different situations and not putting conditionals everywhere, for example:

macro_rules! println_on_debug {
    ($($args:tt)*) => {
        if std::env::var("DEBUG").unwrap_or_else(|_| { "false".to_string() }) == "true" ||
           env!("CARGO_PKG_VERSION").ends_with("dev") {
               println!("DEBUG-MSG: {}", format!{ $($args)* })
        }
    };
}

I'd like to make that macro print objects with the same formatting as pprint, the problem is that pprint is not a macro that sends args to a formatter and redirects the output to stdout, pprint is a function that just formats & prints a complex structure. My suggestion consists in creating a function that instead of printing the formatted structure to stdout, just returns it, so I can call that macro like this:

println_on_debug!("struct: {}", pformat(complex_structure));

use debug2::{pprint, Debug};

use insta::assert_snapshot;

macro_rules! check {
    ($e:expr) => {
        assert_snapshot!(pprint($e))
    };
}

#[derive(Debug)]
enum LinkedList<T> {
    Empty,
    Node(T, Box<LinkedList<T>>),
}

#[test]
fn linked_list_reasonable_time() {
    let mut list = LinkedList::Empty;
    for i in 0..10000 {
        list = LinkedList::Node(i, Box::new(list));
    }
    check!(list);
}

thread 'linked_list_reasonable_time' has overflowed its stack