The macro recursion limit is too low

[carllerche]

I started hitting the recursion limit with the following macro (which isn't actually complete yet either)

    use nixtest::assert_errno_eq;
    use libc::c_int;

    macro_rules! check_errno {
        ($errno:ident) => {{
            assert_errno_eq(stringify!($errno), $errno as c_int);
        }};

        ($errno:ident, $($rest:ident),+) => {{
            check_errno!($errno);
            check_errno!($($rest),*);
        }};
    }

    check_errno!(
        EPERM,
        ENOENT,
        ESRCH,
        EINTR,
        EIO,
        ENXIO,
        E2BIG,
        ENOEXEC,
        EBADF,
        ECHILD,
        EAGAIN,
        ENOMEM,
        EACCES,
        EFAULT,
        ENOTBLK,
        EBUSY,
        EEXIST,
        EXDEV,
        ENODEV,
        ENOTDIR,
        EISDIR,
        EINVAL,
        ENFILE,
        EMFILE,
        ENOTTY,
        ETXTBSY,
        EFBIG,
        ENOSPC,
        ESPIPE,
        EROFS,
        EMLINK,
        EPIPE,
        EDOM,
        ERANGE,
        EDEADLK,
        ENAMETOOLONG,
        ENOLCK,
        ENOSYS,
        ENOTEMPTY,
        ELOOP,
        ENOMSG,
        EIDRM,
        ECHRNG,
        EL2NSYNC,
        EL3HLT,
        EL3RST,
        ELNRNG,
        EUNATCH,
        ENOCSI,
        EL2HLT,
        EBADE,
        EBADR,
        EXFULL,
        ENOANO,
        EBADRQC,
        EBADSLT,
        EBFONT,
        ENOSTR,
        ENODATA,
        ETIME,
        ENOSR,
        ENONET,
        ENOPKG,
        EREMOTE,
        ENOLINK,
        EADV,
        ESRMNT,
        ECOMM,
        EPROTO,
        EMULTIHOP,
        EDOTDOT,
        EBADMSG,
        EOVERFLOW,
        ENOTUNIQ,
        EBADFD,
        EREMCHG,
        ELIBACC,
        ELIBBAD,
        ELIBSCN,
        ELIBMAX,
        ELIBEXEC,
        EILSEQ,
        ERESTART,
        ESTRPIPE,
        EUSERS,
        ENOTSOCK,
        EDESTADDRREQ,
        EMSGSIZE,
        EPROTOTYPE,
        ENOPROTOOPT,
        EPROTONOSUPPORT,
        ESOCKTNOSUPPORT,
        EOPNOTSUPP,
        EPFNOSUPPORT

[eddyb]

You don't need recursion for this macro, was your original usecase more complicated? Also, the example is missing at least a closing ).

[carllerche]

Yeah, it got truncated when copying :P How would I implement the macro without recursion?

[kennytm]

@carllerche

macro_rules! check_errno {
    ($($errno:ident),+) => {{
        $( assert_errno_eq(stringify!($errno), $errno as c_int); )+
    }}
}

[steveklabnik]

/cc @rust-lang/lang, @rust-lang/compiler is there any discussion on what this limit should be?

[eddyb]

@steveklabnik IMO we should implement macro expansion without recursing in the compiler. That would mean it could potentially never terminate, but at least we wouldn't have errors like this. It could also be much more efficient, by reusing ranges of tokens instead of copying them over and over again.

[jonas-schievink]

@eddyb Isn't the recursion limit supposed to prevent never-terminating compilations? Being able to stack-overflow the compiler with a #![recursion_limit = "5000"] sounds bad. (I always thought it limits recursive user code, be it macros, monomorphizations, etc. and not rustc-internal recursion)

That said, the recursion limit isn't doing a very good job if this is what it's supposed to do:

macro_rules! m {
    ( $( $i:ident )* ) => {
        m!( $($i)* $($i)* )
    };
}

fn main() {
    let i = m!(m);
}

This hangs, but doesn't cause a stack overflow. Memory usage slowly increases starting at about 100 MB.

[eddyb]

Macros and monomorphization both recurse in the Rust compiler and would otherwise end up in stack overflows.

[Mark-Simulacrum]

This has been encountered in the wild once more (#38541).

I'm going to mark this as E-easy since I think the primary thing that needs to change is just this https://github.com/rust-lang/rust/blob/master/src/libsyntax/ext/expand.rs#L1041 line, which needs to be updated to a larger value. I'm going to suggest 1024.

@jseyfried Do you agree with the easy designation?

[sirideain]

I would be interested in taking this one as a good way to figure out the process of contributing.

[Mark-Simulacrum]

@sirideain Great! Let us know if you need assistance.

[eddyb]

cc @nikomatsakis Another good use for stacker!