[bug]attempted to repeat an expression containing no syntax variables matched as repeating at this depth

[monchin]

Hello, thank you for this project!

I found that this example fails:

/// main.rs
use macro_rules_rt::{Matcher, Rule, Transcriber};

/// Demonstrates that `macro_rules_rt` 0.1.1 rejects a transcriber where a
/// non-repeating metavariable (`$m`) is referenced inside a `$(...)*` block.
///
/// The equivalent `macro_rules!` is valid Rust and expands correctly — only
/// the runtime expansion library rejects it.

fn main() {
    // ── Case 1: OK — only repeating `$cls` inside $(...)* ──────────────
    let m1: Matcher = "register!($($cls:ty),*)".parse().expect("parse m1");
    let t1: Transcriber = "$(add_class::<$cls>()?;)*".parse().expect("parse t1");
    match Rule::new(m1, t1) {
        Ok(rule) => {
            let result = rule.apply("register!(a, b, c)").expect("apply case 1");
            println!("Case 1 (OK):   apply result: {result}");
        }
        Err(e) => eprintln!("Case 1 (FAIL): unexpected error: {e}"),
    }

    // ── Case 2: FAIL — non-repeating `$m` inside $(...)* ──────────────
    let m2: Matcher = "register!($m:expr, $($cls:ty),*)".parse().expect("parse m2");
    let t2: Transcriber = "$($m.add_class::<$cls>()?;)*".parse().expect("parse t2");
    match Rule::new(m2, t2) {
        Ok(rule) => {
            let result = rule.apply("register!(obj, a, b, c)").expect("apply case 2");
            println!("Case 2 (OK):   apply result: {result}");
        }
        Err(e) => eprintln!("Case 2 (FAIL): {e}"),
    }

    // ── Equivalent standard macro_rules! (works fine) ─────────────────
    //
    // macro_rules! register {
    //     ($m:expr, $($cls:ty),*) => {
    //         $($m.add_class::<$cls>()?;)*
    //     };
    // }
    //
    // This is valid Rust. `cargo expand` confirms it expands as expected.
}

# Cargo.toml
[package]
name = "macro_rules_rt_issue"
version = "0.1.0"
edition = "2024"

[workspace]

[dependencies]
macro-rules-rt = "0.1"
# macro-rules-rt = { path = "../macro-rules-rt" }

case 2 is expected to be parsed successfully, but it failed with message "attempted to repeat an expression containing no syntax variables matched as repeating at this depth“

[frozenlib]

Thanks for the report!

I looked into the behavior of macro_rules_rt based on your example, and it turns out this uncovered a broader issue than the specific case you reported.

In particular, there is another problematic pattern around cases where multiple repetitions in the matcher are combined into a single repetition in the transcriber, for example:

$($a:ident)* + $($b:ident),* => "$($a + $b),*"

This made me realize that my mental model of how macro_rules! handles repetition depth and metavariable binding was not quite accurate. Fixing this properly will probably require a fairly broad change to the current implementation rather than a small local patch.

I’d like to fix this when I have time. Thanks again for the clear reproduction — it was very helpful.