lightbeam panicked when either run or wasm2obj

[qinxk-inter]

- What are the steps to reproduce the issue?
t.wasm is wat2wasm from below text:

(module
  (func $add (param $lhs i32) (param $rhs i32) (result i32)
    get_local $lhs
    get_local $rhs
    i32.add)
  (export "add" (func $add))
)

./wasmtime run --lightbeam t.wasm 1 2
./wasmtime wasm2obj --lightbeam t.wasm t.obj
- What do you expect to happen? What does actually happen? Does it panic, and
if so, with which assertion?
Expect to be run without any issue. Got error with the output:

thread 'main' panicked at 'index out of bounds: the len is 0 but the index is 0', /PATH/wasmtime/crates/lightbeam/src/backend.rs:637:35
note: run with RUST_BACKTRACE=1 environment variable to display a backtrace

- Which Wasmtime version / commit hash / branch are you using?
main latest by now, commit 12e658a
- If relevant, can you include some extra information about your environment?
(Rust version, operating system, architecture...)
rustc 1.45.2, MacOS

[bjorn3]

The panic happens at

wasmtime/crates/lightbeam/src/backend.rs

Line 637 in 12e658a

let func_start = &mut self.func_starts[func_idx as usize];

func_starts should be filled at

wasmtime/crates/lightbeam/src/backend.rs

Lines 610 to 612 in 12e658a

	let func_starts = iter::repeat_with(|| (None, assembler.new_dynamic_label()))
	.take(func_count as usize)
	.collect::<Vec<_>>();

func_count gets it's value from

wasmtime/crates/lightbeam/wasmtime/src/lib.rs

Lines 38 to 42 in 77827a4

	let mut codegen_session: CodeGenSession<_> = CodeGenSession::new(
	translation.function_body_inputs.len() as u32,
	&env,
	lightbeam::microwasm::I32,
	);

All elements in function_body_inputs are taken just before at

wasmtime/crates/jit/src/compiler.rs

Line 131 in 77827a4

let functions = mem::take(&mut translation.function_body_inputs);

leaving the count at zero.

[alexcrichton]

The lightbeam backend is known to currently have many bugs in it and is not tested on CI. It's not recommended for widespread usage. Thanks for the report, though, this is good to know when someone gets around to fixing lightbeam issues!

[daubaris]

Are the lengths of translation.function_body_input and translation.native_signatures always the same or can they be different in some situations?

[alexcrichton]

Lightbeam was removed in #3390 as explained in RFC 14, so I'm going to close this.