wasm32-unknown-unknown compiling panics intermittently

[daboross]

This doesn't occur on every build, but it's happened twice so far. I'm not the one personally experiencing it, but I can get any diagnostic information needed.

Build log:

O:\Code\PrimalFury\Machiavellian> $env:RUST_BACKTRACE=1 ; cargo build
  Compiling machiavellian v0.1.0 (file:///O:/Code/PrimalFury/Machiavellian)
warning: Hard linking files in the incremental compilation cache failed. Copying files instead. Consider moving the cache directory to a file system which supports hard linking in session dir `\\?\O:\Code\PrimalFury\Machiavellian\target\wasm32-unknown-unknown\debug\incremental\machiavellian-2jdc269m535ub\s-ew19ggigy3-1nx6tpl-working`

warning: unused import: `std::option`
// ... various style warnings ...
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

thread '<unnamed>' panicked at 'must be trans'ing at least one module', src\libcore\option.rs:874:4
stack backtrace:
  0: <std::thread::ThreadId as core::fmt::Debug>::fmt
  1: <std::sync::condvar::Condvar as core::fmt::Debug>::fmt
  2: std::panicking::Location::column
  3: std::panicking::Location::column
  4: std::panicking::rust_panic_with_hook
  5: std::panicking::begin_panic_fmt
  6: std::panicking::begin_panic_fmt
  7: rust_begin_unwind
  8: core::panicking::panic_fmt
  9: core::option::expect_failed
 10: <rustc_trans::back::lto::ThinData as core::ops::drop::Drop>::drop
 11: <rustc_trans::ModuleLlvm as core::fmt::Debug>::fmt
 12: <rustc_trans::ModuleLlvm as core::fmt::Debug>::fmt
 13: _rust_maybe_catch_panic
 14: <rustc_trans::back::lto::ThinBuffer as core::ops::drop::Drop>::drop
 15: <std::sync::condvar::Condvar as core::fmt::Debug>::fmt
 16: std::sys::windows::thread::Thread::new
 17: BaseThreadInitThunk
error: Error during translation/LLVM phase.

error: Could not compile `machiavellian`.

To learn more, run the command again with --verbose.

[alexcrichton]

I think the LLVM backend currently has a number of deficiencies at opt-level=0 in that if you turn asserts on then asserts will probably get tripped all over the place. Does this work if you crank up the opt-level? If so this may just be blocked on upgrading LLVM.

[alexcrichton]

(tracking issue for that)

[alexcrichton]

I'm gonna close this in favor of #46367 which I believe is the same iss as this, in general wasm at opt-level=0 isn't working right now.