os x / make check-fast bug that seems to effect invoke-external-native

[boggle]

https://gist.github.com/1500793

looks like this is concurrency related, cannot reproduce this safely on the machine where i see it and am logging this here

make check-fast CFG_DISABLE_VALGRIND=1

run-pass [stage2]: test/run-pass/invoke-external-native.rs
rust: fatal, 'false' failed, ../src/rt/rust_upcall.cpp:63 Native code threw an exception
make: *** [x86_64-apple-darwin/test/run_pass_stage2_driver-x86_64-apple-darwin.out] Abort trap: 6

make check-fast CFG_DISABLE_VALGRIND=1 DEBUG=1 # seems to trigger different bug

/opt/rust/src/rust/src/test/run-pass/while-with-break.rs:12 [1, 2, 3, 4, 5]; // we check that it is freed by break
rust: "smallintmap::get(): key not present"
rust: upcall fail 'explicit failure', ../src/libstd/smallintmap.rs:59
rust: domain main @0x7fa7ea01d600 root task failed
make: *** [x86_64-apple-darwin/stage2/lib/rustc/x86_64-apple-darwin/lib/librun_pass_stage2.dylib] Error 101

[boggle]

is anyone else seing this?

[brson]

On Linux I don't see the 'native exception' error. I do see the DEBUG=1 error on Linux, which might be related to the new debug info - I noticed test failures after it was added.

[graydon]

Still reproducible?

[brson]

The debug thing is a separate issue and I think is probably covered by other debuginfo bugs. I can't speak to the 'native code threw an exception' error.

[catamorphism]

@boggle : Can you still reproduce this?

[boggle]

No, I can't - but I also updated this machine quite a bit since I last checked this. Still, if no one else is seeing this, it is probably best to close this issue for now.