Re: java on cygwin - 2082 tests pass

[Tom Tromey <tromey at redhat dot com>]

>>>>> "David" == Billinghurst, David (CRTS) <David.Billinghurst@riotinto.com> writes:

David>  - CYGWIN=check_case:strict
David>  - rebuilt binutils with 16-byte alignment
David>  - (and a small testsuite cheat) 

I assume you are building with threads enabled?
As I understand it the GC currently doesn't support threads on Cygwin,
though Hans has a pending patch.

David> FAIL: Array_3 execution from source compiled test

This tests null pointer exception creation and handling.

David> FAIL: Divide_1 execution from source compiled test

This tests divide by zero exception creation and handling.

David> FAIL: KeepInline output from source compiled test

>From the test:

    // Demonstrate that private methods can be reflected even if they are
    // not referenced at compile-time (i.e. -fkeep-inline-functions works).

It is odd that this one fails.  What happens?

David> FAIL: SyncTest execution from bytecode->native test
David> FAIL: Thread_Interrupt output from source compiled test
David> FAIL: Thread_Wait_Interrupt output from bytecode->native test

These all require threads.

David> FAIL: Throw_2 execution from source compiled test

This is another null pointer exception creation and handling test.

Overall this is pretty good.  To me it says:

* You need threads to work
* The signal-to-exception code isn't working.  That's not too
  surprising.  Maybe Cygwin needs to use the
  SetUnhandledExceptionFilter code from win32.cc?
  (I'm assuming it doesn't already.)

KeepInline is the only mystery.

Tom