Object alignment, was: cygwin failures - assertion "!(addr &F LAGS)" failed:

[Bryce McKinlay]

Boehm, Hans wrote:

>Locking a statically
>allocated plain java.lang.Object may be an issue, since I think it's too
>small.
>

I guess thats okay, provided the compiler never emits static 
java.lang.Objects. I'm pretty sure it doesn't now.

>I think we concluded that the cygwin port currently doesn't support threads
>anyway (since the GC port hasn't been integrated yet), so I don't think
>we're that close to making this really useful.
>  
>

So, what is the attraction of a cygwin port anyway when we have a 
perfectly good win32 native port? To allow developers on win32 to test 
the posix parts of libgcj, perhaps?

regards

Bryce.