M3 stabilization requests for 2009/4/30

[Paul Hohensee]

The fix is actually more trivial than it looks.  Most of the change is just
moving source code slightly so that native code gets generated into the
code cache rather than into a C++ static byte array.  I'd like it accepted.

Paul

Mark Reinhold wrote:
> Today we have:
>
>   6762511  anthony.petrov   Translucency is not working on Linux using
>                             Metacity
>
>   6794764  artem.ananiev    Translucent windows are completely repainted
>                             on every paint event, on Windows
>
>   6812298  anthony.petrov   Dynamic GraphicsConfig changes don't work on
>                             X11 platforms
>
> These first three are nontrivial changes, and none seems critical for M3.
> Phil, Vita -- what say you?
>
>   6834177  vladimir.kozlov  Running jsynprog on Solaris Nevada can cause
>                             JVM crash
>
> VM crash on 64-bit sparc.  The fix isn't trivial but it's smaller than
> it looks, and VM crashes are bad.  I'm leaning toward accepting it.
> Paul? Vita?
>
>   6834246  alan.bateman     (ch) AsynchronousSocketChannel#write
>                             completes with wrong number of bytes written
>                             under load (win)
>
> Not critical for M3 but the fix is limited to the Windows asynchronous
> socket-channel code, so if it's wrong then that's the only thing that
> would break.  I'd say "yes" to this one.
>
> - Mark
>