Re: java.net.*Socket merge with libgcj

[Mark Wielaard]

Hi,

On Mon, 2002-01-14 at 03:46, Brian Jones wrote:
> The ChangeLog seems pretty clear on the matter, I could see but one
> change to fix a Mauve regression dated 2001-11-25.
> 
> cvs diff -r 1.9 -r 1.10 DatagramSocket.java
> 
> Essentially we didn't have this functionality implemented in the
> native library where getOption is defined and the test was failing.  I
> determined it was stupid to bother with the native method call since
> the value can be stored locally in class private data (and it never
> changes).

Thanks for explaining. I thought that it was something subtle about
SO_BINDADDR. Seems like doing it completely in Java is a smart thing and
it probably solves the FIXME in the libgcj sources.

I still have some problems with Orp and have not yet setup the mauve
regression tests. I will do this soon.

Cheers,

Mark