> >I recently imported current into our cross development tree, and 
> >tried to move our compilers forward to the same version of egcs as
> >in current, and am getting an internal compiler error in a few places.
> >
> >Do people see this in native builds?  I've sent in a report to egcs-bugs,
> >but have not heard anything back.  Trying this on the latest egcs snapshot
> >also fails.
> >
> >Here's an example of the code which kills it.
> >
> >	Regards,
> >	Castor
> >
> >	struct biggy {
> >		char	buf[32800];	/* len further than 16 bit offset */
> >		char    len  [10 ];
> >	};
> >
> >	static
> >	void f ( struct biggy * s, int v)
> >	{
> >		 s->len[v] = 13 ;
> >	}

> Is the above snippet derived code in netbsd's tree, or proprietary
> code of your own?

This bug will occur building libbz2 and compress (or is it gzip?)  Somehow
they all share the property of having really big structures.

I was able to build the rest of the netbsd tree, in cross compilation
(after a few tweaks because I'm using a big-endian mips platform).  The
kernel seems to be reasonably stable, but there are still bugs which I
haven't pinned down.

Incidentally, I've been very happy cross-compiling with egcs-1.0.2 to
mips.  All my previous code was built using 1.0.2, which I settled upon,
because gcc-2.7.x on Intel had many problems with long long support.

Since this bug occurs when just doing the build, there's probably
no real need for a regression test.   Hopefully there won't be others.

	-castor