libgcj and the NPTL posix threads implementation

[Andrew Haley]

Anthony Green writes:
 > On Mon, 2004-03-01 at 06:12, Andrew Haley wrote:
 > > Anthony Green writes:
 > >  > On Mon, 2004-03-01 at 03:02, Andrew Haley wrote:
 > >  > > What happens with this test case if you use LD_ASSUME_KERNEL=2.3?
 > >  > 
 > >  > Segmentation fault.  Do you think this means we shouldn't use this?
 > > 
 > > Yes, I think it does.  We shouldn't crash with LD_ASSUME_KERNEL=2.3.
 > 
 > I think we should push this problem over to Linux distro package
 > maintainers.  Isn't the right way to do this is to have a libgcj in
 > /usr/lib/tls and another in /usr/lib, just like libc?

That would be my guess, yes.

Eventually the LD_ASSUME_KERNEL kludge will go away, but until then we
have to deal with the possibility that people will use gcj with
libraries that need to use LD_ASSUME_KERNEL=2.3.  We don't want to
break libgcj in such situations.

 > gcc/libjava should once again configure and build outside of the GCC
 > tree for this to happen.

Eehhh...  Is it really worth it?  It's easier to have a runtime
switch, at least until pre-NPTL glibc disappears from Linux distros.

Andrew.