Upstreaming the glibc Hurd port

[Florian Weimer <fweimer at redhat dot com>]

As far as I understand it, the Hurd port of glibc is currently mostly a 
downstream affair, maintained in the Debian archive.  The current 
upstream tree does not build, even natively on the Hurd itself.

What would be necessary to get it into a state that in can be built with 
a cross-compiler, perhaps even using build-many-glibcs.py?  Are FSF 
copyright assignments in place for all the relevant patches?  Is it just 
a matter of someone doing the work of upstreaming all the existing patches?

I'm asking because we may want to implement rather far-reaching dynamic 
linker and a thread library integration changes, and having a buildable 
in-tree port for Hurd would make it at least possible to ensure that it 
keeps building, and that we do not make choices which are very difficult 
to mirror on the Hurd side.

If we cannot resolve this is in the coming months, I think we should 
seriously consider a removal of the Hurd port, I'm afraid.

Thanks,
Florian