Re: Status of the Hurd parts in glibc

> The Hurd parts of glibc need maintenance!

I keep saying I'll do this, and it's hard for anyone to do it without my help.
In the longer term, most of the Hurd parts of libc need a wholesale rewrite.

Keeping track of the small stuff, and the too-hairy-for-most stuff like TLS
details, I would do much better if I did more regular builds.  I have a
Hurd cross-compiler setup and I just need to be whacked in the head from
time to time to remember to do that build and fix up trivia.  More useful
work requires that I have a reliably useful Hurd testing setup at home,
which I have not managed to keep going any time recently.  (I have adequate
machine and bandwidth resources handy, but the time and effort to bother
setting things up has been more of a barrier.)

> What also could / should (Roland?) be done is to back-port the following
> from GNU Mach's trunk (a.k.a. OSKit-Mach) to the gnumach-1-branch:

AFAIK, just do it.  If you don't understand the implementation details when
you look at the code, I can explain it.  I'm not at all sure the oskit-mach
code ever worked perfectly, but I know the principle is sound and I don't
think there should be any rocket science in getting an implementation of it
working in whichever code base.


Thanks,
Roland