[Gnewsense-dev] Re: Hurd

[Yavor Doganov]

Tekk wrote:
> I was just wondering how hard it would be for us to port Gnewsense over 
> to Hurd when it's done, can anybody give some information on that?

Very hard (close to impossible) even if gNS moves to deriving from
Debian.

The problem with the GNU/Hurd port [*] is that most upstreams do not
consider it a viable platform.  Even if some of them acccept patches,
they quickly forget that e.g. unconditional use of PATH_MAX is Posix
incompatibility, and the next version of their software relies on the
definition of this symbol somewhere else (which of course leads to
FTBFS, and yet another porting effort in another area of the same
package).

So the sad story is that the GNU/Hurd porters are doomed to fix such
issues constantly, package after package, and in packages where such a
problem was _reported and fixed_ some time ago.  I am almost sure that
gNewSense will not have the available (>= 1/2) *cabaple* developers in
the near future, and it is questionable if it is a good idea in
general the divert the porting effort off Debian (which is, as you may
know, the only distro that cares about GNU/Hurd).

[*] Of course the ultimate problem is that the kernel is not yet ready
for production use, and never will be.  Well, never say never, but
many people rightfully subscribe to this statement.