Re: Linux/Hurd compatibility
On Mon, May 10, 1999 at 10:35:12PM +0100, Adam Sampson wrote:
> On Mon, May 10, 1999 at 01:35:20AM -0400, David Waite wrote:
> > > I would tend to call non-portability a bug.
> > Unfortunately non-portability is a bug that the programmer usually can't
> > test (as they only have one system). =)
>
> But hey, what's free software for? :)
Exactly. Don't be discouraged. Most upstream authors are well meaning and
accept patches.
> > This would be what I would vote for. I feel that both /dev and /proc really
> > aren't that great, /proc is nice for users but things move and change format
> > so often that it breaks all sorts of programs.
>
> The trouble is that /proc is inconsistent even within itself (file formats
> and the like). However, the only programs that use /proc tend to be
> reasonably system-specific anyway, and could certainly be improved on a
> GNU/Hurd system. I'd like a "top" that's more like OpenBSD's systat program.
I have less problems with programs that use /proc than with programs which
rely on the slight POSIX incompleteness/incompatibility of Linux. For
example, it is a bit easier to program if you know PATH_MAX is an integer...
but if it is not defined, you need to check with sysconf and cope with
unlimited paths. On Linux, PATH_MAX is defined, so most people simply rely
on it.
If someone wants to work on such simple incompatibilities, I am happy to
help with distributing tasks.
BTW, /dev/ is almost a non-issue. Most stuff there is the same between
unices, like "random", "zero", "null" etc.
> > [...] I think that the Hurd project (I try not to say we* since all I have
> > *done so far is planned to write code, not contribute anything) needs to
> > get more developers and to get things like PPP working before this is an
> > option.
>
> Well, yes; it probably would make sense to have something to ship before we
> start shipping. :)
Of course we need more developers. How about an "Adam" or a "David"? :)
> Why is ppp support taking such as long time?
Nobody works on it. Even me, who has only a PPP connection at home, thinks
that PPP is not highest priority. But nobody will stand in your way if you
think it is high priority for you and want to do this!
> Perhaps it
> would make sense to implement the simpler SLIP protocol first?
I don't think so. PPP is much more common. However, implementing PPP or SLIP
should not differ too much. It is not the protocols complexity standing in
the way (there are fine example implementations, for example under Linux).
It just needs somebody know has time and knows or wants to know about Hurd
network translators.
Thanks,
Marcus
--
`Rhubarb is no Egyptian god.' Debian http://www.debian.org finger brinkmd@
Marcus Brinkmann GNU http://www.gnu.org master.debian.org
Marcus.Brinkmann@ruhr-uni-bochum.de for public PGP Key
http://homepage.ruhr-uni-bochum.de/Marcus.Brinkmann/ PGP Key ID 36E7CD09
Reply to: