Re: Using Hurd features (was: Re: Hurdish applications for persistence)

[Bas Wijnen]

On Wed, Oct 12, 2005 at 04:52:04AM +0200, address@hidden wrote:
> Hi,
> 
> > A problem here is that programs aren't and shouldn't be written solely
> > for the Hurd.
> 
> Yes they should! What's the use in all those nice Hurd features, if we
> aren't allowed to use them??

We are allowed to use them, but only programs which are meaningless on other
systems should use the features directly.

> Compatibility with legacy applications, and familar interfaces, are very
> important -- that's why we take quite a lot of trouble to implement
> POSIX. But ultimately, we definitely *do* want to have native, hurdish
> applications.

Not really.  We want to have native, hurdish libraries.  But they should
emulate their work (probably slower, perhaps less secure, maybe even less
functional) on other systems.

> For (Hurd-aware) authors caring about portability to other platforms, I
> suggest not restricting themselfs to POSIX, but instead doing it the
> other way round: Using native Hurd features, and a layer emulating those
> on top of other systems.

Right.  That is, creating a library, so it is at least easy to port (only the
library needs to be ported).  Glib is an excellent example of such a library
(which doesn't do specific hurdish things of course).

Any time I need to use #include <linux/...> I'm annoyed.  That's because I
know I'm writing code which can only be used with a linux kernel.  And there's
often no real need for it.  For example, <linux/ppdev.h> just defines a way to
access the parallel port.  Why would that be impossible on other systems?
Agreed, it does so through ioctl, but I consider that a bug: the header file
should define inline functions which make the ioctl calls.

Thanks,
Bas

-- 
I encourage people to send encrypted e-mail (see http://www.gnupg.org).
If you have problems reading my e-mail, use a better reader.
Please send the central message of e-mails as plain text
   in the message body, not as HTML and definitely not as MS Word.
Please do not use the MS Word format for attachments either.
For more information, see http://129.125.47.90/e-mail.html

signature.asc
Description: Digital signature