Re: glibc; introducing slashpackage-foreign

[Thomas Schwinge]

On Thu, Mar 10, 2005 at 01:17:32AM +0100, Alfred M. Szmidt wrote:
>    >    I have to use './configure [...] --prefix=[...]
>    >    --with-headers=[...]'  because I have every package installed
>    >    into its own hierarchy of
>    > 
>    > Then your system is broken.
> 
>    That's an interesting definition of "broken" you're using here.
> 
> The need for that flag alone makes your systen broken.

Why?  I wouldn't call a system broken that it perfectly working for me.

>    >    Everyone, feel free to ask questions if you want to know more
>    >    about that system.
>    > 
>    > It all seems like a basterdised version of stowfs/packagefs.
> 
>    If that is the niveau you decide to use to discuss methods that are
>    new to you, different from those you're already used to, then there
>    is no use in continuing this "discussion" with you.
> 
> Now you are being really silly... Can't you even take some simple
> criticisim without having to resort to these silly things?

Of course I can.

But I do not want to have this system declared "broken" without getting
an explanation why it is broken.  I'm using that system and it's
working for me (and others), so it's your turn now, if you want to
convince me (and others) that it is broken.

I could declare stowfs/packagefs broken, because it doesn't work on any
other system than the GNU one, but I don't intend to do that, because
it's in fact not broken, but just different.  The same applies to
slashpackage(-foreign).

> What you
> presented is indeed a silly and basterdised version of
> stowfs/packagefs.

Could you please try to discuss on a more technical niveau instead of
merely claiming that the system I use is broken?

Why is it "broken"?  Why is it "silly"?  Why is it "basterdised"?

Configuring glibc with '--prefix=[...] --with-headers=[...]' doesn't
make a system broken, but instead is a valid use of these existing
flags.

>    I was under the assumption that the Hurd people (I'm aware that
>    only Alfed replied so far) would be interested in replacing --
>    well, in some way -- deadlocked UNIX techniques with superior ones.
> 
> Yes, and we do that, stowfs/packagefs is a example.

So, I'll have a deeper look at those and I'll try to use the "best"
parts of everything.

> What you
> presented was a basteredised version of it that is based on the same
> old "deadlocked UNIX techniques" but pretending to be "new" in some
> obscure way.

Why "obscure"?  Because it doesn't have the need for any translators,
but simply uses the filesystem?


Regards,
 Thomas