[MiNT] Sparemint and FHS

[Guido Flohr <gufl0000@stud.uni-sb.de>]

Hi,

On Sat, Sep 04, 1999 at 08:28:33AM +0000, Martin-Eric Racine wrote:
> > Sparemint is rather a software archive.  
> 
> And this is where I stop:
> 
> Several docs on your site imply FHS compliance, even though
> most packages follow the older FSSTND standard.  MAN is one
> good example of this:  you put pages in /usr/man and cache 
> them in /var/catman instead of the correct /usr/share/man
> and /var/cache/man.

I would rather say "recommended" than "correct".

> IMHO, if someone is gonna upgrade the filesystem hierarchy, 
> one might as well be 5 years ahead of others rather than 5 
> years behind, so why not go the FHS way, all the way?

It's all explained in the docs.

I would really prefer not to lead that FHS discussion again.
I have the FHS printed out before me, I have read it and
I already explained where and why Sparemint deviates from 
it.

Anybody who cares to make existing Sparemint packages FHS
compliant is welcome to do so.  But please keep it consistent.
Inconsistency is the main reason why Sparemint (and Redhat
Linux, SuSe Linux AFAIK, Slackware, NetBSD, FreeBSD, ...)
is not yet FHS compliant.

Take your manpage example: If you want to put manpages into
/usr/share/man and not /usr/man, you also have to take care
that all manual readers actually search that directory for
manual pages.  Assuring this would mean that we would have
to modify a lot of other packages, too, like startup scripts,
user home directory skeletons, shell init files ...
Most of these packages have not even been built.

Another point: utmp and wtmp files.  Yes, speaking with FHS
they are currently mislocated.  Who cares?  They are not
human-readable, programs have to find them.  Apart from the
fact that there are very many programs that have to be changed
at once, including programs that we don't have the sources
of, there is another important point: The format of our utmp
files is obsoleted, some important fields (e. g. lastlog)
are missing.

The wrong location is in my eyes only mildly repulsive.  The
missing features are more of a nuisance to me.  If we stick
with the "wrong" location for now, we have a good chance to
add the missing features.  Old programs that do not support
the new features will always read/write the old format files
in the wrong places and new programs will read/write the
new format files in the "right" places.  But we first have to
agree on a new format for utmp files.  Changing the location
of them now would be stupid in my opinion, because that
would lead to chaos.

In brief: I would really like to work with emacs under MiNT
(if this is ever possible).  If there was a full-featured
emacs I wouldn't give a ... (f-word omitted) whether it gets
installed in /usr/emacs, /opt/emacs, or even "c:\windows\emacs"
as long as I can work with it.

So, can we stop that discussion?

Thanks

Guido
-- 
http://stud.uni-sb.de/~gufl0000/
mailto:gufl0000@stud.uni-sb.de