[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: gcc-2.6.3

>>>>> "Waldi" == Waldi Ravens <waldi@moacs.indiv.nl.net> writes:

Waldi> Christian Lynbech wrote:

Waldi> You have to be a little patient with the FSF guys, GNU is not a full
Waldi> OS yet. But lately they seem to have adopted the 4.4 BSD filesystem
Waldi> layout.

Ok, I didn't know where GNU was heading, I had my head buried in what
it is today (and this is also why I thought libexec was non-standard).

Waldi> I suppose the most appropriate prefix for a distribution based on
Waldi> MiNT would be /usr. AFAIK we don't have any vendor supplied tools,
Waldi> so any freely distributable distribution will contain either GNU
Waldi> or BSD tools in /usr/bin. If a distribution build around the GNU
Waldi> tools were to install those tools in /usr/gnu/bin, what would be
Waldi> lying around in /usr/bin?

Waldi> I believe the most appropriate locations for distributions primarily
Waldi> build around GNU software are /usr/bin for the GNU stuff, and possibly
Waldi> /usr/ucb for any additional BSD software. For distributions primarily
Waldi> based on BSD tools, I would suggest /usr/bin for the BSD stuff and
Waldi> /usr/gnu for the GNU binaries (if any).

But is it completely obvious that every MiNT user, uses GCC? Isn't
Lattice and sozobon also viable option? Not that we cannot decide that
GCC is the standard compiler for MiNT and thus belongs in /usr, but I
for one think we shouldn't make that tight a coupling between MiNT and
GCC. In other words, I am not convinced that we should promote a
GNUish MiNT, even if that is what many of us here on list uses (I
certainly do), as I think MiNT can exist and be usefull without the
GNU stuff.

My concern is that if we here define /usr to be the prefix to use, and
thus implicitly puts down a lot of conventions in the /usr hierarchy,
we lose some of the potential impact of an agreed upon standard if we
are the only one who has agreed, i.e. /usr is the target for many
peoples attention and I would guess that we then need some rather
widespread acceptance. If we use /usr/gnu (or some equivalent) we can
be quite sure (I hope) that we have that corner to ourselves (not a
very precise statement, I know, but I hope you get the intention).

Anyway, I now feel confident that we largely agrees on goeing with the
GNU conventions, with the argument more being on the precise value of
the prefix. I mistook the talk about libexec as being something
completely unGNUish and that was what made me jump originally.

Christian Lynbech               | Hit the philistines three times over the 
office: R0.33 (phone: 3217)	| head with the Elisp reference manual.
email: lynbech@daimi.aau.dk	|        - petonic@hal.com (Michael A. Petonic)