Re: [9fans] Can't start multiple copies of acme

[Blake McBride <blake@mcbride.name>]

[-- Attachment #1: Type: text/plain, Size: 1850 bytes --]

On Sun, Dec 15, 2013 at 10:00 AM, erik quanstrom <quanstro@quanstro.net>wrote:

> On Sun Dec 15 10:23:13 EST 2013, blake@mcbride.name wrote:
>
> > I checked.  The following shell script does the trick:
> >
> >
> > #
> > mkdir /tmp/acme-$$
> > NAMESPACE=/tmp/acme-$$ acme "$@"
>
> this is swimming up stream.  acme's model is to run 1 copy of acme,
> and edit all files in it.  many things, such as plumbing, will work
> better with 1 copy of acme.
>

I agree completely.  I would, in general, only run one copy because that is
all that would generally be needed.  However, there are a couple
of circumstances where starting more than one copy does make sense.  I
detail those two cases in my first post in this thread.



>
> by the way, this limitation is p9p specific.  but still it's no
> fun to have the same file get plumbed to every acme.
>

As you can tell by all my posts.  All of this is new to me, and I am tring
to understand it all.  I like what I see so far, and, in fact, I have some
ideas that are germinating in my mind.  I am planning a future post about
it.

Focusing on sam & acme, I don't yet understand the plumber except in the
most vague respect.  Specifically what one can do with the plumber on p9p
is unknown to me.

On a semi-different note, I understand the great advancement Plan-9 brings
to the table with respect to making all operations part of the file system.
 On the flip side, I do not understand the benefit p9p brings to the table
with bind and friends.  It is too much of a tack-on IMO.  I deeply
appreciate native sam & acme, and would appreciate an even more native port
of same.  And, not to dispriage the true benefits of Plan-9, I would love
to see a POSIX implementation of those ideas.  (A topic of a future post.)

Thanks.

Blake



>
> - erik
>
>

[-- Attachment #2: Type: text/html, Size: 2994 bytes --]