Copying .emacs.d

[Rusi]

Is it ok to copy over .emacs.d from one machine to another (assume similar OS)?

I am specifically thinking of packages and elpa directory -- is there some kind
of fingerprinting or is it legitimate?

Context: I'd like to setup one minimal box for my students and then
they can copy over .emacs.d from there to their laptops

[Emanuel Berg]

There are several things to consider.

With respect to technology:

1) If the packages save data in .emacs.d
there is no saying how specific that
data is.

2) The software is compiled - someone else
can answer, if the Emacs byte compiler
take external factors into account or if
it is boxed to an Emacs version.

3) The state (metadata) of local package
manager will at best refer to events
that took place on another computer.

But more importantly:

If this has to do with students, wouldn't it be
a good thing for them to learn how to install
a program from the Emacs software archives?
Because it is very easy, and useful, not just
for your specific task but for future
Emacs use!

--
underground experts united .... http://user.it.uu.se/~embe8573
Emacs Gnus Blogomatic ......... http://user.it.uu.se/~embe8573/blogomatic
- so far: 40 Blogomatic articles -

[Rusi]

On Thursday, May 26, 2016 at 1:10:24 PM UTC+5:30, Emanuel Berg wrote:

> Rusi writes:
>
> > Is it ok to copy over .emacs.d from one
> > machine to another (assume similar OS)?
> >
> > I am specifically thinking of packages and
> > elpa directory -- is there some kind of
> > fingerprinting or is it legitimate?
> >
> > Context: I'd like to setup one minimal box
> > for my students and then they can copy over
> > .emacs.d from there to their laptops
>
> There are several things to consider.
>
> With respect to technology:
>
> 1) If the packages save data in .emacs.d
> there is no saying how specific that
> data is.
>
> 2) The software is compiled - someone else
> can answer, if the Emacs byte compiler
> take external factors into account or if
> it is boxed to an Emacs version.
>
> 3) The state (metadata) of local package
> manager will at best refer to events
> that took place on another computer.

[For anyone else who finds this thread...]

I subsequently found http://emacs.stackexchange.com/questions/10577/export-emacs-installed-packages
which suggests (byte-recompile-directory "~/.emacs.d/elpa" 0 t)

[Eric Brown]

An example of a package that might not transfer is pdf-tools, as
it has
a compiled binary.

[Emanuel Berg]

This sounds almost as the beginning of not
Emacs forks, but distributions!

Just think about it - people can use the exact
same software, only add a couple of packages on
top, rebrand the whole thing, and then there
can be forks/redistributions of that
distribution, and so on, and it can all be
hailed as software multitude and proliferation!

Now, if computer science students can't be
trusted to install a bunch of packages from
ELPA, we should all give up, and put the
Russian ten thousand ton atomic submarines to
nuke the entire good-for-nothing university
world from the face of the earth. Only perhaps
we should keep LaTeX...

--
underground experts united .... http://user.it.uu.se/~embe8573
Emacs Gnus Blogomatic ......... http://user.it.uu.se/~embe8573/blogomatic

- so far: 42 Blogomatic articles -

[Stefan Monnier]

> Is it ok to copy over .emacs.d from one machine to another (assume similar OS)?

Yes,


Stefan

[Emanuel Berg]

Stefan Monnier <mon...@iro.umontreal.ca>
writes:

>> Is it ok to copy over .emacs.d from one
>> machine to another (assume similar OS)?
>
> Yes

Is it also within the ordo and omega of reason
that CS students get supposedly sandboxed Emacs
handed out as they aren't entrusted to use ELPA
and the package manager themselves, all of
which you probably wrote, and by the way is
super easy to use?

This is the stupidest thing I ever heard.
The OP would be better off going to Crooks"RUs,
buy a bag of transistors, and throw it at his
students, and have them work something sensible
out of that. The can install Emacs in their
spare time.

--
underground experts united .... http://user.it.uu.se/~embe8573
Emacs Gnus Blogomatic ......... http://user.it.uu.se/~embe8573/blogomatic

- so far: 43 Blogomatic articles -

[Emanuel Berg]

Stefan Monnier <mon...@iro.umontreal.ca>
writes:

>> Is it ok to copy over .emacs.d from one
>> machine to another (assume similar OS)?
>
> Yes

Two more things on this. It is obviously a good
thing that it is portable like this. If it
wasn't, that would be bad. So it is good that
it is movable in terms of technology, for sure.

I have moved my zsh, Emacs, tmux, mplayer, etc.
init (initialization), config, and extention
files between several machines by now.
At first, it was a heck of a job. Now, it goes
much smoother. Total portability tho I haven't
seen. There are always many small things that
break. With Emacs most recently, when I moved
from Debian to Raspbian, I had some problems
with biber (biblatex) and Gnus TLS. The Gnus
problem remains, but the biber was an upgrade
away. I think both problems were related to the
software on Debian was more upgraded.
With actual Emacs, all the Elisp, I don't
remember having any problems.

The best way I found so far is to have
a directory with all the stuff. With rsync(1)
you put it somewhere. On the new machine, you
get it, and then create symlinks where the
programs look for their files.

To put the actual files there is a lot of work.
And then even more so to put them somewhere,
with rsync, as you're likely to continue edit
and refine them, on the new machine! It is too
much diaspora.

The second thing is: universities having Emacs
on their systems and including it in their
education. This also is very good.
Before I attended thrice-accursed computer
science school, I used nano(1). At that time,
I didn't know an editor could be an interface
to the entire system, programmable at that.
But the first course at the university had
a part where we were required to do basic stuff
in Emacs. I don't remember this showing any of
the programmable/all-interface stuff. Still, it
was enough to convince me to start using it, and
I have ever since (even writing this mail/post,
otherwise mailing would be unbearable).

So it is good .emacs.d is portable and it is
good that universities expose students to
Emacs. But I'm still right about the bag of
transistors etc.!

Now, ssh-ing to my school's SunOS 5.10, I see
there is no Emacs binary anymore. Nor nano for
that matter. There is only vi.

What are we going to do about it?