Re: frameset-to-register

[Juanma Barranquero <lekktu@gmail.com>]

On Mon, Aug 5, 2013 at 10:35 PM, Drew Adams <drew.adams@oracle.com> wrote:

> Why does this not just replace `frame-configuration-to-register', i.e.,
> C-x r f?  What are the differences?  In particular, what, if anything,
> does `frame-configuration-to-register' let you do that
> `frameset-to-register' does not let you do?

I don't know, and that's the point. I've used
frame-configuration-to-register all of perhaps ten times, eight or so
just to test something. I think it makes more sense to leave both
options and let the user choose.

> I understand that a frameset is Lisp-readable, so it can be persisted.

Caveat user. The frameset I'm using for frameset-to-register is not
really "Lisp-readable", because I've opted to filter out as few frame
parameters as possible. That means that non-readable parameters, like
buffer-list, are kept. Why? Because frameset-to-register, like
frame-configuration-to-register, is an in-session command, and keeping
links to live (or dead) objects provides a better in-session user
experience. If someone someday decides to save to disk the frameset
contained in a register, we'll have to add code to filter out
additional parameters (should be trivial, once the use case is clearly
defined).

> But it's not clear why one would use `frame-configuration-to-register'
> if `frameset-to-register' is now available.

Perhaps you don't want dead frames to come back to life.

> I can guess that `frame-configuration-to-register' might be quicker,
> but what are the real, user-level differences that would mean that a
> user might want both to be available (and bound to keys)?

You try and tell me ;-)

   J