Re: Finding packages to enable by default

[Stefan Monnier]

>> I think you should filter out the ones that appear in `features' after
>> "emacs -Q".
> This will hardly make a dent in the results, becase there 6000 items
> in the list and emacs with -Q returns only 110 features for me.

Of course, "emacs -Q" has different features depending on the particular
Emacs you use, so we can trim it down a bit more this way, but if we
look at your sample:

> (x . 268)
> (x-dnd . 267)
> (x-win . 267)
> (x-toolkit . 264)
> (newcomment . 264)
> (ring . 259)
> (advice . 257)
> (font-render-setting . 253)
> (macroexp . 252)
> (dbusbind . 249)
> (cl . 246)
> (comint . 243)
> (tabulated-list . 241)
> (easy-mmode . 239)
> (kmacro . 236)
> (edmacro . 232)
> (gtk . 231)
> (byte-compile . 230)
> (bytecomp . 229)
> (byte-opt . 229)
> ...

We see that things like `x-win', `gtk', `macroexp', and `newcomment'
might be trimmed with a more careful handling of "emacs -Q".  But the
bulk of the "spurious" features that remain is made of packages which
"can't be *enabled*".

> These seem fairly basic stuff.

Indeed.

> I should mention that I use 24.1.1. Is it possible that the current
> emacs has a much longer features list with -Q?

Longer, yes, much longer, no.

I can't think of an easy to way to characterize a feature which can be
"enabled".  At least, not a way amenable to mechanization.  So I think
a human will have to go through the list and try to classify the
various features by hand.

Tho, maybe we could start by focusing on minor modes: go through the
list of packages, see if they correspond to an Elisp file, search the
file with a `define-minor-mode'.  It wouldn't have found uniquify, and
it will find packages which simply use minor-modes "internally", but
that might be a start.


        Stefan