Re: [PATCH] Re: package.el changes before the feature freeze

[Stefan Monnier <monnier@iro.umontreal.ca>]

> The CL info pages refer to them as slots, so that's what I figured I'd
> use.  There could be a better name for a package defstruct, but "field"
> "property" and "attribute" are overdone :)

Sure, use whatever suits you.

> I agree that the slot name could definitely be improved. `name' does
> imply a string to me, but I think that it is good for the "primary key"
> of the alists to be a symbol. Something like `canonical-name' perhaps?
> `id' maybe? I'm not terribly attached to any particular slot name.

`name' is no worse than the others, so just use that.

> This was intended to be related to `define-package'; it turns the
> `define-package' call into a `package-desc' struct.

I understand the motivation, but the use of a "package-" prefix is
more important, I think.

> I haven't looked at lexical binding in earnest. Should I reference
> the non-prefixed form in the body of `define-package'?

The symbols prefixed by "_" are not treated specially by the
compiler/evaluator.  The use of _ only affects the warnings you might
get (where a `_foo' that is used or a `foo' that is not used will signal
a warning) and the way things are displayed in *Help* (where the
underscore is stripped, since whether or not an argument is used is an
implementation detail).


        Stefan