[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [STUMP] interactive command in symbol plist
From: |
Howard Yeh |
Subject: |
Re: [STUMP] interactive command in symbol plist |
Date: |
Sun, 4 May 2008 23:32:24 -0700 |
On 5/4/08, Shawn Betts <address@hidden> wrote:
> > Using symbol-plist means it's possible to take advantage of the
> > package system. Using hash, we are effectively back to a single
> > namespace. As mechanisms, there's no difference between the two. But
> > with package it's much easier for library writers to avoid naming
> > conflict.
> >
> > Else we'd have to prefix every defnitions with the package name, as in
> > Emacs. Which sucks...
>
>
> command names are symbols, so they do belong to a package. But when
> you use them interactively they're converted to a string. If you mean
> that the package would prefix the command name then the only
> difference would be a : vs a -. I fail to see how using the
> symbol-plist would benefit. It seems it would be slower than a hash
> because you'd have to search all packages and all symbols
I think you got what I mean backward. Suppose the homepackage is
stupwm. To invoke stumpwm:foo interactively, user types "foo" (i.e.
(intern "foo") ). To invoke emacs:foo, user types "emacs:foo". Vice
versa.
This is not just pedantic quibbling... I hope not. I am thinking about
some kind of mode framework, and not using the package system might
make it more clumsy than necessary.
So my point isn't that "stumpwm:foo" rather than "stumpwm-foo" is
somehow easier for M-x.
> it would mean reimplementing something that works fine already.
That's the rub. I can always do it on an experimental branch.
But I should take a closer look at it.