First, I'm not against incremental changes, I'm for all the way.
Second, we're not talking about entire clos, but a very specific problem. This problem is not unique for clos, but just heavily used by it.
What I'm afraid of is there is a more fundamental issue here and this patch makes it even more hidden.
BTW, ANSI standard does not allow &aux in generic functions (sec 3.4.2), but this version of clos treats it as a legitimate citizen.
Question: how do you explain why the patch does not catch reinitialize-instance.error.1? How does this case differ from the other 3?
 
> Date: Tue, 28 Jul 2009 01:37:13 +0300
> Subject: Re: [armedbear-devel] A clos patch for review
> From: ville.voutilainen@gmail.com
> To: ptsenter@hotmail.com
> CC: armedbear-devel@common-lisp.net
>
> 2009/7/28 Peter Tsenter <ptsenter@hotmail.com>:
> > The patch seems to produce a result as claimed.
> > But it's not generic enough. It does not catch, e.g., defgeneric.error.20
> > nor reinitialize-instance.error.1, which follow the same pattern: supply
> > bogus arguments in initargs. Introducing this patch might create an
> > impression that such cases handled properly always.
>
> Well, I'd rather do these fixes piecemeal than attempt to produce some perfect
> megapatch that fixes everything clos-related in a single sweep.

---

Bing™ brings you maps, menus, and reviews organized in one place. Try it now.