[Sbcl-devel] merged patch (CSR "Mostly types")

[William H. N. <wil...@ai...>]

On Thu, Mar 07, 2002 at 07:27:40PM +0000, Christophe Rhodes wrote:
> I think I've nailed some more type bugs :)
> 
> In particular, now
> 
> * (subtypep '(integer (0) (0)) 'nil) -> T, T [ was error ]
> * (subtypep '(satisfies foo) 'nil) -> NIL, NIL [ was NIL, T -- to see
>   why this is wrong, consider (defun foo (x) nil) ]
> 
> Although I haven't done it, one piece of unsightliness remaining is
> the remaining special-casing of *EMPTY-TYPE* in CSUBTYPEP
> (src/code/late-type.lisp) -- it seems conceptually wrong to be doing
> it at that level, though that special case is correct where the
> previous one wasn't. I'd much rather have it drop through to the NAMED
> type methods, if only because that will make it clearer. I think it's
> fair for CSUBTYPEP to special-case *WILD-TYPE* given that that's not
> something we want to be hanging around the system for too much longer
> anyway.
> 
> Also in this patch are follow-ons from the fix to make malformed
> format-strings throw errors -- a test and deletion of the BUGS entry.

Thanks, I've merged this in sbcl-0.7.1.34.

-- 
William Harold Newman <wil...@ai...>
"1. If you know what to type, type. 2. If you don't know what to type,
take a shower." -- http://www.c2.com/cgi/wiki?UngarMethod
3. If you didn't say the right thing, don't hit the send key.
PGP key fingerprint 85 CE 1C BA 79 8D 51 8C  B9 25 FB EE E0 C3 E5 7C