Re: [Sbcl-devel] Re: Three "interesting" PCL bugs

[Christophe R. <cs...@ca...>]

Gerd Moellmann <ger...@t-...> writes:

> Christophe Rhodes <cs...@ca...> writes:
>=20
> > The first, as reported by Stig Sand=F8 on #lisp IRC, is an annoyance
> > rather than a critical bug.  But it's quite an annoyance, as can be
> > seen from the following transcript:
> >=20
> >   * (defclass foo () ())
> >=20=20=20
> >   #<STANDARD-CLASS FOO>
>=20
> [...]
>=20
> > These code deletion notes only occur when a variable has the same name
>=20
> I'm afraid I don't see that here.  I've tried at the REPL and
> with compiling a file.  I'm using my PCL.

Damn.  OK.  On more reflection (and some tracing), it doesn't look
like the %CLASS declaration can be the culprit, given that
CAN-OPTIMIZE-ACCESS1 (which seems to be the only place that cares
about said %CLASS declaration) doesn't get called in the fragments
that trigger the note.  The fact remains that SPLIT-DECLARATIONS is
splitting the declarations wrongly, which might cause some non-optimal
slot access later on.  (I don't know how often that code path is
taken)

As for these notes, ugh.  Even after sleeping on it, I don't have much
idea.  Walking two DEFMETHOD forms, one with a variable name the same
as a class name, one without, yields identical expansions apart from
the position of the %CLASS declaration for that argument.

Thanks,

Christophe
--=20
http://www-jcsu.jesus.cam.ac.uk/~csr21/       +44 1223 510 299/+44 7729 383=
 757
(set-pprint-dispatch 'number (lambda (s o) (declare (special b)) (format s =
b)))
(defvar b "~&Just another Lisp hacker~%")    (pprint #36rJesusCollegeCambri=
dge)