|
From: Stas B. <sta...@gm...> - 2026-04-29 21:13:43
|
That accidentally reverted a previous change. But it's untestable. These stack scs are never selected (that's why it's been there for 10 years). On Thu, Apr 30, 2026 at 12:08 AM Christophe Rhodes via Sbcl-commits <sbc...@li...> wrote: > > stassats via Sbcl-commits <sbc...@li...> writes: > > > (defun eval-eval-when (body env) > > - (program-destructuring-bind ((&rest situation) &body body) body > > - ;; FIXME: check that SITUATION only contains valid situations > > - (if (or (member :execute situation) > > - (member 'eval situation)) > > - (eval-progn body env)))) > > + (program-destructuring-bind ((&rest situations) &body body) body > > + (let (execute) > > + (loop for situation in situations > > I wonder if we should also check SITUATIONS for properness? Or at least > non-dottedness? > > > diff --git a/src/compiler/arm64/float.lisp b/src/compiler/arm64/float.lisp > > index acbb7aaa0..7d20ce1f0 100644 > > --- a/src/compiler/arm64/float.lisp > > +++ b/src/compiler/arm64/float.lisp > > @@ -704,7 +704,7 @@ > > (signed-stack > > (sc-case res > > (single-reg > > - (loadw res (current-nfp-tn vop) (tn-offset bits))) > > + (loadw res (current-nfp-tn vop) (tn-offset res))) > > (single-stack > > (unless (location= bits res) > > (loadw temp (current-nfp-tn vop) (tn-offset bits)) > > This doesn't look right to me. Even if it's right, it doesn't belong in > this change, and if it is a correct change (fixing a breakage in > MAKE-SINGLE-FLOAT?) it should come with a test. > > Christophe > > > _______________________________________________ > Sbcl-commits mailing list > Sbc...@li... > https://lists.sourceforge.net/lists/listinfo/sbcl-commits |