master: Delete obsolete comments
snuglas via Sbcl-commits <[email protected]> Sun, 05 Jul 2026 23:23:50 +0000
| Newsgroups | gmane.lisp.steel-bank.cvs |
|---|---|
| Message-ID | <[email protected]> |
The branch "master" has been updated in SBCL:
via 9104b7ce06b159064916c9c4bce7a30381099a3b (commit)
from de09ec440d283ce3222a7da0f44f33c1a29a83fc (commit)
- Log -----------------------------------------------------------------
commit 9104b7ce06b159064916c9c4bce7a30381099a3b
Author: Douglas Katzman <[email protected]>
Date: Sun Jul 5 23:22:56 2026 +0000
Delete obsolete comments
---
src/compiler/constraint.lisp | 5 +----
1 file changed, 1 insertion(+), 4 deletions(-)
diff --git a/src/compiler/constraint.lisp b/src/compiler/constraint.lisp
index f02144087..47de67f93 100644
--- a/src/compiler/constraint.lisp
+++ b/src/compiler/constraint.lisp
@@ -201,8 +201,6 @@
;; Bit-vectors win over lightweight hashes for copy, union,
;; intersection, difference, but lose for iteration if you iterate
;; over the whole vector. Tracking extrema helps a bit.
- ;; [Well, it _should_ help if you do it right. We need sentinel
- ;; values that are not mistakable for actual indices]
(min 0 :type fixnum)
(max 0 :type fixnum))
@@ -518,8 +516,7 @@
(declare (optimize speed))
;; Notice that in the algorithm based on CTZ we don't actually have to restrict the scan
;; strictly between CONSET-MIN and CONSET-MAX, because those are just hints where to find
- ;; nonzero bits. Interestingly CONSET-MIN seems to have a flaw- we don't maintain it
- ;; properly so it is stuck at 0 thus entirely defeating that optimization.
+ ;; nonzero bits.
#+(and (not sb-xc-host) (or arm64 x86-64)) ; ctz is not implemented everywhere
(let ((,minword (floor (conset-min ,conset) sb-vm:n-word-bits))
(,maxword (floor (1- (conset-max ,conset)) sb-vm:n-word-bits)))
-----------------------------------------------------------------------
hooks/post-receive
--
SBCL