Fixes for release

* Restore 0-fill of DX arrays

* Restore non-merging of pathnames in self-build.
  (Make the code match the comment)

* Unbreak #+msan
This commit is contained in:
Douglas Katzman 2021-05-30 11:26:03 -04:00
parent 1f4d583988
commit be1773175f
4 changed files with 4 additions and 7 deletions

4
NEWS
View file

@ -4,10 +4,6 @@ changes relative to sbcl-2.1.4:
* optimization: file compilation now produces smaller fasls for files which
reference package literals.
* optimization: faster function calls on ARM64.
* incompatible change: on certain platforms (currently just x86-64),
dynamic-extent arrays specialized on character and numeric types and
created without either :INITIAL-ELEMENT or :INITIAL-CONTENTS will reflect
previous contents of the stack instead of #\null (or 0) in all elements.
* bug fix: on x86-64, fix instruction encoding for TEST on RIP-relative
addresses. (lp#1925808, reported by Shinmera on #sbcl, thanks also to 3b)

View file

@ -916,7 +916,8 @@
(t
;; otherwise, reading an element can't cause an invalid bit pattern
;; to be observed, but the bits could be random.
data-alloc-form))))))
;; KLUDGE: backward-compatibile 0-fill
`(sb-vm::splat ,data-alloc-form nwords 0)))))))
;; Case (3) - constant :INITIAL-CONTENTS and LENGTH
((and c-length

View file

@ -843,7 +843,7 @@ necessary, since type inference may take arbitrarily long to converge.")
;;; The normal build sets it to T in make-target-2, despite that I think most people would
;;; prefer the nonstandard behavior. The standard behavior makes stored pathnames all wrong
;;; when files are physically moved. (Same problem as SBCL_HOME embedded into C pretty much)
(defglobal *merge-pathnames* t)
(defglobal *merge-pathnames* nil)
;;; Given a pathname, return a SOURCE-INFO structure.
(defun make-file-source-info (file external-format &optional form-tracking-p)

View file

@ -550,7 +550,7 @@
(inst xor rdi rsp-tn) ; compute shadow address
(zeroize rax)
(inst rep)
(inst stos rax)))))
(inst stos :qword)))))
;;; ALLOCATE-LIST
(macrolet ((calc-size-in-bytes (length answer)