Commits (1)
......@@ -276,7 +276,8 @@
(is (equal (eval (eval (rq expression))) value)))
t)
#| ;;; This test from dougk's 2014 patch to sbcl's backquote doesn't pass with strict append semantics
;;; This test is from dougk's 2014 patch to sbcl's backquote
#-(or quasiquote-strict-append quasiquote-passes-literals quasiquote-at-macro-expansion-time)
(deftest test-nested-backquote-readable-bogosity ()
(eval (rq "(defmacro broken-macro (more-bindings)
`(macrolet ((with-bindings (&body body)
......@@ -310,7 +311,6 @@
`(let ((thing1 :something) ,(cl:quote . frob)) ,@body)))
(with-bindings (thing)))"
'(broken-macro frob))))
|#
(deftest preserving-inner-backquotes ()
(flet ((e (s v)
......@@ -371,6 +371,7 @@
#+quasiquote-strict-append "`(foo ,.x)"))
(is (equal (prq "`(foo (,x))") "`(foo (,x))")))
;;;; One more test from the SBCL test suite:
;;; more backquote printing brokenness, fixed quasi-randomly by CSR.
;;; NOTE KLUDGE FIXME: because our backquote optimizes at read-time,
;;; these assertions, like the ones above, are fragile. Likewise, it
......
......@@ -14,7 +14,9 @@
;;(pushnew :quasiquote-strict-append *features*)
;;(pushnew :quasiquote-passes-literals *features*)
;;(pushnew :quasiquote-at-macro-expansion-time *features*)
)
(eval-when (:compile-toplevel :load-toplevel :execute)
;;; Functions that actually build data structures.
;; Note that we want our own tokens for decompilation reasons,
;; but as functions they must evaluate the usual way.
......