0.8.7.3

[sbcl.git] / BUGS

diff --git a/BUGS b/BUGS
index 0101f27..0b8a677 100644 (file)
--- a/BUGS
+++ b/BUGS
@@ -264,6 +264,11 @@ WORKAROUND:
   GC, so that thereafter memory usage can never be reduced below that
   level.
 
+  (As of 0.8.7.3 it's likely that the latter half of this bug is fixed.
+  The interaction between gencgc and the variables used by
+  save-lisp-and-die is still nonoptimal, though, so no respite from
+  big core files yet)
+
 98:
   In sbcl-0.6.11.41 (and in all earlier SBCL, and in CMU
   CL), out-of-line structure slot setters are horribly inefficient
@@ -324,6 +329,15 @@ WORKAROUND:
   time trying to GC afterwards. Surely there's some more economical
   way to implement (ROOM T).
 
+  Daniel Barlow doesn't know what fixed this, but observes that it 
+  doesn't seem to be the case in 0.8.7.3 any more.  Instead, (ROOM T)
+  in a fresh SBCL causes
+
+    debugger invoked on a SB-INT:BUG in thread 5911:
+        failed AVER: "(SAP= CURRENT END)"
+
+  unless a GC has happened beforehand.
+
 117:
   When the compiler inline expands functions, it may be that different
   kinds of return values are generated from different code branches.
@@ -1213,3 +1227,18 @@ WORKAROUND:
   This program violates "unknown values LVAR stack discipline": if INT
   returns, values returned by (EXT) must be removed from under that of
   (INT).
+
+299: (aka PFD MISC.186)
+    * (defun foo ()
+        (declare (optimize (debug 1)))
+        (multiple-value-call #'list
+          (if (eval t) (eval '(values :a :b :c)) nil) ; (*)
+          (catch 'foo (throw 'foo (values :x :y)))))
+    FOO
+    * (foo)
+    (:X :Y)
+
+  Operator THROW is represented with a call of a not returning funny
+  function %THROW, unknown values stack after the call is empty, so
+  the unknown values LVAR (*) is considered to be dead after the call
+  and, thus, before it and is popped by the stack analysis.