[Ecls-list] Feedback needed for next release

[Alexander Gavrilov]

> Apart from the mingw32, which I will only be able to deal with when I
> return to Madrid in some time, did people notice any other
> inconveniences, instabilities, etc? Otherwise I will produce a new
> release in one or two days.

This fixes a typo in the binary comparison macro, and
allows properly disabling new IEEE floating point features:


diff --git a/src/c/main.d b/src/c/main.d
index 990e4a4..3678740 100644
--- a/src/c/main.d
+++ b/src/c/main.d
@@ -90,7 +90,9 @@ static char stdout_buf[BUFSIZ];
 
 static const char *feature_names[] = {
         "ECL", "COMMON", ECL_ARCHITECTURE, "FFI", "PREFIXED-API",
+#ifdef ECL_IEEE_FP
         "IEEE-FLOATING-POINT",
+#endif
 #ifdef WITH_GMP
         "COMMON-LISP",
         "ANSI-CL",
diff --git a/src/c/predicate.d b/src/c/predicate.d
index 9e6beaa..8860586 100644
--- a/src/c/predicate.d
+++ b/src/c/predicate.d
@@ -275,7 +275,7 @@ cl_eq(cl_object x, cl_object y)
                 case 6: if (c[5] != d[5]) return 0;    \
                 case 5: if (c[4] != d[4]) return 0;    \
                 case 4: if (c[3] != d[3]) return 0;    \
-                case 3: if (c[1] != d[2]) return 0;    \
+                case 3: if (c[2] != d[2]) return 0;    \
                 case 2: if (c[1] != d[1]) return 0;    \
                 default: return (c[0] == d[0]);        \
                 } }
-- 
1.6.3.2.13.g94af7


By the way, I'm using a 'ecl-readline' library, and after switching to
latest ecl, I had to add the following assignment to its initialization
code, because without it readline wasn't used in the topmost loop.
Is it the expected behavior?


   (c-use-symbol-completer)
   (setf system::*tpl-prompt-hook* #'prompt)
   (setf system::*terminal-io*
         (make-two-way-stream (make-instance 'input-stream) 
                              (two-way-stream-output-stream *terminal-io*)))
+  (setf system::*standard-input* system::*terminal-io*)


(I load the library from my .eclrc)

Alexander