fxbit-set? properly handles negative i

CHANGELOG.md

@@ -10,6 +10,7 @@ Features
 Bug Fixes
 
 - @nmeum fixed `(scheme repl)` to flush the output port prior to writing the prompt, guaranteeing the prompt is written at the correct time. 
+- Fixed `fxbit-set?` to properly handle negative values of `i`.
 - Avoid unnecessary renaming of identifiers when the interpreter performs macro expansion.
 - When allocating a large vector we now guarantee all vector elements are initialized before the major collector can trace those elements. This avoids the potential for a race condition which could lead to a segmentation fault.
 - Ensure atomic objects are properly traced by the major garbage collector.

srfi/143.sld

@@ -50,7 +50,7 @@
     fxarithmetic-shift 
     fxarithmetic-shift-left fxarithmetic-shift-right
     fxbit-count 
-    fxif fxbit-set? fxcopy-bit
+    fxif fxcopy-bit
     fxfirst-set-bit 
     fxbit-field
     mask
@@ -170,7 +170,12 @@
       (fxior (fxand mask n0)
              (fxand (fxnot mask) n1)))
 
-    (define-c fxbit-set?
+    (define (fxbit-set? index i)
+      (or (%fxbit-set? index i)
+          (and (negative? i)
+               (>= index (fxlength i)))))
+
+    (define-c %fxbit-set?
        "(void* data, int argc, closure _, object k, object index, object i)"
        " Cyc_check_fixnum(data, index);
          Cyc_check_fixnum(data, i);