WH>А еще ты, наверное, веришь в вечный двигатель и Кощея бессмертного.

CL-USER 1 > (defun wolf-hound-hobby ()
              :nemerle)
WOLF-HOUND-HOBBY

CL-USER 2 > (disassemble 'wolf-hound-hobby)
21BE5822:
       0:      89E6             move  esi, esp
       2:      81CEFCFF0F00     or    esi, FFFFC
       8:      3966D0           cmp   [esi-30], esp
      11:      7310             jnb   L1
      13:      80FD00           cmpb  ch, 0
      16:      750B             jne   L1
      18:      55               push  ebp
      19:      89E5             move  ebp, esp
      21:      B8570D7121       move  eax, 21710D57    ; :NEMERLE
      26:      FD               std   
      27:      C9               leave 
      28:      C3               ret   
L1:   29:      E83E4A53FE       call  2011A282         ; #<Function RUNTIME:BAD-ARGS-OR-STACK 2011A282>

CL-USER 3 > (defun wolf-hound-hobby-p (hobby)
              (eql hobby :nemerle))
WOLF-HOUND-HOBBY-P

CL-USER 4 > (disassemble 'wolf-hound-hobby-p)
219C8EEA:
       0:      89E6             move  esi, esp
       2:      81CEFCFF0F00     or    esi, FFFFC
       8:      3966D0           cmp   [esi-30], esp
      11:      7317             jnb   L2
      13:      80FD01           cmpb  ch, 1
      16:      7512             jne   L2
      18:      55               push  ebp
      19:      89E5             move  ebp, esp
      21:      3D570D7121       cmp   eax, 21710D57    ; :NEMERLE
      26:      750D             jne   L3
      28:      B803400020       move  eax, 20004003    ; T
L1:   33:      FD               std   
      34:      C9               leave 
      35:      C3               ret   
L2:   36:      E86F1375FE       call  2011A282         ; #<Function RUNTIME:BAD-ARGS-OR-STACK 2011A282>
L3:   41:      B856000000       move  eax, 56
      46:      EBF1             jmp   L1
      48:      90               nop   
      49:      90               nop   
NIL

D>Так как раз eval и может работать через компилятор (см. пост про REPL). В чем проблема?

(defun my-eval (code)
               (eval code)
)

Z>;(defun my-eval (code)
Z>;               (eval code)
Z>;)
Z>;

* (defun my-eval (code) (eval code))

MY-EVAL
* (disassemble 'my-eval)

; disassembly for MY-EVAL
; 245C8C0F:       850500000021     TEST EAX, [#x21000000]     ; no-arg-parsing entry point
;       15:       8BD6             MOV EDX, ESI
;       17:       8B05D88B5C24     MOV EAX, [#x245C8BD8]      ; #<FDEFINITION object for EVAL>
;       1D:       B904000000       MOV ECX, 4
;       22:       FF7504           PUSH DWORD PTR [EBP+4]
;       25:       FF6005           JMP DWORD PTR [EAX+5]
;       28:       CC0A             BREAK 10                   ; error trap
;       2A:       02               BYTE #X02
;       2B:       18               BYTE #X18                  ; INVALID-ARG-COUNT-ERROR
;       2C:       4F               BYTE #X4F                  ; ECX
NIL

(defun my-eval-2 (code)
  (funcall 
    (compile nil `(lambda () ,code))))

* (my-eval-2 '(+ 1 2 3))

6

The DECLARE form in INT-ADD1 includes two kinds of advice. (FIXNUM N) declares that the function parameter N is a small integer. (The range depends upon your Lisp implementation, but you'll typically get 29-bit fixnums on a 32-bit processor; the remaining three bits are often used by the Lisp system to encode type information.)

Contrary to other implementations of the S language, R has adopted an evaluation model in which nested function definitions are lexically scoped. This is analogous to the evaluation model in Scheme.

the underlying implementation and semantics are derived from Scheme. See What are the differences between R and S?, for further details.

Contrary to other implementations of the S language, R has adopted an evaluation model in which nested function definitions are lexically scoped. This is analogous to the evaluation model in Scheme.