gcl-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Gcl-devel] Re: [Maxima] Requesting input on some possible low-level


From: Camm Maguire
Subject: Re: [Gcl-devel] Re: [Maxima] Requesting input on some possible low-level changes
Date: 09 Sep 2003 15:08:50 -0400
User-agent: Gnus/5.09 (Gnus v5.9.0) Emacs/21.2

Greetings, and thanks for this!  Problem is, that for the purposes of
the optimization in question, something can still be usefully done in
the (apply #'aref ...) et.al. cases (eg. add implicit (the fixnum ..)
to indices).  I guess this cannot be completely implemented via
compiler-macros -- I'm wondering if there is a better way?

Take care,

Steve Haflich <address@hidden> writes:

>    From: Camm Maguire <address@hidden>
>    
>    In any case, our define-compiler-macro appears to be lacking in at
>    least one respect, namely it doesn't skip over 'funcall as required by
>    the spec.  I.e. (funcall #'foo ...) is not expanded.  I have a fix to
>    this which seems to be working.  What I'd like to know, however, is if
>    'apply, 'map, and any other functions taking a function descriptor as
>    an argument is also supposed to be expanded.
> 
> No, the ANS only specifies that funcall be handled.  Compiler macros
> should not be invoked for apply or mapping forms.
> 
> A plausible reason for this is that for a compiler macro to be useful,
> in many situations it needs to know the number of arguments the
> operator will receive at run time, and for keyword args, it often
> wants to see the actual keyword arguments at compile time.  (Compiler
> macros typically will punt if this information is not available to
> them.)  There is no reasonable way to destructure an apply or map form
> against the `macro' lambda list of a compiler macro.  For funcall it
> is quite practical.
> 
> Note that the compiler macro receives the funcall form as its &whole
> (so that it can return it if it decides to decline to expand) but that
> the compiler is responsible for unravelling the funcall form so the
> compiler-macro's macro argument list is properly bound.
> 
> 
> _______________________________________________
> Gcl-devel mailing list
> address@hidden
> http://mail.gnu.org/mailman/listinfo/gcl-devel
> 
> 
> 

-- 
Camm Maguire                                            address@hidden
==========================================================================
"The earth is but one country, and mankind its citizens."  --  Baha'u'llah




reply via email to

[Prev in Thread] Current Thread [Next in Thread]