cl

cl

[Katsumi Yamaoka]

[-- Attachment #1: Type: text/plain, Size: 205 bytes --]

This is the patch for eliminating run-time cl.

1999-12-14 04:14:44  Katsumi Yamaoka  <yamaoka@jpl.org>

	* dgnushack.el (last, mapcon, member-if, union): New compiler
	macros for emulating cl functions.


[-- Attachment #2: dgnushack.el.diff --]
[-- Type: application/x-patch, Size: 2015 bytes --]

[-- Attachment #3: Type: text/plain, Size: 365 bytes --]

It has been done in T-gnus 6.14.  If you are interested in it, please
checkout it from the chamonix open CVS server as follows.

% cvs -d :pserver:anonymous@chamonix.jaist.ac.jp:/hare/cvs/root login
CVS password: [CR] # NULL string

% cvs -d :pserver:anonymous@chamonix.jaist.ac.jp:/hare/cvs/root checkout\
-r t-gnus-6_14 gnus
-- 
Katsumi Yamaoka <yamaoka@jpl.org>

Re: cl

[Hrvoje Niksic]

Katsumi Yamaoka <yamaoka@jpl.org> writes:

> This is the patch for eliminating run-time cl.
> 
> 1999-12-14 04:14:44  Katsumi Yamaoka  <yamaoka@jpl.org>
> 
> 	* dgnushack.el (last, mapcon, member-if, union): New compiler
> 	macros for emulating cl functions.

Could you please code these things so that they are not called under
XEmacs, where cl.el is "official"?

Re: cl

[Katsumi Yamaoka]

[-- Attachment #1: Type: text/plain, Size: 499 bytes --]

>>>>> In <9t9ln6xrdkt.fsf@mraz.iskon.hr> 
>>>>>	Hrvoje Niksic <hniksic@iskon.hr> wrote:

>> This is the patch for eliminating run-time cl.
>> 
>> 1999-12-14 04:14:44  Katsumi Yamaoka  <yamaoka@jpl.org>
>> 
>> 	* dgnushack.el (last, mapcon, member-if, union): New compiler
>> 	macros for emulating cl functions.

Hrvoje> Could you please code these things so that they are not called
Hrvoje> under XEmacs, where cl.el is "official"?

Okay, thanks for the advice.  This is a replacement of the patch.

[-- Attachment #2: dgnushack.el.diff --]
[-- Type: application/x-patch, Size: 2143 bytes --]

Re: cl

[Jan Vroonhof]

Katsumi Yamaoka <yamaoka@jpl.org> writes:

> Hrvoje> Could you please code these things so that they are not called
> Hrvoje> under XEmacs, where cl.el is "official"?
> 
> Okay, thanks for the advice.  This is a replacement of the patch.
> 
> 
> --- dgnushack.el~	Mon Dec  6 02:04:45 1999
> +++ dgnushack.el	Tue Dec 14 11:34:55 1999
> @@ -30,6 +30,83 @@
>  
>  (require 'cl)
>  
> +(define-compiler-macro last (&whole form x &optional n)
> +  (if (or (featurep 'xemacs)
> +	  (and (fboundp 'last)

Why not wrap the whole thing an
(unless (featurep 'xemacs) ..)

Jan

Re: cl

[Katsumi Yamaoka]

[-- Attachment #1: Type: text/plain, Size: 329 bytes --]

>>>>> In <byd7s9uzc0.fsf@bolzano.math.ethz.ch> 
>>>>>	Jan Vroonhof <vroonhof@math.ethz.ch> wrote:

>> +(define-compiler-macro last (&whole form x &optional n)
>> +  (if (or (featurep 'xemacs)
>> +	  (and (fboundp 'last)

Jan> Why not wrap the whole thing an
Jan> (unless (featurep 'xemacs) ..)

This is the third time lucky. :-)

[-- Attachment #2: dgnushack.el.diff --]
[-- Type: application/x-patch, Size: 2060 bytes --]

cl

[Lars Magne Ingebrigtsen]

Gowrsh.  I really should re-read CLtL2.

(let ((vars (some-ting))
      var)
  (while (setq var (pop vars))
    ...))

which I usually write, vs

(dolist (var (some-thing))
   ...)

And the latter is even a macro in cl-macs that expands to nice
Emacs-lisp bytecodes.

(And the first form even assumes that there are no non-nil elements in
the list.  Which is true everywhere I use it, but that makes the
dolist thing even more elegant.)

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@ifi.uio.no * Lars Magne Ingebrigtsen

Re: cl

[Jan Vroonhof]

Lars Magne Ingebrigtsen <larsi@ifi.uio.no> writes:

> 
> (dolist (var (some-thing))
>    ...)

Wouldn't it have been much nicer if the map syntax had been

(mapc list func) so you could just write the nice and pure

(mapc some-thing
   (lambda (var)
   ....))

Jan

Re: cl

[Justin Sheehy]

Jan Vroonhof <vroonhof@math.ethz.ch> writes:

> Wouldn't it have been much nicer if the map syntax had been
> 
> (mapc list func) so you could just write the nice and pure
> 
> (mapc some-thing
>    (lambda (var)
>    ....))

What's wrong with the current mapc syntax?  Is there something that
you would like other than to have the parameter order reversed?

-Justin, confused
 

Re: cl

[Lars Magne Ingebrigtsen]

Justin Sheehy <justin@linus.mitre.org> writes:

> What's wrong with the current mapc syntax?  Is there something that
> you would like other than to have the parameter order reversed?

No, but I think that's a good reason.  Virtually all the Lisp "control 
structures" look like

(special-form variables-or-list-or-something
  (do-all-the-work-down-here))

Now, `mapcar' isn't a special form, but even so, I always forget whether 
the predicate is the first or second parameter, due in part to `sort'
working the opposite way.

(mapcar FUNCTION LIST)

vs.

(sort LIST PREDICATE)

(Heh.  I actually wrote the `sort' form with the parameters in the
wrong order, but then I tapped `C-h f', as I always do when I write
`sort' or `mapcar', and it told me I was wrong, as usual.  :-)

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: cl

[Hrvoje Niksic]

Lars Magne Ingebrigtsen <larsi@gnus.org> writes:

> Now, `mapcar' isn't a special form, but even so, I always forget
> whether the predicate is the first or second parameter, due in part
> to `sort' working the opposite way.
> 
> (mapcar FUNCTION LIST)

In Common Lisp it's:

    (mapcar FUNCTION LIST &rest OTHER-LISTS)

If you specify more than one list, FUNCTION will be called with that
many arguments.  In Emacs, the CL version of `mapcar' is called
`mapcar*'.

BW, this fact is a good way of remembering that FUNCTION comes first,
and why.