[PATCH] Support for multi-arg arithmetic comparisons
25 years, 12 months
Hrvoje Niksic
Hrvoje Niksic
This patch completes the work started by Martin some time ago.
Basically, it allows functions `<', `>', `<=', `>=', `=' and `/=' to
accept any number of arguments, while losing no speed for the
two-argument versions, and preserving full bytecode compatibility.
(< 1 2 3)
=> t
(< 1 2 2)
=> nil
(<= 1 2 2)
=> t
The compiler handles the functions specially by generating a constant
(t) for one-argument, the appropriate bytecode for two-argument
version, and a normal funcall for >2 arguments. Here are several
examples:
(disassemble (lambda () (< a b c)))
0 constant <
1 varref a
2 varref b
3 varref c
4 call 3
5 return
However:
(disassemble (lambda () (< a b)))
0 varref a
1 varref b
2 lss
3 return
Also:
(disassemble (lambda () (< a)))
0 constant t
1 return
I've taken care to preserve the special optimization for two-argument
version of /=, where (/= A B) is optimized as (not (= A B)).
(disassemble (lambda () (/= a b)))
0 varref a
1 varref b
2 eqlsign
3 not
4 return
This optimization does not hold with more than two arguments:
(disassemble (lambda () (/= a b c)))
0 constant /=
1 varref a
2 varref b
3 varref c
4 call 3
5 return
With this patch, the comparison functions are consistent with
arithmetic functions, such as +, -, * and others. I have given this
some test, and it runs fine. I think this will make a nice feature
for XEmacs 21.
C ChangeLog:
1998-05-01 Hrvoje Niksic <hniksic(a)srce.hr>
* fileio.c (Fcar_less_than_car): Fix Flss caller.
(Fcdr_less_than_cdr): Ditto.
* lisp.h: Fix declarations.
* data.c: Enable many arguments versions of <, >, <=, >= and /=.
* bytecode.c (Fbyte_code): Use arithcompare.
* data.c (arithcompare): Make non-static.
Lisp ChangeLog:
1998-05-01 Hrvoje Niksic <hniksic(a)srce.hr>
* byte-optimize.el: Don't attempt to optimize /=.
* bytecomp.el (byte-compile-one-ore-more-args): New function.
(byte-compile-/=): Ditto.
--- etc/NEWS.orig Fri May 1 06:44:50 1998
+++ etc/NEWS Fri May 1 06:54:21 1998
＠＠ -201,6 +201,14 ＠＠
- LDAP SDK 1.0 from Netscape Corp.
(http://developer.netscape.com)
+** The arithmetic comparison functions <, >, =, /= now accept 1 or
+more arguments.
+
+This means that if you want to test whether A < B < C, you can write
+it as (< A B C) instead of (and (< A B) (< B C)). Likewise,
+(apply #'> LIST) now tests if LIST is monotonously increasing -- and
+so on.
+
** The XEmacs hashtables now have a consistent read/print syntax.
This means that a hashtable will be readably printed in a
structure-like form:
--- src/fileio.c.orig Fri May 1 06:25:05 1998
+++ src/fileio.c Fri May 1 06:25:46 1998
＠＠ -3401,7 +3401,7 ＠＠
*/
(a, b))
{
- return Flss (Fcar (a), Fcar (b));
+ return arithcompare (Fcar (a), Fcar (b), arith_less);
}
/* Heh heh heh, let's define this too, just to aggravate the person who
＠＠ -3411,7 +3411,7 ＠＠
*/
(a, b))
{
- return Flss (Fcdr (a), Fcdr (b));
+ return arithcompare (Fcdr (a), Fcdr (b), arith_less);
}
/* Build the complete list of annotations appropriate for writing out
--- src/lisp.h.orig Fri May 1 05:50:40 1998
+++ src/lisp.h Fri May 1 06:19:02 1998
＠＠ -2004,6 +2004,16 ＠＠
Lisp_Object wrong_type_argument (Lisp_Object, Lisp_Object);
DECLARE_DOESNT_RETURN (dead_wrong_type_argument (Lisp_Object, Lisp_Object));
void check_int_range (int, int, int);
+
+enum arith_comparison {
+ arith_equal,
+ arith_notequal,
+ arith_less,
+ arith_grtr,
+ arith_less_or_equal,
+ arith_grtr_or_equal };
+Lisp_Object arithcompare (Lisp_Object, Lisp_Object, enum arith_comparison);
+
Lisp_Object word_to_lisp (unsigned int);
unsigned int lisp_to_word (Lisp_Object);
＠＠ -2557,7 +2567,7 ＠＠
EXFUN (Fforward_line, 2);
EXFUN (Ffset, 2);
EXFUN (Ffuncall, MANY);
-EXFUN (Fgeq, 2);
+EXFUN (Fgeq, MANY);
EXFUN (Fget, 3);
EXFUN (Fget_buffer_process, 1);
EXFUN (Fget_coding_system, 1);
＠＠ -2566,7 +2576,7 ＠＠
EXFUN (Fgethash, 3);
EXFUN (Fgettext, 1);
EXFUN (Fgoto_char, 2);
-EXFUN (Fgtr, 2);
+EXFUN (Fgtr, MANY);
EXFUN (Fhashtablep, 1);
EXFUN (Findent_to, 3);
EXFUN (Findirect_function, 1);
＠＠ -2583,10 +2593,10 ＠＠
EXFUN (Flax_plist_get, 3);
EXFUN (Flax_plist_remprop, 2);
EXFUN (Flength, 1);
-EXFUN (Fleq, 2);
+EXFUN (Fleq, MANY);
EXFUN (Flist, MANY);
EXFUN (Flistp, 1);
-EXFUN (Flss, 2);
+EXFUN (Flss, MANY);
EXFUN (Fmake_byte_code, MANY);
EXFUN (Fmake_coding_system, 4);
EXFUN (Fmake_glyph_internal, 1);
--- src/data.c.orig Fri May 1 05:45:25 1998
+++ src/data.c Fri May 1 06:18:07 1998
＠＠ -974,10 +974,9 ＠＠
/* Arithmetic functions */
/**********************************************************************/
-enum comparison { equal, notequal, less, grtr, less_or_equal, grtr_or_equal };
-
-static Lisp_Object
-arithcompare (Lisp_Object num1, Lisp_Object num2, enum comparison comparison)
+Lisp_Object
+arithcompare (Lisp_Object num1, Lisp_Object num2,
+ enum arith_comparison comparison)
{
CHECK_INT_OR_FLOAT_COERCE_CHAR_OR_MARKER (num1);
CHECK_INT_OR_FLOAT_COERCE_CHAR_OR_MARKER (num2);
＠＠ -990,152 +989,95 ＠＠
switch (comparison)
{
- case equal: return f1 == f2 ? Qt : Qnil;
- case notequal: return f1 != f2 ? Qt : Qnil;
- case less: return f1 < f2 ? Qt : Qnil;
- case less_or_equal: return f1 <= f2 ? Qt : Qnil;
- case grtr: return f1 > f2 ? Qt : Qnil;
- case grtr_or_equal: return f1 >= f2 ? Qt : Qnil;
+ case arith_equal: return f1 == f2 ? Qt : Qnil;
+ case arith_notequal: return f1 != f2 ? Qt : Qnil;
+ case arith_less: return f1 < f2 ? Qt : Qnil;
+ case arith_less_or_equal: return f1 <= f2 ? Qt : Qnil;
+ case arith_grtr: return f1 > f2 ? Qt : Qnil;
+ case arith_grtr_or_equal: return f1 >= f2 ? Qt : Qnil;
}
}
#endif /* LISP_FLOAT_TYPE */
switch (comparison)
{
- case equal: return XINT (num1) == XINT (num2) ? Qt : Qnil;
- case notequal: return XINT (num1) != XINT (num2) ? Qt : Qnil;
- case less: return XINT (num1) < XINT (num2) ? Qt : Qnil;
- case less_or_equal: return XINT (num1) <= XINT (num2) ? Qt : Qnil;
- case grtr: return XINT (num1) > XINT (num2) ? Qt : Qnil;
- case grtr_or_equal: return XINT (num1) >= XINT (num2) ? Qt : Qnil;
+ case arith_equal: return XINT (num1) == XINT (num2) ? Qt : Qnil;
+ case arith_notequal: return XINT (num1) != XINT (num2) ? Qt : Qnil;
+ case arith_less: return XINT (num1) < XINT (num2) ? Qt : Qnil;
+ case arith_less_or_equal: return XINT (num1) <= XINT (num2) ? Qt : Qnil;
+ case arith_grtr: return XINT (num1) > XINT (num2) ? Qt : Qnil;
+ case arith_grtr_or_equal: return XINT (num1) >= XINT (num2) ? Qt : Qnil;
}
abort ();
return Qnil; /* suppress compiler warning */
}
-DEFUN ("=", Feqlsign, 2, 2, 0, /*
-Return t if two args, both numbers, characters or markers, are equal.
-*/
- (num1, num2))
-{
- return arithcompare (num1, num2, equal);
-}
-
-DEFUN ("<", Flss, 2, 2, 0, /*
-Return t if first arg is less than second arg.
-Both must be numbers, characters or markers.
-*/
- (num1, num2))
-{
- return arithcompare (num1, num2, less);
-}
-
-DEFUN (">", Fgtr, 2, 2, 0, /*
-Return t if first arg is greater than second arg.
-Both must be numbers, characters or markers.
-*/
- (num1, num2))
-{
- return arithcompare (num1, num2, grtr);
-}
-
-DEFUN ("<=", Fleq, 2, 2, 0, /*
-Return t if first arg is less than or equal to second arg.
-Both must be numbers, characters or markers.
-*/
- (num1, num2))
-{
- return arithcompare (num1, num2, less_or_equal);
-}
-
-DEFUN (">=", Fgeq, 2, 2, 0, /*
-Return t if first arg is greater than or equal to second arg.
-Both must be numbers, characters or markers.
-*/
- (num1, num2))
-{
- return arithcompare (num1, num2, grtr_or_equal);
-}
-
-DEFUN ("/=", Fneq, 2, 2, 0, /*
-Return t if first arg is not equal to second arg.
-Both must be numbers, characters or markers.
-*/
- (num1, num2))
-{
- return arithcompare (num1, num2, notequal);
-}
-
-#if 0
-/* I tried implementing Common Lisp multi-arg comparison functions,
- but failed because the byte-compiler needs to be hacked as well. */
-
static Lisp_Object
-arithcompare_many (enum comparison comparison, int nargs, Lisp_Object *args)
+arithcompare_many (enum arith_comparison comparison,
+ int nargs, Lisp_Object *args)
{
REGISTER int argnum;
for (argnum = 1; argnum < nargs; argnum++)
- if (EQ (arithcompare ( args[argnum-1], args[argnum], comparison), Qnil))
+ if (EQ (arithcompare (args[argnum-1], args[argnum], comparison), Qnil))
return Qnil;
return Qt;
}
-xxxDEFUN ("=", Feqlsign, 1, MANY, 0, /*
+DEFUN ("=", Feqlsign, 1, MANY, 0, /*
Return t if all the arguments are equal.
The arguments may be numbers, characters or markers.
*/
(int nargs, Lisp_Object *args))
{
- return arithcompare (equal, nargs, args);
+ return arithcompare_many (arith_equal, nargs, args);
}
-xxxDEFUN ("<", Flss, 1, MANY, 0, /*
+DEFUN ("<", Flss, 1, MANY, 0, /*
Return t if the sequence of arguments is monotonically increasing.
The arguments may be numbers, characters or markers.
*/
(int nargs, Lisp_Object *args))
{
- return arithcompare (less, nargs, args);
+ return arithcompare_many (arith_less, nargs, args);
}
-xxxDEFUN (">", Fgtr, 1, MANY, 0, /*
+DEFUN (">", Fgtr, 1, MANY, 0, /*
Return t if the sequence of arguments is monotonically decreasing.
The arguments may be numbers, characters or markers.
*/
(int nargs, Lisp_Object *args))
{
- return arithcompare (grtr, nargs, args);
+ return arithcompare_many (arith_grtr, nargs, args);
}
-xxxDEFUN ("<=", Fleq, 1, MANY, 0, /*
+DEFUN ("<=", Fleq, 1, MANY, 0, /*
Return t if the sequence of arguments is monotonically nondecreasing.
The arguments may be numbers, characters or markers.
*/
(int nargs, Lisp_Object *args))
{
- return arithcompare (less_or_equal, nargs, args);
+ return arithcompare_many (arith_less_or_equal, nargs, args);
}
-xxxDEFUN (">=", Fgeq, 1, MANY, 0, /*
+DEFUN (">=", Fgeq, 1, MANY, 0, /*
Return t if the sequence of arguments is monotonically nonincreasing.
The arguments may be numbers, characters or markers.
*/
(int nargs, Lisp_Object *args))
{
- return arithcompare_many (grtr_or_equal, nargs, args);
+ return arithcompare_many (arith_grtr_or_equal, nargs, args);
}
-xxxDEFUN ("/=", Fneq, 1, MANY, 0, /*
+DEFUN ("/=", Fneq, 1, MANY, 0, /*
Return t if the sequence of arguments is monotonically increasing.
The arguments may be numbers, characters or markers.
*/
(int nargs, Lisp_Object *args))
{
- return arithcompare_many (notequal, nargs, args);
+ return arithcompare_many (arith_notequal, nargs, args);
}
-#endif /* 0 - disabled for now */
DEFUN ("zerop", Fzerop, 1, 1, 0, /*
Return t if NUMBER is zero.
--- src/bytecode.c.orig Fri May 1 05:53:17 1998
+++ src/bytecode.c Fri May 1 05:56:06 1998
＠＠ -811,22 +811,22 ＠＠
case Bgtr:
v1 = POP;
- TOP = Fgtr (TOP, v1);
+ TOP = arithcompare (TOP, v1, arith_grtr);
break;
case Blss:
v1 = POP;
- TOP = Flss (TOP, v1);
+ TOP = arithcompare (TOP, v1, arith_less);
break;
case Bleq:
v1 = POP;
- TOP = Fleq (TOP, v1);
+ TOP = arithcompare (TOP, v1, arith_less_or_equal);
break;
case Bgeq:
v1 = POP;
- TOP = Fgeq (TOP, v1);
+ TOP = arithcompare (TOP, v1, arith_grtr_or_equal);
break;
case Bdiff:
--- lisp/bytecomp.el.orig Fri May 1 05:25:10 1998
+++ lisp/bytecomp.el Fri May 1 06:11:10 1998
＠＠ -2889,11 +2889,12 ＠＠
(byte-defop-compiler20 old-memq 2)
(byte-defop-compiler cons 2)
(byte-defop-compiler aref 2)
-(byte-defop-compiler (= byte-eqlsign) 2)
-(byte-defop-compiler (< byte-lss) 2)
-(byte-defop-compiler (> byte-gtr) 2)
-(byte-defop-compiler (<= byte-leq) 2)
-(byte-defop-compiler (>= byte-geq) 2)
+(byte-defop-compiler (= byte-eqlsign) byte-compile-one-or-more-args)
+(byte-defop-compiler (< byte-lss) byte-compile-one-or-more-args)
+(byte-defop-compiler (> byte-gtr) byte-compile-one-or-more-args)
+(byte-defop-compiler (<= byte-leq) byte-compile-one-or-more-args)
+(byte-defop-compiler (>= byte-geq) byte-compile-one-or-more-args)
+(byte-defop-compiler /= byte-compile-/=)
(byte-defop-compiler get 2+1)
(byte-defop-compiler nth 2)
(byte-defop-compiler substring 2-3)
＠＠ -3103,6 +3104,21 ＠＠
(byte-defop-compiler nconc)
(byte-defop-compiler-1 beginning-of-line)
+(defun byte-compile-one-or-more-args (form)
+ (let ((len (length form)))
+ (cond ((= len 1) (byte-compile-subr-wrong-args form "1 or more"))
+ ((= len 2) (byte-compile-constant t))
+ ((= len 3) (byte-compile-two-args form))
+ (t (byte-compile-normal-call form)))))
+
+(defun byte-compile-/= (form)
+ (let ((len (length form)))
+ (cond ((= len 1) (byte-compile-subr-wrong-args form "1 or more"))
+ ((= len 2) (byte-compile-constant t))
+ ;; optimize (/= X Y) to (not (= X Y))
+ ((= len 3) (byte-compile-form-do-effect `(not (= ,＠(cdr form)))))
+ (t (byte-compile-normal-call form)))))
+
(defun byte-compile-buffer-substring (form)
(let ((len (length form)))
;; buffer-substring used to take exactly two args, but now takes 0-3.
＠＠ -3539,18 +3555,18 ＠＠
(byte-compile-out 'byte-unbind (length (car (cdr form))))))
-(byte-defop-compiler-1 /= byte-compile-negated)
+;;(byte-defop-compiler-1 /= byte-compile-negated)
(byte-defop-compiler-1 atom byte-compile-negated)
(byte-defop-compiler-1 nlistp byte-compile-negated)
-(put '/= 'byte-compile-negated-op '=)
+;;(put '/= 'byte-compile-negated-op '=)
(put 'atom 'byte-compile-negated-op 'consp)
(put 'nlistp 'byte-compile-negated-op 'listp)
(defun byte-compile-negated (form)
(byte-compile-form-do-effect (byte-compile-negation-optimizer form)))
-;; Even when optimization is off, /= is optimized to (not (= ...)).
+;; Even when optimization is off, atom is optimized to (not (consp ...)).
(defun byte-compile-negation-optimizer (form)
;; an optimizer for forms where <form1> is less efficient than (not <form2>)
(list 'not
--- lisp/byte-optimize.el.orig Fri May 1 06:08:11 1998
+++ lisp/byte-optimize.el Fri May 1 06:08:20 1998
＠＠ -1029,7 +1029,7 ＠＠
(put 'while 'byte-optimizer 'byte-optimize-while)
;; byte-compile-negation-optimizer lives in bytecomp.el
-(put '/= 'byte-optimizer 'byte-compile-negation-optimizer)
+;(put '/= 'byte-optimizer 'byte-compile-negation-optimizer)
(put 'atom 'byte-optimizer 'byte-compile-negation-optimizer)
(put 'nlistp 'byte-optimizer 'byte-compile-negation-optimizer)
--
Hrvoje Niksic <hniksic(a)srce.hr> | Student at FER Zagreb, Croatia
--------------------------------+--------------------------------
"Memory is like an orgasm. It's a lot better if you don't have to
fake it." -- Seymour Cray, on virtual memory
- 1 participants
- 0 comments
RE: [PATCH] Fix for possible `make install' bug in beta38-pre2
25 years, 12 months
Damon Lipparelli
Damon Lipparelli
Actually, what I found was that I had to run "autoconf" to re-generate
"configure".
-lipp
>>>> -----Original Message-----
>>>> From: SL Baur [mailto:steve＠xemacs.org]
>>>> Sent: Wednesday, April 29, 1998 9:05 PM
>>>> To: xemacs-beta(a)xemacs.org
>>>> Subject: Re: [PATCH] Fix for possible `make install' bug
>>>> in beta38-pre2
>>>>
>>>>
>>>> Tetsuya HOYANO <hoyano(a)ari.bekkoame.ne.jp> writes:
>>>>
>>>> > ----Next_Part(Thu_Apr_30_13:00:42_1998_349)--
>>>> > Content-Type: Text/Plain; charset=us-ascii
>>>> > Content-Transfer-Encoding: 7bit
>>>>
>>>> >>>>>> "sb" == SL Baur <steve(a)xemacs.org> writes:
>>>>
>>>> sb> There's confusion between configure's `progname' and
>>>> the `progname'
>>>> sb> used by the Makefiles. This patch makes everything
>>>> consistent. After
>>>> sb> applying this patch, the generated src/paths.h looks
>>>> cleaner anyway.
>>>>
>>>> > beta38-pre2 built fine, then `./config.status --recheck',
>>>>
>>>> I'm sorry. You must rerun configure from scratch.
>>>>
>>>> > `make beta' and `make install' install them into
>>>> > /usr/local/lib/-21.0-b38/ not into
>>>> /usr/local/lib/xemacs-21.0-b38/
>>>> > after applied this patch.
>>>>
>>>> > I am confusing...
>>>>
>>>> That sounds about right, unfortunately. Rerun configure
>>>> from scratch
>>>> and things should be O.K.
>>>>
- 2 participants
- 1 comments
[PATCH] doc patchlet
25 years, 12 months
Hrvoje Niksic
Hrvoje Niksic
--- lisp/frame.el.orig Fri May 1 03:17:34 1998
+++ lisp/frame.el Fri May 1 03:22:06 1998
＠＠ -516,14 +516,15 ＠＠
(defun other-frame (arg)
"Select the ARG'th different visible frame, and raise it.
All frames are arranged in a cyclic order.
-This command selects the frame ARG steps away in that order,
-regardless of the value of `focus-follows-mouse'.
-A negative ARG moves in the opposite order."
+This command selects the frame ARG steps away in that order.
+A negative ARG moves in the opposite order.
+
+This command ignores the value of `focus-follows-mouse'."
(interactive "p")
(let ((frame (selected-frame))
+ (old-focus-follows-mouse focus-follows-mouse)
;; Allow selecting another frame even when
;; focus-follows-mouse is true.
- (focus-policy focus-follows-mouse)
(focus-follows-mouse nil))
(while (> arg 0)
(setq frame (next-frame frame 'visible-nomini))
＠＠ -533,8 +534,10 ＠＠
(setq arg (1+ arg)))
(raise-frame frame)
(select-frame frame)
- (if focus-policy ;; this hack allows the focus change to be
- (sit-for 0)) ;; processed while focus-follows-mouse is nil.
+ ;; Allow the focus change to be processed while
+ ;; focus-follows-mouse is nil.
+ (and old-focus-follows-mouse
+ (sit-for 0))
;this is a bad idea; you should in general never warp the
;pointer unless the user asks for this. Furthermore,
;our version of `set-mouse-position' takes a window,
--- etc/NEWS.orig Fri May 1 03:21:04 1998
+++ etc/NEWS Fri May 1 03:21:04 1998
＠＠ -33,17 +33,6 ＠＠
* Changes in XEmacs 21.0
========================
-** When the variable focus-follows-mouse is non-nil, it is no longer
-possible for select-frame to permanently select a different frame.
-The frame selection is temporary and is reverted when the current
-command terminates, much like the buffer selected by `set-buffer'.
-In order to effect a permanent focus change in this case, bind
-focus-follows-mouse to nil, select the frame you want, and do a
-(sit-for 0) within the scope of the binding. The function
-other-frame (normally bound to C-x 5 o) is the only command provided
-which does not respect the setting of focus-follows-mouse, and *will*
-allow one to select a different frame even when set non-nil.
-
** XEmacs has been unbundled into constituent installable packages.
See the file `etc/PACKAGES' in the distribution for a full
description.
＠＠ -196,6 +185,13 ＠＠
machines.) Configure with --use-minimal-tagbits to test. With this
change, the maximum buffer size on 32-bit machines is bumped from 128M
to 1G. This setting will be the default in a future XEmacs version.
+
+** When the variable focus-follows-mouse is non-nil, `select-frame' no
+longer permanently selects a different frame. The frame selection is
+temporary and is reverted when the current command terminates, much
+like the buffer selected by `set-buffer'. The `other-frame' command
+(`C-x 5 o') is unaffected by `focus-follows-mouse', and its behaviour
+is unchanged.
** It is now possible to build XEmacs with LDAP support
You need to install a LDAP library first. The following have been
--
Hrvoje Niksic <hniksic(a)srce.hr> | Student at FER Zagreb, Croatia
--------------------------------+--------------------------------
What is mind? No matter.
What is matter? Never mind.
- 1 participants
- 0 comments
Re: giflib, what version and WHERE?
25 years, 12 months
P. E. Jareth Hein
P. E. Jareth Hein
Jens Lautenbacher <jens(a)tellux.de> writes:
> OK, it seems XEmacs wants version 3.1, but I've been unable to find
> this somewhere... all I see are rpms for 3.04 or such...
>
> any pointers would be welcome..
giflib 3.1 is mine, but I'm getting cold feet... Unisys is apparently
getting frisky (again) about enforcing patents, and my publicly
releasing giflib sounds like I could be opening myself up for a world
of hurt. For now, you can find the latest version on
http://www.camelot-soft.com/~jareth but I'm not sure how long it'll be
there. Anybody have any experience with patent law and how that
applies to public, freely available software?
--
Jareth Hein | jareth(a)camelot-soft.com | ハイン ジェラス
Toolsmith & Program lead | http://www.camelot-soft.com | 開発部チーフ
Camelot Software, Ltd. | |（株）キャメロット
"Contrary to popular belief, penguins are not the salvation of modern
technology. Neither do they throw parties for the urban proletariat."
- 3 participants
- 2 comments
Re: [Success] XEmacs 21.0-b38 "Repartida-pre2" on i686-linux
25 years, 12 months
SL Baur
SL Baur
toge <toge(a)accad1.kek.jp> writes:
> BTW, gnus in the X-window mode barks at me unless I also install w3.
> The reason is that gnus wants images.el which is part of w3.
> I guess there are cases where the user wants to use gnus but not w3.
> Can images.el be part of xemacs-base or something, or am I too picky?
Could you please get and post a lisp stack backtrace? Thanks.
(setq debug-on-error t) and recreate the problem.
- 2 participants
- 1 comments
Re: select-window focus problem [was: Re: minor compile.el bug]
25 years, 12 months
Hrvoje Niksic
Hrvoje Niksic
Martin Buchholz <martin(a)xemacs.org> writes:
> >>>>> "Hrvoje" == Hrvoje Niksic <hniksic(a)srce.hr> writes:
>
> Hrvoje> BTW, are you aware that FSF 20.3 sets focus-follows-mouse to t by
> Hrvoje> default? Maybe we should copy what they do?
>
> I've had the vague feeling that our focus-handling has been bogus for
> a long time. Apparently it's something that is hard to get right, and
> some past patches have been switching between different behaviors,
> none of which make everyone happy.
This is true, but, as I said before, I am very satisfied with Greg's
latest patch, except for the `C-x 5 o' part. Note, though, that I
have *not* read the relevant X docs, so I cannot guarantee that this
behaviour is X-ly correct.
Maybe someone who knows X should look at it and comment? Glynn,
maybe?
--
Hrvoje Niksic <hniksic(a)srce.hr> | Student at FER Zagreb, Croatia
--------------------------------+--------------------------------
Jone's law: The man who smiles when things go wrong has thought of
someone to blame it on.
- 3 participants
- 2 comments
Re: Lstreams, coding and md5 (was: Re: [PATCH] menubar selection in mswindows)
25 years, 12 months
Hrvoje Niksic
Hrvoje Niksic
Pete Ware <ware(a)pc-786a.cis.ohio-state.edu> writes:
> > Before coding systems came into the picture, Fmd5 was an exemplary
> > of Lisp elegance:
> >
> > `md5' is a built-in function
> > (md5 OBJECT &optional START END)
> >
> > Documentation:
> > Return the MD5 message digest of OBJECT, a buffer or string.
> >
> > Optional arguments START and END denote positions for computing the
> > digest of a portion of OBJECT.
>
>
> I'm a MOLE (MOnoLingual Emacs) so take what I have with a shalt shaker
> but shouldn't OBJECT be in the already appropriate coding system for
> whatever the message digest is being computed?
Nope. OBJECT is a buffer or string in internal XEmacs format, and we
don't want to calculate a digest of the internal representation
(that's what the old non-Mule-ized code did), but of the external
representation, as saved to disk etc.
> To put it another way, shouldn't there be a set of functions that
> given an OBJECT in one coding system return another OBJECT in a
> different coding system?
It would be very wasteful, I think.
--
Hrvoje Niksic <hniksic(a)srce.hr> | Student at FER Zagreb, Croatia
--------------------------------+--------------------------------
- Now what did we learn from this?
- I learned what my liver looks like!
- 2 participants
- 1 comments
Re: Lstreams, coding and md5 (was: Re: [PATCH] menubar selection in mswindows)
25 years, 12 months
Hrvoje Niksic
Hrvoje Niksic
"Stephen J. Turnbull" <turnbull(a)sk.tsukuba.ac.jp> writes:
> Hrvoje> No, the idea is fine. But is it the best thing to infest
> Hrvoje> random primitives with CODING-SYSTEM arguments? Isn't
> Hrvoje> there a more general variable that can be set? Please
> Hrvoje> bear with me -- I know very little about coding systems
> Hrvoje> and stuff, so maybe I'm just raving.
>
> Sigh. You _should_ rave. Heaven knows, Erik Naggum does.
:-)
> In the current implementation of Mule, coding systems are a property
> of ... you can't really be sure. And their semantics are ... rather
> ad hoc.
This looks like a major problem, in major need of fixing.
> In the particular case of Fmd5 and the uses of it that I know about,
> it probably doesn't need it; proper care in implementation should
> have Fmd5 use the output coding-system (but this must be specified
> by the mailer, and most mailers do the translation to standard
> Internet mail encodings as a post-processing step). But then, the
> original implementor of md5 didn't consider that Steve Baur was
> going to use XEmacs to md5 checksum packages' external
> representation; he thought it would be OK as long as the buffer
> representation would be checksummed correctly. So I wouldn't bet on
> it.
For the record: I *do* note that md5 checksums are useful for things
other than the internal buffer representation. But I don't think that
`md5' *function* should have that additional CODING-SYSTEM argument.
Couldn't it pick the coding system from a generally useful variable,
like coding-system-for-write? Am I making any sense?
> The problem is that often textual source which you would like to
> combine in a buffer have varied encodings to start with. So some
> choice of buffer representation and mechanisms for translation are
> needed. Note that "textual source" not only means files that are
> encoded differently, but also you may have process and user input
> in yet other encodings.
Well, this is true, but the case of `md5' calculating a digest is
surely no harder than the case of `C-x C-s' saving the buffer, if you
see what I mean...
> The problems that we are currently discussing stem, in my opinion,
> from the lack of an good extent abstraction in Mule. This means that
> source and target encodings must be defined either as properties of
> characters or as a property of a whole stream. The former is ugly,
> time-consuming, and mistake-prone; the latter is simply not
> fine-grained enough. So you end up with coding-system specifications
> in operators. Yuck.
Yes, I think you have a point.
> Concretely, this would be like the current situation with no-mule
> XEmacs where you can change a buffer displayed as junk to legible
> Croatian by changing the face from a *-iso8859-1 to a *-iso8859-2
> font.[4]
That would be cool. Even cooler would be to be able to specify the
default latin* character set, so that when I encounter Š anywhere in
a file, I want to see it as Š, not as the copyright symbol. Sort of
"trus me pal, I know it's latin2."
This is one big advantage of non-Mule XEmacs over Mule XEmacs for
Croatian (and similar languages.)
--
Hrvoje Niksic <hniksic(a)srce.hr> | Student at FER Zagreb, Croatia
--------------------------------+--------------------------------
Personifiers Unite! You have nothing to lose but Mr. Dignity!
- 3 participants
- 3 comments
Re: Lstreams, coding and md5 (was: Re: [PATCH] menubar selection in mswindows)
25 years, 12 months
Hrvoje Niksic
Hrvoje Niksic
"Stephen J. Turnbull" <turnbull(a)sk.tsukuba.ac.jp> writes:
> But if you're actually going to stuff it into a TCP/IP stream via a
> process buffer, you need to use coding-system-for-process-output (or
> whatever the appropriate variable is named).
OK, this is the crux of your argument, and I see what you mean.
Still, how often does coding-system-for-write differ from
coding-system-for-process-output? If such case is degenerated, then
using a coding-system-for-md5 variable might not be such a bad idea,
because it would be used only very rarely.
> It seems to me that in any case where external and internal
> encodings differ, Fmd5 needs to know how you plan to use its value.
> In a world where the external encodings are chosen by oracle, you
> need an argument to Fmd5. You can either make it explicit,
> optional, and default it as Jareth has done
It's not as simple as that. Remember that we are creating, among
other things, a programming language. Before coding systems came into
the picture, Fmd5 was an exemplary of Lisp elegance:
`md5' is a built-in function
(md5 OBJECT &optional START END)
Documentation:
Return the MD5 message digest of OBJECT, a buffer or string.
Optional arguments START and END denote positions for computing the
digest of a portion of OBJECT.
This is as nice as it can get. However, with Mule, it suddenly got 10
additional lines of docstring and two more optional arguments that
have *nothing* to do with the process of md5-ing a stream of bytes,
and that are very (Mule)-implementation specific.
These uglifications are not to be taken lightly. Imagine what would
happen if we wanted to, say, add another "real" argument to Fmd5, say
GRANULARITY, or whatever? For compatibility, it would have to go
*after* the Mule arguments, and callers would have to write:
(md5 foo nil nil nil nil 10000)
The number of useless arguments is obscene. This example is, of
course, purely hypothetical (there is no such thing as MD5
granularity), and the point I'm trying to make is that interface to
user-visible Lisp functions should be well thought-out, or problems
will arise.
--
Hrvoje Niksic <hniksic(a)srce.hr> | Student at FER Zagreb, Croatia
--------------------------------+--------------------------------
- Now what did we learn from this?
- I learned what my liver looks like!
- 3 participants
- 3 comments
[OK] 21.0 "Repartida-pre2" XEmacs Lucid (beta38) on i586-pc-linux
25 years, 12 months
Larry Ayers
Larry Ayers
Works fine, built with no errors. Switched from egcs to pgcc, libc.5.4.44.
If someone wants to try an excellent TrueType screen font for
XEmacs, I recommend Georgia, which was commissioned by Microsoft
and is available from
http://www.microsoft.com/typography/fontpack/default.htm. It's
the sole MS product on this machine, I swear! (just kidding).
> XEmacs Build Report as generated
> by build-report-version 1.35 follows:
> Contents of /usr/local/src/xemacs-21.0-b37/Installation:
> (Output from most recent run of ./configure)
uname -a: Linux liatris 2.0.33 #1 Sun Apr 19 16:35:48 CDT 1998 i586 unknown
./configure '--prefix=/usr/local/xemacs-beta' '--with-site-lisp' '--package-path=/usr/local/xemacs-beta/lib/xemacs/packages' '--with-gif=no' '--with-png=no' '--with-tiff=no' '--with-jpeg' '--with-xpm' '--cflags=-O6 -mpentium -fno-exceptions -fno-peep-spills -fno-risc' '--with-sound=native' '--with-toolbars=no' '--error-checking=none' '--with-pop' '--gung-ho' '--verbose'
XEmacs 21.0-b38 "Repartida-pre2" configured for `i586-pc-linux'.
Where should the build process find the source code? /usr/local/src/xemacs-21.0-b37
What installation prefix should install use? /usr/local/xemacs-beta
What operating system and machine description files should XEmacs use?
`s/linux.h' and `m/intel386.h'
What compiler should XEmacs be built with? gcc -O6 -mpentium -fno-exceptions -fno-peep-spills -fno-risc
Should XEmacs use the GNU version of malloc? yes
(Using Doug Lea's new malloc from the Linux C Library.)
Should XEmacs use the relocating allocator for buffers? yes
What window system should XEmacs use? x11
Where do we find X Windows header files? /usr/X11/include
Where do we find X Windows libraries? /usr/X11/lib
Compiling in support for XAUTH.
Compiling in support for XPM images.
Compiling in support for JPEG image conversion.
Compiling in native sound support.
Compiling in support for Berkeley DB.
Compiling in support for GNU DBM.
Compiling in support for ncurses.
Compiling in support for GPM (General Purpose Mouse).
Compiling in support for OffiX.
Compiling in support for proper session-management.
Using Lucid menubars.
Using Lucid scrollbars.
Using Motif dialog boxes.
Compiling in DLL support.
movemail will use "dot-locking" for locking mail spool files.
Using POP for mail access
Using Lisp_Objects with minimal tagbits.
Using indexed lrecord implementation.
Compiling in extra code for debugging.
Compiling in code for checking XEmacs memory usage.
> Contents of /usr/local/src/xemacs-21.0-b37/beta.err
> keeping lines matching
> "^--\[\[\|\]\]$\|make\[\|error\|warn\|pure.*\(space\|size\)\|hides\b\|strange\|shadowings\|^Compilation\|not\s-+found"
> and then deleting lines matching
> "confl.*with.*auto-inlining\|/usr/local/src/xemacs-21\.0-b37/lisp/[^
]+ hides "
Purespace usage: 773292 of 1507600 (51% -- 717k wasted).
XEmacs: Pure size adjusted, Don't Panic! I will restart the `make'
make[1]: [xemacs] Error 1 (ignored)
Purespace usage: 773292 of 773292 (100%).
/usr/local/xemacs-beta/lib/xemacs/site-lisp/time-stamp hides /usr/local/xemacs-beta/lib/xemacs/packages/lisp/xemacs-base/time-stamp
/usr/local/xemacs-beta/lib/xemacs/packages/lisp/efs/passwd hides /usr/local/xemacs-beta/lib/xemacs/packages/lisp/xemacs-base/passwd
/usr/local/xemacs-beta/lib/xemacs/packages/lisp/efs/auto-save hides /usr/local/src/xemacs-21.0-b36/lisp/auto-save
3 Emacs Lisp load-path shadowings were foundCleaning out of date .elcs in directory `lisp'...
- 1 participants
- 0 comments