From gracin at tel.fer.hr  Thu Apr 26 20:04:08 2007
From: gracin at tel.fer.hr (Josip Gracin)
Date: Thu, 26 Apr 2007 22:04:08 +0200
Subject: [cl-containers-devel] Problems with priority queues
Message-ID: <463105B8.5000806@tel.fer.hr>

Hello!

I'm having problems with priority queue in cl-containers (installed by 
asdf-install, SBCL 1.0.4.109).  Here's what I'm trying:

======================================================================
* (defparameter *s* (make-container 'priority-queue-on-container))

*S*
* (insert-item *s* 1.0)

#<BINARY-SEARCH-TREE {BEFBF61}>
#<METABANG.CL-CONTAINERS::BST-NODE 1.0>
* (first-item *s*)

debugger invoked on a SIMPLE-ERROR in thread #<THREAD "initial thread" 
{B7DC899}>:
   There is no applicable method for the generic function
     #<STANDARD-GENERIC-FUNCTION ELEMENT (6)>
   when called with arguments
     (1.0).
======================================================================

I've checked the sources of cl-containers and if I delete all calls to 
function ELEMENT, priority queues begin to work as expected.

Am I doing something wrong in the above example?



From gwking at metabang.com  Fri Apr 27 02:30:35 2007
From: gwking at metabang.com (Gary King)
Date: Thu, 26 Apr 2007 22:30:35 -0400
Subject: [cl-containers-devel] Problems with priority queues
In-Reply-To: <463105B8.5000806@tel.fer.hr>
References: <463105B8.5000806@tel.fer.hr>
Message-ID: <3CFB2150-4BD2-470B-853C-F67C3C7EFF68@metabang.com>

Hi Josip,

Thanks for your clear instructions on how to recreate this bug (and  
it's definitely a CL-Containers problem, not anything that you're  
doing wrong).

It looks to me as if the first-element methods for priority-queue-on- 
container are screwed up. Since the first-element call to the  
container of the priority-queue-on-container will return an element  
(not a node), it makes no sense to call element again. In short,  
instead of

> (defmethod first-element ((q priority-queue-on-container))
>   (element (first-element (container q))))

it should be

> (defmethod first-element ((q priority-queue-on-container))
>   (first-element (container q)))
>
> (defmethod (setf first-element) (value (q priority-queue-on- 
> container))
>   (setf (first-element (container q)) value))

I should be able to get this out sometime tomorrow.

thanks again,
--
Gary Warren King, metabang.com
Cell: (413) 885 9127
Fax: (206) 338-4052
gwkkwg on Skype * garethsan on AIM






From gracin at tel.fer.hr  Fri Apr 27 04:51:10 2007
From: gracin at tel.fer.hr (Josip Gracin)
Date: Fri, 27 Apr 2007 06:51:10 +0200
Subject: [cl-containers-devel] Problems with priority queues
In-Reply-To: <3CFB2150-4BD2-470B-853C-F67C3C7EFF68@metabang.com>
References: <463105B8.5000806@tel.fer.hr>
	<3CFB2150-4BD2-470B-853C-F67C3C7EFF68@metabang.com>
Message-ID: <4631813E.30001@tel.fer.hr>

Gary King wrote:
> I should be able to get this out sometime tomorrow.

Great!  Thanks for great work!


From gwking at metabang.com  Sun Apr 29 22:13:21 2007
From: gwking at metabang.com (Gary King)
Date: Sun, 29 Apr 2007 18:13:21 -0400
Subject: [cl-containers-devel] Problems with priority queues
In-Reply-To: <4631813E.30001@tel.fer.hr>
References: <463105B8.5000806@tel.fer.hr>
	<3CFB2150-4BD2-470B-853C-F67C3C7EFF68@metabang.com>
	<4631813E.30001@tel.fer.hr>
Message-ID: <B9B22ACB-9F0D-4DB0-853E-11DDFE67EE56@metabang.com>

Hi Josip,

This should now be corrected; you may need to update metatilities too.

Please let me know,


On Apr 27, 2007, at 12:51 AM, Josip Gracin wrote:

> Gary King wrote:
>> I should be able to get this out sometime tomorrow.
>
> Great!  Thanks for great work!

--
Gary Warren King, metabang.com
Cell: (413) 885 9127
Fax: (206) 338-4052
gwkkwg on Skype * garethsan on AIM






From gracin at tel.fer.hr  Mon Apr 30 08:35:14 2007
From: gracin at tel.fer.hr (Josip Gracin)
Date: Mon, 30 Apr 2007 10:35:14 +0200
Subject: [cl-containers-devel] Problems with priority queues
In-Reply-To: <B9B22ACB-9F0D-4DB0-853E-11DDFE67EE56@metabang.com>
References: <463105B8.5000806@tel.fer.hr>
	<3CFB2150-4BD2-470B-853C-F67C3C7EFF68@metabang.com>
	<4631813E.30001@tel.fer.hr>
	<B9B22ACB-9F0D-4DB0-853E-11DDFE67EE56@metabang.com>
Message-ID: <4635AA42.1030000@tel.fer.hr>

Hi!

Gary King wrote:
> This should now be corrected; you may need to update metatilities too.

I've updated metatilities but now I have a problem compiling them.  I 
don't know if its SBCL (1.0.5.2) or metatilities problem.  Here's what I 
get:

; compiling (DEFUN NICELY-FORMAT-FILENAME ...)
; file: /home/gracin/.sbcl/site/metatilities/dev/utilities/files.lisp
; in: DEFUN NICELY-FORMAT-FILENAME
;     (FORMAT STREAM "~(~a~@[.~A~] {~:[...:~]~{~a:~}}~)"
;           (PATHNAME-NAME METABANG.UTILITIES::FILE)
;           (AND METABANG.UTILITIES::SHOW-TYPE?
;                (PATHNAME-TYPE METABANG.UTILITIES::FILE))
;           (AND METABANG.UTILITIES::INITIAL-ELLIPSIS?
;                (<= (LENGTH METABANG.UTILITIES::DIRECTORIES)
;                    METABANG.UTILITIES::DEPTH))
;           (LAST METABANG.UTILITIES::DIRECTORIES 
METABANG.UTILITIES::DEPTH))
; --> FORMAT
; ==>
;   (FORMATTER "~(~a~@[.~A~] {~:[...:~]~{~a:~}}~)")
;
; caught ERROR:
;   (in macroexpansion of (FORMATTER "~(~a~@[.~A~] {~:[...:~]~{~a:~}}~)"))
;   (hint: For more precise location, try *BREAK-ON-SIGNALS*.)
;   error in FORMAT: must specify exactly two sections
;     ~(~a~@[.~A~] {~:[...:~]~{~a:~}}~)
;                     ^

; ==>
;   (FORMAT SB-C::DEST (FORMATTER "~(~a~@[.~A~] {~:[...:~]~{~a:~}}~)") #:G65
;           #:G66 #:G67 #:G68)
;
; note: The second argument never returns a value.


From gracin at tel.fer.hr  Mon Apr 30 14:59:43 2007
From: gracin at tel.fer.hr (Josip Gracin)
Date: Mon, 30 Apr 2007 16:59:43 +0200
Subject: [cl-containers-devel] Problems with priority queues
In-Reply-To: <B9B22ACB-9F0D-4DB0-853E-11DDFE67EE56@metabang.com>
References: <463105B8.5000806@tel.fer.hr>
	<3CFB2150-4BD2-470B-853C-F67C3C7EFF68@metabang.com>
	<4631813E.30001@tel.fer.hr>
	<B9B22ACB-9F0D-4DB0-853E-11DDFE67EE56@metabang.com>
Message-ID: <4636045F.70607@tel.fer.hr>

Gary King wrote:
> This should now be corrected; you may need to update metatilities too.
> 
> Please let me know,

Gary, after checking the source code, here's another contrived example 
which does not work for me:

(let ((c (make-container 'priority-queue-on-container)))
   (insert-item c 2)
   (insert-item c 1)
   (insert-item c 3)
   (delete-element c 2))



From gwking at metabang.com  Mon Apr 30 16:06:16 2007
From: gwking at metabang.com (Gary King)
Date: Mon, 30 Apr 2007 12:06:16 -0400
Subject: [cl-containers-devel] Problems with priority queues
In-Reply-To: <4635AA42.1030000@tel.fer.hr>
References: <463105B8.5000806@tel.fer.hr>
	<3CFB2150-4BD2-470B-853C-F67C3C7EFF68@metabang.com>
	<4631813E.30001@tel.fer.hr>
	<B9B22ACB-9F0D-4DB0-853E-11DDFE67EE56@metabang.com>
	<4635AA42.1030000@tel.fer.hr>
Message-ID: <D526CF40-7CA1-4536-BB59-E896E01B4B72@metabang.com>

Hi Josip,

This hasn't changed in metatilities and it looks to me as if SBCL is  
confused. The error message looks like it thinks I'm using ~:[. As in:

> * (format t "~:[A~;B~]" nil)
> A
> NIL
> * (format t "~:[A~;B~]" t)
> B
> NIL
> * (format t "~:[AB~]" t)
>
> debugger invoked on a SB-FORMAT:FORMAT-ERROR:
>   error in FORMAT: must specify exactly two sections
>   ~:[AB~]
>     ^


But ~@[ should be OK. What does

> (format t "~@[.~a~]" 23)

give you?





On Apr 30, 2007, at 4:35 AM, Josip Gracin wrote:

> ; compiling (DEFUN NICELY-FORMAT-FILENAME ...)
> ; file: /home/gracin/.sbcl/site/metatilities/dev/utilities/files.lisp
> ; in: DEFUN NICELY-FORMAT-FILENAME
> ;     (FORMAT STREAM "~(~a~@[.~A~] {~:[...:~]~{~a:~}}~)"
> ;           (PATHNAME-NAME METABANG.UTILITIES::FILE)
> ;           (AND METABANG.UTILITIES::SHOW-TYPE?
> ;                (PATHNAME-TYPE METABANG.UTILITIES::FILE))
> ;           (AND METABANG.UTILITIES::INITIAL-ELLIPSIS?
> ;                (<= (LENGTH METABANG.UTILITIES::DIRECTORIES)
> ;                    METABANG.UTILITIES::DEPTH))
> ;           (LAST METABANG.UTILITIES::DIRECTORIES  
> METABANG.UTILITIES::DEPTH))
> ; --> FORMAT
> ; ==>
> ;   (FORMATTER "~(~a~@[.~A~] {~:[...:~]~{~a:~}}~)")
> ;
> ; caught ERROR:
> ;   (in macroexpansion of (FORMATTER "~(~a~@[.~A~] {~:[...:~]~ 
> {~a:~}}~)"))
> ;   (hint: For more precise location, try *BREAK-ON-SIGNALS*.)
> ;   error in FORMAT: must specify exactly two sections
> ;     ~(~a~@[.~A~] {~:[...:~]~{~a:~}}~)
> ;                     ^
>
> ; ==>
> ;   (FORMAT SB-C::DEST (FORMATTER "~(~a~@[.~A~] {~:[...:~]~{~a:~}} 
> ~)") #:G65
> ;           #:G66 #:G67 #:G68)
> ;
> ; note: The second argument never returns a value.
--
Gary Warren King, metabang.com
Cell: (413) 885 9127
Fax: (206) 338-4052
gwkkwg on Skype * garethsan on AIM






From gracin at tel.fer.hr  Mon Apr 30 16:23:21 2007
From: gracin at tel.fer.hr (Josip Gracin)
Date: Mon, 30 Apr 2007 18:23:21 +0200
Subject: [cl-containers-devel] Problems with priority queues
In-Reply-To: <D526CF40-7CA1-4536-BB59-E896E01B4B72@metabang.com>
References: <463105B8.5000806@tel.fer.hr>
	<3CFB2150-4BD2-470B-853C-F67C3C7EFF68@metabang.com>
	<4631813E.30001@tel.fer.hr>
	<B9B22ACB-9F0D-4DB0-853E-11DDFE67EE56@metabang.com>
	<4635AA42.1030000@tel.fer.hr>
	<D526CF40-7CA1-4536-BB59-E896E01B4B72@metabang.com>
Message-ID: <463617F9.7020408@tel.fer.hr>

Gary King wrote:
> But ~@[ should be OK. What does
> 
>> (format t "~@[.~a~]" 23)
> 
> give you?

 >> (format t "~@[.~a~]" 23)
.23
NIL



From gwking at metabang.com  Mon Apr 30 18:12:07 2007
From: gwking at metabang.com (Gary King)
Date: Mon, 30 Apr 2007 14:12:07 -0400
Subject: [Sbcl-help] [cl-containers-devel] Problems with priority queues
In-Reply-To: <87hcqx7r1r.fsf@vasara.proghammer.com>
References: <463105B8.5000806@tel.fer.hr>
	<3CFB2150-4BD2-470B-853C-F67C3C7EFF68@metabang.com>
	<4631813E.30001@tel.fer.hr>
	<B9B22ACB-9F0D-4DB0-853E-11DDFE67EE56@metabang.com>
	<4635AA42.1030000@tel.fer.hr>
	<D526CF40-7CA1-4536-BB59-E896E01B4B72@metabang.com>
	<87hcqx7r1r.fsf@vasara.proghammer.com>
Message-ID: <CDF68651-AC16-4B43-B891-6AD8FCE8A7BC@metabang.com>

I used to be confused about why I was confused but it took too much  
work.

Thanks Juho...

SBCL keeps getter better and smarter!



On Apr 30, 2007, at 12:41 PM, Juho Snellman wrote:

> Gary King <gwking at metabang.com> writes:
>> Hi Josip,
>>
>> This hasn't changed in metatilities and it looks to me as if SBCL is
>> confused. The error message looks like it thinks I'm using ~:[. As  
>> in:
>>
>>> * (format t "~:[A~;B~]" nil)
>>> A
>>> NIL
>>> * (format t "~:[A~;B~]" t)
>>> B
>>> NIL
>>> * (format t "~:[AB~]" t)
>>>
>>> debugger invoked on a SB-FORMAT:FORMAT-ERROR:
>>>   error in FORMAT: must specify exactly two sections
>>>   ~:[AB~]
>>>     ^
>
> I'm confused about why you think SBCL is confused, since you *are*  
> using
> using ~:[ :-)
>
>>> ;     (FORMAT STREAM "~(~a~@[.~A~] {~:[...:~]~{~a:~}}~)"
>                                         ^^^^^^^^^
>
> -- 
> Juho Snellman
>

--
Gary Warren King, metabang.com
Cell: (413) 885 9127
Fax: (206) 338-4052
gwkkwg on Skype * garethsan on AIM






From jsnell at iki.fi  Mon Apr 30 16:41:52 2007
From: jsnell at iki.fi (Juho Snellman)
Date: 30 Apr 2007 19:41:52 +0300
Subject: [Sbcl-help] [cl-containers-devel] Problems with priority queues
In-Reply-To: <D526CF40-7CA1-4536-BB59-E896E01B4B72@metabang.com>
References: <463105B8.5000806@tel.fer.hr>
	<3CFB2150-4BD2-470B-853C-F67C3C7EFF68@metabang.com>
	<4631813E.30001@tel.fer.hr>
	<B9B22ACB-9F0D-4DB0-853E-11DDFE67EE56@metabang.com>
	<4635AA42.1030000@tel.fer.hr>
	<D526CF40-7CA1-4536-BB59-E896E01B4B72@metabang.com>
Message-ID: <87hcqx7r1r.fsf@vasara.proghammer.com>

Gary King <gwking at metabang.com> writes:
> Hi Josip,
> 
> This hasn't changed in metatilities and it looks to me as if SBCL is  
> confused. The error message looks like it thinks I'm using ~:[. As in:
> 
> > * (format t "~:[A~;B~]" nil)
> > A
> > NIL
> > * (format t "~:[A~;B~]" t)
> > B
> > NIL
> > * (format t "~:[AB~]" t)
> >
> > debugger invoked on a SB-FORMAT:FORMAT-ERROR:
> >   error in FORMAT: must specify exactly two sections
> >   ~:[AB~]
> >     ^

I'm confused about why you think SBCL is confused, since you *are* using
using ~:[ :-)
 
> > ;     (FORMAT STREAM "~(~a~@[.~A~] {~:[...:~]~{~a:~}}~)"
                                        ^^^^^^^^^                             

-- 
Juho Snellman