From tim at fractaldragon.net  Tue Jan  4 07:45:35 2005
From: tim at fractaldragon.net (Tim Lavoie)
Date: Tue, 4 Jan 2005 01:45:35 -0600
Subject: [Bese-devel] UCW 0.3.4 - browser issue?
Message-ID: <20050104074534.GA2984@fractaldragon.net>

Hello,

I've just started tinkering with UCW 0.3.4, and have what seems like a
peculiar problem. I have it installed (Linux/SBCL/mod_lisp), and can
browse the example pages.

The odd bit is that while I can do it without difficulty using the w3m
text browser, I get the following on the REPL on most pages with Opera
or Firefox. I have disabled the usual proxies etc., and cookies should
be accepted OK. Any ideas?

	Thanks,
	Tim

=================== error follows ===============

Condition IT.BESE.UCW::INEXISTENT-ENTRY-POINT was signalled.
   [Condition of type IT.BESE.UCW::INEXISTENT-ENTRY-POINT]

Restarts:
  0: [SERVER-ERROR] Send the client an internal server error page.
  1: [DESTROY-THREAD] Destroy this thread (2669)

Backtrace:
  0: ((SWANK-BACKEND:CALL-WITH-DEBUGGING-ENVIRONMENT (T)) #<unused argument> #<unused argument> #<FUNCTION "#'(LAMBDA NIL (SWANK::SLDB-LOOP SWANK::*SLDB-LEVEL*))" {91F8FE5}>)
  1: (SWANK::DEBUG-IN-EMACS #<IT.BESE.UCW::INEXISTENT-ENTRY-POINT {A2A9CB1}>)
  2: ("#'(LAMBDA NIL (LET # #))")
  3: (SWANK::CALL-WITH-REDIRECTED-IO #<SWANK::CONNECTION {A0A1419}> #<FUNCTION "CLOSURE" {A2A9EC5}>)
  4: (SWANK:SWANK-DEBUGGER-HOOK #<IT.BESE.UCW::INEXISTENT-ENTRY-POINT {A2A9CB1}> #<unavailable argument>)
  5: ("#'(LAMBDA (IT.BESE.UCW::C) (WHEN IT.BESE.UCW::*DEBUG-ON-ERROR* #) ...)" #<IT.BESE.UCW::INEXISTENT-ENTRY-POINT {A2A9CB1}>)
  6: ("XEP for #'(LAMBDA (IT.BESE.UCW::C) (WHEN IT.BESE.UCW::*DEBUG-ON-ERROR* #) ...)" 1 #<IT.BESE.UCW::INEXISTENT-ENTRY-POINT {A2A9CB1}>)[:EXTERNAL]
  7: (SIGNAL 1 #<IT.BESE.UCW::INEXISTENT-ENTRY-POINT {A2A9CB1}>)[:EXTERNAL]
  8: (ERROR 1 IT.BESE.UCW::INEXISTENT-ENTRY-POINT)[:EXTERNAL]
  9: ("LABELS IT.BESE.UCW::SERVICE-AS-ENTRY-POINT")
 10: ((IT.BESE.UCW::SERVICE (IT.BESE.UCW::STANDARD-SESSION-FRAME IT.BESE.UCW::REQUEST-CONTEXT)) #<unavailable argument> #<unavailable argument> #<IT.BESE.UCW::STANDARD-SESSION-FRAME {A2A6CB9}> #<IT.BESE.UCW::COOKIE-SESSION-REQUEST-CONTEXT {A29D711}>)
 11: ("#'(LAMBDA (SB-PCL::.PV-CELL. SB-PCL::.NEXT-METHOD-CALL. SB-PCL::.ARG0. ...) (DECLARE #) ...)" #<unavailable argument> #<unavailable argument> #<IT.BESE.UCW::STANDARD-SESSION-FRAME {A2A6CB9}> #<IT.BESE.UCW::COOKIE-SESSION-REQUEST-CONTEXT {A29D711}>)
 12: ("#'(LAMBDA (SB-PCL::.PV-CELL. SB-PCL::.NEXT-METHOD-CALL. SB-PCL::.ARG0. ...) (DECLARE #) ...)" #<unavailable argument> #<unavailable argument> #<IT.BESE.UCW::STANDARD-SESSION 42 #<IT.BESE.UCW::STANDARD-SESSION-FRAME {A2A6CB9}> {A3F9439}> #<IT.BESE.UCW::COOKIE-SESSION-REQUEST-CONTEXT {A29D711}>)
 13: ("#'(LAMBDA (SB-PCL::.PV-CELL. SB-PCL::.NEXT-METHOD-CALL. SB-PCL::.ARG0. ...) (DECLARE #) ...)" #<unavailable argument> #<unavailable argument> #<IT.BESE.UCW::COOKIE-SESSION-APPLICATION /ucw/examples/ 4> #<IT.BESE.UCW::COOKIE-SESSION-REQUEST-CONTEXT {A29D711}>)
 14: ((IT.BESE.UCW::HANDLE-REQUEST (STANDARD-SERVER REQUEST RESPONSE)) #<unavailable argument> #<unavailable argument> #<STANDARD-SERVER MOD-LISP-BACKEND 2 {9DC8909}> #<IT.BESE.UCW::MOD-LISP-REQUEST :RAW-URL "/ucw/examples/stylesheet.css"> #<IT.BESE.UCW::MOD-LISP-RESPONSE {A299691}>)
 15: (IT.BESE.UCW::MOD-LISP-SERVER-LOOP #<MOD-LISP-BACKEND {9C4CD91}>)
 16: ("XEP for IT.BESE.UCW::MOD-LISP-SERVER-LOOP" 1 #<MOD-LISP-BACKEND {9C4CD91}>)[:EXTERNAL]
 17: ("#'(LAMBDA NIL (IT.BESE.UCW::MOD-LISP-SERVER-LOOP IT.BESE.UCW::BACKEND))")
 18: ("XEP for #'(LAMBDA NIL (IT.BESE.UCW::MOD-LISP-SERVER-LOOP IT.BESE.UCW::BACKEND))" 0)[:EXTERNAL]
 19: ("#'(LAMBDA NIL (LET # # ...) ...)")
 20: ("foreign function: call_into_lisp")
 21: ("foreign function: funcall0")



From randall at randallsquared.com  Tue Jan  4 10:52:56 2005
From: randall at randallsquared.com (Randall Randall)
Date: Tue, 4 Jan 2005 05:52:56 -0500
Subject: [Bese-devel] UCW 0.3.4 - browser issue?
In-Reply-To: <20050104074534.GA2984@fractaldragon.net>
References: <20050104074534.GA2984@fractaldragon.net>
Message-ID: <CA91FC0A-5E3E-11D9-9EB3-000A95A0F1E8@randallsquared.com>


On Jan 4, 2005, at 2:45 AM, Tim Lavoie wrote:
> I've just started tinkering with UCW 0.3.4, and have what seems like a
> peculiar problem. I have it installed (Linux/SBCL/mod_lisp), and can
> browse the example pages.
>
> The odd bit is that while I can do it without difficulty using the w3m
> text browser, I get the following on the REPL on most pages with Opera
> or Firefox. I have disabled the usual proxies etc., and cookies should
> be accepted OK. Any ideas?

This may be Opera and Firefox looking for a favicon in
the current folder.  More information about this behavior
can be found at http://en.wikipedia.org/wiki/Favicon .

--
Randall Randall <randall at randallsquared.com>
Property law should use #'EQ , not #'EQUAL .



From mb at bese.it  Fri Jan 14 16:55:56 2005
From: mb at bese.it (Marco Baringer)
Date: Fri, 14 Jan 2005 17:55:56 +0100
Subject: [Bese-devel] UCW 0.3.4 - browser issue?
References: <20050104074534.GA2984@fractaldragon.net>
	<CA91FC0A-5E3E-11D9-9EB3-000A95A0F1E8@randallsquared.com>
Message-ID: <m2k6qglzkz.fsf@soma.local>

Randall Randall <randall at randallsquared.com> writes:

> On Jan 4, 2005, at 2:45 AM, Tim Lavoie wrote:
>> I've just started tinkering with UCW 0.3.4, and have what seems like a
>> peculiar problem. I have it installed (Linux/SBCL/mod_lisp), and can
>> browse the example pages.
>>
>> The odd bit is that while I can do it without difficulty using the w3m
>> text browser, I get the following on the REPL on most pages with Opera
>> or Firefox. I have disabled the usual proxies etc., and cookies should
>> be accepted OK. Any ideas?
>
> This may be Opera and Firefox looking for a favicon in
> the current folder.  More information about this behavior
> can be found at http://en.wikipedia.org/wiki/Favicon .

ucw's default handling of 404s is to drop to a debugger (if
*debug-on-error* is T), this could probably be changed to something
else (just returning a 404 would be a good idea). However, what i'd
generally suggest doing is making apache more selective about which
requests get sent to ucw and which don't. As opposed to having
everything under a certain directory be sent to ucw you probably want
images and static html files to be sevrved by apache, something like
this in your httpdconf:

<LocationMatch ".*\.ucw">
  SetHandler lisp
</LocationMatch>

assuming all your entry-points end in ".ucw"

what we really need is a way for the applications to be able to
communicate with the backend about which urls they should pass to ucw
and which should be dealt with "normally". (though this may only be
applicable to aserve and araneida)

-- 
-Marco
Ring the bells that still can ring.
Forget the perfect offering.
There is a crack in everything.
That's how the light gets in.
	-Leonard Cohen


From tim at fractaldragon.net  Fri Jan 14 17:26:03 2005
From: tim at fractaldragon.net (Tim Lavoie)
Date: Fri, 14 Jan 2005 11:26:03 -0600
Subject: [Bese-devel] UCW 0.3.4 - browser issue?
In-Reply-To: <m2k6qglzkz.fsf@soma.local> (Marco Baringer's message of "Fri,
	14 Jan 2005 17:55:56 +0100")
References: <20050104074534.GA2984@fractaldragon.net>
	<CA91FC0A-5E3E-11D9-9EB3-000A95A0F1E8@randallsquared.com>
	<m2k6qglzkz.fsf@soma.local>
Message-ID: <87y8evrkgk.fsf@theasylum.dyndns.org>

Hi Marco, thanks for the good ideas. Being more specific helps with
app-related images too, which of course Apache does just fine.

   Cheers,
   Tim



From mb at bese.it  Wed Jan 19 17:38:21 2005
From: mb at bese.it (Marco Baringer)
Date: Wed, 19 Jan 2005 18:38:21 +0100
Subject: [Bese-devel] clisp and araneida (and windows)
Message-ID: <m2brbl8gky.fsf@soma.local>


in the past few todays i've commited patches (patch-184 through
patch-190) allowing ucw to run on clisp (latest cvs only) and using
araneida (0.9) as a backend. what this means (and the reason i did it)
is that you can now deploy your ucw apps on windows servers.

the code works but hasn't been tested much [just enough to pass the
demo i did today with ucw + araneida + clisp on win2k]. supporting
clisp was fairly painless (after figuring out that clisp deals with
some corner cases of the mop in interesting ways). supporting araneida
required a fairly large change to the server vs. backend
protocol. previously we simply registered an app with a server and
that was that, now apps are no longer registered but each individual
entry-point is. a nice concequence of this is that we now have a
publish-directory function (only implemnted for araneida) and GET
/favicon.ico non longer bombs ucw (on aserve and araneida at
least). this change also fixed a bug inthe aserve backend whereby
entry-points were required to end in ".ucw". nothing has changed as
far as mod_lisp is concerned.

p.s. - if anyone is interested in testing ucw on windows and doesn't
want to build clisp from cvs (which took me 30 minutes + 100MB (not
counting cygwin)) i have a 2.6MB tar ball containing clisp.exe and the
bare minimum cygwin dlls required to run it.

-- 
-Marco
Ring the bells that still can ring.
Forget the perfect offering.
There is a crack in everything.
That's how the light gets in.
	-Leonard Cohen


From edgar.goncalves at tagus.ist.utl.pt  Fri Jan 21 11:10:36 2005
From: edgar.goncalves at tagus.ist.utl.pt (=?iso-8859-1?q?Edgar_Gon=E7alves?=)
Date: Fri, 21 Jan 2005 11:10:36 +0000
Subject: [Bese-devel] clisp and araneida (and windows)
Message-ID: <usm4vavgz.fsf@tagus.ist.utl.pt>


Hi,

> in the past few todays i've commited patches (patch-184 through
> patch-190) allowing ucw to run on clisp (latest cvs only) and using
> araneida (0.9) as a backend. what this means (and the reason i did it)
> is that you can now deploy your ucw apps on windows servers.

I've come to some problems when trying to use ucw in the clisp version that you
sent me. After a few system configuration issues, the ucw compilation process
stopped in the following function (in
src/rerl/standard-component/standard-component-class.lisp): 


(defun initialize-component-class-after (class initargs next-method)
  ":after initialization function for standard-component-class objects.

Setup the proper values in component-class.backtrack-slots and
component-class.component-slots in CLASS based on the effective
slots of CLASS."
  (declare (ignore initargs next-method))
  (iterate
    (for slot in (mopp:class-slots class))
    (when (typep slot 'standard-component-effective-slot)
      (when (component-slot.backtrack slot)
        (push (cons (mopp:slot-definition-name slot)
                    (component-slot.backtrack slot))
              (component-class.backtrack-slots class)))
      (when (component-slot.component slot)
        (push (cons (mopp:slot-definition-name slot)
                    (component-slot.component slot))
              (component-class.component-slots class)))))
  (setf (component-class.place-slot class) (first (component-class.place-slot class))))

The error is:

Iterate, in (LET ((#1=G32723 #))
 (SETF (COMPONENT-CLASS.BACKTRACK-SLOTS CLASS) #1#)):
The form ((SETF COMPONENT-CLASS.BACKTRACK-SLOTS) #1# CLASS) is not a valid Lisp expression
   [Condition of type SIMPLE-ERROR]


As I haven't got enough commonlisp knowhow to figure out the problem, I can only
assume it's related to those pushes. These macroexpand to forms like:

(LET*
 ((#:G8373
   (CONS (CLOS:SLOT-DEFINITION-NAME SLOT) (COMPONENT-SLOT.BACKTRACK SLOT))))
 (LET* ((#:G8372 CLASS))
  ((SETF COMPONENT-CLASS.BACKTRACK-SLOTS)
   (CONS #:G8373 (COMPONENT-CLASS.BACKTRACK-SLOTS #:G8372)) #:G8372)))


Any ideas on what may be going on?

By the way, if you plan to make ucw a bit more userfriendly to windows
developers, arch shouldn't be the way to go, as it creates too long filenames
and makes the folder unusable... Right now I have to use a linux machine, remove
the arch directories and copy the result to windows!

Thanks!
  
-- 
Edgar Gon?alves
Portugal


From edgar.goncalves at tagus.ist.utl.pt  Fri Jan 21 10:44:25 2005
From: edgar.goncalves at tagus.ist.utl.pt (=?iso-8859-1?q?Edgar_Gon=E7alves?=)
Date: Fri, 21 Jan 2005 10:44:25 +0000
Subject: [Bese-devel] clisp and araneida (and windows)
Message-ID: <uu0pbawom.fsf@tagus.ist.utl.pt>


Hi,

> in the past few todays i've commited patches (patch-184 through
> patch-190) allowing ucw to run on clisp (latest cvs only) and using
> araneida (0.9) as a backend. what this means (and the reason i did it)
> is that you can now deploy your ucw apps on windows servers.

I've come to some problems when trying to use ucw in the clisp version that you
sent me. After a few system configuration issues, the ucw compilation process
stopped in the following function (in
src/rerl/standard-component/standard-component-class.lisp): 


(defun initialize-component-class-after (class initargs next-method)
  ":after initialization function for standard-component-class objects.

Setup the proper values in component-class.backtrack-slots and
component-class.component-slots in CLASS based on the effective
slots of CLASS."
  (declare (ignore initargs next-method))
  (iterate
    (for slot in (mopp:class-slots class))
    (when (typep slot 'standard-component-effective-slot)
      (when (component-slot.backtrack slot)
        (push (cons (mopp:slot-definition-name slot)
                    (component-slot.backtrack slot))
              (component-class.backtrack-slots class)))
      (when (component-slot.component slot)
        (push (cons (mopp:slot-definition-name slot)
                    (component-slot.component slot))
              (component-class.component-slots class)))))
  (setf (component-class.place-slot class) (first (component-class.place-slot class))))

The error is:

Iterate, in (LET ((#1=G32723 #))
 (SETF (COMPONENT-CLASS.BACKTRACK-SLOTS CLASS) #1#)):
The form ((SETF COMPONENT-CLASS.BACKTRACK-SLOTS) #1# CLASS) is not a valid Lisp expression
   [Condition of type SIMPLE-ERROR]


As I haven't got enough commonlisp knowhow to figure out the problem, I can only
assume it's related to those pushes. These macroexpand to forms like:

(LET*
 ((#:G8373
   (CONS (CLOS:SLOT-DEFINITION-NAME SLOT) (COMPONENT-SLOT.BACKTRACK SLOT))))
 (LET* ((#:G8372 CLASS))
  ((SETF COMPONENT-CLASS.BACKTRACK-SLOTS)
   (CONS #:G8373 (COMPONENT-CLASS.BACKTRACK-SLOTS #:G8372)) #:G8372)))


Any ideas on what may be going on?

By the way, if you plan to make ucw a bit more userfriendly to windows
developers, arch shouldn't be the way to go, as it creates too long filenames
and makes the folder unusable... Right now I have to use a linux machine, remove
the arch directories and copy the result to windows!

Thanks!
  
-- 
Edgar Gon?alves
Portugal


From mb at bese.it  Fri Jan 21 12:31:29 2005
From: mb at bese.it (Marco Baringer)
Date: Fri, 21 Jan 2005 13:31:29 +0100
Subject: [Bese-devel] clisp and araneida (and windows)
In-Reply-To: <usm4vavgz.fsf@tagus.ist.utl.pt> (
	=?iso-8859-1?q?Edgar_Gon=E7alves's_message_of?= "Fri,
	21 Jan 2005 11:10:36 +0000")
References: <usm4vavgz.fsf@tagus.ist.utl.pt>
Message-ID: <m2llan6k0u.fsf@soma.local>

edgar.goncalves at tagus.ist.utl.pt (Edgar Gon?alves) writes:

> Hi,
>
>> in the past few todays i've commited patches (patch-184 through
>> patch-190) allowing ucw to run on clisp (latest cvs only) and using
>> araneida (0.9) as a backend. what this means (and the reason i did it)
>> is that you can now deploy your ucw apps on windows servers.
>
> I've come to some problems when trying to use ucw in the clisp version that you
> sent me. After a few system configuration issues, the ucw compilation process
> stopped in the following function (in
> src/rerl/standard-component/standard-component-class.lisp): 

[snip]

> The error is:
>
> Iterate, in (LET ((#1=G32723 #))
>  (SETF (COMPONENT-CLASS.BACKTRACK-SLOTS CLASS) #1#)):
> The form ((SETF COMPONENT-CLASS.BACKTRACK-SLOTS) #1# CLASS) is not a valid Lisp expression
>    [Condition of type SIMPLE-ERROR]

CLISP has the "interesting" characteristic of expanding this:

(setf (foo X) Y)

to this:

((setf foo) Y X)

The CPS transformer breaks violently on this. You should update the
arnesi lib (which contains the cps transformer) since i've only
recently taught it to handle these forms. [the easiest way to update
ucw (and arnesi and yaclml and iterate) is to run the bin/update.sh
script.]

> By the way, if you plan to make ucw a bit more userfriendly to windows
> developers, arch shouldn't be the way to go, as it creates too long filenames
> and makes the folder unusable... Right now I have to use a linux machine, remove
> the arch directories and copy the result to windows!

ouch! have you tried this (unpack it and put the resulting files in
you cygwin path):

http://ct.radiology.uiowa.edu/~jfmeinel/tla-short-path/

if you use that to get ucw the file names are shortened client side to
avoid those too-long-filename errors (nto that i've tried).

p.s. - did you resolve that MOP issue?

-- 
-Marco
Ring the bells that still can ring.
Forget the perfect offering.
There is a crack in everything.
That's how the light gets in.
	-Leonard Cohen


From edgar.goncalves at tagus.ist.utl.pt  Fri Jan 21 14:36:25 2005
From: edgar.goncalves at tagus.ist.utl.pt (=?iso-8859-1?q?Edgar_Gon=E7alves?=)
Date: Fri, 21 Jan 2005 14:36:25 +0000
Subject: [Bese-devel] clisp and araneida (and windows)
In-Reply-To: <m2llan6k0u.fsf@soma.local> (Marco Baringer's message of "Fri,
	21 Jan 2005 13:31:29 +0100")
References: <usm4vavgz.fsf@tagus.ist.utl.pt> <m2llan6k0u.fsf@soma.local>
Message-ID: <uoefic0ie.fsf@tagus.ist.utl.pt>


This afternoon, Marco Baringer wrote:
[snip]
> The CPS transformer breaks violently on this. You should update the
> arnesi lib (which contains the cps transformer) since i've only
> recently taught it to handle these forms. [the easiest way to update
> ucw (and arnesi and yaclml and iterate) is to run the bin/update.sh
> script.]

Hmm, I have just updated both ucw--dev--0.3 and support--dists, but the script
says I have the trees up to date (both ucw's and the arnesi, yaclml and iterate
library's). I must be doing something wrong - I'm just running the update script
you mentioned, and trying to load ucw again, but the error remains the same, as
there were no updates (obviously).

Is the update already uploaded? 

>
> http://ct.radiology.uiowa.edu/~jfmeinel/tla-short-path/
>
> if you use that to get ucw the file names are shortened client side to
> avoid those too-long-filename errors (nto that i've tried).
Great! Thanks, that worked like a charm!

> p.s. - did you resolve that MOP issue?

Yes, it was my fault, solely - I was using a different clisp, due to a wrong
environment variable with the clisp path! (*shame*)

Thanks.

-- 
Edgar Gon?alves
Software Engineering Group @ INESC-ID
Portugal


From mb at bese.it  Fri Jan 21 16:04:26 2005
From: mb at bese.it (Marco Baringer)
Date: Fri, 21 Jan 2005 17:04:26 +0100
Subject: [Bese-devel] clisp and araneida (and windows)
In-Reply-To: <uoefic0ie.fsf@tagus.ist.utl.pt> (
	=?iso-8859-1?q?Edgar_Gon=E7alves's_message_of?= "Fri,
	21 Jan 2005 14:36:25 +0000")
References: <usm4vavgz.fsf@tagus.ist.utl.pt> <m2llan6k0u.fsf@soma.local>
	<uoefic0ie.fsf@tagus.ist.utl.pt>
Message-ID: <m2acr2vkdx.fsf@soma.local>

edgar.goncalves at tagus.ist.utl.pt (Edgar Gon?alves) writes:

> Hmm, I have just updated both ucw--dev--0.3 and support--dists, but the script
> says I have the trees up to date (both ucw's and the arnesi, yaclml and iterate
> library's). I must be doing something wrong - I'm just running the update script
> you mentioned, and trying to load ucw again, but the error remains the same, as
> there were no updates (obviously).

strange. try evalutaing this:

(asdf:oos 'asdf:load-op :arnesi)

(defun (setf foo) (v) v)

(arnesi:to-cps '(setf (foo) 5))

if that returns something like this:

(IT.BESE.ARNESI::DRIVE-CPS
 (LAMBDA NIL
  (FUNCALL #'IT.BESE.ARNESI::TOPLEVEL-K
   (FUNCALL '|(SETF COMMON-LISP-USER:FOO)| 5))))

then you're using the latest arnesi and we need to look into the
matter more (are you ever on #lisp?). if that errors then for some
reason you've got a stale arnesi lying around.

> Is the update already uploaded? 

you want the patch named arnesi--dev--patch-36, use 'tla log -S' in
the arnesi tree to see if you have it.

-- 
-Marco
Ring the bells that still can ring.
Forget the perfect offering.
There is a crack in everything.
That's how the light gets in.
	-Leonard Cohen


From edgar.goncalves at tagus.ist.utl.pt  Fri Jan 21 23:23:44 2005
From: edgar.goncalves at tagus.ist.utl.pt (=?iso-8859-1?q?Edgar_Gon=E7alves?=)
Date: Fri, 21 Jan 2005 23:23:44 +0000
Subject: [Bese-devel] clisp and araneida (and windows)
In-Reply-To: <m2acr2vkdx.fsf@soma.local> (Marco Baringer's message of "Fri,
	21 Jan 2005 17:04:26 +0100")
References: <usm4vavgz.fsf@tagus.ist.utl.pt> <m2llan6k0u.fsf@soma.local>
	<uoefic0ie.fsf@tagus.ist.utl.pt> <m2acr2vkdx.fsf@soma.local>
Message-ID: <uk6q6bc3j.fsf@tagus.ist.utl.pt>

This afternoon, Marco Baringer wrote:
> strange. try evalutaing this:
>
> (asdf:oos 'asdf:load-op :arnesi)
> (defun (setf foo) (v) v)
> (arnesi:to-cps '(setf (foo) 5))
>
> if that returns something like this:
>
> (IT.BESE.ARNESI::DRIVE-CPS
>  (LAMBDA NIL
>   (FUNCALL #'IT.BESE.ARNESI::TOPLEVEL-K
>    (FUNCALL '|(SETF COMMON-LISP-USER:FOO)| 5))))
>
> then you're using the latest arnesi and we need to look into the
> matter more (are you ever on #lisp?). if that errors then for some
> reason you've got a stale arnesi lying around.
>
Yes, this is the result I get.
>
> you want the patch named arnesi--dev--patch-36, use 'tla log -S' in
> the arnesi tree to see if you have it.
Yes, I've got that patch, and that's why the code you suggested works just
fine.

So I gess we could have a talk. I was never on #lisp, but I may be there, just
tell me what's the irc server you're on. Or else, reach me at:
MSN: edy_in_carpathia [a-t] hotmail [d-o-t] com
ICQ: 73218277

-- 
Edgar Gon?alves
Software Engineering Group @ INESC-ID
Portugal


From mb at bese.it  Sat Jan 22 15:05:26 2005
From: mb at bese.it (Marco Baringer)
Date: Sat, 22 Jan 2005 16:05:26 +0100
Subject: [Bese-devel] clisp and araneida (and windows)
In-Reply-To: <uk6q6bc3j.fsf@tagus.ist.utl.pt> (
	=?iso-8859-1?q?Edgar_Gon=E7alves's_message_of?= "Fri,
	21 Jan 2005 23:23:44 +0000")
References: <usm4vavgz.fsf@tagus.ist.utl.pt> <m2llan6k0u.fsf@soma.local>
	<uoefic0ie.fsf@tagus.ist.utl.pt> <m2acr2vkdx.fsf@soma.local>
	<uk6q6bc3j.fsf@tagus.ist.utl.pt>
Message-ID: <m2zmz1a4i1.fsf@soma.local>

edgar.goncalves at tagus.ist.utl.pt (Edgar Gon?alves) writes:

> So I gess we could have a talk. I was never on #lisp, but I may be there, just
> tell me what's the irc server you're on. Or else, reach me at:
> MSN: edy_in_carpathia [a-t] hotmail [d-o-t] com
> ICQ: 73218277

you, and anyone else, can find me on #lisp at irc.freenode.net under the
nickname segv.

-- 
-Marco
Ring the bells that still can ring.
Forget the perfect offering.
There is a crack in everything.
That's how the light gets in.
	-Leonard Cohen


From edgar.goncalves at tagus.ist.utl.pt  Sun Jan 23 03:55:52 2005
From: edgar.goncalves at tagus.ist.utl.pt (=?iso-8859-1?q?Edgar_Gon=E7alves?=)
Date: Sun, 23 Jan 2005 03:55:52 +0000
Subject: [Bese-devel] clisp and araneida (and windows)
In-Reply-To: <m2zmz1a4i1.fsf@soma.local> (Marco Baringer's message of "Sat,
	22 Jan 2005 16:05:26 +0100")
References: <usm4vavgz.fsf@tagus.ist.utl.pt> <m2llan6k0u.fsf@soma.local>
	<uoefic0ie.fsf@tagus.ist.utl.pt> <m2acr2vkdx.fsf@soma.local>
	<uk6q6bc3j.fsf@tagus.ist.utl.pt> <m2zmz1a4i1.fsf@soma.local>
Message-ID: <umzv0lrxz.fsf@tagus.ist.utl.pt>


Thanks for the help. The patch you made for the iterate package did the trick.

Yet another thing on ucw on windows: I tried using ucw with clisp and araneida
(instead of portable aserve, because I haven't compiled the latest clisp with
support for syscalls, yet), and I noticed that, although the server starts ok,
the examples (and admin) apps don't get loaded to the due paths. I'm pointing to
http://localhost:8080/ucw/examples/index.ucw and the result was the Araneida
sorry page (Was asked for URL http://localhost:8080/ucw/examples/index.ucw , but
it didn't happen for us. Sorry)

Being single threaded, clisp does not allow me to mess around with the server
running, but it should be allright (ie, I can see the hello page if I load the
example from araneida, and the ucw loads without errors...)

Thanks again - I feel I'm close to see a ucw screen on windows!! :)

-- 
Edgar Gon?alves
Software Engineering Group @ INESC-ID
Portugal


From mb at bese.it  Mon Jan 24 18:50:57 2005
From: mb at bese.it (Marco Baringer)
Date: Mon, 24 Jan 2005 19:50:57 +0100
Subject: [Bese-devel] ucw an allegro (windows + aserve)
Message-ID: <m2k6q24q5q.fsf@soma.local>


thanks to the help of Edgar Gon?alves i have commited initial
(compiles and loads) support for allegro. (you'll need the latest ucw
and the latest arnesi).

i remeber someone had asked about allegro support [sorry if i don't
remeber who you are], i don't (and don't plan on having) an allegro
implementation to test this on so if anyone else wants to try ucw and
see what goes wrong that'd be great (and open the road for a 0.3.5
release).

-- 
-Marco
Ring the bells that still can ring.
Forget the perfect offering.
There is a crack in everything.
That's how the light gets in.
	-Leonard Cohen


From mb at bese.it  Wed Jan 26 16:25:18 2005
From: mb at bese.it (Marco Baringer)
Date: Wed, 26 Jan 2005 17:25:18 +0100
Subject: [Bese-devel] ucw an allegro
Message-ID: <m2sm4of98x.fsf@soma.local>


Thanks to Edgar Gon?alves (and much trial and error) UCW now supports
Allegro Common Lisp using the aserve backend (mod_lisp and araneida
should work as well but haven't been tested).

Here's a brief summary of the platforms UCW runs on:

                  MacOS X Windows Linux
CLISP             +       +       ?     [araneida backend only]
OpenMCL           +       -       -     [mod_lisp and araneida backend]
SBCL              +       -       +     [mod_lisp and araneida backend]
CMUCL             +       -       +     [all backends]
ACL               ?       +       ?     [aserve backend]

I would like to point out that, thanks to SLIME, the MOPP package and
the various implementation's adherence to the standard, UCW code will
run _unchanged_ on the various platforms.

not bad heh?
-- 
-Marco
Ring the bells that still can ring.
Forget the perfect offering.
There is a crack in everything.
That's how the light gets in.
	-Leonard Cohen


From ats at acm.org  Fri Jan 28 17:38:32 2005
From: ats at acm.org (Alan Shutko)
Date: Fri, 28 Jan 2005 11:38:32 -0600
Subject: [Bese-devel] Re: clisp and araneida (and windows)
References: <umzv0lrxz.fsf@tagus.ist.utl.pt>
Message-ID: <87d5vp1mjr.fsf@vera.springies.com>


>  I'm pointing to http://localhost:8080/ucw/examples/index.ucw and
> the result was the Araneida sorry page (Was asked for URL
> http://localhost:8080/ucw/examples/index.ucw , but it didn't happen
> for us. Sorry)

I was seeing the same thing on SBCL, and after a bit of tracing, found
that I needed to point to http://127.0.0.1:8080/ucw/examples/index.ucw 

If you want to change that, change 

(defparameter *backend* (make-instance 'ucw:araneida-backend))

to something like

(defparameter *backend* (make-instance 'ucw:araneida-backend :host "localhost"))


-- 
Alan Shutko <ats at acm.org> - I am the rocks.
Anyone stupid enough to be caught by the police is probably guilty.


From mb at bese.it  Fri Jan 28 18:30:30 2005
From: mb at bese.it (Marco Baringer)
Date: Fri, 28 Jan 2005 19:30:30 +0100
Subject: [Bese-devel] Re: clisp and araneida (and windows)
In-Reply-To: <87d5vp1mjr.fsf@vera.springies.com> (Alan Shutko's message of
	"Fri, 28 Jan 2005 11:38:32 -0600")
References: <umzv0lrxz.fsf@tagus.ist.utl.pt>
	<87d5vp1mjr.fsf@vera.springies.com>
Message-ID: <m2wttxjtix.fsf@soma.local>

Alan Shutko <ats at acm.org> writes:

>>  I'm pointing to http://localhost:8080/ucw/examples/index.ucw and
>> the result was the Araneida sorry page (Was asked for URL
>> http://localhost:8080/ucw/examples/index.ucw , but it didn't happen
>> for us. Sorry)
>
> I was seeing the same thing on SBCL, and after a bit of tracing, found
> that I needed to point to http://127.0.0.1:8080/ucw/examples/index.ucw 
>
> If you want to change that, change 
>
> (defparameter *backend* (make-instance 'ucw:araneida-backend))
>
> to something like
>
> (defparameter *backend* (make-instance 'ucw:araneida-backend :host "localhost"))

i don't know much about araneida at all, does anyone know what we can
do to avoid this? (something akin to binding apache to 0.0.0.0)?

-- 
-Marco
Ring the bells that still can ring.
Forget the perfect offering.
There is a crack in everything.
That's how the light gets in.
	-Leonard Cohen


From rm at fabula.de  Fri Jan 28 19:16:04 2005
From: rm at fabula.de (rm at fabula.de)
Date: Fri, 28 Jan 2005 20:16:04 +0100
Subject: [Bese-devel] Re: clisp and araneida (and windows)
In-Reply-To: <m2wttxjtix.fsf@soma.local>
References: <umzv0lrxz.fsf@tagus.ist.utl.pt>
	<87d5vp1mjr.fsf@vera.springies.com> <m2wttxjtix.fsf@soma.local>
Message-ID: <20050128191604.GB20337@seid-online.de>

On Fri, Jan 28, 2005 at 07:30:30PM +0100, Marco Baringer wrote:
> Alan Shutko <ats at acm.org> writes:
> 
> >>  I'm pointing to http://localhost:8080/ucw/examples/index.ucw and
> >> the result was the Araneida sorry page (Was asked for URL
> >> http://localhost:8080/ucw/examples/index.ucw , but it didn't happen
> >> for us. Sorry)
> >
> > I was seeing the same thing on SBCL, and after a bit of tracing, found
> > that I needed to point to http://127.0.0.1:8080/ucw/examples/index.ucw 
> >
> > If you want to change that, change 
> >
> > (defparameter *backend* (make-instance 'ucw:araneida-backend))
> >
> > to something like
> >
> > (defparameter *backend* (make-instance 'ucw:araneida-backend :host "localhost"))
> 
> i don't know much about araneida at all, does anyone know what we can
> do to avoid this? (something akin to binding apache to 0.0.0.0)?

Not really much, iirc. But, thinking of it, this shouldn't be fixed. After
all there _is_ a differnce between the URL 'http://127.0.0.1/' and 'http://localhost/'.
Remember, the net location is put into the 'Host' header by all but the most out-
dated and stupid browsers (HTTP/1.1 requires this). So ther first results in an:

 GET / HTTP/1.1
 Host: 127.0.0.1
 ....

the second in:

 GET / HTTP/1.1
 Host: localhost

While this might look arguably silly it _is_ important in cases where the same
host (same IP address, to be more precise) hosts several domains (NamedVirtualServers
in Apache parlance).

 Cheers   RalfD
> -- 
> -Marco
> Ring the bells that still can ring.
> Forget the perfect offering.
> There is a crack in everything.
> That's how the light gets in.
> 	-Leonard Cohen
> _______________________________________________
> bese-devel mailing list
> bese-devel at common-lisp.net
> http://common-lisp.net/cgi-bin/mailman/listinfo/bese-devel


From ats at acm.org  Fri Jan 28 23:36:32 2005
From: ats at acm.org (Alan Shutko)
Date: Fri, 28 Jan 2005 17:36:32 -0600
Subject: [Bese-devel] Araneida: doesn't register CSS
Message-ID: <87zmytyvlr.fsf@vera.springies.com>

I notice that stylesheets don't work on Araneida.  Trying to retrieve
one gives a .css gives a 404.  I don't think that the stylesheet is
being registered with araneida.

Should UCW be registering external files with the backend server, or
should I be managing that myself?

-- 
Alan Shutko <ats at acm.org> - I am the rocks.
But doc, I eat lots of fruit - 3 cherries in every Hollywood!


From drewc at tech.coop  Sat Jan 29 06:04:16 2005
From: drewc at tech.coop (Drew Crampsie)
Date: Fri, 28 Jan 2005 22:04:16 -0800
Subject: [Bese-devel] Introduction to UCW.
In-Reply-To: <87d5vp1mjr.fsf@vera.springies.com>
References: <umzv0lrxz.fsf@tagus.ist.utl.pt>
	<87d5vp1mjr.fsf@vera.springies.com>
Message-ID: <41FB2760.4000101@tech.coop>

I've put a page up at *hyper-cliki* that introduces continuations and 
UCW. Should serve as a good introduction for newbies, and reduce the 
influx of email i've been getting somewhat :)

Looking for comments and copy-editors. I hope somebody finds it helpful  :)

http://lisp.tech.coop/Web%2FContinuation

drewc


From mb at bese.it  Sat Jan 29 12:24:39 2005
From: mb at bese.it (Marco Baringer)
Date: Sat, 29 Jan 2005 13:24:39 +0100
Subject: [Bese-devel] Araneida: doesn't register CSS
In-Reply-To: <87zmytyvlr.fsf@vera.springies.com> (Alan Shutko's message of
	"Fri, 28 Jan 2005 17:36:32 -0600")
References: <87zmytyvlr.fsf@vera.springies.com>
Message-ID: <m2acqs7794.fsf@soma.local>

Alan Shutko <ats at acm.org> writes:

> I notice that stylesheets don't work on Araneida.  Trying to retrieve
> one gives a .css gives a 404.  I don't think that the stylesheet is
> being registered with araneida.
>
> Should UCW be registering external files with the backend server, or
> should I be managing that myself?

ucw sholud definetely be doing it, in the mean time here's what i have
in another app:

(publish-directory
 (ucw::server.backend ucw:*default-server*)
 (make-pathname :directory `(:absolute ,@(cdr (pathname-directory *load-truename*)) :up "wwwroot"))
 "/")

you may need to play with the loacl path and the published path to
make sure that http://host/ucw/examples/stylesheet.css maps to
ucw/examples/tal/stylesheet.css

publish-directory is part of the backend protocol (but is only
implemented for araneida), server.backend should probably be exported
but isn't at the moment.

-- 
-Marco
Ring the bells that still can ring.
Forget the perfect offering.
There is a crack in everything.
That's how the light gets in.
	-Leonard Cohen


From mb at bese.it  Sun Jan 30 11:14:21 2005
From: mb at bese.it (Marco Baringer)
Date: Sun, 30 Jan 2005 12:14:21 +0100
Subject: [Bese-devel] my GPG key
Message-ID: <m2wttvb242.fsf@soma.local>


as some of you my have noticed recent ucw patches have been signed
with a new GPG key (id 6450AA06). on new year's day my hard disk
crashed and took my secret gpg key with it... 

attached is my new key, hopefully this will be the _last_ time i have
to mess with gpg keys for a while. this key is not signed by any keys
other than itself (kinda hard to sign your new key when you've lost
the old one).

i believe, but i'm not sure, that this key is available from
common-lisp.net's keyring.asc.

-----BEGIN PGP PUBLIC KEY BLOCK-----
Version: GnuPG v1.4.0 (Darwin)
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=RY2A
-----END PGP PUBLIC KEY BLOCK-----

-- 
-Marco
Ring the bells that still can ring.
Forget the perfect offering.
There is a crack in everything.
That's how the light gets in.
	-Leonard Cohen


From ats at acm.org  Mon Jan 31 18:44:38 2005
From: ats at acm.org (Alan Shutko)
Date: Mon, 31 Jan 2005 12:44:38 -0600
Subject: [Bese-devel] Araneida: doesn't register CSS
In-Reply-To: <m2acqs7794.fsf@soma.local> (Marco Baringer's message of "Sat, 29
	Jan 2005 13:24:39 +0100")
References: <87zmytyvlr.fsf@vera.springies.com> <m2acqs7794.fsf@soma.local>
Message-ID: <87mzupa161.fsf@vera.springies.com>

"Marco Baringer" <mb at bese.it> writes:

> you may need to play with the loacl path and the published path to
> make sure that http://host/ucw/examples/stylesheet.css maps to
> ucw/examples/tal/stylesheet.css

Thanks, that's working for me.  I'd like to give a shot at making
defapplication do this, with a :www-roots option similar to
:tal-roots, except that it would just publish the directories rather
than making a generator.  Is that the right place to put it?

-- 
Alan Shutko <ats at acm.org> - I am the rocks.
If it ain't broke...give it to a liberal to fix.