From achambers.home at googlemail.com  Fri Aug 14 13:52:26 2009
From: achambers.home at googlemail.com (Andy Chambers)
Date: Fri, 14 Aug 2009 14:52:26 +0100
Subject: [cells-devel] bug in celtk
Message-ID: <a70cc98c0908140652n6171bf0eq772e10b9af5c9634@mail.gmail.com>

Hey All,

There seems to be a strange bug in celtk.  I'm looking into it but
wondered if any of you guys had any clues.

Celtk seems to be delaying the evaluation of an "on-command" procedure
until the next "on-command" event
is generated.  You can see this in the example ltktest-ci.  If you
start that, and click the "Hallo" button, then
the "Wald" button, then the "Hallo" button, you'll see what I'm talking about.

I've made a few changes to my local copy so it might just be me but it
would be interesting to find out if
anyone else can confirm this behaviour.

Cheers,
Andy



From kentilton at gmail.com  Fri Aug 14 14:07:54 2009
From: kentilton at gmail.com (Kenneth Tilton)
Date: Fri, 14 Aug 2009 10:07:54 -0400
Subject: [cells-devel] bug in celtk
In-Reply-To: <a70cc98c0908140652n6171bf0eq772e10b9af5c9634@mail.gmail.com>
References: <a70cc98c0908140652n6171bf0eq772e10b9af5c9634@mail.gmail.com>
Message-ID: <4A856FBA.80109@gmail.com>

I do not have that code over here. Maybe FG committed something. Maybe 
if you can describe it better I can figure out what must be going on. I 
am not aware of any problems at the core level, but then again I do not 
have that code so maybe something new is going on.

kt

Andy Chambers wrote:
> Hey All,
> 
> There seems to be a strange bug in celtk.  I'm looking into it but
> wondered if any of you guys had any clues.
> 
> Celtk seems to be delaying the evaluation of an "on-command" procedure
> until the next "on-command" event
> is generated.  You can see this in the example ltktest-ci.  If you
> start that, and click the "Hallo" button, then
> the "Wald" button, then the "Hallo" button, you'll see what I'm talking about.
> 
> I've made a few changes to my local copy so it might just be me but it
> would be interesting to find out if
> anyone else can confirm this behaviour.
> 
> Cheers,
> Andy
> 
> _______________________________________________
> cells-devel site list
> cells-devel at common-lisp.net
> http://common-lisp.net/mailman/listinfo/cells-devel
> 
> 
> ------------------------------------------------------------------------
> 
> 
> No virus found in this incoming message.
> Checked by AVG - www.avg.com 
> Version: 8.5.392 / Virus Database: 270.13.54/2300 - Release Date: 08/13/09 06:11:00
> 

-- 

http://thelaughingstockatpngs.com/
http://www.facebook.com/pages/The-Laughingstock/115923141782?ref=nf



From achambers.home at googlemail.com  Fri Aug 14 15:12:26 2009
From: achambers.home at googlemail.com (Andy Chambers)
Date: Fri, 14 Aug 2009 16:12:26 +0100
Subject: [cells-devel] bug in celtk
In-Reply-To: <4A857ABC.8090209@gmail.com>
References: <a70cc98c0908140652n6171bf0eq772e10b9af5c9634@mail.gmail.com>
	<4A856FBA.80109@gmail.com>
	<a70cc98c0908140730t65a2e52fk9a0d99b21e0b937b@mail.gmail.com>
	<4A857ABC.8090209@gmail.com>
Message-ID: <a70cc98c0908140812w5920c645x1c6a2a036310bf7b@mail.gmail.com>

On Fri, Aug 14, 2009 at 3:54 PM, Kenneth Tilton<kentilton at gmail.com> wrote:

> If you now click some place at random do you see "hitting button C"? You
> might have to create a checkbox or something else that is enabled and
> handles mouse events.

No but it does eventually catch up when you close the app down.  I added a tcl
trace to the "do-on-command" procedure and tcl does indeed seem to be calling
that procedure at the correct time.

I have a theory that altering defcommand to make it use a command created by
TclCreateObjCommand (as opposed to TclCreateCommand) might make it
work.  I'll send a patch if it does.

Cheers,
Andy





-- 
----
Andy Chambers
Formedix Ltd



From frgo at me.com  Fri Aug 14 14:56:23 2009
From: frgo at me.com (Frank Goenninger)
Date: Fri, 14 Aug 2009 16:56:23 +0200
Subject: [cells-devel] bug in celtk
In-Reply-To: <4A856FBA.80109@gmail.com>
References: <a70cc98c0908140652n6171bf0eq772e10b9af5c9634@mail.gmail.com>
	<4A856FBA.80109@gmail.com>
Message-ID: <0E19AC1A-DC92-4B8B-985C-DEB9229DD845@me.com>

Hi Kenny, hi Andy,

Am 14.08.2009 um 16:07 schrieb Kenneth Tilton:

> I do not have that code over here. Maybe FG committed something.

I did. But that just extended the number of Mouse Button Down handlers  
(on-mouse-...). Also I do not see this behaviour in my Celtk copy. I  
always get on-command fired immediately.

> Maybe
> if you can describe it better I can figure out what must be going  
> on. I
> am not aware of any problems at the core level, but then again I do  
> not
> have that code so maybe something new is going on.

May well be. Do you have a small example, Andy ?

Cheers
    Frank




From achambers.home at googlemail.com  Fri Aug 14 15:38:07 2009
From: achambers.home at googlemail.com (Andy Chambers)
Date: Fri, 14 Aug 2009 16:38:07 +0100
Subject: [cells-devel] bug in celtk
In-Reply-To: <0E19AC1A-DC92-4B8B-985C-DEB9229DD845@me.com>
References: <a70cc98c0908140652n6171bf0eq772e10b9af5c9634@mail.gmail.com>
	<4A856FBA.80109@gmail.com>
	<0E19AC1A-DC92-4B8B-985C-DEB9229DD845@me.com>
Message-ID: <a70cc98c0908140838s7636c88bgc75abf49a213d067@mail.gmail.com>

On Fri, Aug 14, 2009 at 3:56 PM, Frank Goenninger<frgo at me.com> wrote:
> Hi Kenny, hi Andy,
>
> Am 14.08.2009 um 16:07 schrieb Kenneth Tilton:
>
>> I do not have that code over here. Maybe FG committed something.
>
> I did. But that just extended the number of Mouse Button Down handlers
> (on-mouse-...). Also I do not see this behaviour in my Celtk copy. I always
> get on-command fired immediately.

What version of tcl/tk are you using?  If you're on 8.4, this could confirm my
theory.

My own copy has diverged somewhat from that in CVS so it's equally likely this
is a problem of my own making :-)

I sent an example earlier but forgot to CC the list.  I'll re-attach it here...

(defmd test-buttons (window)
 (title$ "My Buttons")
 (kids (c? (the-kids
            (mk-stack (:packing (c?pack-self))
              (mk-stack ()
                (loop for button-id in '(a b c)
                   collect
                   (let ((id button-id))
                     (mk-button-ex ((symbol-name id) (format t
"hitting button: ~a" id)))))))))))

If you click "A", then "B", then "C", the console will print...

;;nothing when you click "A"
"hitting button: A" ;;when you click "B"
"hitting button: B" ;;when you click "C"



From achambers.home at googlemail.com  Fri Aug 14 15:45:48 2009
From: achambers.home at googlemail.com (Andy Chambers)
Date: Fri, 14 Aug 2009 16:45:48 +0100
Subject: [cells-devel] bug in celtk
In-Reply-To: <4A8580A5.2090701@gmail.com>
References: <a70cc98c0908140652n6171bf0eq772e10b9af5c9634@mail.gmail.com>
	<4A856FBA.80109@gmail.com>
	<a70cc98c0908140730t65a2e52fk9a0d99b21e0b937b@mail.gmail.com>
	<4A857ABC.8090209@gmail.com>
	<a70cc98c0908140812w5920c645x1c6a2a036310bf7b@mail.gmail.com>
	<4A8580A5.2090701@gmail.com>
Message-ID: <a70cc98c0908140845y4a9f3c89j708982bcf43135bc@mail.gmail.com>

On Fri, Aug 14, 2009 at 4:20 PM, Kenneth Tilton<kentilton at gmail.com> wrote:

> That would just leave the problem extant, by finding some other way to
> dispatch commands. I may DL and try myself later. Did you grab from CVS
> recently? Or, better: zip up your copy and send me that so we are looking at
> the same code for sure.

Yeah but I think that may be the cause of the problem.  Consider this
bit from the
Tcl docs

> Tcl_CreateCommand defines a new command in interp and associates it with procedure proc such that whenever cmdName is invoked as a Tcl command __(via a call to Tcl_Eval)__ the Tcl interpreter will call proc to process the command.

vs

> Tcl_CreateObjCommand defines a new command in interp and associates it with procedure proc such that whenever name is invoked as a Tcl command (e.g., via a call to Tcl_EvalObjEx) the Tcl interpreter will call proc to process the command.

How is Tcl going to know to invoke do-on-command using Tcl_Eval rather
than Tcl_EvalObjEx?

You can grab a zip of my current working directory at the link below.

http://github.com/cddr/celtk/zipball/master

Cheers,
Andy



From frgo at me.com  Fri Aug 14 15:50:01 2009
From: frgo at me.com (Frank Goenninger)
Date: Fri, 14 Aug 2009 17:50:01 +0200
Subject: [cells-devel] bug in celtk
In-Reply-To: <a70cc98c0908140838s7636c88bgc75abf49a213d067@mail.gmail.com>
References: <a70cc98c0908140652n6171bf0eq772e10b9af5c9634@mail.gmail.com>
	<4A856FBA.80109@gmail.com>
	<0E19AC1A-DC92-4B8B-985C-DEB9229DD845@me.com>
	<a70cc98c0908140838s7636c88bgc75abf49a213d067@mail.gmail.com>
Message-ID: <D78AF593-9CED-4288-ADCB-882E18E2C4A0@me.com>

Am 14.08.2009 um 17:38 schrieb Andy Chambers:

> On Fri, Aug 14, 2009 at 3:56 PM, Frank Goenninger<frgo at me.com> wrote:
>> Hi Kenny, hi Andy,
>>
>> Am 14.08.2009 um 16:07 schrieb Kenneth Tilton:
>>
>>> I do not have that code over here. Maybe FG committed something.
>>
>> I did. But that just extended the number of Mouse Button Down  
>> handlers
>> (on-mouse-...). Also I do not see this behaviour in my Celtk copy.  
>> I always
>> get on-command fired immediately.
>
> What version of tcl/tk are you using?  If you're on 8.4, this could  
> confirm my
> theory.


Mac OS X 10.5.8 with Tcl/Tk 8.5.6 (as downloaded from ActiveState).


> My own copy has diverged somewhat from that in CVS so it's equally  
> likely this
> is a problem of my own making :-)
>
> I sent an example earlier but forgot to CC the list.  I'll re-attach  
> it here...
>
> (defmd test-buttons (window)
> (title$ "My Buttons")
> (kids (c? (the-kids
>            (mk-stack (:packing (c?pack-self))
>              (mk-stack ()
>                (loop for button-id in '(a b c)
>                   collect
>                   (let ((id button-id))
>                     (mk-button-ex ((symbol-name id) (format t
> "hitting button: ~a" id)))))))))))
>
> If you click "A", then "B", then "C", the console will print...
>
> ;;nothing when you click "A"
> "hitting button: A" ;;when you click "B"
> "hitting button: B" ;;when you click "C"


Using your code above I see the following here:

Click on "A" ->
0> defcommand >  ^on-command b5 nilhitting button: A

This comes before any other button is pressed ...

So, again, sorry to say: I can't duplicate it here.

Oh - btw - I discovered a "bug" in your on-hover patch: The hover- 
timer slot in tk.object is defined as ":reader hover-timer" but should  
be :accessor ... I'm sure you have it ok in your code base...

Cheers
    Frank




> _______________________________________________
> cells-devel site list
> cells-devel at common-lisp.net
> http://common-lisp.net/mailman/listinfo/cells-devel

--
   Frank Goenninger

   Cell: 	+49 175 4321058
   E-Mail: 	frgo at me.com








From achambers.home at googlemail.com  Fri Aug 14 15:56:29 2009
From: achambers.home at googlemail.com (Andy Chambers)
Date: Fri, 14 Aug 2009 16:56:29 +0100
Subject: [cells-devel] bug in celtk
In-Reply-To: <D78AF593-9CED-4288-ADCB-882E18E2C4A0@me.com>
References: <a70cc98c0908140652n6171bf0eq772e10b9af5c9634@mail.gmail.com>
	<4A856FBA.80109@gmail.com>
	<0E19AC1A-DC92-4B8B-985C-DEB9229DD845@me.com>
	<a70cc98c0908140838s7636c88bgc75abf49a213d067@mail.gmail.com>
	<D78AF593-9CED-4288-ADCB-882E18E2C4A0@me.com>
Message-ID: <a70cc98c0908140856t68cfd4ep35041f0620c9b2cf@mail.gmail.com>

On Fri, Aug 14, 2009 at 4:50 PM, Frank Goenninger<frgo at me.com> wrote:

> Using your code above I see the following here:
>
> Click on "A" ->
> 0> defcommand > ?^on-command b5 nilhitting button: A

OK Thanks for checking that out Frank.  I must have something else
wrong over here.

> This comes before any other button is pressed ...
>
> So, again, sorry to say: I can't duplicate it here.
>
> Oh - btw - I discovered a "bug" in your on-hover patch: The hover-timer slot
> in tk.object is defined as ":reader hover-timer" but should be :accessor ...
> I'm sure you have it ok in your code base...

Hadn't come across that as being a problem but it makes sense.  Thanks

Cheers,
Andy



From achambers.home at googlemail.com  Sat Aug 15 07:40:12 2009
From: achambers.home at googlemail.com (Andy Chambers)
Date: Sat, 15 Aug 2009 08:40:12 +0100
Subject: [cells-devel] bug in celtk
In-Reply-To: <7DA143FD-3462-47A0-95E6-A8F2BDA1BBD7@consequor.de>
References: <a70cc98c0908140652n6171bf0eq772e10b9af5c9634@mail.gmail.com>
	<4A856FBA.80109@gmail.com>
	<0E19AC1A-DC92-4B8B-985C-DEB9229DD845@me.com>
	<a70cc98c0908140838s7636c88bgc75abf49a213d067@mail.gmail.com>
	<D78AF593-9CED-4288-ADCB-882E18E2C4A0@me.com>
	<a70cc98c0908140856t68cfd4ep35041f0620c9b2cf@mail.gmail.com>
	<7DA143FD-3462-47A0-95E6-A8F2BDA1BBD7@consequor.de>
Message-ID: <a70cc98c0908150040k303ce93dq9c833c26d7e93a8e@mail.gmail.com>

On Fri, Aug 14, 2009 at 5:21 PM, Frank Goenninger - Consequor
Consulting AG<frank.goenninger at consequor.de> wrote:
> Am 14.08.2009 um 17:56 schrieb Andy Chambers:
>
>> OK Thanks for checking that out Frank. ?I must have something else
>> wrong over here.

I think the problem was actually with my Lisp.  I changed the test to
have the button click reflected in the gui
and it happened immediately.  My new theory is that because I'm using
a single threaded lisp, it doesn't get
round to executing the print statement until the next set of events
are being processed.

If I just just setf the new value into a cell, it all seems to work as expected.

--
Andy