From matthias.berth at googlemail.com  Wed Aug  1 19:47:48 2007
From: matthias.berth at googlemail.com (Matthias Berth)
Date: Wed, 1 Aug 2007 19:47:48 +0200
Subject: WARenderLoopConfiguration vs WARenderConfiguration
Message-ID: <9aa6ae270708011047t3bb500ddq19c1655d6dba75a6@mail.gmail.com>

Hello,


I was updating my Pier installation (using Pier-Seaside-lr.182 and
others), it stumbled at

PRPieronfiguration>>ancestors
       ^ Array with: WARenderConfiguration new

not knowing the class WARenderConfiguration. Replacing this with
WARenderLoopConfiguration works OK so far.

I am running seaside 2.7b1. Looks like the ancestors method comes from
Pier-Seaside-lr.182. Maybe WARenderLoopConfiguration got renamed to
WARenderConfiguration in Seaside 2.8?

Is it OK to use WARenderLoopConfiguration here? If not, where can I
get WARenderConfiguration ?


best regards

Matthias

From renggli at iam.unibe.ch  Wed Aug  1 20:42:50 2007
From: renggli at iam.unibe.ch (Lukas Renggli)
Date: Wed, 1 Aug 2007 20:42:50 +0200
Subject: WARenderLoopConfiguration vs WARenderConfiguration
In-Reply-To: <9aa6ae270708011047t3bb500ddq19c1655d6dba75a6@mail.gmail.com>
References: <9aa6ae270708011047t3bb500ddq19c1655d6dba75a6@mail.gmail.com>
Message-ID: <FC4B8965-F02E-4A9B-84FD-909528A1AD69@iam.unibe.ch>

Thank you for reporting this issue. This renaming in Seaside 2.8 was  
not intentional and I reverted it to the old name. The latest version  
of Pier should be compatible to Seaside 2.7 again.

Name: Pier-All-lr.221
Author: lr
Time: 1 August 2007, 8:38:28 pm
UUID: 491ee006-70b0-47ed-9a95-841f5ca60a20
Ancestors: Pier-All-lr.220
Dependencies: Pier-Model-lr.156, Pier-Tests-lr.72, Pier-Seaside-lr. 
183, Pier-OmniBrowser-lr.25, Pier-Security-lr.82, Pier-Blog-lr.56

- renamed WARenderConfiguration back to WARenderLoopConfiguration to  
make Pier compatible with Seaside 2.7
- requires to re-register the pier application


On 01.08.2007, at 19:47, Matthias Berth wrote:

> Hello,
>
>
> I was updating my Pier installation (using Pier-Seaside-lr.182 and
> others), it stumbled at
>
> PRPieronfiguration>>ancestors
>        ^ Array with: WARenderConfiguration new
>
> not knowing the class WARenderConfiguration. Replacing this with
> WARenderLoopConfiguration works OK so far.
>
> I am running seaside 2.7b1. Looks like the ancestors method comes from
> Pier-Seaside-lr.182. Maybe WARenderLoopConfiguration got renamed to
> WARenderConfiguration in Seaside 2.8?
>
> Is it OK to use WARenderLoopConfiguration here? If not, where can I
> get WARenderConfiguration ?
>
>
> best regards
>
> Matthias
>
> _______________________________________________
> SmallWiki, Magritte, Pier and Related Tools ...
> https://www.iam.unibe.ch/mailman/listinfo/smallwiki

-- 
Lukas Renggli
http://www.lukas-renggli.ch

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 3774 bytes
Desc: not available
Url : http://www.iam.unibe.ch/pipermail/smallwiki/attachments/20070801/e0652a99/smime.bin

From matthias.berth at googlemail.com  Thu Aug  2 12:40:40 2007
From: matthias.berth at googlemail.com (Matthias Berth)
Date: Thu, 2 Aug 2007 12:40:40 +0200
Subject: WARenderLoopConfiguration vs WARenderConfiguration
In-Reply-To: <FC4B8965-F02E-4A9B-84FD-909528A1AD69@iam.unibe.ch>
References: <9aa6ae270708011047t3bb500ddq19c1655d6dba75a6@mail.gmail.com>
	<FC4B8965-F02E-4A9B-84FD-909528A1AD69@iam.unibe.ch>
Message-ID: <9aa6ae270708020340k9548b95u55f5dfd0afe06f56@mail.gmail.com>

Thanks, works without problems now :-)

Matthias

On 8/1/07, Lukas Renggli <renggli at iam.unibe.ch> wrote:
> Thank you for reporting this issue. This renaming in Seaside 2.8 was
> not intentional and I reverted it to the old name. The latest version
> of Pier should be compatible to Seaside 2.7 again.

From keith_hodges at yahoo.co.uk  Thu Aug  2 20:02:46 2007
From: keith_hodges at yahoo.co.uk (Keith Hodges)
Date: Thu, 02 Aug 2007 19:02:46 +0100
Subject: Repository down?
Message-ID: <46B21C46.4040900@yahoo.co.uk>

cheers

Keith

From renggli at iam.unibe.ch  Thu Aug  2 20:08:41 2007
From: renggli at iam.unibe.ch (Lukas Renggli)
Date: Thu, 2 Aug 2007 20:08:41 +0200
Subject: Repository down?
In-Reply-To: <46B21C46.4040900@yahoo.co.uk>
References: <46B21C46.4040900@yahoo.co.uk>
Message-ID: <63FE9CE8-0628-4153-9AC6-0714ED9C5CA1@iam.unibe.ch>

Thanks for reporting. Unfortunately Apache didn't want to restart  
today due to a GCC update on my Gentoo Linux. Some libraries are now  
recompiling ...

Sorry for the inconvenience.

Lukas

-- 
Lukas Renggli
http://www.lukas-renggli.ch


From renggli at iam.unibe.ch  Thu Aug  2 22:28:39 2007
From: renggli at iam.unibe.ch (Lukas Renggli)
Date: Thu, 2 Aug 2007 22:28:39 +0200
Subject: [Fwd: Help needed to run widgets in other contexts]
In-Reply-To: <46B23C74.6030000@yahoo.co.uk>
References: <46B23C74.6030000@yahoo.co.uk>
Message-ID: <82071D44-2E1C-4EBF-A54E-05443DC4D52B@iam.unibe.ch>

Yes, the code I gave you at that time won't work in Seaside 2.8.  
WARenderLoop has been completely rewritten to avoid the need to store  
many unnecessary continuations. This greatly reduces the memory  
requirements per session.

What I suggest to do is that you implement a decoration that you can  
add around your component. Inside this decoration you store the  
context-holder and wrap the rendering and callback processing into

	PRCurrentContext use: contextHolder during: [ ... ]

HTH,
Lukas

On 02.08.2007, at 22:20, Keith Hodges wrote:

> Hello Lukas,
>
> a while ago you gave me this code. Now I am using the latest  
> seaisde 2.8 it doesnt work as I would like. The widgit works but  
> its callbacks fail. Any thoughts on what may have changed?
>
> cheers
>
> Keith
>
>
> From: Keith Hodges <keith_hodges at yahoo.co.uk>
> Date: 27. Juli 2007 04:30:25 GMT+02:00
> To: "small >> \"SmallWiki, Magritte, Pier and Related Tools ...\""  
> <smallwiki at iam.unibe.ch>
> Subject: Help needed to run widgets in other contexts
> Reply-To: "Magritte, Pier and Related Tools ..."  
> <smallwiki at iam.unibe.ch>
>
>
> Help!
>
> Hi All,
>
> A while ago Lukas gave me the following code for opening a pier widget
> so that it would be able to access its context even though we are not
> even in a pier application!
>
> self - is my control panel which runs in the WADispatcherEditor's  
> session.
> widget is PUUsers or PUGroups
>
> ControlPanel-callWidget: widget  titled: title
>
>     PRCurrentContext use: self during: [
>         WARenderLoop new call: (widget
>             addDecoration: (WAWindowDecoration new
>                 title: title;
>                 yourself);
>             yourself) ]
>
>
> This works to the extent that the widget displays correctly,  
> however the
> buttons do not work due to PRCurrentContext value being nil.
>
> I am using the latest everything in the development universes.
>
> Does anyone have any ideas?
>
> thanks in advance
>
> Keith
>
>
> _______________________________________________
> SmallWiki, Magritte, Pier and Related Tools ...
> https://www.iam.unibe.ch/mailman/listinfo/smallwiki
>
>
>

-- 
Lukas Renggli
http://www.lukas-renggli.ch


From renggli at iam.unibe.ch  Thu Aug  2 22:32:12 2007
From: renggli at iam.unibe.ch (Lukas Renggli)
Date: Thu, 2 Aug 2007 22:32:12 +0200
Subject: Repository down?
In-Reply-To: <63FE9CE8-0628-4153-9AC6-0714ED9C5CA1@iam.unibe.ch>
References: <46B21C46.4040900@yahoo.co.uk>
	<63FE9CE8-0628-4153-9AC6-0714ED9C5CA1@iam.unibe.ch>
Message-ID: <14B2B41E-AA96-473E-B1C5-F1E3DE5BDE53@iam.unibe.ch>

Ok, Apache is finally up and running again. The kernel will also need  
to be rebuilt and restarted, but that will hopefully only cause a  
short outage later tonight. If only Linux was written in Smalltalk ;-)

Thanks for your patience.

Lukas


On 02.08.2007, at 20:08, Lukas Renggli wrote:

> Thanks for reporting. Unfortunately Apache didn't want to restart
> today due to a GCC update on my Gentoo Linux. Some libraries are now
> recompiling ...
>
> Sorry for the inconvenience.
>
> Lukas
>
> -- 
> Lukas Renggli
> http://www.lukas-renggli.ch
>
>
> _______________________________________________
> SmallWiki, Magritte, Pier and Related Tools ...
> https://www.iam.unibe.ch/mailman/listinfo/smallwiki

-- 
Lukas Renggli
http://www.lukas-renggli.ch


From keith_hodges at yahoo.co.uk  Fri Aug  3 03:00:46 2007
From: keith_hodges at yahoo.co.uk (Keith Hodges)
Date: Fri, 03 Aug 2007 02:00:46 +0100
Subject: [bug]
Message-ID: <46B27E3E.2070802@yahoo.co.uk>

I found that pasting a 'long dash' character into a pier page results in 
the contents being rendered like so

?

<???????x???m???l??? 
???v???e???r???s???i???o???n???=???"???1???.???0???"??? 
???e???n???c???o???d???i???n???g???=???"???U???T???F???-???8???"???????>???<???!???D???O???C???T???Y???P???E??? 
???h???t???m???l??? ???P???U???B???L???I???C??? 
???"???-???/???/???W???3???C???/???/???D???T???D??? ???X???H???T???M???L???

I am using the standard WAKom.

best regards

Keith

==
fyi: using pier-beta in 3.10 dev image

From philippe.marschall at gmail.com  Fri Aug  3 08:28:53 2007
From: philippe.marschall at gmail.com (Philippe Marschall)
Date: Fri, 3 Aug 2007 08:28:53 +0200
Subject: [bug]
In-Reply-To: <46B27E3E.2070802@yahoo.co.uk>
References: <46B27E3E.2070802@yahoo.co.uk>
Message-ID: <66666f210708022328g1262dc79gbab4a27ce8970ade@mail.gmail.com>

2007/8/3, Keith Hodges <keith_hodges at yahoo.co.uk>:
> I found that pasting a 'long dash' character into a pier page results in
> the contents being rendered like so
>
> ?
>
> <???????x???m???l???
> ???v???e???r???s???i???o???n???=???"???1???.???0???"???
> ???e???n???c???o???d???i???n???g???=???"???U???T???F???-???8???"???????>???<???!???D???O???C???T???Y???P???E???
> ???h???t???m???l??? ???P???U???B???L???I???C???
> ???"???-???/???/???W???3???C???/???/???D???T???D??? ???X???H???T???M???L???
>
> I am using the standard WAKom.
>
> best regards
>
> Keith
>
> ==
> fyi: using pier-beta in 3.10 dev image

Are yo using Kom from SqueakMap? If you use Kom from SqueakMap then
your only option for Squeak 3.9+ is WAKomEncoded39.

Cheers
Philippe


From renggli at iam.unibe.ch  Fri Aug  3 08:42:20 2007
From: renggli at iam.unibe.ch (Lukas Renggli)
Date: Fri, 3 Aug 2007 08:42:20 +0200
Subject: [bug]
In-Reply-To: <66666f210708022328g1262dc79gbab4a27ce8970ade@mail.gmail.com>
References: <46B27E3E.2070802@yahoo.co.uk>
	<66666f210708022328g1262dc79gbab4a27ce8970ade@mail.gmail.com>
Message-ID: <623A30CA-3474-4E5D-8029-F26E722E7BC0@iam.unibe.ch>

Pier from SqueakMap comes with a patched version of Kom. With this  
configuration I cannot reproduce your bug. I can even past any of  
these characters into a page, display and re-edit:

	http://jshin.net/i18n/korean/hunmin.html

I assume that you are using WAKom and an unpatched version of Kom. In  
this case you have to use WAKomEncoded39 or the patched version of Kom.

Pier doesn't do any encoding conversion. It just stores and displays  
what it gets from your client/server/seaside instance.

Lukas

On 03.08.2007, at 08:28, Philippe Marschall wrote:

> 2007/8/3, Keith Hodges <keith_hodges at yahoo.co.uk>:
>> I found that pasting a 'long dash' character into a pier page  
>> results in
>> the contents being rendered like so
>>
>> ?
>>
>> <???????x???m???l???
>> ???v???e???r???s???i???o??? 
>> n???=???"???1???.???0???"?? 
>> ?
>> ???e???n???c???o???d???i??? 
>> n???g???=???"???U???T???F?? 
>> ?-???8???"???????>???<???! 
>> ???D???O???C???T???Y???P??? 
>> E???
>> ???h???t???m???l???  
>> ???P???U???B???L???I???C???
>> ???"???-???/???/ 
>> ???W???3???C???/???/ 
>> ???D???T???D???  
>> ???X???H???T???M???L???
>>
>> I am using the standard WAKom.
>>
>> best regards
>>
>> Keith
>>
>> ==
>> fyi: using pier-beta in 3.10 dev image
>
> Are yo using Kom from SqueakMap? If you use Kom from SqueakMap then
> your only option for Squeak 3.9+ is WAKomEncoded39.
>
> Cheers
> Philippe
>
>
> _______________________________________________
> SmallWiki, Magritte, Pier and Related Tools ...
> https://www.iam.unibe.ch/mailman/listinfo/smallwiki
>

-- 
Lukas Renggli
http://www.lukas-renggli.ch



From keith_hodges at yahoo.co.uk  Fri Aug  3 08:57:18 2007
From: keith_hodges at yahoo.co.uk (Keith Hodges)
Date: Fri, 03 Aug 2007 07:57:18 +0100
Subject: [bug]
In-Reply-To: <66666f210708022328g1262dc79gbab4a27ce8970ade@mail.gmail.com>
References: <46B27E3E.2070802@yahoo.co.uk>
	<66666f210708022328g1262dc79gbab4a27ce8970ade@mail.gmail.com>
Message-ID: <46B2D1CE.1090205@yahoo.co.uk>


> Are yo using Kom from SqueakMap? If you use Kom from SqueakMap then
> your only option for Squeak 3.9+ is WAKomEncoded39.
>
> Cheers
> Philippe
>
>   
thanks for the advice.

I am using whatever version of Kom is loaded when you load seaside-beta 
from the 'development' universe. Is this the patched one?

Keith

From norbert at hartl.name  Fri Aug  3 12:07:10 2007
From: norbert at hartl.name (Norbert Hartl)
Date: Fri, 03 Aug 2007 12:07:10 +0200
Subject: Pier and Albatross versions
Message-ID: <1186135630.7442.3.camel@localhost>

Hi,

I'm trying to see the albatross tests running. These
require pier to run. I installed pier from universes
(production and beta) but had no success using the
newest version from squeaksource.

In pier-beta I get

Internal Error
MessageNotUnderstood: UndefinedObject>>default

[] in WAApplication>>updateRoot: {[:each | each default updateRoot:
anHtmlRoot]}

Which versions should I use?

thanks,

Norbert


From norbert at hartl.name  Fri Aug  3 12:10:35 2007
From: norbert at hartl.name (Norbert Hartl)
Date: Fri, 03 Aug 2007 12:10:35 +0200
Subject: Pier and Albatross versions
In-Reply-To: <1186135630.7442.3.camel@localhost>
References: <1186135630.7442.3.camel@localhost>
Message-ID: <1186135835.7442.6.camel@localhost>


On Fri, 2007-08-03 at 12:07 +0200, Norbert Hartl wrote:
> Hi,
> 
> I'm trying to see the albatross tests running. These
> require pier to run. I installed pier from universes
> (production and beta) but had no success using the
> newest version from squeaksource.
> 
> In pier-beta I get
> 
> Internal Error
> MessageNotUnderstood: UndefinedObject>>default
> 
> [] in WAApplication>>updateRoot: {[:each | each default updateRoot:
> anHtmlRoot]}
> 
Sorry, too less information. I get this error while
accessing http://localhost:8080/seaside/albatross

Running the test APierTest fails while there is no response
from the browser. Is there a browser opened while running
the tests?

Norbert


From matthias.berth at googlemail.com  Fri Aug  3 13:59:16 2007
From: matthias.berth at googlemail.com (Matthias Berth)
Date: Fri, 3 Aug 2007 13:59:16 +0200
Subject: Pier: Invoking a PRCommand from outside, on-the-fly-login,
	redirect to users page
Message-ID: <9aa6ae270708030459g25e78cb7wb4cec21c6fc8049@mail.gmail.com>

Hi,

I have tried various ways to accomplish this, so far with limited
success, any pointers would be appreciated :-)

Here is my scenario:

- a user fills in a web form on a non-Pier web site

- the form is submitted via HTTP post to a URL on the Pier site, like so:

 http://the.piersite.net/seaside/pier?command=DoSomething

- DoSomething is a PRCommand that requires a login

- the user logs in

- Pier shows a confirmation form that summarizes the data from the original form

- user clicks OK

- DoSomething is executed

- Pier shows the home page of the user
 http://the.piersite.net/seaside/pier/accounts/joeuser?command=PRViewCommand


I can have DoSomething execute successfully, but I cannot make the
change to the users home page.

Maybe my approach is wrong:

- I use Seaside 2.7b with the latest Pier, Magritte etc

- I use Pier Unix security

- DoSomething is a PRCommand that implements isValidIn: aContext on
the class side as aContext isLoggedIn. This, together with configuring
PUPierMain as the main class will invoke the on-the-fly login.

- I have a made a patch to PRPierMain>>commandFrom: such that the
newly created PRCCommand will have a chance to read additional data
from the external HTTP Request (inside DoSomething>>processRequest:)

- DoSomething answers a custom subclass of WAComponent (say,
ConfirmView) in its asComponent method

- the ConfirmView knows the DoSomething command so it can display the
users data, with a button "Confirm". Clicking the Confirm button will
send the message confirm to the view

ConfirmView>>confirm
	command user: self context user..
	command execute

- I can see that the command executes successfully. Inside
DoSomething>>execute I have
	self
		answer: (self context
                              structure: self user account "account is
a PRStructure"
                              view: UserAccountView)

- I have inspected the command at this point, and answer gets the
right contents.

- However, when I click on the button, the Confirm view just stays there.

Do you see any chance to get this working? I have tried to look at
Pier Blog and others, but they all seem to use #answer: without
problems.

I'd be thankful for any hints, or alternative ways to achieve the page
flow I described.

Cheers

Matthias

From renggli at iam.unibe.ch  Fri Aug  3 14:26:45 2007
From: renggli at iam.unibe.ch (Lukas Renggli)
Date: Fri, 3 Aug 2007 14:26:45 +0200
Subject: Pier and Albatross versions
In-Reply-To: <1186135630.7442.3.camel@localhost>
References: <1186135630.7442.3.camel@localhost>
Message-ID: <BB8C4F1C-C3F6-43D3-82A2-9A8363026C42@iam.unibe.ch>

> Which versions should I use?

I am sure Damien knows more.

Otherwise you might want to try to load the version from SqueakMap.  
It should work on 3.8 and 3.9.

Lukas

-- 
Lukas Renggli
http://www.lukas-renggli.ch


From renggli at iam.unibe.ch  Fri Aug  3 14:31:17 2007
From: renggli at iam.unibe.ch (Lukas Renggli)
Date: Fri, 3 Aug 2007 14:31:17 +0200
Subject: Pier and Albatross versions
In-Reply-To: <1186135835.7442.6.camel@localhost>
References: <1186135630.7442.3.camel@localhost>
	<1186135835.7442.6.camel@localhost>
Message-ID: <42A4ADE9-F3B6-414B-A971-9174AFB60BA8@iam.unibe.ch>

> Running the test APierTest fails while there is no response
> from the browser. Is there a browser opened while running
> the tests?

As far as I know Albatros it always opens a browser while running the  
tests.

Lukas

-- 
Lukas Renggli
http://www.lukas-renggli.ch


From keith_hodges at yahoo.co.uk  Fri Aug  3 15:20:48 2007
From: keith_hodges at yahoo.co.uk (Keith Hodges)
Date: Fri, 03 Aug 2007 14:20:48 +0100
Subject: Pier: Invoking a PRCommand from outside, on-the-fly-login,
	redirect to users page
In-Reply-To: <9aa6ae270708030459g25e78cb7wb4cec21c6fc8049@mail.gmail.com>
References: <9aa6ae270708030459g25e78cb7wb4cec21c6fc8049@mail.gmail.com>
Message-ID: <46B32BB0.9000901@yahoo.co.uk>

Matthias,

I tried some stuff in the PRPierControlPanel which operates on pier, yet 
runs in the context of the seaside/config

For code which does this check out the Magma-PersistencyManager and it 
depends upon Seaside28Jetsam.

This is the crucial bit non trivial bit for me, where self responds to 
#context

ControlPanel-callWidget: widget  titled: title

    PRCurrentContext use: self during: [
        WARenderLoop new call: (widget
            addDecoration: (WAWindowDecoration new
                title: title;
                yourself);
            yourself) ]

FYI: This was working in 2.7, but no longer works in 2.8

best regards

Keith



From matthias.berth at googlemail.com  Fri Aug  3 16:08:36 2007
From: matthias.berth at googlemail.com (Matthias Berth)
Date: Fri, 3 Aug 2007 16:08:36 +0200
Subject: Pier: Invoking a PRCommand from outside, on-the-fly-login,
	redirect to users page
In-Reply-To: <46B32BB0.9000901@yahoo.co.uk>
References: <9aa6ae270708030459g25e78cb7wb4cec21c6fc8049@mail.gmail.com>
	<46B32BB0.9000901@yahoo.co.uk>
Message-ID: <9aa6ae270708030708x1cf0f23chabfc73f30e8a7354@mail.gmail.com>

Keith,

thanks, interesting. So this snippet lets you render the Pier widget
from another Seaside app, and all the interactions etc work as
expected? Can I interpret it as embedding parts of Pier in a Seaside
app?

PRCurrentContext use: self during: [...] looks pretty mysterious to me.

I'm not sure, but I think that everything is already set up more or
less to run inside Pier once I have logged in. The login executes a
Pier command, and Confirm is another Pier command, both of them are
successful, it's only that Confirm does not go where I want it to go.


Cheers

Matthias

On 8/3/07, Keith Hodges <keith_hodges at yahoo.co.uk> wrote:
> Matthias,
>
> I tried some stuff in the PRPierControlPanel which operates on pier, yet
> runs in the context of the seaside/config
>
> For code which does this check out the Magma-PersistencyManager and it
> depends upon Seaside28Jetsam.
>
> This is the crucial bit non trivial bit for me, where self responds to
> #context
>
> ControlPanel-callWidget: widget  titled: title
>
>     PRCurrentContext use: self during: [
>         WARenderLoop new call: (widget
>             addDecoration: (WAWindowDecoration new
>                 title: title;
>                 yourself);
>             yourself) ]
>
> FYI: This was working in 2.7, but no longer works in 2.8
>
> best regards
>
> Keith
>
>
>
> _______________________________________________
> SmallWiki, Magritte, Pier and Related Tools ...
> https://www.iam.unibe.ch/mailman/listinfo/smallwiki
>

From norbert at hartl.name  Fri Aug  3 16:18:03 2007
From: norbert at hartl.name (Norbert Hartl)
Date: Fri, 03 Aug 2007 16:18:03 +0200
Subject: Pier and Albatross versions
In-Reply-To: <BB8C4F1C-C3F6-43D3-82A2-9A8363026C42@iam.unibe.ch>
References: <1186135630.7442.3.camel@localhost>
	<BB8C4F1C-C3F6-43D3-82A2-9A8363026C42@iam.unibe.ch>
Message-ID: <1186150683.7442.57.camel@localhost>


On Fri, 2007-08-03 at 14:26 +0200, Lukas Renggli wrote:
> > Which versions should I use?
> 
> I am sure Damien knows more.
> 

> Otherwise you might want to try to load the version from SqueakMap.  
> It should work on 3.8 and 3.9.
> 
Albatross is not on squeakmap. Pier from squeakmap loads seaside 
2.8 and albatross is using registerObjectForBacktracking.

Norbert



From renggli at iam.unibe.ch  Fri Aug  3 18:54:33 2007
From: renggli at iam.unibe.ch (Lukas Renggli)
Date: Fri, 3 Aug 2007 18:54:33 +0200
Subject: Pier: Invoking a PRCommand from outside, on-the-fly-login,
	redirect to users page
In-Reply-To: <9aa6ae270708030708x1cf0f23chabfc73f30e8a7354@mail.gmail.com>
References: <9aa6ae270708030459g25e78cb7wb4cec21c6fc8049@mail.gmail.com>
	<46B32BB0.9000901@yahoo.co.uk>
	<9aa6ae270708030708x1cf0f23chabfc73f30e8a7354@mail.gmail.com>
Message-ID: <671C4759-D59D-45AE-B459-2F4E92A14B2A@iam.unibe.ch>

I slightly refactored your changes so that they are in the right  
package and that they match the way Seaside implements these methods.  
PRCommand now understands #updateUrl: and #initialRequest: (the  
former #processRequest:).

These changes makes some of the Pier code much simpler and generally  
more powerful. Thanks for providing the initial implementation. To  
see an example of the interplay of these methods have a look at  
PRViewCommand that implements #updateUrl: and #initialRequest:.

HTH,
Lukas

On 03.08.2007, at 16:08, Matthias Berth wrote:

> Keith,
>
> thanks, interesting. So this snippet lets you render the Pier widget
> from another Seaside app, and all the interactions etc work as
> expected? Can I interpret it as embedding parts of Pier in a Seaside
> app?
>
> PRCurrentContext use: self during: [...] looks pretty mysterious to  
> me.
>
> I'm not sure, but I think that everything is already set up more or
> less to run inside Pier once I have logged in. The login executes a
> Pier command, and Confirm is another Pier command, both of them are
> successful, it's only that Confirm does not go where I want it to go.
>
>
> Cheers
>
> Matthias
>
> On 8/3/07, Keith Hodges <keith_hodges at yahoo.co.uk> wrote:
>> Matthias,
>>
>> I tried some stuff in the PRPierControlPanel which operates on  
>> pier, yet
>> runs in the context of the seaside/config
>>
>> For code which does this check out the Magma-PersistencyManager  
>> and it
>> depends upon Seaside28Jetsam.
>>
>> This is the crucial bit non trivial bit for me, where self  
>> responds to
>> #context
>>
>> ControlPanel-callWidget: widget  titled: title
>>
>>     PRCurrentContext use: self during: [
>>         WARenderLoop new call: (widget
>>             addDecoration: (WAWindowDecoration new
>>                 title: title;
>>                 yourself);
>>             yourself) ]
>>
>> FYI: This was working in 2.7, but no longer works in 2.8
>>
>> best regards
>>
>> Keith
>>
>>
>>
>> _______________________________________________
>> SmallWiki, Magritte, Pier and Related Tools ...
>> https://www.iam.unibe.ch/mailman/listinfo/smallwiki
>>
>
> _______________________________________________
> SmallWiki, Magritte, Pier and Related Tools ...
> https://www.iam.unibe.ch/mailman/listinfo/smallwiki

-- 
Lukas Renggli
http://www.lukas-renggli.ch


From damien.pollet at gmail.com  Mon Aug  6 23:58:39 2007
From: damien.pollet at gmail.com (Damien Pollet)
Date: Mon, 6 Aug 2007 23:58:39 +0200
Subject: Changing the title of the Header component ?
Message-ID: <34b4844b0708061458l519519abh7b5052b06ae0e21e@mail.gmail.com>

I'm trying to make a custom header as a subclass of PRHeaderWidget.
I'd like it to have a title and a subtitle so I made
subtitleDescription, defaults, and accessor inspired by the ones for
title in PRHeaderWidget.

Changing the subtitle from the Settings command works, but not the
title. As far as I understand it, the PRComponent holds the state to
configure instances of the PRHeaderWidget. When changing its settings,
it will update that state, unless it has fields with the same name.
In this case, #title from PRComponent is masking the one in
PRHeaderWidget, so when I change the Title from the web, I'm changing
the structure instead of the seaside component configuration... is
this really wanted? shouldn't the Edit command modify the structure
and Settings the component settings? Or was the idea that Edits are
persisted but not Settings ?

-- 
Damien Pollet
type less, do more [ | ] http://typo.cdlm.fasmz.org

From keith_hodges at yahoo.co.uk  Tue Aug  7 00:45:01 2007
From: keith_hodges at yahoo.co.uk (Keith Hodges)
Date: Mon, 06 Aug 2007 23:45:01 +0100
Subject: Changing the title of the Header component ?
In-Reply-To: <34b4844b0708061458l519519abh7b5052b06ae0e21e@mail.gmail.com>
References: <34b4844b0708061458l519519abh7b5052b06ae0e21e@mail.gmail.com>
Message-ID: <46B7A46D.1070502@yahoo.co.uk>

Damien Pollet wrote:
> I'm trying to make a custom header as a subclass of PRHeaderWidget.
> I'd like it to have a title and a subtitle so I made
> subtitleDescription, defaults, and accessor inspired by the ones for
> title in PRHeaderWidget.
>
> Changing the subtitle from the Settings command works, but not the
> title. As far as I understand it, the PRComponent holds the state to
> configure instances of the PRHeaderWidget. When changing its settings,
> it will update that state, unless it has fields with the same name.
> In this case, #title from PRComponent is masking the one in
> PRHeaderWidget, so when I change the Title from the web, I'm changing
> the structure instead of the seaside component configuration... is
> this really wanted? shouldn't the Edit command modify the structure
> and Settings the component settings? Or was the idea that Edits are
> persisted but not Settings ?
>
>   
Dear Damien,

from what I remember the title in the header widget is set to the name 
of the PRKernel which you are using. To rename this. I think that 
(PRKernel instanceNamed: 'pier') name: 'new name'. should do it. In the 
Pier-PersistencyManager package I provide a control panel UI for 
renaming and adding/removing kernels. I havent had a chance to fully get 
this working in Seaside 2.8, but if you are only using a non-magma 
persistency it should be ok.

Keith
 

From damien.pollet at gmail.com  Tue Aug  7 01:54:30 2007
From: damien.pollet at gmail.com (Damien Pollet)
Date: Tue, 7 Aug 2007 01:54:30 +0200
Subject: Changing the title of the Header component ?
In-Reply-To: <46B7A46D.1070502@yahoo.co.uk>
References: <34b4844b0708061458l519519abh7b5052b06ae0e21e@mail.gmail.com>
	<46B7A46D.1070502@yahoo.co.uk>
Message-ID: <34b4844b0708061654t2638df8fxaa36f1080b03266a@mail.gmail.com>

On 07/08/07, Keith Hodges <keith_hodges at yahoo.co.uk> wrote:
> from what I remember the title in the header widget is set to the name
> of the PRKernel which you are using.

Yes, the default value for the header title is %k, which gets expanded
to the kernel name. You can have the root or current page title also,
see PRMacroExpander.

So the PRComponent gets %k as a default title and you can change it
from the web with both Edit and Settings commands. The problem is that
the new title is stored in the PRComponent, not the PRHeaderWidget,
and the latter doesn't have access to the new value so it always uses
its default.


-- 
Damien Pollet
type less, do more [ | ] http://typo.cdlm.fasmz.org

From damien.pollet at gmail.com  Tue Aug  7 02:18:49 2007
From: damien.pollet at gmail.com (Damien Pollet)
Date: Tue, 7 Aug 2007 02:18:49 +0200
Subject: Small changes to the URL updating
Message-ID: <34b4844b0708061718h46336106x2299a3d765fad34c@mail.gmail.com>

Hi,

I committed Pier-Model-cdlm.159 and Pier-Seaside-cdlm.186.
I tweaked #updateUrl: to not add the view and command parameters if
the values are the default ones that would be guessed by
PRPiermain>>commandFrom:

Hopefully the patch makes sense, please review :)

-- 
Damien Pollet
type less, do more [ | ] http://typo.cdlm.fasmz.org

From cappadonza at hotmail.com  Thu Aug  9 14:10:35 2007
From: cappadonza at hotmail.com (isaiah)
Date: Thu, 9 Aug 2007 05:10:35 -0700 (PDT)
Subject: help with Magritte tutorial
Message-ID: <12071267.post@talk.nabble.com>


hey i working through the magritte tutorial excercises. need ome help with
exercise 18. it says "Add a class-instance variable to the MAPersonModel
called CustomDescription. initialize it with an emptyinstance of
maContainer".
i did the following to add a class-instance variable
MAPersonModel class
	instanceVariableNames: 'CustomDescription'

how do i initialize this variable
if i override the initialize method on the class

i get the follwoing warning from system browser "initialize is used in the
existing class system, overriding can cause serious problem is this really
what you want to do ?"
am i doing something wrong, can some one help me ?
im not sure what the difference between class variable and class-instance
variable

-- 
View this message in context: http://www.nabble.com/help-with-Magritte-tutorial-tf4242237.html#a12071267
Sent from the SmallWiki mailing list archive at Nabble.com.


From renggli at iam.unibe.ch  Thu Aug  9 15:15:11 2007
From: renggli at iam.unibe.ch (Lukas Renggli)
Date: Thu, 9 Aug 2007 15:15:11 +0200
Subject: help with Magritte tutorial
In-Reply-To: <12071267.post@talk.nabble.com>
References: <12071267.post@talk.nabble.com>
Message-ID: <F1833C4C-43AC-4838-9CFF-C746CC5A1709@iam.unibe.ch>

> i get the follwoing warning from system browser "initialize is used  
> in the
> existing class system, overriding can cause serious problem is this  
> really
> what you want to do ?"

I don't know why Squeak is complaining here. You can safely ignore  
this warning.

> am i doing something wrong, can some one help me ?

No, everything correct.

> im not sure what the difference between class variable and class- 
> instance
> variable

 From the Smalltalk FAQ [1]: Class instance variables are similar to  
class variables, except that they are created for eachsubclass of the  
defining class. When a class declares a class instance variable, a  
new variable is created for each subclass of that class. Each  
subclass then has its own instance of the variable and retains its  
own value for the variable, but each subclass has a variable with the  
same name. Only class methods of a class and its subclasses can refer  
to class instance variables; instance methods cannot.

Cheers,
Lukas

[1] http://www.faqs.org/faqs/smalltalk-faq/

-- 
Lukas Renggli
http://www.lukas-renggli.ch


From renggli at iam.unibe.ch  Thu Aug  9 15:27:03 2007
From: renggli at iam.unibe.ch (Lukas Renggli)
Date: Thu, 9 Aug 2007 15:27:03 +0200
Subject: Small changes to the URL updating
In-Reply-To: <34b4844b0708061718h46336106x2299a3d765fad34c@mail.gmail.com>
References: <34b4844b0708061718h46336106x2299a3d765fad34c@mail.gmail.com>
Message-ID: <95439713-1C47-4853-AE2F-C8D728F1362F@iam.unibe.ch>

> I committed Pier-Model-cdlm.159 and Pier-Seaside-cdlm.186.
> I tweaked #updateUrl: to not add the view and command parameters if
> the values are the default ones that would be guessed by
> PRPiermain>>commandFrom:
>
> Hopefully the patch makes sense, please review :)

It looks great. Thanks for the contribution.

Cheers,
Lukas

-- 
Lukas Renggli
http://www.lukas-renggli.ch


From renggli at iam.unibe.ch  Thu Aug  9 15:37:24 2007
From: renggli at iam.unibe.ch (Lukas Renggli)
Date: Thu, 9 Aug 2007 15:37:24 +0200
Subject: Changing the title of the Header component ?
In-Reply-To: <34b4844b0708061654t2638df8fxaa36f1080b03266a@mail.gmail.com>
References: <34b4844b0708061458l519519abh7b5052b06ae0e21e@mail.gmail.com>
	<46B7A46D.1070502@yahoo.co.uk>
	<34b4844b0708061654t2638df8fxaa36f1080b03266a@mail.gmail.com>
Message-ID: <08EF9C40-F053-430A-B6B5-0258F385B5D7@iam.unibe.ch>

Now I understand this problem. This was caused by a conflict caused  
by the merging of meta-descriptions. I changed #title to #heading in  
my latest commit.

Thanks for reporting.

Lukas

On 7 Aug 2007, at 01:54, Damien Pollet wrote:

> On 07/08/07, Keith Hodges <keith_hodges at yahoo.co.uk> wrote:
>> from what I remember the title in the header widget is set to the  
>> name
>> of the PRKernel which you are using.
>
> Yes, the default value for the header title is %k, which gets expanded
> to the kernel name. You can have the root or current page title also,
> see PRMacroExpander.
>
> So the PRComponent gets %k as a default title and you can change it
> from the web with both Edit and Settings commands. The problem is that
> the new title is stored in the PRComponent, not the PRHeaderWidget,
> and the latter doesn't have access to the new value so it always uses
> its default.
>
>
> -- 
> Damien Pollet
> type less, do more [ | ] http://typo.cdlm.fasmz.org
>
> _______________________________________________
> SmallWiki, Magritte, Pier and Related Tools ...
> https://www.iam.unibe.ch/mailman/listinfo/smallwiki

-- 
Lukas Renggli
http://www.lukas-renggli.ch


From norbert at hartl.name  Thu Aug  9 16:32:50 2007
From: norbert at hartl.name (Norbert Hartl)
Date: Thu, 09 Aug 2007 16:32:50 +0200
Subject: validate on ElementComponent
Message-ID: <1186669970.12249.7.camel@localhost>

Hi,

isn't validate supposed to be called on componentClasses?
I have a component generated from asComponent. I added
a componentClass to one of the descriptions. If I call component
validate I would have expected that the validate of my
componentClass is also called. 

Do I misunderstand something or is it likely I did something
wrong?

thanks,

Norbert


From cappadonza at hotmail.com  Thu Aug  9 18:56:45 2007
From: cappadonza at hotmail.com (isaiah)
Date: Thu, 9 Aug 2007 09:56:45 -0700 (PDT)
Subject: help with Magritte tutorial
In-Reply-To: <F1833C4C-43AC-4838-9CFF-C746CC5A1709@iam.unibe.ch>
References: <12071267.post@talk.nabble.com>
	<F1833C4C-43AC-4838-9CFF-C746CC5A1709@iam.unibe.ch>
Message-ID: <12076297.post@talk.nabble.com>


cool thanks lukas
exercise 18 also ays "Override #description on the instance-side, call super
and
compose it with the CustomDescription." what does this mean how do i compose
it


Lukas Renggli-2 wrote:
> 
>> i get the follwoing warning from system browser "initialize is used  
>> in the
>> existing class system, overriding can cause serious problem is this  
>> really
>> what you want to do ?"
> 
> I don't know why Squeak is complaining here. You can safely ignore  
> this warning.
> 
>> am i doing something wrong, can some one help me ?
> 
> No, everything correct.
> 
>> im not sure what the difference between class variable and class- 
>> instance
>> variable
> 
>  From the Smalltalk FAQ [1]: Class instance variables are similar to  
> class variables, except that they are created for eachsubclass of the  
> defining class. When a class declares a class instance variable, a  
> new variable is created for each subclass of that class. Each  
> subclass then has its own instance of the variable and retains its  
> own value for the variable, but each subclass has a variable with the  
> same name. Only class methods of a class and its subclasses can refer  
> to class instance variables; instance methods cannot.
> 
> Cheers,
> Lukas
> 
> [1] http://www.faqs.org/faqs/smalltalk-faq/
> 
> -- 
> Lukas Renggli
> http://www.lukas-renggli.ch
> 
> 
> _______________________________________________
> SmallWiki, Magritte, Pier and Related Tools ...
> https://www.iam.unibe.ch/mailman/listinfo/smallwiki
> 
> 

-- 
View this message in context: http://www.nabble.com/help-with-Magritte-tutorial-tf4242237.html#a12076297
Sent from the SmallWiki mailing list archive at Nabble.com.


From damien.cassou at gmail.com  Thu Aug  9 20:53:55 2007
From: damien.cassou at gmail.com (Damien Cassou)
Date: Thu, 9 Aug 2007 20:53:55 +0200
Subject: help with Magritte tutorial
In-Reply-To: <12076297.post@talk.nabble.com>
References: <12071267.post@talk.nabble.com>
	<F1833C4C-43AC-4838-9CFF-C746CC5A1709@iam.unibe.ch>
	<12076297.post@talk.nabble.com>
Message-ID: <6ac749c10708091153q258af333t92193a271a9ac188@mail.gmail.com>

2007/8/9, isaiah <cappadonza at hotmail.com>:
>
> cool thanks lukas
> exercise 18 also ays "Override #description on the instance-side, call super
> and
> compose it with the CustomDescription." what does this mean how do i compose
> it

I guess it means something like:

MyObject>>description
  ^ super description copy
         add: self customDescription;
         yourself

I'm not sure, but I think #copy is important after the call to super.

-- 
Damien Cassou

From ramon.leon at allresnet.com  Thu Aug  9 20:59:00 2007
From: ramon.leon at allresnet.com (Ramon Leon)
Date: Thu, 9 Aug 2007 11:59:00 -0700
Subject: help with Magritte tutorial
In-Reply-To: <F1833C4C-43AC-4838-9CFF-C746CC5A1709@iam.unibe.ch>
References: <12071267.post@talk.nabble.com>
	<F1833C4C-43AC-4838-9CFF-C746CC5A1709@iam.unibe.ch>
Message-ID: <011c01c7dab7$588303c0$f000a8c0@hq.allresnet.com>

>  From the Smalltalk FAQ [1]: Class instance variables are 
> similar to class variables, except that they are created for 
> eachsubclass of the defining class. When a class declares a 
> class instance variable, a new variable is created for each 
> subclass of that class. Each subclass then has its own 
> instance of the variable and retains its own value for the 
> variable, but each subclass has a variable with the same 
> name. Only class methods of a class and its subclasses can 
> refer to class instance variables; instance methods cannot.
> 
> Cheers,
> Lukas

Or to put it simply, classes are objects too, and all objects have instance
variables.

Ramon Leon
http://onsmalltalk.com


From keith_hodges at yahoo.co.uk  Fri Aug 10 05:56:11 2007
From: keith_hodges at yahoo.co.uk (Keith Hodges)
Date: Fri, 10 Aug 2007 04:56:11 +0100
Subject: Customizing magritte/pier
Message-ID: <46BBE1DB.90407@yahoo.co.uk>

I am looking for some tips/info on how to customize the look of 
magritte/pier components.

My first excursion found this quite difficult. Since it appears that 
textInput's always have class='text'. I solved this by wrapping them in 
a div class='agent-code' in a custom component and using the following 
to effect the styling.

 .contents form .agent-code input.text {
    width : auto;
}

What would be the correct way to change the pier login buttons to 
'Login' 'Cancel' rather than 'Save' 'Cancel'. Do I subclass PULogin, and 
change its asComponent to take the std component and search for its 
FormDecoration for the button label in order to change it?

many thanks in advance

Keith

From renggli at iam.unibe.ch  Fri Aug 10 08:32:54 2007
From: renggli at iam.unibe.ch (Lukas Renggli)
Date: Fri, 10 Aug 2007 08:32:54 +0200
Subject: Customizing magritte/pier
In-Reply-To: <46BBE1DB.90407@yahoo.co.uk>
References: <46BBE1DB.90407@yahoo.co.uk>
Message-ID: <2FD65884-A4DA-4E77-ACA5-9D41195BCE93@iam.unibe.ch>

> My first excursion found this quite difficult. Since it appears  
> that textInput's always have class='text'.

This is a Seaside feature coming from the fact that IE doesn't  
support CSS selectors like input[type=text].

> What would be the correct way to change the pier login buttons to
> 'Login' 'Cancel' rather than 'Save' 'Cancel'. Do I subclass  
> PULogin, and
> change its asComponent to take the std component and search for its
> FormDecoration for the button label in order to change it?

I guess that's the easiest way to do, yes. Relabeling with CSS could  
also work, but again IE will fail to do properly.

Note that Pier also puts a class into the body representing the  
current structure. It does not do so for different commands/views  
though, so maybe we should add that?

Lukas

-- 
Lukas Renggli
http://www.lukas-renggli.ch


From cappadonza at hotmail.com  Fri Aug 10 11:09:18 2007
From: cappadonza at hotmail.com (i p)
Date: Fri, 10 Aug 2007 09:09:18 +0000
Subject: help with Magritte tutorial
Message-ID: <BAY130-W2491A737DF7334B97B6492D2E10@phx.gbl>


thanks guys
the folowing worked. i realised i had to use addAll not add
MyObject>>description ^ super description copy addAll: self customDescription; yourself> Date: Thu, 9 Aug 2007 20:53:55 +0200> From: damien.cassou at gmail.com> To: smallwiki at iam.unibe.ch> Subject: Re: help with Magritte tutorial> > 2007/8/9, isaiah <cappadonza at hotmail.com>:> >> > cool thanks lukas> > exercise 18 also ays "Override #description on the instance-side, call super> > and> > compose it with the CustomDescription." what does this mean how do i compose> > it> > I guess it means something like:> > MyObject>>description> ^ super description copy> add: self customDescription;> yourself> > I'm not sure, but I think #copy is important after the call to super.> > -- > Damien Cassou> > _______________________________________________> SmallWiki, Magritte, Pier and Related Tools ...> https://www.iam.unibe.ch/mailman/listinfo/smallwiki
_________________________________________________________________
Explore the seven wonders of the world
http://search.msn.com/results.aspx?q=7+wonders+world&mkt=en-US&form=QBRE
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.iam.unibe.ch/pipermail/smallwiki/attachments/20070810/cf3d5fab/attachment-0001.html

From keith_hodges at yahoo.co.uk  Fri Aug 10 12:31:00 2007
From: keith_hodges at yahoo.co.uk (Keith Hodges)
Date: Fri, 10 Aug 2007 11:31:00 +0100
Subject: Customizing magritte/pier
In-Reply-To: <2FD65884-A4DA-4E77-ACA5-9D41195BCE93@iam.unibe.ch>
References: <46BBE1DB.90407@yahoo.co.uk>
	<2FD65884-A4DA-4E77-ACA5-9D41195BCE93@iam.unibe.ch>
Message-ID: <46BC3E64.4010106@yahoo.co.uk>

 
>> What would be the correct way to change the pier login buttons to
>> 'Login' 'Cancel' rather than 'Save' 'Cancel'. Do I subclass  
>> PULogin, and
>> change its asComponent to take the std component and search for its
>> FormDecoration for the button label in order to change it?
>>     
>
> I guess that's the easiest way to do, yes. 
Ok - it's not pretty, but for the record:

MyLoginSubclass-asComponent

    | component |
   
    component := super asComponent.
   
    component allDecorationsDo: [ :dec | dec class = MAFormDecoration 
ifTrue: [ dec buttons first value: 'Login'  ]].
   
    ^ component
> Relabeling with CSS could  
> also work, but again IE will fail to do properly.
>   
I am afraid IE is main customer in this case.
> Note that Pier also puts a class into the body representing the  
> current structure. It does not do so for different commands/views  
> though, so maybe we should add that?
>   
Good idea I think

Keith


From keith_hodges at yahoo.co.uk  Fri Aug 10 16:23:13 2007
From: keith_hodges at yahoo.co.uk (Keith Hodges)
Date: Fri, 10 Aug 2007 15:23:13 +0100
Subject: Users
Message-ID: <46BC74D1.4050906@yahoo.co.uk>

Hello,

I have 13000+ users validation records in a database, so I thought I 
would be clever. I define a pier user, role-dbuser, and use this as a 
template for creating a "User whose validation is certified by the 
database" on the fly.

Since it is created on the fly and is not known by PRKernel-users, nor 
is the User-Groups relation maintained, so the groups are only aware of 
the template user 'role-dbuser'.

This works fine until the check on groups permissions in 
PUSecurity-validateCommand: which reads;

(self group notNil
        and: [ (self group includes: aCommand context user)

So I turned it around to read the following, and implemented PUUser 
includesGroup:

(self group notNil
        and: [ (aCommand context user includesGroup: self group)

I thought that you might consider this tweak useful.

best regards

Keith

From keith_hodges at yahoo.co.uk  Fri Aug 10 18:00:48 2007
From: keith_hodges at yahoo.co.uk (Keith Hodges)
Date: Fri, 10 Aug 2007 17:00:48 +0100
Subject: Users
In-Reply-To: <46BC74D1.4050906@yahoo.co.uk>
References: <46BC74D1.4050906@yahoo.co.uk>
Message-ID: <46BC8BB0.2030509@yahoo.co.uk>


> I thought that you might consider this tweak useful.
>
> best regards
>
> Keith
>
>   
Just to say that there is more code to change to get this idea working, 
I just wondered if anyone had any thoughts on the idea in principle.

Keith


From renggli at iam.unibe.ch  Fri Aug 10 18:42:00 2007
From: renggli at iam.unibe.ch (Lukas Renggli)
Date: Fri, 10 Aug 2007 18:42:00 +0200
Subject: Users
In-Reply-To: <46BC74D1.4050906@yahoo.co.uk>
References: <46BC74D1.4050906@yahoo.co.uk>
Message-ID: <71A64B96-990D-41F9-9F52-A8389FAFDCA6@iam.unibe.ch>

> I have 13000+ users validation records in a database, so I thought I
> would be clever. I define a pier user, role-dbuser, and use this as a
> template for creating a "User whose validation is certified by the
> database" on the fly.

That sounds like an exciting project ;-)

Seems to be something essential to do authentication trough LDAP or  
an external for example.

> Since it is created on the fly and is not known by PRKernel-users, nor
> is the User-Groups relation maintained, so the groups are only  
> aware of
> the template user 'role-dbuser'.

What about creating a special group subclass? It could then override  
#includes: to check for the external user.

> So I turned it around to read the following, and implemented PUUser
> includesGroup:
>
> (self group notNil
>         and: [ (aCommand context user includesGroup: self group)

aCommand context user could be nil.

Cheers,
Lukas

-- 
Lukas Renggli
http://www.lukas-renggli.ch


From keith_hodges at yahoo.co.uk  Sat Aug 11 01:32:03 2007
From: keith_hodges at yahoo.co.uk (Keith Hodges)
Date: Sat, 11 Aug 2007 00:32:03 +0100
Subject: Users
In-Reply-To: <71A64B96-990D-41F9-9F52-A8389FAFDCA6@iam.unibe.ch>
References: <46BC74D1.4050906@yahoo.co.uk>
	<71A64B96-990D-41F9-9F52-A8389FAFDCA6@iam.unibe.ch>
Message-ID: <46BCF573.2010906@yahoo.co.uk>


> aCommand context user could be nil.
>
> Cheers,
> Lukas
>
>   
Yes good point, everything worked until I hit the logout button!

Keith


From keith_hodges at yahoo.co.uk  Sat Aug 11 01:37:16 2007
From: keith_hodges at yahoo.co.uk (Keith Hodges)
Date: Sat, 11 Aug 2007 00:37:16 +0100
Subject: Users
In-Reply-To: <71A64B96-990D-41F9-9F52-A8389FAFDCA6@iam.unibe.ch>
References: <46BC74D1.4050906@yahoo.co.uk>
	<71A64B96-990D-41F9-9F52-A8389FAFDCA6@iam.unibe.ch>
Message-ID: <46BCF6AC.2020408@yahoo.co.uk>

Following your suggestion the simplest fix to make this work:

PUGroup-includes: anUser

    anUser ifNil: [ ^ false ].
   
    ^ anUser groups includes: self

regards

Keith

From thams at thams.com  Sat Aug 11 01:53:32 2007
From: thams at thams.com (Kurt Thams)
Date: Fri, 10 Aug 2007 16:53:32 -0700
Subject: CSS style on Magritte generated component
Message-ID: <00e701c7dba9$a8a01d50$cd01a8c0@ararauna>

I have a seaside call like this:

result := self call: (SurveyResponse new asComponent addValidatedForm;
yourself).

I can't figure out how to get the css to be applied to this component.
(Having SurveryReponse implement #style doesn't do it, nor does having it
defined in the class making the call do it).



From renggli at iam.unibe.ch  Sat Aug 11 08:35:54 2007
From: renggli at iam.unibe.ch (Lukas Renggli)
Date: Sat, 11 Aug 2007 08:35:54 +0200
Subject: Customizing magritte/pier
In-Reply-To: <46BC3E64.4010106@yahoo.co.uk>
References: <46BBE1DB.90407@yahoo.co.uk>
	<2FD65884-A4DA-4E77-ACA5-9D41195BCE93@iam.unibe.ch>
	<46BC3E64.4010106@yahoo.co.uk>
Message-ID: <05312543-E0E8-4256-BCFF-3216FDDB04C4@iam.unibe.ch>

>> Note that Pier also puts a class into the body representing the
>> current structure. It does not do so for different commands/views
>> though, so maybe we should add that?
>
> Good idea I think

Name: Pier-All-lr.224
Author: lr
Time: 11 August 2007, 8:33:38 am
UUID: 5779ace6-07cf-499f-84b8-c8c06d78cf03
Ancestors: Pier-All-lr.223
Dependencies: Pier-Model-cdlm.159, Pier-Tests-lr.72, Pier-Seaside-lr. 
188, Pier-Security-lr.85, Pier-Blog-lr.57

- not only add the type of the structure, but also the command and  
view as css classes to the body-tag

Lukas

-- 
Lukas Renggli
http://www.lukas-renggli.ch


From renggli at iam.unibe.ch  Sat Aug 11 08:39:36 2007
From: renggli at iam.unibe.ch (Lukas Renggli)
Date: Sat, 11 Aug 2007 08:39:36 +0200
Subject: CSS style on Magritte generated component
In-Reply-To: <00e701c7dba9$a8a01d50$cd01a8c0@ararauna>
References: <00e701c7dba9$a8a01d50$cd01a8c0@ararauna>
Message-ID: <531E5EC8-BF2C-4D20-A04A-E014429AED7B@iam.unibe.ch>

> I have a seaside call like this:
>
> result := self call: (SurveyResponse new asComponent addValidatedForm;
> yourself).
>
> I can't figure out how to get the css to be applied to this component.
> (Having SurveryReponse implement #style doesn't do it, nor does  
> having it
> defined in the class making the call do it).

I guess SurveyResponse is a model class, right?

#style has to be implemented in a subclass of WAComponent. Since you  
cannot modify the component you get from Magritte (of course you can,  
but that would probably be overkill in your case) it is the simples  
if you put #style into the WAComponent/WATask you call it from.

Another possibility would be to use WAFileLibrary. Check out the  
class comment.

Cheers,
Lukas

-- 
Lukas Renggli
http://www.lukas-renggli.ch


From keith_hodges at yahoo.co.uk  Sat Aug 11 21:37:37 2007
From: keith_hodges at yahoo.co.uk (Keith Hodges)
Date: Sat, 11 Aug 2007 20:37:37 +0100
Subject: PULogin command to show 'Login' button
In-Reply-To: <46BCF6AC.2020408@yahoo.co.uk>
References: <46BC74D1.4050906@yahoo.co.uk>	<71A64B96-990D-41F9-9F52-A8389FAFDCA6@iam.unibe.ch>
	<46BCF6AC.2020408@yahoo.co.uk>
Message-ID: <46BE1001.30402@yahoo.co.uk>

The cleanest approach I have found to achieve this is:

PULogin-asComponent
    ^ (self description asComponentOn: self)
        addValidatedForm: (Array with: #save -> 'Login' with: #cancel);
         yourself

It might be worth including this in Pier, since users have often 
commented to me about it

regards

Keith

From keith_hodges at yahoo.co.uk  Sat Aug 11 21:48:53 2007
From: keith_hodges at yahoo.co.uk (Keith Hodges)
Date: Sat, 11 Aug 2007 20:48:53 +0100
Subject: Magritte-ComponentDecoration s
Message-ID: <46BE12A5.2070801@yahoo.co.uk>

I just committed this to http://mc.lukas-renggli.ch/pieraddons

Magritte-ComponentDecorations-kph.1.mcz (note it has a pier extension in 
it too)

This package allows a magritte description to specify decorationClasses:
These are added at component creation time and they pick up their 
parameters from the description's properties (unless I can think of a 
better way).

example:

WSLogin-descriptionUsername
    ^ (super descriptionUsername)
        decorationClasses: (Array with: MAAnnotationDecoration ) ;
        propertyAt: #postAnnotation put: '(6-12 letters)';
        yourself
        
pier wiki text annotation

WSLogin-descriptionUsername
    ^ (super descriptionUsername)
        decorationClasses: (Array with: PRAnnotationDecoration ) ;
        propertyAt: #postAnnotation put: 'or 
*Register>/Environment/RegistrationComponent*';
        yourself

best regards

Keith

From renggli at iam.unibe.ch  Sat Aug 11 22:02:38 2007
From: renggli at iam.unibe.ch (Lukas Renggli)
Date: Sat, 11 Aug 2007 22:02:38 +0200
Subject: PULogin command to show 'Login' button
In-Reply-To: <46BE1001.30402@yahoo.co.uk>
References: <46BC74D1.4050906@yahoo.co.uk>	<71A64B96-990D-41F9-9F52-A8389FAFDCA6@iam.unibe.ch>
	<46BCF6AC.2020408@yahoo.co.uk> <46BE1001.30402@yahoo.co.uk>
Message-ID: <F2CC1C9C-BA68-432D-BE1C-FD543F42A872@iam.unibe.ch>

Very good idea. Fixed in:

Name: Pier-All-lr.225
Author: lr
Time: 11 August 2007, 10:01:33 pm
UUID: 2c90b9ce-6ed0-4587-9495-0e550263c1ce
Ancestors: Pier-All-lr.224
Dependencies: Pier-Model-lr.160, Pier-Tests-lr.72, Pier-Seaside-lr. 
189, Pier-Security-lr.86, Pier-Blog-lr.57

- display proper action lables (add, copy, move, apply, login) in  
commands, instead of the default save

On 11 Aug 2007, at 21:37, Keith Hodges wrote:

> The cleanest approach I have found to achieve this is:
>
> PULogin-asComponent
>     ^ (self description asComponentOn: self)
>         addValidatedForm: (Array with: #save -> 'Login' with:  
> #cancel);
>          yourself
>
> It might be worth including this in Pier, since users have often
> commented to me about it
>
> regards
>
> Keith
>
> _______________________________________________
> SmallWiki, Magritte, Pier and Related Tools ...
> https://www.iam.unibe.ch/mailman/listinfo/smallwiki

-- 
Lukas Renggli
http://www.lukas-renggli.ch


From keith_hodges at yahoo.co.uk  Sat Aug 11 22:11:15 2007
From: keith_hodges at yahoo.co.uk (Keith Hodges)
Date: Sat, 11 Aug 2007 21:11:15 +0100
Subject: PULogin command to show 'Login' button
In-Reply-To: <F2CC1C9C-BA68-432D-BE1C-FD543F42A872@iam.unibe.ch>
References: <46BC74D1.4050906@yahoo.co.uk>	<71A64B96-990D-41F9-9F52-A8389FAFDCA6@iam.unibe.ch>	<46BCF6AC.2020408@yahoo.co.uk>
	<46BE1001.30402@yahoo.co.uk>
	<F2CC1C9C-BA68-432D-BE1C-FD543F42A872@iam.unibe.ch>
Message-ID: <46BE17E3.4000802@yahoo.co.uk>

Lukas Renggli wrote:
> Very good idea. Fixed in:
>
> Name: Pier-All-lr.225
>   
Thats what I call a fast turn around!

Keith

From keith_hodges at yahoo.co.uk  Sat Aug 11 23:13:00 2007
From: keith_hodges at yahoo.co.uk (Keith Hodges)
Date: Sat, 11 Aug 2007 22:13:00 +0100
Subject: [Bug] Subscripts at beginning of line?
Message-ID: <46BE265C.1000204@yahoo.co.uk>

If you put

@@tiny writing@@ at the beginning of a line, it thinks that you are 
defining an anchor.

cheers

Keith

From keith_hodges at yahoo.co.uk  Sun Aug 12 04:30:54 2007
From: keith_hodges at yahoo.co.uk (Keith Hodges)
Date: Sun, 12 Aug 2007 03:30:54 +0100
Subject: Customizing magritte/pier
In-Reply-To: <46BBE1DB.90407@yahoo.co.uk>
References: <46BBE1DB.90407@yahoo.co.uk>
Message-ID: <46BE70DE.7060309@yahoo.co.uk>

 
> My first excursion found this quite difficult. Since it appears that 
> textInput's always have class='text'. I solved this by wrapping them in 
> a div class='agent-code' in a custom component and using the following 
> to effect the styling.
>
>  .contents form .agent-code input.text {
>     width : auto;
> }
>
>   
I was finding it tricky to target specific input.text elements within a 
custom subclass of MATextInputComponent.

how about this change?

MAComponentRenderer-classFor: aDescription

    ^ String streamContents: [ :stream |
"<mod>"
        stream nextPutAll: ((self childAt: aDescription) ifNil: [ 
component ]) cssClass.
"</mod>"
        aDescription isReadonly
            ifTrue: [ stream space; nextPutAll: 'readonly' ].
        aDescription isRequired
            ifTrue: [ stream space; nextPutAll: 'required' ].
        (self hasError: aDescription)
            ifTrue: [ stream space; nextPutAll: 'error' ] ]

together with

MAComponent-cssClass
    "Answer a CSS class that should be used when this component is 
rendered."
   
    ^ ((self class fullName readStream) next:2;  upToAll: 'Component') 
asLowercase
 
best regards

Keith

From renggli at iam.unibe.ch  Sun Aug 12 09:20:05 2007
From: renggli at iam.unibe.ch (Lukas Renggli)
Date: Sun, 12 Aug 2007 09:20:05 +0200
Subject: [Bug] Subscripts at beginning of line?
In-Reply-To: <46BE265C.1000204@yahoo.co.uk>
References: <46BE265C.1000204@yahoo.co.uk>
Message-ID: <5C152AC9-A02B-4EED-BD99-4FF7319BD87C@iam.unibe.ch>

> If you put
>
> @@tiny writing@@ at the beginning of a line, it thinks that you are
> defining an anchor.

That's like it is for all overlapping token. As a workaround you can  
put a space in front, like:

   @@tiny writing@@

Alternatively, if you don't use anchors you can also remove the  
anchor handler from the parser.

Lukas

-- 
Lukas Renggli
http://www.lukas-renggli.ch


From keith_hodges at yahoo.co.uk  Tue Aug 14 17:55:00 2007
From: keith_hodges at yahoo.co.uk (Keith Hodges)
Date: Tue, 14 Aug 2007 16:55:00 +0100
Subject: [bug/fix] PRContext-copy 
Message-ID: <46C1D054.8070306@yahoo.co.uk>

I discovered that the referential integrity of PRContext <-> PRCommand 
gets broken when you assign a user to the context because #user: does a 
copy, but the instance of PRCommand's context is still the pre-copy one.

This is fixed by

--
PRContext-postCopy

    self setCommand: command copy.

--
cheers

Keith

 


From rbb at techgame.net  Tue Aug 14 18:23:43 2007
From: rbb at techgame.net (Brian Brown)
Date: Tue, 14 Aug 2007 10:23:43 -0600
Subject: Question with ICal-Magritte
Message-ID: <300506B2-9C05-46B8-8371-F41C448BADD2@techgame.net>

I've started using the really excellent ICal package (thanks  
Phillipe!), and just loaded the ICal Magritte package.

On ICCalendarEntry class>>descriptionUrl, the definition is:

descriptionUrl
	^(NAUrlDescription selector: #url label: 'Link' priority: 1000)
		beEditable;
		yourself


NAUrlDescription is not defined anywhere - I would expect it to be  
named MAUrlDescription, but that isn't defined in my image either ;)

Any ideas where I can get this class, or what else should be in here?

Thanks!

Brian


From cnantais at gmail.com  Tue Aug 14 19:19:17 2007
From: cnantais at gmail.com (Chad Nantais)
Date: Tue, 14 Aug 2007 10:19:17 -0700
Subject: Wizard-type form sequences
Message-ID: <3de931600708141019w5150531bo153597ac60b70bd3@mail.gmail.com>

I can't seem to find anything in the mailing list archives about this,
so here goes...

Is there a way to present forms as a sequence of input requests with a
'Next' submit button rather than the default 'Save' and 'Cancel'?

Thanks,

Chad

-- 
Chad Nantais

http://myspace.com/chadnantais

From philippe.marschall at gmail.com  Tue Aug 14 19:22:41 2007
From: philippe.marschall at gmail.com (Philippe Marschall)
Date: Tue, 14 Aug 2007 19:22:41 +0200
Subject: Question with ICal-Magritte
In-Reply-To: <300506B2-9C05-46B8-8371-F41C448BADD2@techgame.net>
References: <300506B2-9C05-46B8-8371-F41C448BADD2@techgame.net>
Message-ID: <66666f210708141022g31f01d61u2dc72ac4a4ef4b7b@mail.gmail.com>

2007/8/14, Brian Brown <rbb at techgame.net>:
> I've started using the really excellent ICal package (thanks
> Phillipe!), and just loaded the ICal Magritte package.
>
> On ICCalendarEntry class>>descriptionUrl, the definition is:
>
> descriptionUrl
>         ^(NAUrlDescription selector: #url label: 'Link' priority: 1000)
>                 beEditable;
>                 yourself
>
>
> NAUrlDescription is not defined anywhere - I would expect it to be
> named MAUrlDescription, but that isn't defined in my image either ;)

You need the NetworkAddress package from:
http://source.lukas-renggli.ch/magritteaddons.html
I haven't looked at it in a while so stuff might be broken. There are
however some tests which you can check out.

Cheers
Philippe

> Any ideas where I can get this class, or what else should be in here?
>
> Thanks!
>
> Brian
>
>
> _______________________________________________
> SmallWiki, Magritte, Pier and Related Tools ...
> https://www.iam.unibe.ch/mailman/listinfo/smallwiki
>

From j.squeak at cyberhaus.us  Tue Aug 14 23:11:13 2007
From: j.squeak at cyberhaus.us (Jimmie Houchin)
Date: Tue, 14 Aug 2007 16:11:13 -0500
Subject: Pier v. Plone
Message-ID: <46C21A71.9020204@cyberhaus.us>

Hello,

I am exploring my options for a couple of websites. I am a big fan of
Squeak, Seaside and Pier. But I have been exploring and learning about
Plone.

There is a tremendous amount of energy and work behind Plone and
products available for it. Plone 3.0 is about to be released. They have
put a lot of work into UI usability and into a lot of areas which are
nice to not necessarily implement yourself.

That said, I don't really have a grasp as to how far Pier is from
Plone-like abilities for both the end user and the developer.

One thing I am looking at is the Plone4Artist product which implements a
very sweet package for multimedia and Plone.
http://www.plone4artists.com/

I do very much like Squeak over Python.
In general I like Seaside/Pier over Zope/Plone.

I know that Lukas has used Zope/Plone, but I don't know how current his
knowledge is.

How far is Pier from offering a competitive experience to end users and
developers with regard to something like Plone?

Is there any interest in the community of competing in that sphere?

Or do most here happily use Seaside/Pier as a toolbox and the bigger
application type CMS isn't on the radar?

And initially I speak as to a basic Plone 3.0 out-of-the-box
capabilities and user/developer experience. Not necessarily the plethora
of add-on products. Those can come later if the out-of-the-box
experience is compelling enough.

Naively it seems that content types are easier to develop in Pier as
opposed to Plone.

I just don't know how much I could build on Pier towards a Plone-like
experience in the time it takes to do some things in Plone. And I really
don't know how much of the machinery that makes Plone what it is, is in
Pier. Or the effort to get there.

Comments regarding my questions above or most anything relevant to Pier
v. Plone 3.0 are greatly appreciated.

Thanks.

Jimmie

From renggli at iam.unibe.ch  Tue Aug 14 23:54:05 2007
From: renggli at iam.unibe.ch (Lukas Renggli)
Date: Tue, 14 Aug 2007 23:54:05 +0200
Subject: [bug/fix] PRContext-copy 
In-Reply-To: <46C1D054.8070306@yahoo.co.uk>
References: <46C1D054.8070306@yahoo.co.uk>
Message-ID: <C29EF236-B751-4860-96B5-FB6CB6B61EFC@iam.unibe.ch>

> I discovered that the referential integrity of PRContext <-> PRCommand
> gets broken when you assign a user to the context because #user:  
> does a
> copy, but the instance of PRCommand's context is still the pre-copy  
> one.

Nice catch. Thanks for reporting.

This is tested and fixed in:

Name: Pier-All-lr.226
Author: lr
Time: 14 August 2007, 11:53:01 pm
UUID: a1d051f0-d8f0-4ecb-a8b7-3ee3584a13a3
Ancestors: Pier-All-lr.225
Dependencies: Pier-Model-lr.161, Pier-Tests-lr.73, Pier-Seaside-lr. 
189, Pier-Security-lr.86, Pier-Blog-lr.57

fixes and tests "I discovered that the referential integrity of  
PRContext <-> PRCommand gets broken when you assign a user to the  
context because #user: does a copy, but the instance of PRCommand's  
context is still the pre-copy one." (Keith Hodges)

-- 
Lukas Renggli
http://www.lukas-renggli.ch


From renggli at iam.unibe.ch  Tue Aug 14 23:57:28 2007
From: renggli at iam.unibe.ch (Lukas Renggli)
Date: Tue, 14 Aug 2007 23:57:28 +0200
Subject: Wizard-type form sequences
In-Reply-To: <3de931600708141019w5150531bo153597ac60b70bd3@mail.gmail.com>
References: <3de931600708141019w5150531bo153597ac60b70bd3@mail.gmail.com>
Message-ID: <D51FC21F-E639-4A86-AB08-38EC1721A3AE@iam.unibe.ch>

> I can't seem to find anything in the mailing list archives about  
> this, so here goes...
>
> Is there a way to present forms as a sequence of input requests  
> with a 'Next' submit button rather than the default 'Save' and  
> 'Cancel'?

Use #addValidatedForm: to customize the buttons around your form.

     self call: (aModel asComponent
         addValidatedForm: (Array with: #save -> 'Next');
         yourself).
     self call: (aModel ...

Cheers,
Lukas

-- 
Lukas Renggli
http://www.lukas-renggli.ch


From rbb at techgame.net  Wed Aug 15 00:08:48 2007
From: rbb at techgame.net (Brian Brown)
Date: Tue, 14 Aug 2007 16:08:48 -0600
Subject: Question with ICal-Magritte
In-Reply-To: <66666f210708141022g31f01d61u2dc72ac4a4ef4b7b@mail.gmail.com>
References: <300506B2-9C05-46B8-8371-F41C448BADD2@techgame.net>
	<66666f210708141022g31f01d61u2dc72ac4a4ef4b7b@mail.gmail.com>
Message-ID: <F259CB72-7AE9-4F0E-B07D-F2240542322C@techgame.net>

Thank you, so far it works fine :)


On Aug 14, 2007, at 11:22 AM, Philippe Marschall wrote:

> 2007/8/14, Brian Brown <rbb at techgame.net>:
>> I've started using the really excellent ICal package (thanks
>> Phillipe!), and just loaded the ICal Magritte package.
>>
>> On ICCalendarEntry class>>descriptionUrl, the definition is:
>>
>> descriptionUrl
>>         ^(NAUrlDescription selector: #url label: 'Link' priority:  
>> 1000)
>>                 beEditable;
>>                 yourself
>>
>>
>> NAUrlDescription is not defined anywhere - I would expect it to be
>> named MAUrlDescription, but that isn't defined in my image either ;)
>
> You need the NetworkAddress package from:
> http://source.lukas-renggli.ch/magritteaddons.html
> I haven't looked at it in a while so stuff might be broken. There are
> however some tests which you can check out.
>
> Cheers
> Philippe
>
>> Any ideas where I can get this class, or what else should be in here?
>>
>> Thanks!
>>
>> Brian
>>
>>
>> _______________________________________________
>> SmallWiki, Magritte, Pier and Related Tools ...
>> https://www.iam.unibe.ch/mailman/listinfo/smallwiki
>>
>
> _______________________________________________
> SmallWiki, Magritte, Pier and Related Tools ...
> https://www.iam.unibe.ch/mailman/listinfo/smallwiki


From renggli at iam.unibe.ch  Wed Aug 15 00:13:49 2007
From: renggli at iam.unibe.ch (Lukas Renggli)
Date: Wed, 15 Aug 2007 00:13:49 +0200
Subject: Pier v. Plone
In-Reply-To: <46C21A71.9020204@cyberhaus.us>
References: <46C21A71.9020204@cyberhaus.us>
Message-ID: <1AFDAFB5-43E5-40D6-B807-6DA0AAC6F03C@iam.unibe.ch>

> I know that Lukas has used Zope/Plone, but I don't know how current  
> his
> knowledge is.

My knowledge of Zope and Plone dates back to 2002. I maintained some  
legacy Zope projects in 2003, but as I discovered Smalltalk and  
Seaside around the same time I never went back to Python again.

> How far is Pier from offering a competitive experience to end users  
> and
> developers with regard to something like Plone?

I have no idea in what state Zope/Plone is today. Maybe I should give  
another try ...

> Is there any interest in the community of competing in that sphere?

I am interested, but I can't compete against hundreds of companies  
and thousands of developers all alone.

> Or do most here happily use Seaside/Pier as a toolbox and the bigger
> application type CMS isn't on the radar?

I am using Seaside/Pier for all my web related projects.

> And initially I speak as to a basic Plone 3.0 out-of-the-box
> capabilities and user/developer experience. Not necessarily the  
> plethora
> of add-on products. Those can come later if the out-of-the-box
> experience is compelling enough.

To make the out-of-the-box experience of Pier (especially for those  
not knowing Smalltalk and Seaside) quite some work is required.

> Naively it seems that content types are easier to develop in Pier as
> opposed to Plone.

This is certainly the case, even if Zope/Plone improved a lot over  
the past 5 years.

Lukas

-- 
Lukas Renggli
http://www.lukas-renggli.ch


From admin at moedae.com  Wed Aug 15 00:44:53 2007
From: admin at moedae.com (Taun)
Date: Tue, 14 Aug 2007 18:44:53 -0400
Subject: Pier v. Plone
In-Reply-To: <46C21A71.9020204@cyberhaus.us>
References: <46C21A71.9020204@cyberhaus.us>
Message-ID: <46C23065.9080702@moedae.com>

Jimmie, I have been agonizing over the same choice for months. For now, I am sticking with
Plone for my clients.

I have worked with Smalltalk since the late 80's and Plone since it's inception on top of
CMF for Zope.

Plone Cons -

   Past performance - Historically, I have had major performance problems with Plone
including memory leaks leading to repeated server restarts on an almost hourly basis under
heavy loads. Performance has improved but I am still concerned for the site I am currently
creating using Plone.

   Software Stack - This is my biggest complaint. The Plone stack is too dependent on C
libraries and python modules and zope adapters. For example, connecting Plone to postresql
requires a postgresql adapter which consists of a stack in C, python and Zope. Like most
OpenSource, all the versions have to be compatible. Just the out of the box stack is
python C virtual machine binary -> python libraries -> Zope -> CMF -> Plone -> Plone
Standard Products. Addon products such as PIL (imaging library) add a 3 layer stack of C
module, python module and Zope module.

   Development environment doesn't hold a candle to Seaside.

Plone Pros -

   Very active community.
   Tons of interesting Products.
   Workflow and security model very easy to customise.
   Easy to add custom products with ArchGenXML and ArgoUML.
   Very Easy to add custom content types with ArchGenXML and ArgoUML.
   **Visual Tools like ArgoUML/ArchGenXML are what are really missing from Squeak. You use
ArgoUML to create a UML class diagram for your product or content type. You then run
ArchGenXML to create a Plone product from the class diagram. I am currently using this
process to create a site with 8 custom content types with cross references and it is very
easy.
http://plone.org/documentation/tutorial/archgenxml-getting-started


Squeak Cons -

   The flip side of all of the Plone Pros.

Squeak Pros -

   I much prefer Smalltalk to Python.
   I prefer the Smalltalk environment to any Python IDE.
   I think people write better Smalltalk tools/objects than Python programs.
   I prefer the architecture and tools of Seaside to Plone.
   Much shorter stack - Smalltalk VM -> Smalltalk source -> Seaside -> Magritte -> Pier.
   Fewer C module dependencies for add-on products/objects.


Personal Conclusion -

   Zope/Plone feels like a framework where you are locked into the frame.
   Smalltalk/Seaside feels like a great tool box to help you build whatever you want.
Unfortunately, I would have to build Plone when that is about all I need.

   If one only needs the functionality provided by Zope, Seaside is the better choice for
me. I would be more productive and have a better end product with Seaside versus Zope.

   I chose painful but functional Plone over Fun but less functional Pier based on the
Plone Workflow model, ArchGenXML for products and many available 3rd party products.

   I really hope to be able to make a different choice in the future.




Taun


Jimmie Houchin wrote:
> Hello,
> 
> I am exploring my options for a couple of websites. I am a big fan of
> Squeak, Seaside and Pier. But I have been exploring and learning about
> Plone.
> 
> There is a tremendous amount of energy and work behind Plone and
> products available for it. Plone 3.0 is about to be released. They have
> put a lot of work into UI usability and into a lot of areas which are
> nice to not necessarily implement yourself.
> 
> That said, I don't really have a grasp as to how far Pier is from
> Plone-like abilities for both the end user and the developer.
> 
> One thing I am looking at is the Plone4Artist product which implements a
> very sweet package for multimedia and Plone.
> http://www.plone4artists.com/
> 
> I do very much like Squeak over Python.
> In general I like Seaside/Pier over Zope/Plone.
> 
> I know that Lukas has used Zope/Plone, but I don't know how current his
> knowledge is.
> 
> How far is Pier from offering a competitive experience to end users and
> developers with regard to something like Plone?
> 
> Is there any interest in the community of competing in that sphere?
> 
> Or do most here happily use Seaside/Pier as a toolbox and the bigger
> application type CMS isn't on the radar?
> 
> And initially I speak as to a basic Plone 3.0 out-of-the-box
> capabilities and user/developer experience. Not necessarily the plethora
> of add-on products. Those can come later if the out-of-the-box
> experience is compelling enough.
> 
> Naively it seems that content types are easier to develop in Pier as
> opposed to Plone.
> 
> I just don't know how much I could build on Pier towards a Plone-like
> experience in the time it takes to do some things in Plone. And I really
> don't know how much of the machinery that makes Plone what it is, is in
> Pier. Or the effort to get there.
> 
> Comments regarding my questions above or most anything relevant to Pier
> v. Plone 3.0 are greatly appreciated.
> 
> Thanks.
> 
> Jimmie
> 
> _______________________________________________
> SmallWiki, Magritte, Pier and Related Tools ...
> https://www.iam.unibe.ch/mailman/listinfo/smallwiki


From tblanchard at mac.com  Wed Aug 15 04:40:19 2007
From: tblanchard at mac.com (Todd Blanchard)
Date: Tue, 14 Aug 2007 19:40:19 -0700
Subject: Pier v. Plone
In-Reply-To: <46C23065.9080702@moedae.com>
References: <46C21A71.9020204@cyberhaus.us> <46C23065.9080702@moedae.com>
Message-ID: <94EA985D-6DE0-4ACC-9F44-BB03C6C6E911@mac.com>

FWIW, I'm using drupal for a musician's website I'm doing - just  
configure and tweak css - no coding necessary - tons of modules.

On Aug 14, 2007, at 3:44 PM, Taun wrote:

> Jimmie, I have been agonizing over the same choice for months. For  
> now, I am sticking with
> Plone for my clients.
>
> I have worked with Smalltalk since the late 80's and Plone since  
> it's inception on top of
> CMF for Zope.
>
> Plone Cons -
>
>    Past performance - Historically, I have had major performance  
> problems with Plone
> including memory leaks leading to repeated server restarts on an  
> almost hourly basis under
> heavy loads. Performance has improved but I am still concerned for  
> the site I am currently
> creating using Plone.
>
>    Software Stack - This is my biggest complaint. The Plone stack  
> is too dependent on C
> libraries and python modules and zope adapters. For example,  
> connecting Plone to postresql
> requires a postgresql adapter which consists of a stack in C,  
> python and Zope. Like most
> OpenSource, all the versions have to be compatible. Just the out of  
> the box stack is
> python C virtual machine binary -> python libraries -> Zope -> CMF - 
> > Plone -> Plone
> Standard Products. Addon products such as PIL (imaging library) add  
> a 3 layer stack of C
> module, python module and Zope module.
>
>    Development environment doesn't hold a candle to Seaside.
>
> Plone Pros -
>
>    Very active community.
>    Tons of interesting Products.
>    Workflow and security model very easy to customise.
>    Easy to add custom products with ArchGenXML and ArgoUML.
>    Very Easy to add custom content types with ArchGenXML and ArgoUML.
>    **Visual Tools like ArgoUML/ArchGenXML are what are really  
> missing from Squeak. You use
> ArgoUML to create a UML class diagram for your product or content  
> type. You then run
> ArchGenXML to create a Plone product from the class diagram. I am  
> currently using this
> process to create a site with 8 custom content types with cross  
> references and it is very
> easy.
> http://plone.org/documentation/tutorial/archgenxml-getting-started
>
>
> Squeak Cons -
>
>    The flip side of all of the Plone Pros.
>
> Squeak Pros -
>
>    I much prefer Smalltalk to Python.
>    I prefer the Smalltalk environment to any Python IDE.
>    I think people write better Smalltalk tools/objects than Python  
> programs.
>    I prefer the architecture and tools of Seaside to Plone.
>    Much shorter stack - Smalltalk VM -> Smalltalk source -> Seaside  
> -> Magritte -> Pier.
>    Fewer C module dependencies for add-on products/objects.
>
>
> Personal Conclusion -
>
>    Zope/Plone feels like a framework where you are locked into the  
> frame.
>    Smalltalk/Seaside feels like a great tool box to help you build  
> whatever you want.
> Unfortunately, I would have to build Plone when that is about all I  
> need.
>
>    If one only needs the functionality provided by Zope, Seaside is  
> the better choice for
> me. I would be more productive and have a better end product with  
> Seaside versus Zope.
>
>    I chose painful but functional Plone over Fun but less  
> functional Pier based on the
> Plone Workflow model, ArchGenXML for products and many available  
> 3rd party products.
>
>    I really hope to be able to make a different choice in the future.
>
>
>
>
> Taun
>
>
> Jimmie Houchin wrote:
>> Hello,
>>
>> I am exploring my options for a couple of websites. I am a big fan of
>> Squeak, Seaside and Pier. But I have been exploring and learning  
>> about
>> Plone.
>>
>> There is a tremendous amount of energy and work behind Plone and
>> products available for it. Plone 3.0 is about to be released. They  
>> have
>> put a lot of work into UI usability and into a lot of areas which are
>> nice to not necessarily implement yourself.
>>
>> That said, I don't really have a grasp as to how far Pier is from
>> Plone-like abilities for both the end user and the developer.
>>
>> One thing I am looking at is the Plone4Artist product which  
>> implements a
>> very sweet package for multimedia and Plone.
>> http://www.plone4artists.com/
>>
>> I do very much like Squeak over Python.
>> In general I like Seaside/Pier over Zope/Plone.
>>
>> I know that Lukas has used Zope/Plone, but I don't know how  
>> current his
>> knowledge is.
>>
>> How far is Pier from offering a competitive experience to end  
>> users and
>> developers with regard to something like Plone?
>>
>> Is there any interest in the community of competing in that sphere?
>>
>> Or do most here happily use Seaside/Pier as a toolbox and the bigger
>> application type CMS isn't on the radar?
>>
>> And initially I speak as to a basic Plone 3.0 out-of-the-box
>> capabilities and user/developer experience. Not necessarily the  
>> plethora
>> of add-on products. Those can come later if the out-of-the-box
>> experience is compelling enough.
>>
>> Naively it seems that content types are easier to develop in Pier as
>> opposed to Plone.
>>
>> I just don't know how much I could build on Pier towards a Plone-like
>> experience in the time it takes to do some things in Plone. And I  
>> really
>> don't know how much of the machinery that makes Plone what it is,  
>> is in
>> Pier. Or the effort to get there.
>>
>> Comments regarding my questions above or most anything relevant to  
>> Pier
>> v. Plone 3.0 are greatly appreciated.
>>
>> Thanks.
>>
>> Jimmie
>>
>> _______________________________________________
>> SmallWiki, Magritte, Pier and Related Tools ...
>> https://www.iam.unibe.ch/mailman/listinfo/smallwiki
>
>
> _______________________________________________
> SmallWiki, Magritte, Pier and Related Tools ...
> https://www.iam.unibe.ch/mailman/listinfo/smallwiki


From franzl at fjkonrad.de  Wed Aug 15 12:09:09 2007
From: franzl at fjkonrad.de (Franz Josef)
Date: Wed, 15 Aug 2007 12:09:09 +0200
Subject: Pier v. Plone
In-Reply-To: <46C23065.9080702@moedae.com>
References: <46C21A71.9020204@cyberhaus.us> <46C23065.9080702@moedae.com>
Message-ID: <46C2D0C5.8020702@fjkonrad.de>

Some remarks to the statements of Taun.

- rg performance - I don't see performance problems with a well
configured Plone 2.5.x environment when using CacheFu with Apache. If
this isn't enough you can use Squid or Varnish. A prerequisite is of
course enough RAM (>= 1 GB). Of course Plone is known, not being the
fastest.
- rg Software Stack - with 2.5.3 and the upcoming 3.0 release you have
the universal installer that takes care about all the dependencies of
your libraries. Anyway, I never run in problems on Debian based systems
with simply install with apt-get what you read in the install files.

I don't see that Pier can reach the the functionality that Plone
delivers today in near or mid future. If you make a list with
requirements what you need for your app and Plone covers most of them
*out of the box* then go with Plone. If you have, let's say 50% what is
not covered out of the box then it's a different picture. Of course you
can do everything with Plone and some developers do, but this doesn't
make always sense.

My idea is to use *both*. Plone and Seaside. Use Plone for everything
that's easy to use out of the box and use Seaside for requirements that
would need special python programming. It should be possible to bring
both together with either http://plone.org/products/htmlproxy (a proxy
based integration) or http://plone.org/products/windowz/releases (iFrame
based integration).
I haven't tried this approach but it's on my todo list.

regards,
Franz Josef



Taun schrieb:
> Jimmie, I have been agonizing over the same choice for months. For now, I am sticking with
> Plone for my clients.
>
> I have worked with Smalltalk since the late 80's and Plone since it's inception on top of
> CMF for Zope.
>
> Plone Cons -
>
>    Past performance - Historically, I have had major performance problems with Plone
> including memory leaks leading to repeated server restarts on an almost hourly basis under
> heavy loads. Performance has improved but I am still concerned for the site I am currently
> creating using Plone.
>
>    Software Stack - This is my biggest complaint. The Plone stack is too dependent on C
> libraries and python modules and zope adapters. For example, connecting Plone to postresql
> requires a postgresql adapter which consists of a stack in C, python and Zope. Like most
> OpenSource, all the versions have to be compatible. Just the out of the box stack is
> python C virtual machine binary -> python libraries -> Zope -> CMF -> Plone -> Plone
> Standard Products. Addon products such as PIL (imaging library) add a 3 layer stack of C
> module, python module and Zope module.
>
>    Development environment doesn't hold a candle to Seaside.
>
> Plone Pros -
>
>    Very active community.
>    Tons of interesting Products.
>    Workflow and security model very easy to customise.
>    Easy to add custom products with ArchGenXML and ArgoUML.
>    Very Easy to add custom content types with ArchGenXML and ArgoUML.
>    **Visual Tools like ArgoUML/ArchGenXML are what are really missing from Squeak. You use
> ArgoUML to create a UML class diagram for your product or content type. You then run
> ArchGenXML to create a Plone product from the class diagram. I am currently using this
> process to create a site with 8 custom content types with cross references and it is very
> easy.
> http://plone.org/documentation/tutorial/archgenxml-getting-started
>
>
> Squeak Cons -
>
>    The flip side of all of the Plone Pros.
>
> Squeak Pros -
>
>    I much prefer Smalltalk to Python.
>    I prefer the Smalltalk environment to any Python IDE.
>    I think people write better Smalltalk tools/objects than Python programs.
>    I prefer the architecture and tools of Seaside to Plone.
>    Much shorter stack - Smalltalk VM -> Smalltalk source -> Seaside -> Magritte -> Pier.
>    Fewer C module dependencies for add-on products/objects.
>
>
> Personal Conclusion -
>
>    Zope/Plone feels like a framework where you are locked into the frame.
>    Smalltalk/Seaside feels like a great tool box to help you build whatever you want.
> Unfortunately, I would have to build Plone when that is about all I need.
>
>    If one only needs the functionality provided by Zope, Seaside is the better choice for
> me. I would be more productive and have a better end product with Seaside versus Zope.
>
>    I chose painful but functional Plone over Fun but less functional Pier based on the
> Plone Workflow model, ArchGenXML for products and many available 3rd party products.
>
>    I really hope to be able to make a different choice in the future.
>
>
>
>
> Taun
>
>
> Jimmie Houchin wrote:
>   
>> Hello,
>>
>> I am exploring my options for a couple of websites. I am a big fan of
>> Squeak, Seaside and Pier. But I have been exploring and learning about
>> Plone.
>>
>> There is a tremendous amount of energy and work behind Plone and
>> products available for it. Plone 3.0 is about to be released. They have
>> put a lot of work into UI usability and into a lot of areas which are
>> nice to not necessarily implement yourself.
>>
>> That said, I don't really have a grasp as to how far Pier is from
>> Plone-like abilities for both the end user and the developer.
>>
>> One thing I am looking at is the Plone4Artist product which implements a
>> very sweet package for multimedia and Plone.
>> http://www.plone4artists.com/
>>
>> I do very much like Squeak over Python.
>> In general I like Seaside/Pier over Zope/Plone.
>>
>> I know that Lukas has used Zope/Plone, but I don't know how current his
>> knowledge is.
>>
>> How far is Pier from offering a competitive experience to end users and
>> developers with regard to something like Plone?
>>
>> Is there any interest in the community of competing in that sphere?
>>
>> Or do most here happily use Seaside/Pier as a toolbox and the bigger
>> application type CMS isn't on the radar?
>>
>> And initially I speak as to a basic Plone 3.0 out-of-the-box
>> capabilities and user/developer experience. Not necessarily the plethora
>> of add-on products. Those can come later if the out-of-the-box
>> experience is compelling enough.
>>
>> Naively it seems that content types are easier to develop in Pier as
>> opposed to Plone.
>>
>> I just don't know how much I could build on Pier towards a Plone-like
>> experience in the time it takes to do some things in Plone. And I really
>> don't know how much of the machinery that makes Plone what it is, is in
>> Pier. Or the effort to get there.
>>
>> Comments regarding my questions above or most anything relevant to Pier
>> v. Plone 3.0 are greatly appreciated.
>>
>> Thanks.
>>
>> Jimmie
>>
>> _______________________________________________
>> SmallWiki, Magritte, Pier and Related Tools ...
>> https://www.iam.unibe.ch/mailman/listinfo/smallwiki
>>     
>
>
> _______________________________________________
> SmallWiki, Magritte, Pier and Related Tools ...
> https://www.iam.unibe.ch/mailman/listinfo/smallwiki
>   


From renggli at iam.unibe.ch  Wed Aug 15 13:37:16 2007
From: renggli at iam.unibe.ch (Lukas Renggli)
Date: Wed, 15 Aug 2007 13:37:16 +0200
Subject: Pier v. Plone
In-Reply-To: <46C2D0C5.8020702@fjkonrad.de>
References: <46C21A71.9020204@cyberhaus.us> <46C23065.9080702@moedae.com>
	<46C2D0C5.8020702@fjkonrad.de>
Message-ID: <E0845684-C2D6-4837-9B83-71D804F15F7E@iam.unibe.ch>

> My idea is to use *both*. Plone and Seaside. Use Plone for everything
> that's easy to use out of the box and use Seaside for requirements  
> that
> would need special python programming. It should be possible to bring
> both together with either http://plone.org/products/htmlproxy (a proxy
> based integration) or http://plone.org/products/windowz/releases  
> (iFrame
> based integration). I haven't tried this approach but it's on my  
> todo list.

I've done several proxy based integrations with Zope and Seaside.  
This works nicely and is completely transparent, as both application  
servers are mostly run behind Apache anyway. A visitor can't tell the  
difference of the Zope and Seaside parts, except by looking at the  
URL parameters.

Eventually I moved everything to SmallWiki (the Pier at that time).

> I don't see that Pier can reach the the functionality that Plone
> delivers today in near or mid future.

What are the missing parts in Seaside (vs. Zope) and Pier (vs. Plone)  
then?

Lukas

-- 
Lukas Renggli
http://www.lukas-renggli.ch


From matthias.berth at googlemail.com  Wed Aug 15 18:45:55 2007
From: matthias.berth at googlemail.com (Matthias Berth)
Date: Wed, 15 Aug 2007 18:45:55 +0200
Subject: PRCommandsWidget: problem with collecting available commands?
Message-ID: <9aa6ae270708150945i50a5b214kee4e9e5321d9a75f@mail.gmail.com>

Hello,

I have made a custom subclass of PULogin which will direct users to
their homepage after they login. The Commands widget in Pier then
shows both PULogin and my CustomLogin command. I'd like to hide
PULogin, so I thought I make PULogin invalid (isValidIn: aContext
returns false). Problem is that now I get CustomLogin listed twice in
the widget.

Here is the method, with some comments about what I think is going on

PRommandsWidget>>items
	| commands command |
	commands := self context commands. "These are all allowed commands"
	^ Array
		streamContents: [:stream | self commandClasses
				"These are all possible commands"
				do: [:class |
					command := commands
								detect: [:each | each = class
										or: [each inheritsFrom: class]]
								ifNone: [].
					"command is an allowed command that is a subclass of any of
								the command classes."
					command isNil
						ifFalse: [stream nextPut: command]]].

So it iterates over all possible commands (subclasses of PRCommand)
and adds allowed commands if they inherit from one of them.

My questions are:

- What's the point of the inheritance check?

- Shouldn't the list of commands in the widget be a subset of the
commands allowed in the context?

- Why shouldnt the widget just show the commands that the context allows?

I'd like to understand this before I patch it :-)

Cheers

Matthias

From j.squeak at cyberhaus.us  Wed Aug 15 20:54:47 2007
From: j.squeak at cyberhaus.us (Jimmie Houchin)
Date: Wed, 15 Aug 2007 13:54:47 -0500
Subject: Pier v. Plone
In-Reply-To: <46C23065.9080702@moedae.com>
References: <46C21A71.9020204@cyberhaus.us> <46C23065.9080702@moedae.com>
Message-ID: <46C34BF7.5040309@cyberhaus.us>

Hello Taun,

Thanks for the reply.

Taun wrote:
> Jimmie, I have been agonizing over the same choice for months. For now, I am sticking with
> Plone for my clients.

Me too. That's why I decided rather than just jumping into full bore 
with Zope/Plone purely do to the size of community and a great end-user 
experience to post a question here. I am glad I did.

> I have worked with Smalltalk since the late 80's and Plone since it's inception on top of
> CMF for Zope.
> 
> Plone Cons -
>    Past performance - [snip]
>    Software Stack - [snip]
>
>    Development environment doesn't hold a candle to Seaside.

Oh absolutely. One of my big issues when doing Python. I've never 
integrated an editor into my brain. :)

I was shocked the first time I installed Plone on my Ubuntu laptop. 
AMD64 3200, 1.25gb ram, 100gb hard disk. It took about 30 minutes to 
compile and install and was absolutely huge.

> Plone Pros -
>    Very active community.
>    Tons of interesting Products.
>    Workflow and security model very easy to customise.
>    Easy to add custom products with ArchGenXML and ArgoUML.
>    Very Easy to add custom content types with ArchGenXML and ArgoUML.
>    **Visual Tools like ArgoUML/ArchGenXML are what are really missing from Squeak. You use
> ArgoUML to create a UML class diagram for your product or content type. You then run
> ArchGenXML to create a Plone product from the class diagram. I am currently using this
> process to create a site with 8 custom content types with cross references and it is very
> easy.
> http://plone.org/documentation/tutorial/archgenxml-getting-started

Looks reasonable. But I haven't yet learned how that works with Plone 
3.0 and Zope 3 technologies. I do not know nor desire to Zope 2.


> Squeak Cons -
>    The flip side of all of the Plone Pros.

Yes. But that is why I posted. Does the community desire to over time 
work in such a direction. If core Pier could move to and become =+ Plone 
core, then we can start talking. Add-ons and products can come when the 
machinery supports the basic usability and building blocks equal to and 
better than Plone.

> Squeak Pros -
>    I much prefer Smalltalk to Python.
>    I prefer the Smalltalk environment to any Python IDE.
>    I think people write better Smalltalk tools/objects than Python programs.
>    I prefer the architecture and tools of Seaside to Plone.
>    Much shorter stack - Smalltalk VM -> Smalltalk source -> Seaside -> Magritte -> Pier.
>    Fewer C module dependencies for add-on products/objects.

Oh absolutely. I am working on website that I have to migrate to a 
machine not under my control. Most likely Windows. I am a little 
apprehensive about having him install Plone and me trying to 
support/develop Plone/Python on Windows. I am not a Windows guy.

> Personal Conclusion -
>    Zope/Plone feels like a framework where you are locked into the frame.
>    Smalltalk/Seaside feels like a great tool box to help you build whatever you want.
> Unfortunately, I would have to build Plone when that is about all I need.
> 
>    If one only needs the functionality provided by Zope, Seaside is the better choice for
> me. I would be more productive and have a better end product with Seaside versus Zope.

Cool.

>    I chose painful but functional Plone over Fun but less functional Pier based on the
> Plone Workflow model, ArchGenXML for products and many available 3rd party products.

Fortunately for me. One project I am working on is to replace the 
current website which is done in an ancient version of MS SharePoint. 
(Yuck!) So, surpassing it in usability is a fairly low target. And I 
have the time to make it better after an initial equaling(or better) of 
the capabilities.

The other project I am working on is my own. And for it I am willing to 
work on improving Pier where it may need it, if in the end overall time 
spent is equal to or less than a Plone version.

And for me that time includes, learning Python more sufficiently to get 
my hands dirty in the Python side of Zope3/Plone3. Learn Zope3, Learn 
Plone3. And develop both sites.

The learning curve for me is an advantage to Squeak/Seaside/Pier.

So I think a good question is...

If one is starting from ground zero. An understanding of programming and 
maybe some web development. But no mastery of Python or Smalltalk or 
there libraries. If the project requires doing Python or Smalltalk 
programming in order to fully use the toolset or framework. (ie: not 
simply a user of the application Plone/Pier).
Then which one gets a moderate website up and running the 
fastest/easiest?   I know, very subjective.

> I really hope to be able to make a different choice in the future.

Me too. That's why I wanted to give Pier an opportunity before I stuck 
my head into Plone and disappeared for an indeterminate amount of time. 
If said time could be spent being more productive with better tools and 
improving those tools and accomplishing the goal in similar time frame. 
Then I am all Squeak/Seaside/Pier.

Thanks again for the reply.

Jimmie


[snip original message]

From renggli at iam.unibe.ch  Wed Aug 15 21:03:47 2007
From: renggli at iam.unibe.ch (Lukas Renggli)
Date: Wed, 15 Aug 2007 21:03:47 +0200
Subject: PRCommandsWidget: problem with collecting available commands?
In-Reply-To: <9aa6ae270708150945i50a5b214kee4e9e5321d9a75f@mail.gmail.com>
References: <9aa6ae270708150945i50a5b214kee4e9e5321d9a75f@mail.gmail.com>
Message-ID: <BA61072C-8053-42A6-A568-C20AEC04BD18@iam.unibe.ch>

> - What's the point of the inheritance check?

You are right, the inheritance check is completely useless in the  
current version of Pier. A few years ago the commands to add and edit  
different structure classes were different subclasses of an abstract  
add and edit command. Luckily this is no longer the case and therefor  
this method can be simplified to:

PRCommandsWidget>>items
	| commands |
	commands := self context commands.
	^ self commandClasses select: [ :each | commands includes: each ]

I applied the same refactoring to PRViewsWidget which does the same  
unnecessary and complicated checks.

> - Shouldn't the list of commands in the widget be a subset of the
> commands allowed in the context?

Exactly, this is what the new code is doing. It takes the selected  
classes in the selected order, and then removes all that are not  
useable in the current context.

> - Why shouldnt the widget just show the commands that the context  
> allows?

To allow people to customize (see the settings of the widget) what  
commands to show in what order. For example by using two different  
PRCommandsWidget's you can logically split commands like Login/Logout  
and Add/Edit/Remove page from each other.

Name: Pier-Seaside-lr.191
Author: lr
Time: 15 August 2007, 9:03:35 pm
UUID: 86949e29-2ca2-438d-bb37-9e8faa2cd905
Ancestors: Pier-Seaside-lr.189

- Fixed PRCommandsWidget>>items and PRViewsWidget>>items that was  
doing far too much expecting a different command model that has been  
dropped from pier a long time ago
- Kudos to Matthias Berth that reported this problem

Cheers,
Lukas

-- 
Lukas Renggli
http://www.lukas-renggli.ch


From j.squeak at cyberhaus.us  Wed Aug 15 22:03:21 2007
From: j.squeak at cyberhaus.us (Jimmie Houchin)
Date: Wed, 15 Aug 2007 15:03:21 -0500
Subject: Pier v. Plone
In-Reply-To: <1AFDAFB5-43E5-40D6-B807-6DA0AAC6F03C@iam.unibe.ch>
References: <46C21A71.9020204@cyberhaus.us>
	<1AFDAFB5-43E5-40D6-B807-6DA0AAC6F03C@iam.unibe.ch>
Message-ID: <46C35C09.50508@cyberhaus.us>

Lukas Renggli wrote:
>> I know that Lukas has used Zope/Plone, but I don't know how current  
>> his knowledge is.
> 
> My knowledge of Zope and Plone dates back to 2002. I maintained some  
> legacy Zope projects in 2003, but as I discovered Smalltalk and  
> Seaside around the same time I never went back to Python again.

I can certainly understand that.

>> How far is Pier from offering a competitive experience to end users  
>> and developers with regard to something like Plone?
> 
> I have no idea in what state Zope/Plone is today. Maybe I should give  
> another try ...

Only for comparison, please. :)

>> Is there any interest in the community of competing in that sphere?
> 
> I am interested, but I can't compete against hundreds of companies  
> and thousands of developers all alone.

Super! I understand you can't compete alone. But possibly if we started 
an issue tracker with things that the community would like to see in a 
Pier distribution so that we have a target to work towards. Then little 
by little we can improve Pier to where Pier can be a top contender in 
the CMS marketplace. At least so that those of who choose Smalltalk can 
use the tool of our choice and be able to justify it to any PHB as 
necessary. :)

>> Or do most here happily use Seaside/Pier as a toolbox and the bigger
>> application type CMS isn't on the radar?
> 
> I am using Seaside/Pier for all my web related projects.

As I thought.

>> And initially I speak as to a basic Plone 3.0 out-of-the-box
>> capabilities and user/developer experience. Not necessarily the  
>> plethora of add-on products. Those can come later if the 
 >> out-of-the-box experience is compelling enough.
> 
> To make the out-of-the-box experience of Pier (especially for those  
> not knowing Smalltalk and Seaside) quite some work is required.

It would be nice to start learning what those areas are from those 
experienced in both. Hopefully start adding them to the above issue tracker.

>> Naively it seems that content types are easier to develop in Pier as
>> opposed to Plone.
> 
> This is certainly the case, even if Zope/Plone improved a lot over  
> the past 5 years.

I was watching a presentation on b-org: Creating content types the Plone 
2.5 way. 
http://video.google.com/videoplay?docid=-4512976899640436688&q=plone+duration%3Along&total=16&start=0&num=100&so=1&type=search&plindex=8
It seemed the methodology he was using was also applicable to Plone 3.

One guy during the presentation spoke of having to touch 16 files in 
order to create content types following this model which was spoken of 
as being a best practice.

I would say as far as easily spying out the land, look at the 
"marketing" type materials on Plone.org. Movies.
http://plone.org/about/movies

Most of these movies are short. You could probably watch them in a 
window on one side of your monitor, and do something else on the other.

I would initially skip the Better Web Dev. and the Recovery from 
Addiction unless you want to see how Plone compares to Java,... etc.

Other videos from the 2006 Plone conference.
http://plone.org/events/conferences/seattle-2006/presentations/session-videos

These are longer, but if you have a big enough monitor. You can still 
watch and get something done. :)

The video by Eben Moglen was excellent and not specific to Plone.
The video   What If Our Systems Could do the Rest? - Bob Boiko
was excellent for anybody doing CMS systems. Had nothing specific to Plone.

And then there is the laundry list of features in Plone 3.
http://plone.org/products/plone/features/3.0

That would give you some easy low lying fruit, just to see what people 
exploring the CMS marketplace see when looking at Plone.

How does Pier compare when looking at these items?
What are those things that are appropriate for Seaside/Pier and we would 
like to see in Pier?
What in this list are the low lying fruit?
   Reasonably short time to implement.

Lets start creating a target to shoot for.
What would be good projects for Google SoC?
What about those things that Esug sponsors?
What about students in general?

What can we do to enable and give direction to anybody who would like to 
participate in improving the out-of-box Pier experience. What defines 
Pier core? What do we want in a rich Pier distribution?

So I think the place to start is to see how Pier compares.
Then create a roadmap to get to where we want to be.
Enable the community to step up to the plate and contribute.

We can also pay attention to where we differ philosophically and where 
we can exploit that difference. ie: the web heresy type stuff.

Just some thoughts.

Thanks for your time and for your contribution to the Squeak community. 
It is appreciated.

Jimmie Houchin

From j.squeak at cyberhaus.us  Wed Aug 15 22:21:16 2007
From: j.squeak at cyberhaus.us (Jimmie Houchin)
Date: Wed, 15 Aug 2007 15:21:16 -0500
Subject: Pier v. Plone
In-Reply-To: <46C2D0C5.8020702@fjkonrad.de>
References: <46C21A71.9020204@cyberhaus.us> <46C23065.9080702@moedae.com>
	<46C2D0C5.8020702@fjkonrad.de>
Message-ID: <46C3603C.3020104@cyberhaus.us>

Franz Josef wrote:
> Some remarks to the statements of Taun.
[snip quote]
> I don't see that Pier can reach the the functionality that Plone
> delivers today in near or mid future. If you make a list with
> requirements what you need for your app and Plone covers most of them
> *out of the box* then go with Plone. If you have, let's say 50% what is
> not covered out of the box then it's a different picture. Of course you
> can do everything with Plone and some developers do, but this doesn't
> make always sense.

My projects definitely require Python or Smalltalk programming.

So it seems your saying that if Plone the app can do what you want, then 
use Plone. Because Plone for the user is fine. But if you have to put on 
your developer hat and get dirty, then you are on a sliding scale on 
which at some point things can tip over to Squeak/Seaside/Pier. Correct?

But this is what I was looking for. It seems that I could spend a fair 
amount of time in Zope/Plone creating and editing a plethora of files 
and that I could spend far less time doing similar programming in 
Squeak. Spend the difference on developing the things in Pier I might 
want that Plone has available. Over time that would continue lower the 
point of enter for Seaside/Pier on the above sliding scale. :)

It might not be the near or mid future, but if we create a target and 
start working towards it. It could eventually come to pass that Pier is 
competitive at certain levels and a winner on several. Just depends on 
what your looking for.

I think that the productivity in Squeak/Seaside/Pier could enable us to 
intersect with and compare nicely in the CMS marketplace at some point 
in the future. The Plone people are already making their list for what 
they want in Plone 3.5.

> My idea is to use *both*. Plone and Seaside. Use Plone for everything
> that's easy to use out of the box and use Seaside for requirements that
> would need special python programming. It should be possible to bring
> both together with either http://plone.org/products/htmlproxy (a proxy
> based integration) or http://plone.org/products/windowz/releases (iFrame
> based integration).

Okay. Here your talking about using Plone when you can be a user, and 
using Pier when you need to be a developer. I can see some use cases 
where this could be possible. But I would have to think about it.

Seems like it could become complicated if you have a user/member 
database, authentication and such. Managing permissions to content 
objects across two content systems seems potentially problematic.

Would require some thought.

Thanks for the reply.

Jimmie

[big snip]

From rbb at techgame.net  Wed Aug 15 22:38:15 2007
From: rbb at techgame.net (Brian Brown)
Date: Wed, 15 Aug 2007 14:38:15 -0600
Subject: Pier v. Plone
In-Reply-To: <46C34BF7.5040309@cyberhaus.us>
References: <46C21A71.9020204@cyberhaus.us> <46C23065.9080702@moedae.com>
	<46C34BF7.5040309@cyberhaus.us>
Message-ID: <C92A1BBD-64FC-4117-B10F-A79E83D3CD9F@techgame.net>


On Aug 15, 2007, at 12:54 PM, Jimmie Houchin wrote:

> Hello Taun,
>
> Thanks for the reply.
>
> Taun wrote:
>> Jimmie, I have been agonizing over the same choice for months. For  
>> now, I am sticking with
>> Plone for my clients.
>
> Me too. That's why I decided rather than just jumping into full bore
> with Zope/Plone purely do to the size of community and a great end- 
> user
> experience to post a question here. I am glad I did.
>

I couldn't resist any longer ;)

These are my opinions from my experience, so take them with the  
proverbial grain of salt :)


My company is a Python shop - we have many commercial products on the  
desktop that are all written in Python and it has served us well...  
That being said, all our web development work is in ... Seaside


I have a lot of experience with Zope, I started using it when it was  
a commercial product known as Principia. I built several sites out of  
it in the 2.x timeframe, and have delved a bit into 3.0 to see how  
things had progressed.

The things that killed me over time were some of the main "selling  
points". With  Zope/Plone there is a plethora of third party modules  
and or plugins you can use, with varying levels of configurability or  
extensibility. I used several of these types of Zope plugins and was  
usually bit when the zope platform upgraded and they no longer  
worked, or were not backward compatible because of ZODB being  
underneath everything. I know that Plone mitigates some that risk;  
the Plone layer itself is dealing with it.

I intensely disliked that fact that I could create "products", mostly  
strait python code, use a templating language (ZPT or others), and  
also edit things through the web using the Zope interface. At first  
that all seemed great, but it just ate up brain cycles trying to get  
everything to talk together. Zope has a complicated model and getting  
to what you need can be a pain. Again, Plone may be easier as a layer  
on top.

In Seaside, I write in Smalltalk, whether I am using Magritte, Pier,  
ShoreComponents, Scriptaculous, etc; and a BIG plus for me: since  
this is a smaller community, I can actually hire the developers of  
these technologies to customize, implement, or train and I really  
know what I  am getting. It's not so easy in larger communities.

I am someone who has been using Seaside off and on since 2003, and  
did not come from an OO background, so have had significant hurdles  
in doing things in OOish ways. I can still create just about any kind  
of web app I want and do it much, much faster than any of the other  
ways I have used, including ASP, PHP, cherrypy, albatross, cold  
fusion, etc. No one else might enjoy my code, however ;)

Anyway, my vote is go Pier and Seaside, but your mileage may vary :)

Cheers,

Brian




>> I have worked with Smalltalk since the late 80's and Plone since  
>> it's inception on top of
>> CMF for Zope.
>>
>> Plone Cons -
>>    Past performance - [snip]
>>    Software Stack - [snip]
>>
>>    Development environment doesn't hold a candle to Seaside.
>
> Oh absolutely. One of my big issues when doing Python. I've never
> integrated an editor into my brain. :)
>
> I was shocked the first time I installed Plone on my Ubuntu laptop.
> AMD64 3200, 1.25gb ram, 100gb hard disk. It took about 30 minutes to
> compile and install and was absolutely huge.
>
>> Plone Pros -
>>    Very active community.
>>    Tons of interesting Products.
>>    Workflow and security model very easy to customise.
>>    Easy to add custom products with ArchGenXML and ArgoUML.
>>    Very Easy to add custom content types with ArchGenXML and ArgoUML.
>>    **Visual Tools like ArgoUML/ArchGenXML are what are really  
>> missing from Squeak. You use
>> ArgoUML to create a UML class diagram for your product or content  
>> type. You then run
>> ArchGenXML to create a Plone product from the class diagram. I am  
>> currently using this
>> process to create a site with 8 custom content types with cross  
>> references and it is very
>> easy.
>> http://plone.org/documentation/tutorial/archgenxml-getting-started
>
> Looks reasonable. But I haven't yet learned how that works with Plone
> 3.0 and Zope 3 technologies. I do not know nor desire to Zope 2.
>
>
>> Squeak Cons -
>>    The flip side of all of the Plone Pros.
>
> Yes. But that is why I posted. Does the community desire to over time
> work in such a direction. If core Pier could move to and become =+  
> Plone
> core, then we can start talking. Add-ons and products can come when  
> the
> machinery supports the basic usability and building blocks equal to  
> and
> better than Plone.
>
>> Squeak Pros -
>>    I much prefer Smalltalk to Python.
>>    I prefer the Smalltalk environment to any Python IDE.
>>    I think people write better Smalltalk tools/objects than Python  
>> programs.
>>    I prefer the architecture and tools of Seaside to Plone.
>>    Much shorter stack - Smalltalk VM -> Smalltalk source ->  
>> Seaside -> Magritte -> Pier.
>>    Fewer C module dependencies for add-on products/objects.
>
> Oh absolutely. I am working on website that I have to migrate to a
> machine not under my control. Most likely Windows. I am a little
> apprehensive about having him install Plone and me trying to
> support/develop Plone/Python on Windows. I am not a Windows guy.
>
>> Personal Conclusion -
>>    Zope/Plone feels like a framework where you are locked into the  
>> frame.
>>    Smalltalk/Seaside feels like a great tool box to help you build  
>> whatever you want.
>> Unfortunately, I would have to build Plone when that is about all  
>> I need.
>>
>>    If one only needs the functionality provided by Zope, Seaside  
>> is the better choice for
>> me. I would be more productive and have a better end product with  
>> Seaside versus Zope.
>
> Cool.
>
>>    I chose painful but functional Plone over Fun but less  
>> functional Pier based on the
>> Plone Workflow model, ArchGenXML for products and many available  
>> 3rd party products.
>
> Fortunately for me. One project I am working on is to replace the
> current website which is done in an ancient version of MS SharePoint.
> (Yuck!) So, surpassing it in usability is a fairly low target. And I
> have the time to make it better after an initial equaling(or  
> better) of
> the capabilities.
>
> The other project I am working on is my own. And for it I am  
> willing to
> work on improving Pier where it may need it, if in the end overall  
> time
> spent is equal to or less than a Plone version.
>
> And for me that time includes, learning Python more sufficiently to  
> get
> my hands dirty in the Python side of Zope3/Plone3. Learn Zope3, Learn
> Plone3. And develop both sites.
>
> The learning curve for me is an advantage to Squeak/Seaside/Pier.
>
> So I think a good question is...
>
> If one is starting from ground zero. An understanding of  
> programming and
> maybe some web development. But no mastery of Python or Smalltalk or
> there libraries. If the project requires doing Python or Smalltalk
> programming in order to fully use the toolset or framework. (ie: not
> simply a user of the application Plone/Pier).
> Then which one gets a moderate website up and running the
> fastest/easiest?   I know, very subjective.
>
>> I really hope to be able to make a different choice in the future.
>
> Me too. That's why I wanted to give Pier an opportunity before I stuck
> my head into Plone and disappeared for an indeterminate amount of  
> time.
> If said time could be spent being more productive with better tools  
> and
> improving those tools and accomplishing the goal in similar time  
> frame.
> Then I am all Squeak/Seaside/Pier.
>
> Thanks again for the reply.
>
> Jimmie
>
>
> [snip original message]
>
> _______________________________________________
> SmallWiki, Magritte, Pier and Related Tools ...
> https://www.iam.unibe.ch/mailman/listinfo/smallwiki


From renggli at iam.unibe.ch  Wed Aug 15 23:36:04 2007
From: renggli at iam.unibe.ch (Lukas Renggli)
Date: Wed, 15 Aug 2007 23:36:04 +0200
Subject: Pier v. Plone
In-Reply-To: <46C35C09.50508@cyberhaus.us>
References: <46C21A71.9020204@cyberhaus.us>
	<1AFDAFB5-43E5-40D6-B807-6DA0AAC6F03C@iam.unibe.ch>
	<46C35C09.50508@cyberhaus.us>
Message-ID: <69AA5F07-7549-4B78-9E3B-14997306FB34@iam.unibe.ch>

> Super! I understand you can't compete alone. But possibly if we  
> started
> an issue tracker with things that the community would like to see in a
> Pier distribution so that we have a target to work towards. Then  
> little
> by little we can improve Pier to where Pier can be a top contender in
> the CMS marketplace. At least so that those of who choose Smalltalk  
> can
> use the tool of our choice and be able to justify it to any PHB as
> necessary. :)

Seaside, Magritte and Pier all use the bug tracker of Squeak:

	http://bugs.squeak.org

Select the project 'Squeak Package' and choose Seaside, Magritte or  
Pier as the category.

>> To make the out-of-the-box experience of Pier (especially for those
>> not knowing Smalltalk and Seaside) quite some work is required.
>
> It would be nice to start learning what those areas are from those
> experienced in both. Hopefully start adding them to the above issue  
> tracker.

Please report bugs and other issues there. Or just commit your  
enhancements to the code repositories, there is public write access  
everywhere.

> And then there is the laundry list of features in Plone 3.
> http://plone.org/products/plone/features/3.0

That's a nice feature list. I don't think we are too far away from  
that ...

> How does Pier compare when looking at these items?

Quickly scanning through the list I identify the following features  
as already present:

- LiveSearch
- Human-readable URLs
- Powerful page editor (but that could still be improved)
- Flexible navigation and always-updated site maps
- Adjustable templates on content
- Powerful standard content types
- Content is automatically formatted for printing
- Standards-compliant XHTML and CSS
- Accessibility compliant
- Pervasive RSS feed support
- Rich ecosystem of free add-on products
- Cross-platform
- Microformat support
- "Hot backup" support
- Cut/copy/paste operations on content
- Always-updated table of contents
- Portlets engine
- Wiki support
- Support for the search engine Sitemap protocol
- Support for multiple mark-up formats
- Easy collaboration and sharing

Things like the following points are extremely easy to implement:

- Link and reference integrity checking
- Automatic previous/next navigation

All the points related to versioning require a minor re-design of  
some internal data structures. That kind of stuff was already present  
at some point in the past, but got lost in the process because I was  
not satisfied with the implementation:

- Working Copy support
- Versioning, history and reverting content
- etc.

Other stuff is a bit more complicated, but I don't see anything that  
wouldn't be possible to add within a couple of hours or days of  
development effort.

Cheers,
Lukas

-- 
Lukas Renggli
http://www.lukas-renggli.ch

From jason.johnson.081 at gmail.com  Thu Aug 16 06:13:20 2007
From: jason.johnson.081 at gmail.com (Jason Johnson)
Date: Thu, 16 Aug 2007 06:13:20 +0200
Subject: Pier v. Plone
In-Reply-To: <46C34BF7.5040309@cyberhaus.us>
References: <46C21A71.9020204@cyberhaus.us> <46C23065.9080702@moedae.com>
	<46C34BF7.5040309@cyberhaus.us>
Message-ID: <aa22f0200708152113w1cbb0819vaea99d637fa23ffc@mail.gmail.com>

On 8/15/07, Jimmie Houchin <j.squeak at cyberhaus.us> wrote:
>
> Yes. But that is why I posted. Does the community desire to over time
> work in such a direction. If core Pier could move to and become =+ Plone
> core, then we can start talking. Add-ons and products can come when the
> machinery supports the basic usability and building blocks equal to and
> better than Plone.

And don't just look at Plone.  Joomla is very nice as well with all
sorts of cool features and add ons (e.g. for the less technical
members of your web site team you can set up specific email addresses
that if they send mail to the Joomla website the mail gets turned into
an article with a template you configure to display it)

I can't imagine the answer being anything but "yes of course!".  Pier
can and should compete in the CMS world.

As was mentioned in this thread, with Joomla, and I'm sure Plone and
everything else out there, you can get all kinds of components and
load them into your site to add new features.  But there is always the
issue of trying to integrate them.  And in some cases you have to do
complicated configurations of different software stacks.

In Pier we have all the integration of a homogeneous system.  Pier can
run any component made for Pier *or* Seaside.  So if someone writes
things for base Seaside, or even some other framework under Seaside
Pier can potentially use it.  And with the Pier #call:/#answer:
scheme, as well as normal message sending, composability really can be
a matter of simply sending a message to a component to do it's job.

And for building new components, I don't know what the Joomla or Plone
frameworks look like but it's hard for me to imagine them being easier
to make something new then Seasides, i.e. simply create a normal
model-domain object and put a render method, or just a Magritte
description to display it!

I also personally think that in Seaside we have a much better chance
to avoid the configuration nightmare of some of these things.  No
config files!  One should be able to configure and change anything
from the code or inside the website itself.

We could even have a component that generates apache rewrite rules for
your site. :)

From ducasse at iam.unibe.ch  Thu Aug 16 10:35:42 2007
From: ducasse at iam.unibe.ch (=?ISO-8859-1?Q?st=E9phane_ducasse?=)
Date: Thu, 16 Aug 2007 10:35:42 +0200
Subject: Pier v. Plone
In-Reply-To: <46C35C09.50508@cyberhaus.us>
References: <46C21A71.9020204@cyberhaus.us>
	<1AFDAFB5-43E5-40D6-B807-6DA0AAC6F03C@iam.unibe.ch>
	<46C35C09.50508@cyberhaus.us>
Message-ID: <F7C0B6F5-EA1C-4C91-A414-18C8ECCAE07D@iam.unibe.ch>

good idea.

Stef

On 15 ao?t 07, at 22:03, Jimmie Houchin wrote:

> Super! I understand you can't compete alone. But possibly if we  
> started
> an issue tracker with things that the community would like to see in a
> Pier distribution so that we have a target to work towards. Then  
> little
> by little we can improve Pier to where Pier can be a top contender in
> the CMS marketplace. At least so that those of who choose Smalltalk  
> can
> use the tool of our choice and be able to justify it to any PHB as
> necessary. :)



From franzl at fjkonrad.de  Thu Aug 16 13:11:33 2007
From: franzl at fjkonrad.de (Franz Josef)
Date: Thu, 16 Aug 2007 13:11:33 +0200
Subject: Pier v. Plone
In-Reply-To: <E0845684-C2D6-4837-9B83-71D804F15F7E@iam.unibe.ch>
References: <46C21A71.9020204@cyberhaus.us>
	<46C23065.9080702@moedae.com>	<46C2D0C5.8020702@fjkonrad.de>
	<E0845684-C2D6-4837-9B83-71D804F15F7E@iam.unibe.ch>
Message-ID: <46C430E5.8060701@fjkonrad.de>

Lukas Renggli schrieb:
> What are the missing parts in Seaside (vs. Zope) and Pier (vs. Plone)  
> then?
>
> Lukas
Lukas, I only can speak for Plone not Zope. And I hope my statement
wasn?t understood as an offensive. I would love to see that I?m wrong. I
also have to correct my sentence. I had not in mind the pure
functionality but the big picture with technical and non technical
aspects (Plone vs. Pier). Sorry for not being precise enough.
So when I compare a CMS with Plone I also look at:

- Support/Community - there are 167 companies from 44 countries listed
at plone.net who offer commercial support. I know there are a lot more.
A lot of them help to improve Plone and/or write add ons.
- Maturity - see the number of websites and how prominent some of them
are that run with Plone (<http://plone.net/sites>).
- Documentation - several books are available, you will find about 250
documents alone at <http://plone.org/documentation> (useful and
comprehensive, most of them up2date (what?s not the default for a
community driven Open Source project). You can join an annual conference
(325 attendees registered 2006).

Parts that are more related to functionality where I see a gap:
- Internationalization - Plone supports over 35 languages including
right-to-left languages out of the box when it?s installed. It has a
good i18n framework and with LinguaPlone you can add the same conent in
whatever language you need.
- User Management - use Authentication Service (PAS) in Plone and manage
your users in a database, LDAP, etc. (haven?t seen something similar for
Pier).
- Availability of more than 100 add on products that can be used from an
end user (you don?t need to be a developer). Examples are full featured
web shops, message board systems, registration systems, newsletter
integration (details at <http://plone.org/products>).
- Scalability - you get it out of the box with ZEO (run 1 to n Plone
instances on one content database. Guess this is doable with Magma or
Gemstone, but not today out of the box).
- Full-text Searching - all the content is searchable even Word or
OpenOffice documents and PDF files (with add ons under Linux also for
Excel and PowerPoint and other file types).

It would be nice if someone of the Pier core developer or a more
experienced user than me (I must admit I only played around with Pier in
the web-dev-images. So perhaps I missed some important points of Pier)
could make an entry at <http://www.cmsmatrix.org/>. Then it?s easy to
compare Pier to all the other CMS.

IMHO Pier today is a powerful CMS framework for a Smalltalk/Squeak
programmer. Plone at the other side is also made to be as easy as
possible for an end user. Here I see a difference today. It would be
interesting how other people position Pier against other popular Open
Source CMS today (like Plone, Drupal, Joomla, etc.) and in what
direction they wish Pier will develop in the future.

Regards,
Franz Josef


From matthias.berth at googlemail.com  Thu Aug 16 13:37:33 2007
From: matthias.berth at googlemail.com (Matthias Berth)
Date: Thu, 16 Aug 2007 13:37:33 +0200
Subject: PRCommandsWidget: problem with collecting available commands?
In-Reply-To: <BA61072C-8053-42A6-A568-C20AEC04BD18@iam.unibe.ch>
References: <9aa6ae270708150945i50a5b214kee4e9e5321d9a75f@mail.gmail.com>
	<BA61072C-8053-42A6-A568-C20AEC04BD18@iam.unibe.ch>
Message-ID: <9aa6ae270708160437i385ded72ldf987522a6baa03e@mail.gmail.com>

Lukas,


thanks a lot, this makes it clear.

On 8/15/07, Lukas Renggli <renggli at iam.unibe.ch> wrote:
> > - What's the point of the inheritance check?
>
> You are right, the inheritance check is completely useless in the
> current version of Pier.
[...]
> > - Why shouldnt the widget just show the commands that the context
> > allows?
>
> To allow people to customize (see the settings of the widget) what
> commands to show in what order. For example by using two different
> PRCommandsWidget's you can logically split commands like Login/Logout
> and Add/Edit/Remove page from each other.

OK , got that. Where are the settings of the widget? Are they
accessible through the web somehow?

Matthias

From matthias.berth at googlemail.com  Thu Aug 16 14:03:33 2007
From: matthias.berth at googlemail.com (Matthias Berth)
Date: Thu, 16 Aug 2007 14:03:33 +0200
Subject: The best way to extend Pier's user model?
Message-ID: <9aa6ae270708160503r2ef53d5dhd051a5ef9c7e5c9c@mail.gmail.com>

Hello,


I want to give pier users additional attributes (first name, last
name, email, etc). So far I was creating a subclass of PUUser. Would
it be better to just extend PUUser by a few Magritte descriptions on
the class side?

When I load new versions of Pier-Unix-Security from Monticello, I
guess my additional description methods would be deleted again? And if
I put them into my own category *my-customizations (with a star in the
name), monticello would leave them alone?

And, on a related note, what if I want to change a method in Pier's
core (like PULogin class>>isValidIn:), how do I avoid that the next
load from monticello gives me that old version again?


Cheers

Matthias

From j.squeak at cyberhaus.us  Thu Aug 16 16:08:54 2007
From: j.squeak at cyberhaus.us (Jimmie Houchin)
Date: Thu, 16 Aug 2007 09:08:54 -0500
Subject: Pier v. Plone
In-Reply-To: <46C430E5.8060701@fjkonrad.de>
References: <46C21A71.9020204@cyberhaus.us>	<46C23065.9080702@moedae.com>	<46C2D0C5.8020702@fjkonrad.de>	<E0845684-C2D6-4837-9B83-71D804F15F7E@iam.unibe.ch>
	<46C430E5.8060701@fjkonrad.de>
Message-ID: <46C45A76.1060604@cyberhaus.us>

Hello Franz,

Franz Josef wrote:
> Lukas Renggli schrieb:
>> What are the missing parts in Seaside (vs. Zope) and Pier (vs. Plone)  
>> then?
>>
>> Lukas
> Lukas, I only can speak for Plone not Zope. And I hope my statement
> wasn?t understood as an offensive. I would love to see that I?m wrong. I
> also have to correct my sentence. I had not in mind the pure
> functionality but the big picture with technical and non technical
> aspects (Plone vs. Pier). Sorry for not being precise enough.
> So when I compare a CMS with Plone I also look at:
> 
> - Support/Community - [snip]
> - Maturity - [snip]
> - Documentation - [snip]

Absolutely correct.

This is why I wanted to get a feel for community interest, even if that 
community is small. A small community with a target (goal) can be 
powerful. I was hoping that if Pier was close enough, that we could 
create that roadmap and work towards parity in many areas and 
superiority in others.

As Plone didn't start where they are, neither will we. We have to 
provide a compelling story for the developer and the end user. The end 
user won't necessarily care about the underlying technology. But, they 
might care about piles of dependencies. Thirty minute compile times. 
(Yes, I include Linux users) Thousands of files...  As I watched many of 
the Plone videos it is very apparent that you have to make sure you have 
the just right combinations for things to work. The right Python, Zope, 
Plone, Archetypes, ... I heard many questions of can I use x with ZopeX, 
PloneY, ProductOfQuestionZ? Its anything but simple.

All of Plone's end user experience isn't rosy. And this is an aspect 
which I anticipate will never be pretty. It is the nature of file based 
systems.

Squeak, Seaside, Pier on the other hand if we can deliver a nice end 
user experience. Will always be nicer than most when it comes to 
deployment. On the small side just install the VM, *.Sources, *.image, 
*.changes files and your are ready. You could put them all on a cd and 
run the thing. Almost nothing else can compete here. IMHO. (at least I 
haven't seen anything else)

> Parts that are more related to functionality where I see a gap:
> - Internationalization - Plone supports over 35 languages including
> right-to-left languages out of the box when it?s installed. It has a
> good i18n framework and with LinguaPlone you can add the same conent in
> whatever language you need.

I believe when we have the compelling story that we can sell, that these 
things will come. But the machinery needs to be there to enable such.

> - User Management - use Authentication Service (PAS) in Plone and manage
> your users in a database, LDAP, etc. (haven?t seen something similar for
> Pier).

Don't know.

> - Availability of more than 100 add on products that can be used from an
> end user (you don?t need to be a developer). Examples are full featured
> web shops, message board systems, registration systems, newsletter
> integration (details at <http://plone.org/products>).

This is nice but somewhat deceiving. This argument has been used by 
Windows people to beat on everybody else for years. I am not saying that 
that is what your doing. But in those 100+ add-ons we have many 
duplications. But almost nobody is going to run 3 different blog 
products, 4 different forum products, ... They'll pick the one they want 
and go on. The others are just superfluous.

What we do need is some quality components which can compete in those 
areas. The number being fewer is not significant if we can provide the 
capability and do so well.

> - Scalability - you get it out of the box with ZEO (run 1 to n Plone
> instances on one content database. Guess this is doable with Magma or
> Gemstone, but not today out of the box).

I think here the story would be that on the small end, keep it in the 
image. If you have more data, use Glorp/Postgres. Scaling with 
Glorp/Postgres would actually win a lot of people. Because there are 
lots of people who don't like, prefer or have other reasons not to use 
the ZODB. And with Glorp/Postgres you aren't fighting the way of the 
system in order to use an RDBMS which some people feel like with Zope/Plone.

And then if you want OODB, you probably don't get any better than 
Gemstone. Gemstone is officially supporting Seaside on there web product.

> - Full-text Searching - all the content is searchable even Word or
> OpenOffice documents and PDF files (with add ons under Linux also for
> Excel and PowerPoint and other file types).

Full-text searching. Definitely want that. :)
It would be nice to improve the story on doc management.

> It would be nice if someone of the Pier core developer or a more
> experienced user than me (I must admit I only played around with Pier in
> the web-dev-images. So perhaps I missed some important points of Pier)
> could make an entry at <http://www.cmsmatrix.org/>. Then it?s easy to
> compare Pier to all the other CMS.

Absolutely. Lukas or whomever might do a simple honest comparison. See 
if and what work would want to be done for a release before putting an 
entry on there. Then put it on there and work on improving our story.

> IMHO Pier today is a powerful CMS framework for a Smalltalk/Squeak
> programmer. Plone at the other side is also made to be as easy as
> possible for an end user. Here I see a difference today. It would be
> interesting how other people position Pier against other popular Open
> Source CMS today (like Plone, Drupal, Joomla, etc.) and in what
> direction they wish Pier will develop in the future.

Absolutely. That's why I keep speaking about the end use experience.

Let me clarify that just in case there is any confusion.

Pier the framework  vs.  Pier the application

Pier the framework makes the developers life easier.

Pier the application enables non-developers to take Pier and build 
something interesting without knowing Squeak, Seaside, Pier programming, 
etc. They just have to be a competent computer user.

That is the story that needs to be made compelling.

In the end if Pier the application has a great usability story, then 
developers lives are easier, because we don't have to go out of our way 
to provide those benefits to the end user. They come with the 
application. We provide the content types, business logic, whatever else 
that requires Smalltalk, Python, etc.

In the end, I believe that because of Smalltalk there is much of our 
story which nobody can compete against easily. We just need to catch up 
where we're behind. I believe that we can provide a compelling story for 
the marketplace.

Smalltalk isn't dying. :)

Its time for Smalltalk (Seaside, Pier) to provide the compelling use 
case for the marketplace. We can do this. And as we go along, we will 
pick up new co-laborers who are excited about the vision and want to be 
a part of this story.

Got get busy. :)

Jimmie



From keith_hodges at yahoo.co.uk  Thu Aug 16 17:44:09 2007
From: keith_hodges at yahoo.co.uk (Keith Hodges)
Date: Thu, 16 Aug 2007 16:44:09 +0100
Subject: Pier v. Plone
In-Reply-To: <46C45A76.1060604@cyberhaus.us>
References: <46C21A71.9020204@cyberhaus.us>	<46C23065.9080702@moedae.com>	<46C2D0C5.8020702@fjkonrad.de>	<E0845684-C2D6-4837-9B83-71D804F15F7E@iam.unibe.ch>	<46C430E5.8060701@fjkonrad.de>
	<46C45A76.1060604@cyberhaus.us>
Message-ID: <46C470C9.2070406@yahoo.co.uk>

I think we can compete -eventually.

Pier/seaside
 etc are developed with a minimalist philosophy.
Plone/Zope
etc have an everything and the kitchen sink philosophy.

If you spend time integrating all of the minimalist bits we already have 
then we may start to have a platform which can compete.

e.g 1 x Magma Server to n x Pier clients should work "out of the box". 
Its just a matter of trying it and doing the configuration.

Personally I am a kitchen sink kind of person since it takes me a lot of 
effort to see the path from A to B without concrete examples.

At present building pier/seaside apps are like building a den with lego 
rather than cardboard. You can assemble what you want out of lots of 
little pieces, and it will be quite robust. If you want a custom piece 
you take one you already have and adapt it a bit and carry on building.

With plone/zope, you take your cardboard boxes to build your den,  the 
boxes are bigger so you could build your whole den very easily, but 
adapting boxes is much harder because they are a bit more flimsily 
assembled and the underlying frameworks are not so concise. My limited 
experience of zope a few years back when I did attempt to use it for 
serious things (my ex-company corporate site uses it even today) under 
the hood is that it is awful code/design. Easy things are easy and 
slightly more difficult things are impossible.

In smalltalk/seaside everything is moderately difficult, but everything 
is doable in a moderately difficult kind of way..

When you add pier, it becomes, some things are easy you just configure 
and arrange the components, and the rest is moderately difficult, but 
still everything is doable in a moderately difficult kind of way once 
you understand it.

If you want to build a mansion out of card board boxes then, you will 
probably get going quite fast.
If you want to build a mansion out of lego its going to take some time.

best regards

Keith








From keith_hodges at yahoo.co.uk  Thu Aug 16 17:45:19 2007
From: keith_hodges at yahoo.co.uk (Keith Hodges)
Date: Thu, 16 Aug 2007 16:45:19 +0100
Subject: Another piece
Message-ID: <46C4710F.6060809@yahoo.co.uk>

I put this in the pieraddons repository, it adds some configuratble 
debugging hooks.

Name: Pier-Debugging-kph.1
Author: kph
Time: 16 August 2007, 4:26:56 pm
UUID: 207cf14d-4c0d-11dc-8162-000a95edb42a
Ancestors:

- some debugging hooks with switching in preferences

best regards

Keith

From j.squeak at cyberhaus.us  Thu Aug 16 20:42:51 2007
From: j.squeak at cyberhaus.us (Jimmie Houchin)
Date: Thu, 16 Aug 2007 13:42:51 -0500
Subject: Pier v. Plone
In-Reply-To: <46C470C9.2070406@yahoo.co.uk>
References: <46C21A71.9020204@cyberhaus.us>	<46C23065.9080702@moedae.com>	<46C2D0C5.8020702@fjkonrad.de>	<E0845684-C2D6-4837-9B83-71D804F15F7E@iam.unibe.ch>	<46C430E5.8060701@fjkonrad.de>	<46C45A76.1060604@cyberhaus.us>
	<46C470C9.2070406@yahoo.co.uk>
Message-ID: <46C49AAB.7040100@cyberhaus.us>

Keith Hodges wrote:
> I think we can compete -eventually.
> 
> Pier/seaside
>  etc are developed with a minimalist philosophy.

Is Pier?
I know it currently is somewhat minimalist right now. But I don't if 
that is due to philosophy or due to the current situation of the community.

Lukas seems willing for Pier to compete in the Plone space. But is 
incapable of doing it alone. Very understandable.

> Plone/Zope
> etc have an everything and the kitchen sink philosophy.
> 
> If you spend time integrating all of the minimalist bits we already have 
> then we may start to have a platform which can compete.
 >
> e.g 1 x Magma Server to n x Pier clients should work "out of the box". 
> Its just a matter of trying it and doing the configuration.

I have no experience with Magma. Is Magma ready to scale to millions of 
objects and 10s to 100s of gigabytes of space? I know this isn't the 
Magma list. But one of the two projects I am working on will have 
10+million objects and 10s of gb of space to start with and grow from 
there.

If it is ready. Then fantastic.

If not, we still have an excellent story in Glorp/Postgres or Gemstone.

My other project is small enough to do in image. (I wish the image 
itself was scalable to 10s of millions of objects and 10s to 100s of gb 
space. I have the disk space and many gb of ram. :)

> Personally I am a kitchen sink kind of person since it takes me a lot of 
> effort to see the path from A to B without concrete examples.

Me too. That's why I love the squeak-dev/web images.

If Pier seeks to be minimalist, which I'm still not sure, then I think 
that we need to begin to either build a platform on top Seaside/Pier or 
Seaside which can be the "full stack", "kitchen sink" providing an 
excellent application which provides an excellent end user experience 
and capabilities and provides the developer the abilities to deliver a 
richer customizable experience to their clients.

Just need to ascertain whether that is Pier. Or if that is on top of 
Pier or Seaside. Then start creating the target/roadmap so that over 
time as people and resources become available we have clearly defined 
goals and needs which can be accomplished. Then over time it can and 
will happen. But if we never have direction and everything is just hodge 
podge. It will never happen. We won't live long enough for evolution, we 
need creation. :)

> At present building pier/seaside apps are like building a den with lego 
> rather than cardboard. You can assemble what you want out of lots of 
> little pieces, and it will be quite robust. If you want a custom piece 
> you take one you already have and adapt it a bit and carry on building.
> 
> With plone/zope, you take your cardboard boxes to build your den,  the 
> boxes are bigger so you could build your whole den very easily, but 
> adapting boxes is much harder because they are a bit more flimsily 
> assembled and the underlying frameworks are not so concise. My limited 
> experience of zope a few years back when I did attempt to use it for 
> serious things (my ex-company corporate site uses it even today) under 
> the hood is that it is awful code/design. Easy things are easy and 
> slightly more difficult things are impossible.

I understand. My experience with Zope goes back to when it was Principia 
and Bobo. Long before I knew of Squeak. I'm even credited in the first 
Zope book. I helped proofread and sent back comments.

My problem was when I delved into the Zope source and I tried follow 
source back to its roots to try understand things. I ran into way too 
much magic. My brain could not contain all that. So I through my hands 
up and left Zope looking for better ways.

That is one reason I love Smalltalk/Squeak. Everything is so much more 
discoverable than searching through bazillion text files trying to make 
cognitive sense of the array of interconnecting classes and objects and 
mixins and... and... and...

> In smalltalk/seaside everything is moderately difficult, but everything 
> is doable in a moderately difficult kind of way..

But is it more difficult that doing ZPT, ZCML, .py, .txt, or whatever 
else in the plethora of files per content object.

> When you add pier, it becomes, some things are easy you just configure 
> and arrange the components, and the rest is moderately difficult, but 
> still everything is doable in a moderately difficult kind of way once 
> you understand it.
> 
> If you want to build a mansion out of card board boxes then, you will 
> probably get going quite fast.
> If you want to build a mansion out of lego its going to take some time.

Currently, to me, it seems that Plone wins when its Plone the 
application. The more you go to the framework side, then the advantage 
starts to slide over the Squeak/Seaside/Pier. I would like to see Pier 
the application be able to compete with Plone the application.

I know it will take time. But as time goes on, if we keep making 
progress, the decision "for" Pier becomes much easier. Neither Plone nor 
Pier have "arrived" yet. But in some areas Plone is further down the road.

If we want a target and a time frame. And if equaling/beating Plone on 
some of these is a legitimate goal. Then I would say we could look at 
Plone 3.5. Work towards having a compelling story vis a vis Plone 3. and 
before Plone 3.5 comes out have a compelling story towards it.

I am not necessarily proposing a time-line. But I am proposing that we 
establish processes that move us towards a goal. As processes improve, 
the product improves, the community grows and we hopefully start to 
attain critical mass to achieve what we desire. I don't believe that the 
critical mass (sustainable mass) for Squeak/Seaside/Pier is as large as 
what is required for Python/Zope/Plone or even anywhere close.

I would like to be a part of this process.

Lukas said that http://bugs.squeak.org/ is the issue tracker for Seaside 
and Pier. I haven't learned how to use that yet. But I will.

Is there a place where we establish and document processes for the 
community to create Seaside/Pier the platform? A wiki? I didn't see any 
such on seaside.st? Maybe this needs to be part of the establishing of 
processes. A place to document established processes and goals.

A few more thoughts hopefully towards becoming a productive contributing 
member of this community. I know I am not qualified for many of the core 
areas of Pier, but hopefully I can help with components/products on top 
of Pier core, documentation, etc.

Thanks to all who are participating in this dialog.

Jimmie

From renggli at iam.unibe.ch  Thu Aug 16 21:03:29 2007
From: renggli at iam.unibe.ch (Lukas Renggli)
Date: Thu, 16 Aug 2007 21:03:29 +0200
Subject: PRCommandsWidget: problem with collecting available commands?
In-Reply-To: <9aa6ae270708160437i385ded72ldf987522a6baa03e@mail.gmail.com>
References: <9aa6ae270708150945i50a5b214kee4e9e5321d9a75f@mail.gmail.com>
	<BA61072C-8053-42A6-A568-C20AEC04BD18@iam.unibe.ch>
	<9aa6ae270708160437i385ded72ldf987522a6baa03e@mail.gmail.com>
Message-ID: <9FF82DF2-9A0F-44A0-A196-79649C9D5262@iam.unibe.ch>

> On 8/15/07, Lukas Renggli <renggli at iam.unibe.ch> wrote:
>>> - What's the point of the inheritance check?
>>
>> You are right, the inheritance check is completely useless in the
>> current version of Pier.
> [...]
>>> - Why shouldnt the widget just show the commands that the context
>>> allows?
>>
>> To allow people to customize (see the settings of the widget) what
>> commands to show in what order. For example by using two different
>> PRCommandsWidget's you can logically split commands like Login/Logout
>> and Add/Edit/Remove page from each other.
>
> OK , got that. Where are the settings of the widget? Are they
> accessible through the web somehow?

There is a command called 'settings' available on all structures.  
Widgets usually have a lot of settings to customize their appearance.

Lukas

-- 
Lukas Renggli
http://www.lukas-renggli.ch


From renggli at iam.unibe.ch  Thu Aug 16 21:09:35 2007
From: renggli at iam.unibe.ch (Lukas Renggli)
Date: Thu, 16 Aug 2007 21:09:35 +0200
Subject: The best way to extend Pier's user model?
In-Reply-To: <9aa6ae270708160503r2ef53d5dhd051a5ef9c7e5c9c@mail.gmail.com>
References: <9aa6ae270708160503r2ef53d5dhd051a5ef9c7e5c9c@mail.gmail.com>
Message-ID: <A39E06E5-FB3F-4401-8E7C-3D75BE35E788@iam.unibe.ch>

> I want to give pier users additional attributes (first name, last
> name, email, etc). So far I was creating a subclass of PUUser. Would
> it be better to just extend PUUser by a few Magritte descriptions on
> the class side?

Yes, that's usually easier.

> When I load new versions of Pier-Unix-Security from Monticello, I
> guess my additional description methods would be deleted again? And if
> I put them into my own category *my-customizations (with a star in the
> name), monticello would leave them alone?

Yes, you need to package your extensions properly. If you use a  
category name as you wrote, Monticello will leave your extensions  
alone when loading a new version of Pier-Unix-Security.

> And, on a related note, what if I want to change a method in Pier's
> core (like PULogin class>>isValidIn:), how do I avoid that the next
> load from monticello gives me that old version again?

Try to avoid that at all cost. If there is really no other way you  
can use the category naming *my-customizations-override. There are  
well known drawbacks and probably bugs with this, so better avoid it  
altogether.

Check the archive or ask in the Squeak mailing-list to learn more  
about Monticello and packaging in Squeak.

Cheers,
Lukas

-- 
Lukas Renggli
http://www.lukas-renggli.ch


From renggli at iam.unibe.ch  Thu Aug 16 21:54:45 2007
From: renggli at iam.unibe.ch (Lukas Renggli)
Date: Thu, 16 Aug 2007 21:54:45 +0200
Subject: Pier v. Plone
In-Reply-To: <46C430E5.8060701@fjkonrad.de>
References: <46C21A71.9020204@cyberhaus.us>
	<46C23065.9080702@moedae.com>	<46C2D0C5.8020702@fjkonrad.de>
	<E0845684-C2D6-4837-9B83-71D804F15F7E@iam.unibe.ch>
	<46C430E5.8060701@fjkonrad.de>
Message-ID: <C3A30441-EBAF-48C6-AAB5-D7E3EABD7AA8@iam.unibe.ch>

> - Support/Community - there are 167 companies from 44 countries listed
> at plone.net who offer commercial support. I know there are a lot  
> more.
> A lot of them help to improve Plone and/or write add ons.
> - Maturity - see the number of websites and how prominent some of them
> are that run with Plone (<http://plone.net/sites>).
> - Documentation - several books are available, you will find about 250
> documents alone at <http://plone.org/documentation> (useful and
> comprehensive, most of them up2date (what?s not the default for a
> community driven Open Source project). You can join an annual  
> conference
> (325 attendees registered 2006).

I can hardly do anything about these points, other than using Pier  
for all sites I am building myself (which I already do, of course).  
This is also why Pier exists in the first place and where most  
features come from.

> Parts that are more related to functionality where I see a gap:
> - Internationalization - Plone supports over 35 languages including
> right-to-left languages out of the box when it?s installed. It has a
> good i18n framework and with LinguaPlone you can add the same  
> conent in
> whatever language you need.

If people provide translations that would be certainly a nice addition.

> - Availability of more than 100 add on products that can be used  
> from an
> end user (you don?t need to be a developer). Examples are full  
> featured
> web shops, message board systems, registration systems, newsletter
> integration (details at <http://plone.org/products>).

There is already a big number of extensions. Have a look at my  
proposal [1] for the ESUG Innovation Awards 2007 [2]. And don't  
forget to vote for Pier, this will certainly help to further improve  
it ...

[1] http://www.iam.unibe.ch/~scg/Archive/Reports/Reng07c.pdf
[2] http://www.esug.org/conferences/ 
15thinternationalsmalltalkjointconference2007/ 
innovationtechnologyawards/submissions/

> - Full-text Searching - all the content is searchable even Word or
> OpenOffice documents and PDF files (with add ons under Linux also for
> Excel and PowerPoint and other file types).

There is already a full-text search. It does not included uploaded  
files though. If somebody would implement readers for those strange  
binary formats, that would certainly be an easy thing to add to the  
search results.

Cheers,
Lukas

-- 
Lukas Renggli
http://www.lukas-renggli.ch

From keith_hodges at yahoo.co.uk  Fri Aug 17 03:38:09 2007
From: keith_hodges at yahoo.co.uk (Keith Hodges)
Date: Fri, 17 Aug 2007 02:38:09 +0100
Subject: [Enh] Pier-Shore
Message-ID: <46C4FC01.3060601@yahoo.co.uk>

Enable PierWidgets to make use of ShoreComponents.

All you have to do is configure the application to use PRPierShoreFrame 
as the root component and all those great shore components are yours!

best regards

Keith


p.s In this initial version there might be a very slight css effect from 
the shore css, but I tracked it down yet.




From keith_hodges at yahoo.co.uk  Fri Aug 17 03:48:55 2007
From: keith_hodges at yahoo.co.uk (Keith Hodges)
Date: Fri, 17 Aug 2007 02:48:55 +0100
Subject: [Enh] Pier-ExternalLogin
Message-ID: <46C4FE87.4070906@yahoo.co.uk>

A small framework to enable external validation plugins to be added and 
configured for Pier Login.

I am using this with a plugin (not included yet) which validates the 
user/pass against a MYSQL query, then it logs in this user using an 
internally defined pier-user (or as I call it a "role") as a template.

Finally I perform a second query to obtain a UserData record, and a 
potential invalidation is performed on "any other criteria",  such as an 
account expiry date.

You add your plugin by subclassing PUExternalValidationPlugin, and 
select the plugin to use in the application configuration.

Finally your plugin gets the opportunity to customize the Login Dialog's 
descriptions. This enables you to customize the presentation of the 
dialog if you add Magritte-ComponentDecoration 's. You should also be 
able to add extra fields, though I leave this as an exercise for the reader.

enjoy

Keith

From keith_hodges at yahoo.co.uk  Fri Aug 17 03:58:54 2007
From: keith_hodges at yahoo.co.uk (Keith Hodges)
Date: Fri, 17 Aug 2007 02:58:54 +0100
Subject: [Enh] PRCommandLink
Message-ID: <46C500DE.6090406@yahoo.co.uk>

PRCommandLink allows the following syntax *command:PULogin* to generate 
a link which performs that command. If the user is not authorized the 
link does not show.

Alias's are supported in the normal manner: *Register 
Here>command:PRRegisterUserCommand*
 
As an advanced example you can include command links in the dialogs of 
other commands by adding a PRComponentDecoration's to the description.

e,g.

PUMyLogin-description

       (super description)
             componentDecorations: (Array with: PRComponentDecoration);
             preAnnotation: 'If you do not have an account *Register 
Here>command:PRRegisterUserCommand*';
          yourself.

best regards

Keith



From jason.johnson.081 at gmail.com  Fri Aug 17 07:12:06 2007
From: jason.johnson.081 at gmail.com (Jason Johnson)
Date: Fri, 17 Aug 2007 07:12:06 +0200
Subject: [Enh] Pier-ExternalLogin
In-Reply-To: <46C4FE87.4070906@yahoo.co.uk>
References: <46C4FE87.4070906@yahoo.co.uk>
Message-ID: <aa22f0200708162212w134869ebydcb541f6f34ef9b3@mail.gmail.com>

If you are making a framework for this part then one thing to consider
is passwords.  Most frameworks leave this up to the user, so the user
does the simplest thing they can: they store plain text passwords.
You can tell these sites because when you forget your password and
press the password recovery link they send you your password directly
(ironically, pgp.com does this too!).

Obviously if the password is stored in plain text, then once your site
is hacked all your user's passwords will be known.  This wont make
your users happy.  So it would be nice to have some kind of password
object or something that always does some kind of one-way encryption
on the passwords (e.g. using MD5).

It would be nice as well for something to do the whole password
recovery stage as well, since this always works the same: user clicks
a link, gets an email forwarded that points to a secure site where
they can type in a new password, since passwords are not recoverable.

Just something to think about.

On 8/17/07, Keith Hodges <keith_hodges at yahoo.co.uk> wrote:
> A small framework to enable external validation plugins to be added and
> configured for Pier Login.
>
> I am using this with a plugin (not included yet) which validates the
> user/pass against a MYSQL query, then it logs in this user using an
> internally defined pier-user (or as I call it a "role") as a template.
>
> Finally I perform a second query to obtain a UserData record, and a
> potential invalidation is performed on "any other criteria",  such as an
> account expiry date.
>
> You add your plugin by subclassing PUExternalValidationPlugin, and
> select the plugin to use in the application configuration.
>
> Finally your plugin gets the opportunity to customize the Login Dialog's
> descriptions. This enables you to customize the presentation of the
> dialog if you add Magritte-ComponentDecoration 's. You should also be
> able to add extra fields, though I leave this as an exercise for the reader.
>
> enjoy
>
> Keith
>
> _______________________________________________
> SmallWiki, Magritte, Pier and Related Tools ...
> https://www.iam.unibe.ch/mailman/listinfo/smallwiki
>

From renggli at iam.unibe.ch  Fri Aug 17 07:30:19 2007
From: renggli at iam.unibe.ch (Lukas Renggli)
Date: Fri, 17 Aug 2007 07:30:19 +0200
Subject: [Enh] Pier-ExternalLogin
In-Reply-To: <aa22f0200708162212w134869ebydcb541f6f34ef9b3@mail.gmail.com>
References: <46C4FE87.4070906@yahoo.co.uk>
	<aa22f0200708162212w134869ebydcb541f6f34ef9b3@mail.gmail.com>
Message-ID: <6B94B21F-EE48-4930-8CF5-2D8FD4830404@iam.unibe.ch>

> So it would be nice to have some kind of password
> object or something that always does some kind of one-way encryption
> on the passwords (e.g. using MD5).

Pier never stored passwords as plain text. It is using SHA to  
generate a hash number that is then stored in the user object.

> It would be nice as well for something to do the whole password
> recovery stage as well, since this always works the same: user clicks
> a link, gets an email forwarded that points to a secure site where
> they can type in a new password, since passwords are not recoverable.

That's easily doable (see for example SqueakSource), the only problem  
is that a Pier User doesn't include an e-mail address right now.

Lukas

-- 
Lukas Renggli
http://www.lukas-renggli.ch


From keith_hodges at yahoo.co.uk  Fri Aug 17 07:33:14 2007
From: keith_hodges at yahoo.co.uk (Keith Hodges)
Date: Fri, 17 Aug 2007 06:33:14 +0100
Subject: [bug] Copying Component settings crosstalk
Message-ID: <46C5331A.4040904@yahoo.co.uk>

If you copy a component, it seems that the settings for both components 
are the same instance. There is no longer a ui to change the settings on 
the original (or is it the copy I am not sure which)

Keith

From renggli at iam.unibe.ch  Fri Aug 17 08:07:23 2007
From: renggli at iam.unibe.ch (Lukas Renggli)
Date: Fri, 17 Aug 2007 08:07:23 +0200
Subject: [bug] Copying Component settings crosstalk
In-Reply-To: <46C5331A.4040904@yahoo.co.uk>
References: <46C5331A.4040904@yahoo.co.uk>
Message-ID: <27FA426E-E1A5-42BA-9B15-B236CB218FB6@iam.unibe.ch>

That's a bug then. See if this helps:

Name: Pier-All-lr.228
Author: lr
Time: 17 August 2007, 8:05:13 am
UUID: 2768cfa3-da2b-4dd5-a0e9-875969788c60
Ancestors: Pier-All-lr.227
Dependencies: Pier-Model-lr.162, Pier-Tests-lr.73, Pier-Seaside-lr. 
193, Pier-Security-lr.86, Pier-Blog-lr.58

- fixed several places where copying did not really copy all the  
necessary state (thanks to Keith Hodges)

Lukas

On 17 Aug 2007, at 07:33, Keith Hodges wrote:

> If you copy a component, it seems that the settings for both  
> components
> are the same instance. There is no longer a ui to change the  
> settings on
> the original (or is it the copy I am not sure which)
>
> Keith
>
> _______________________________________________
> SmallWiki, Magritte, Pier and Related Tools ...
> https://www.iam.unibe.ch/mailman/listinfo/smallwiki

-- 
Lukas Renggli
http://www.lukas-renggli.ch


From keith_hodges at yahoo.co.uk  Fri Aug 17 08:22:57 2007
From: keith_hodges at yahoo.co.uk (Keith Hodges)
Date: Fri, 17 Aug 2007 07:22:57 +0100
Subject: [Enh] Pier-ExternalLogin
In-Reply-To: <aa22f0200708162212w134869ebydcb541f6f34ef9b3@mail.gmail.com>
References: <46C4FE87.4070906@yahoo.co.uk>
	<aa22f0200708162212w134869ebydcb541f6f34ef9b3@mail.gmail.com>
Message-ID: <46C53EC1.4030503@yahoo.co.uk>

Jason Johnson wrote:
> If you are making a framework for this part then one thing to consider
> is passwords.  Most frameworks leave this up to the user, so the user
> does the simplest thing they can: they store plain text passwords.
> You can tell these sites because when you forget your password and
> press the password recovery link they send you your password directly
> (ironically, pgp.com does this too!).
>
> Obviously if the password is stored in plain text, then once your site
> is hacked all your user's passwords will be known.  This wont make
> your users happy.  So it would be nice to have some kind of password
> object or something that always does some kind of one-way encryption
> on the passwords (e.g. using MD5).
>
> It would be nice as well for something to do the whole password
> recovery stage as well, since this always works the same: user clicks
> a link, gets an email forwarded that points to a secure site where
> they can type in a new password, since passwords are not recoverable.
>
> Just something to think about.
>
>   
Sure! In the interests of minimal configurable components, this bit only 
provides the wiring to your supplied plugin.

At present using MySql I am storing the passwords in the db 
PASSWORD("blah") format and querying them the same.
However the Cryptography package is capable of generating hash so that 
plaintext does not need to be exchanged with the server.

I haven't got to password reminding schemes, or supervised password 
resetting yet.  I need at least one scheme soon.

thanks for a "requirements statement"

Keith



From keith_hodges at yahoo.co.uk  Fri Aug 17 09:21:54 2007
From: keith_hodges at yahoo.co.uk (Keith Hodges)
Date: Fri, 17 Aug 2007 08:21:54 +0100
Subject: [bug] Copying Component settings crosstalk
In-Reply-To: <27FA426E-E1A5-42BA-9B15-B236CB218FB6@iam.unibe.ch>
References: <46C5331A.4040904@yahoo.co.uk>
	<27FA426E-E1A5-42BA-9B15-B236CB218FB6@iam.unibe.ch>
Message-ID: <46C54C92.8060600@yahoo.co.uk>

Lukas Renggli wrote:
> That's a bug then. See if this helps:
>
>   
This shows up that now if you set the settings for a component, then 
edit it and #save even with no changes the settings are
reset.

Keith

From matthias.berth at googlemail.com  Fri Aug 17 10:41:12 2007
From: matthias.berth at googlemail.com (Matthias Berth)
Date: Fri, 17 Aug 2007 10:41:12 +0200
Subject: Documenting Pier...
In-Reply-To: <763D4DA2-A6F6-481B-AA32-84E2847BEF6C@iam.unibe.ch>
References: <9aa6ae270704170458w2642ed5fy668510c66f948632@mail.gmail.com>
	<763D4DA2-A6F6-481B-AA32-84E2847BEF6C@iam.unibe.ch>
Message-ID: <9aa6ae270708170141p463f9685t6ec924309dcda000@mail.gmail.com>

Hi,

I wanted to add a FAQ to the pier wiki, but the anonymous login
(anonymous/xopu) does not work for me - I get no link to edit the
page. Back in April it worked.

Please help...

Matthias

On 4/17/07, Lukas Renggli <renggli at iam.unibe.ch> wrote:
> > I'd like to collect some snippets of documentation about Pier. Would
> >
> >  http://smallwiki.unibe.ch/smallwiki/pier/
> >
> > be the right place to do so? I was not able to figure out a login to
> > that wiki :-(
>
> The username/password is anonymous/xopu.
>
> Cheers,
> Lukas
>
> --
> Lukas Renggli
> http://www.lukas-renggli.ch
>
>
>
> _______________________________________________
> SmallWiki, Magritte, Pier and Related Tools ...
> https://www.iam.unibe.ch/mailman/listinfo/smallwiki
>

From renggli at iam.unibe.ch  Fri Aug 17 10:47:23 2007
From: renggli at iam.unibe.ch (Lukas Renggli)
Date: Fri, 17 Aug 2007 10:47:23 +0200
Subject: [bug] Copying Component settings crosstalk
In-Reply-To: <46C54C92.8060600@yahoo.co.uk>
References: <46C5331A.4040904@yahoo.co.uk>
	<27FA426E-E1A5-42BA-9B15-B236CB218FB6@iam.unibe.ch>
	<46C54C92.8060600@yahoo.co.uk>
Message-ID: <D9C71F92-CCC1-4B74-BD77-DC19E90DFB7C@iam.unibe.ch>

Ok, that's another unrelated problem. It is fixed in:

Name: Pier-All-lr.229
Author: lr
Time: 17 August 2007, 10:46:46 am
UUID: aa712c72-6578-4f26-8d00-6d8b0d7795fc
Ancestors: Pier-All-lr.228
Dependencies: Pier-Model-lr.162, Pier-Tests-lr.73, Pier-Seaside-lr. 
194, Pier-Security-lr.86, Pier-Blog-lr.58

- fixes: "This shows up that now if you set the settings for a  
component, then edit it and #save even with no changes the settings  
are reset." (thanks to Keith Hodges)


On 17 Aug 2007, at 09:21, Keith Hodges wrote:

> Lukas Renggli wrote:
>> That's a bug then. See if this helps:
>>
>>
> This shows up that now if you set the settings for a component, then
> edit it and #save even with no changes the settings are
> reset.
>
> Keith
>
> _______________________________________________
> SmallWiki, Magritte, Pier and Related Tools ...
> https://www.iam.unibe.ch/mailman/listinfo/smallwiki

-- 
Lukas Renggli
http://www.lukas-renggli.ch


From renggli at iam.unibe.ch  Fri Aug 17 10:48:36 2007
From: renggli at iam.unibe.ch (Lukas Renggli)
Date: Fri, 17 Aug 2007 10:48:36 +0200
Subject: Documenting Pier...
In-Reply-To: <9aa6ae270708170141p463f9685t6ec924309dcda000@mail.gmail.com>
References: <9aa6ae270704170458w2642ed5fy668510c66f948632@mail.gmail.com>
	<763D4DA2-A6F6-481B-AA32-84E2847BEF6C@iam.unibe.ch>
	<9aa6ae270708170141p463f9685t6ec924309dcda000@mail.gmail.com>
Message-ID: <A96D14AF-160D-4215-8046-CB975CBA5563@iam.unibe.ch>

Yes, we broke something when adding Google Analytics to SmallWiki.  
This is rather strange. I will have a look at it.

Cheers,
Lukas

On 17 Aug 2007, at 10:41, Matthias Berth wrote:

> Hi,
>
> I wanted to add a FAQ to the pier wiki, but the anonymous login
> (anonymous/xopu) does not work for me - I get no link to edit the
> page. Back in April it worked.
>
> Please help...
>
> Matthias
>
> On 4/17/07, Lukas Renggli <renggli at iam.unibe.ch> wrote:
>>> I'd like to collect some snippets of documentation about Pier. Would
>>>
>>>  http://smallwiki.unibe.ch/smallwiki/pier/
>>>
>>> be the right place to do so? I was not able to figure out a login to
>>> that wiki :-(
>>
>> The username/password is anonymous/xopu.
>>
>> Cheers,
>> Lukas
>>
>> --
>> Lukas Renggli
>> http://www.lukas-renggli.ch
>>
>>
>>
>> _______________________________________________
>> SmallWiki, Magritte, Pier and Related Tools ...
>> https://www.iam.unibe.ch/mailman/listinfo/smallwiki
>>
>
> _______________________________________________
> SmallWiki, Magritte, Pier and Related Tools ...
> https://www.iam.unibe.ch/mailman/listinfo/smallwiki

-- 
Lukas Renggli
http://www.lukas-renggli.ch


From jason.johnson.081 at gmail.com  Fri Aug 17 11:03:22 2007
From: jason.johnson.081 at gmail.com (Jason Johnson)
Date: Fri, 17 Aug 2007 11:03:22 +0200
Subject: [Enh] Pier-ExternalLogin
In-Reply-To: <6B94B21F-EE48-4930-8CF5-2D8FD4830404@iam.unibe.ch>
References: <46C4FE87.4070906@yahoo.co.uk>
	<aa22f0200708162212w134869ebydcb541f6f34ef9b3@mail.gmail.com>
	<6B94B21F-EE48-4930-8CF5-2D8FD4830404@iam.unibe.ch>
Message-ID: <aa22f0200708170203l64a56192y98b3194253471432@mail.gmail.com>

On 8/17/07, Lukas Renggli <renggli at iam.unibe.ch> wrote:
>
> Pier never stored passwords as plain text. It is using SHA to
> generate a hash number that is then stored in the user object.

Ah good to know.  Most things are nice to leave up to the user, but
things like this are so fundamental that it's best if the
default/template is already using good behavior.  Nice to hear that
Pier already is.  In my mind that is a positive for Pier over many of
these other CMSes.

> That's easily doable (see for example SqueakSource), the only problem
> is that a Pier User doesn't include an e-mail address right now.
>
> Lukas

Good point.  I was thinking more along the model point of view, something like:

PPaswordManager>>changePassword
  self confirmUserWantsChangeElse: [ ^ self ].
  passwordHash := self encrypt: (self askPassword).

The model can be displayed by the different schemes, one way would be
to use mail.  Or maybe one view would do it directly in the site, not
requiring email.

Maybe we could even use the call:/answer: system to ask the password.
That would probably work whether an HTTP get happens between, or an
email conversation, no? :)

From renggli at iam.unibe.ch  Fri Aug 17 11:27:23 2007
From: renggli at iam.unibe.ch (Lukas Renggli)
Date: Fri, 17 Aug 2007 11:27:23 +0200
Subject: Documenting Pier...
In-Reply-To: <A96D14AF-160D-4215-8046-CB975CBA5563@iam.unibe.ch>
References: <9aa6ae270704170458w2642ed5fy668510c66f948632@mail.gmail.com>
	<763D4DA2-A6F6-481B-AA32-84E2847BEF6C@iam.unibe.ch>
	<9aa6ae270708170141p463f9685t6ec924309dcda000@mail.gmail.com>
	<A96D14AF-160D-4215-8046-CB975CBA5563@iam.unibe.ch>
Message-ID: <099AA58F-73A3-4A27-983E-3A1CFAF5FB93@iam.unibe.ch>

It should work again. The Google cookies triggered a subtle bug in  
the cookie parsing code of SmallWiki.

Cheers,
Lukas


On 17 Aug 2007, at 10:48, Lukas Renggli wrote:

> Yes, we broke something when adding Google Analytics to SmallWiki.
> This is rather strange. I will have a look at it.
>
> Cheers,
> Lukas
>
> On 17 Aug 2007, at 10:41, Matthias Berth wrote:
>
>> Hi,
>>
>> I wanted to add a FAQ to the pier wiki, but the anonymous login
>> (anonymous/xopu) does not work for me - I get no link to edit the
>> page. Back in April it worked.
>>
>> Please help...
>>
>> Matthias
>>
>> On 4/17/07, Lukas Renggli <renggli at iam.unibe.ch> wrote:
>>>> I'd like to collect some snippets of documentation about Pier.  
>>>> Would
>>>>
>>>>  http://smallwiki.unibe.ch/smallwiki/pier/
>>>>
>>>> be the right place to do so? I was not able to figure out a  
>>>> login to
>>>> that wiki :-(
>>>
>>> The username/password is anonymous/xopu.
>>>
>>> Cheers,
>>> Lukas
>>>
>>> --
>>> Lukas Renggli
>>> http://www.lukas-renggli.ch
>>>
>>>
>>>
>>> _______________________________________________
>>> SmallWiki, Magritte, Pier and Related Tools ...
>>> https://www.iam.unibe.ch/mailman/listinfo/smallwiki
>>>
>>
>> _______________________________________________
>> SmallWiki, Magritte, Pier and Related Tools ...
>> https://www.iam.unibe.ch/mailman/listinfo/smallwiki
>
> -- 
> Lukas Renggli
> http://www.lukas-renggli.ch
>
>
> _______________________________________________
> SmallWiki, Magritte, Pier and Related Tools ...
> https://www.iam.unibe.ch/mailman/listinfo/smallwiki

-- 
Lukas Renggli
http://www.lukas-renggli.ch


From renggli at iam.unibe.ch  Fri Aug 17 11:32:42 2007
From: renggli at iam.unibe.ch (Lukas Renggli)
Date: Fri, 17 Aug 2007 11:32:42 +0200
Subject: [Enh] Pier-ExternalLogin
In-Reply-To: <aa22f0200708170203l64a56192y98b3194253471432@mail.gmail.com>
References: <46C4FE87.4070906@yahoo.co.uk>
	<aa22f0200708162212w134869ebydcb541f6f34ef9b3@mail.gmail.com>
	<6B94B21F-EE48-4930-8CF5-2D8FD4830404@iam.unibe.ch>
	<aa22f0200708170203l64a56192y98b3194253471432@mail.gmail.com>
Message-ID: <94615F98-78B6-4035-9F11-3EC0DB21EA05@iam.unibe.ch>

> Good point.  I was thinking more along the model point of view,  
> something like:
>
> PPaswordManager>>changePassword
>   self confirmUserWantsChangeElse: [ ^ self ].
>   passwordHash := self encrypt: (self askPassword).
>
> The model can be displayed by the different schemes, one way would be
> to use mail.  Or maybe one view would do it directly in the site, not
> requiring email.
>
> Maybe we could even use the call:/answer: system to ask the password.
> That would probably work whether an HTTP get happens between, or an
> email conversation, no? :)

What I do and what is very easy to do is to create a secret  
continuation key that does the login of the requested user. This URL  
is not displayed in the browser, but sent by mail. Clicking on the  
URL enables the user to login and to change its password. The  
potentially dangerous link eventually  expires when Seaside cleans up  
the Session.

Lukas

-- 
Lukas Renggli
http://www.lukas-renggli.ch


From matthias.berth at googlemail.com  Fri Aug 17 14:57:00 2007
From: matthias.berth at googlemail.com (Matthias Berth)
Date: Fri, 17 Aug 2007 14:57:00 +0200
Subject: The best way to extend Pier's user model?
In-Reply-To: <A39E06E5-FB3F-4401-8E7C-3D75BE35E788@iam.unibe.ch>
References: <9aa6ae270708160503r2ef53d5dhd051a5ef9c7e5c9c@mail.gmail.com>
	<A39E06E5-FB3F-4401-8E7C-3D75BE35E788@iam.unibe.ch>
Message-ID: <9aa6ae270708170557t1c497420kda4262309137186c@mail.gmail.com>

Thanks a lot, Lukas. I've put a note on extending the user model in
the FAQs on the wiki
http://smallwiki.unibe.ch/smallwiki/pier/pierunixsecurity/

Matthias

From cappadonza at hotmail.com  Fri Aug 17 17:48:42 2007
From: cappadonza at hotmail.com (isaiah)
Date: Fri, 17 Aug 2007 08:48:42 -0700 (PDT)
Subject: Error in PUUsersWidget
Message-ID: <12202036.post@talk.nabble.com>


hey i got the latest version of pier, when i added a system users component
to a page, when i click on any thing on the component which invokes a
callback. i get the following error 
"WAComponentsNotFoundError: Components not found while processing callbacks:
an Array(a PUUsersWidget)" .
any ideas how to resolve this
-- 
View this message in context: http://www.nabble.com/Error-in-PUUsersWidget-tf4286525.html#a12202036
Sent from the SmallWiki mailing list archive at Nabble.com.


From matthias.berth at googlemail.com  Fri Aug 17 18:20:58 2007
From: matthias.berth at googlemail.com (Matthias Berth)
Date: Fri, 17 Aug 2007 18:20:58 +0200
Subject: Error in PUUsersWidget
In-Reply-To: <12202036.post@talk.nabble.com>
References: <12202036.post@talk.nabble.com>
Message-ID: <9aa6ae270708170920o566cf314v354a6efd18dfd729@mail.gmail.com>

Hello,

I see the same problem, I have two images, the older one where it
works, and the newer one with the problem:

Working:
  Pier-Security-lr.85
  Pier-Seaside-lr.185
  Pier-Model-lr.158

Not working:
  Pier-Security-lr.86
  Pier-Seaside-lr.192
  Pier-Model-lr.161

I've had a look at the changes in Pier-Security, but I have no idea
what could cause this problem.

Cheers

Matthias


On 8/17/07, isaiah <cappadonza at hotmail.com> wrote:
>
> hey i got the latest version of pier, when i added a system users component
> to a page, when i click on any thing on the component which invokes a
> callback. i get the following error
> "WAComponentsNotFoundError: Components not found while processing callbacks:
> an Array(a PUUsersWidget)" .
> any ideas how to resolve this
> --
> View this message in context: http://www.nabble.com/Error-in-PUUsersWidget-tf4286525.html#a12202036
> Sent from the SmallWiki mailing list archive at Nabble.com.
>
>
> _______________________________________________
> SmallWiki, Magritte, Pier and Related Tools ...
> https://www.iam.unibe.ch/mailman/listinfo/smallwiki
>

From renggli at iam.unibe.ch  Fri Aug 17 18:36:37 2007
From: renggli at iam.unibe.ch (Lukas Renggli)
Date: Fri, 17 Aug 2007 18:36:37 +0200
Subject: Error in PUUsersWidget
In-Reply-To: <9aa6ae270708170920o566cf314v354a6efd18dfd729@mail.gmail.com>
References: <12202036.post@talk.nabble.com>
	<9aa6ae270708170920o566cf314v354a6efd18dfd729@mail.gmail.com>
Message-ID: <4E3D6D15-7F9E-4454-8598-2C6611B02806@iam.unibe.ch>

The problem is fixed in:

Name: Pier-All-lr.230
Author: lr
Time: 17 August 2007, 6:35:49 pm
UUID: 49baa1a6-da58-4dfb-8c98-ec9d1c2a521e
Ancestors: Pier-All-lr.229
Dependencies: Pier-Model-lr.162, Pier-Tests-lr.73, Pier-Seaside-lr. 
194, Pier-Security-lr.87, Pier-Blog-lr.58

- fixes: "when i added a system users component to a page, when i  
click on any thing on the component which invokes a callback. i get  
the following error 'Components not found while processing callbacks:  
an Array(a PUUsersWidget)'. " (thanks to isaiah)

> I see the same problem, I have two images, the older one where it
> works, and the newer one with the problem:
>
> Working:
>   Pier-Security-lr.85
>   Pier-Seaside-lr.185
>   Pier-Model-lr.158
>
> Not working:
>   Pier-Security-lr.86
>   Pier-Seaside-lr.192
>   Pier-Model-lr.161
>
> I've had a look at the changes in Pier-Security, but I have no idea
> what could cause this problem.
>
> Cheers
>
> Matthias
>
>
> On 8/17/07, isaiah <cappadonza at hotmail.com> wrote:
>>
>> hey i got the latest version of pier, when i added a system users  
>> component
>> to a page, when i click on any thing on the component which invokes a
>> callback. i get the following error
>> "WAComponentsNotFoundError: Components not found while processing  
>> callbacks:
>> an Array(a PUUsersWidget)" .
>> any ideas how to resolve this
>> --
>> View this message in context: http://www.nabble.com/Error-in- 
>> PUUsersWidget-tf4286525.html#a12202036
>> Sent from the SmallWiki mailing list archive at Nabble.com.
>>
>>
>> _______________________________________________
>> SmallWiki, Magritte, Pier and Related Tools ...
>> https://www.iam.unibe.ch/mailman/listinfo/smallwiki
>>
>
> _______________________________________________
> SmallWiki, Magritte, Pier and Related Tools ...
> https://www.iam.unibe.ch/mailman/listinfo/smallwiki

-- 
Lukas Renggli
http://www.lukas-renggli.ch

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 3774 bytes
Desc: not available
Url : http://www.iam.unibe.ch/pipermail/smallwiki/attachments/20070817/2f2a8158/smime.bin

From keith_hodges at yahoo.co.uk  Sat Aug 18 04:02:39 2007
From: keith_hodges at yahoo.co.uk (Keith Hodges)
Date: Sat, 18 Aug 2007 03:02:39 +0100
Subject: [Enh] MACustomDescriptionBuilder
Message-ID: <46C6533F.9070502@yahoo.co.uk>

I wrote this because I want my data model classes to have one set of 
#description's to model their data structure and another set of 
#settings to model the view/filter settings that are available for each 
class.

I dont know what the best way to achieve this with magritte would be, 
but I figured being able to have multiple sets of descriptions for 
different unrelated purposes could be very useful.

As an implementation detail, it might be useful if the 
MADescriptionBuilder 'Cache' was a class instance var rather than a 
classVar.

best regards

Keith

p.s. Lukas, I want to thank you for all of the great help you have 
provided, especially in the past few days.


In http://mc.lukas-renggli.ch/pieraddons

Name: Magritte-CustomBuilder-kph.1
Author: kph
Time: 18 August 2007, 2:53:47 am
UUID: dc2d465c-4d2d-11dc-b626-000a95edb42a
Ancestors:

- Variant of MADescriptionBuilder that is not tied to #description as a 
prefix.

usage: (MCCustomDescriptionBuilder prefix: #settings) for: anObject

Notes: Includes a performance enhancement to double speed of description 
building. (selectors of Object are ignored)

before:
[ 1000 timesRepeat: [MADescriptionBuilder default build: (ProductData) ] 
] timeToRun. 21622

after:
[ 1000 timesRepeat: [(MACustomDescriptionBuilder prefix: #description) 
build: (ProductData) ] ] timeToRun.  11075




From keith_hodges at yahoo.co.uk  Sat Aug 18 05:20:43 2007
From: keith_hodges at yahoo.co.uk (Keith Hodges)
Date: Sat, 18 Aug 2007 04:20:43 +0100
Subject: How to have sub-settings
In-Reply-To: <4E3D6D15-7F9E-4454-8598-2C6611B02806@iam.unibe.ch>
References: <12202036.post@talk.nabble.com>	<9aa6ae270708170920o566cf314v354a6efd18dfd729@mail.gmail.com>
	<4E3D6D15-7F9E-4454-8598-2C6611B02806@iam.unibe.ch>
Message-ID: <46C6658B.2040506@yahoo.co.uk>

I am trying to work out how to have a component in pier, one of whose 
settings select a class, and either other settings options are 
determined by which class is chosen, or the particular class is able to 
supply an editor (Internal or external) for the settings which relate to it.

any inspiration would be appreciated

thanks in advance

Keith

From cappadonza at hotmail.com  Sat Aug 18 10:23:04 2007
From: cappadonza at hotmail.com (i p)
Date: Sat, 18 Aug 2007 08:23:04 +0000
Subject: Error in PUUsersWidget
Message-ID: <BAY130-W13B49692071A6BCFE84A3FD2D90@phx.gbl>


thanks lukas
i found that when i click on 'Groups' or 'superuser' column sort on the users. i get the following error. if i click on groups i get "MessageNotUnderstood: OrderedCollection>><="
and if i click on 'superUsers' column i get messageNotUnderstood: True>> <=.
 
is it just a matter of adding <= to Boolean so it can respond to the binary message <= , for orderedcollection not sure 
any ideas ?> From: renggli at iam.unibe.ch> Subject: Re: Error in PUUsersWidget> Date: Fri, 17 Aug 2007 18:36:37 +0200> To: smallwiki at iam.unibe.ch> > The problem is fixed in:> > Name: Pier-All-lr.230> Author: lr> Time: 17 August 2007, 6:35:49 pm> UUID: 49baa1a6-da58-4dfb-8c98-ec9d1c2a521e> Ancestors: Pier-All-lr.229> Dependencies: Pier-Model-lr.162, Pier-Tests-lr.73, Pier-Seaside-lr. > 194, Pier-Security-lr.87, Pier-Blog-lr.58> > - fixes: "when i added a system users component to a page, when i > click on any thing on the component which invokes a callback. i get > the following error 'Components not found while processing callbacks: > an Array(a PUUsersWidget)'. " (thanks to isaiah)> > > I see the same problem, I have two images, the older one where it> > works, and the newer one with the problem:> >> > Working:> > Pier-Security-lr.85> > Pier-Seaside-lr.185> > Pier-Model-lr.158> >> > Not working:> > Pier-Security-lr.86> > Pier-Seaside-lr.192> > Pier-Model-lr.161> >> > I've had a look at the changes in Pier-Security, but I have no idea> > what could cause this problem.> >> > Cheers> >> > Matthias> >> >> > On 8/17/07, isaiah <cappadonza at hotmail.com> wrote:> >>> >> hey i got the latest version of pier, when i added a system users > >> component> >> to a page, when i click on any thing on the component which invokes a> >> callback. i get the following error> >> "WAComponentsNotFoundError: Components not found while processing > >> callbacks:> >> an Array(a PUUsersWidget)" .> >> any ideas how to resolve this> >> --> >> View this message in context: http://www.nabble.com/Error-in- > >> PUUsersWidget-tf4286525.html#a12202036> >> Sent from the SmallWiki mailing list archive at Nabble.com.> >>> >>> >> _______________________________________________> >> SmallWiki, Magritte, Pier and Related Tools ...> >> https://www.iam.unibe.ch/mailman/listinfo/smallwiki> >>> >> > _______________________________________________> > SmallWiki, Magritte, Pier and Related Tools ...> > https://www.iam.unibe.ch/mailman/listinfo/smallwiki> > -- > Lukas Renggli> http://www.lukas-renggli.ch> 
_________________________________________________________________
Explore the seven wonders of the world
http://search.msn.com/results.aspx?q=7+wonders+world&mkt=en-US&form=QBRE
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.iam.unibe.ch/pipermail/smallwiki/attachments/20070818/4032f202/attachment-0001.html

From renggli at iam.unibe.ch  Sat Aug 18 11:03:24 2007
From: renggli at iam.unibe.ch (Lukas Renggli)
Date: Sat, 18 Aug 2007 11:03:24 +0200
Subject: How to have sub-settings
In-Reply-To: <46C6658B.2040506@yahoo.co.uk>
References: <12202036.post@talk.nabble.com>	<9aa6ae270708170920o566cf314v354a6efd18dfd729@mail.gmail.com>
	<4E3D6D15-7F9E-4454-8598-2C6611B02806@iam.unibe.ch>
	<46C6658B.2040506@yahoo.co.uk>
Message-ID: <9ABFE461-2577-451C-B756-D7989F9C2895@iam.unibe.ch>

> I am trying to work out how to have a component in pier, one of whose
> settings select a class, and either other settings options are
> determined by which class is chosen, or the particular class is  
> able to
> supply an editor (Internal or external) for the settings which  
> relate to it.

This is something currently not supported by Magritte. That's why the  
add and edit step are separated. In case of PRComponent there are  
even 3 steps required: (1) choose to add a component, (2) select the  
kind of component class to add, and (3) setup the properties of the  
component class.

Of course it would be nice to combine those steps into one, but as I  
said Magritte is not really designed to support that kind of  
functionality all automatic. There are many different kind of objects  
involved and one can think of many different ways how this  
interaction could happen, which makes it difficult to do meta- 
descriptions.

What I suggest is that you override #asComponent in you command and  
put some glue code that builds different Magritte forms and displays  
them depending on the state of the other Magritte components. That  
works quite well and requires only little manual code to be written.  
Moreover it can be even combined with manual created forms, etc.

Lukas

-- 
Lukas Renggli
http://www.lukas-renggli.ch

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 3774 bytes
Desc: not available
Url : http://www.iam.unibe.ch/pipermail/smallwiki/attachments/20070818/c29854de/smime.bin

From renggli at iam.unibe.ch  Sat Aug 18 11:12:34 2007
From: renggli at iam.unibe.ch (Lukas Renggli)
Date: Sat, 18 Aug 2007 11:12:34 +0200
Subject: [Enh] MACustomDescriptionBuilder
In-Reply-To: <46C6533F.9070502@yahoo.co.uk>
References: <46C6533F.9070502@yahoo.co.uk>
Message-ID: <9F41F462-979A-4C89-A0F8-DE129F06F910@iam.unibe.ch>

> I wrote this because I want my data model classes to have one set of
> #description's to model their data structure and another set of
> #settings to model the view/filter settings that are available for  
> each class.

Usually I tag my descriptions to do that, e.g. #beEditable,  
#beSearchable, #beSetting, etc. This allows to easily define  
(overlapping) sub-sets of all the existing descriptions.

> I dont know what the best way to achieve this with magritte would be,
> but I figured being able to have multiple sets of descriptions for
> different unrelated purposes could be very useful.

Yes, have a look at the senders and implementors in Pier of  
#beEditable, #isEditable, #beSearchable, #isSearchable, ...

> As an implementation detail, it might be useful if the
> MADescriptionBuilder 'Cache' was a class instance var rather than a
> classVar.

The idea was only to have one builder active in the image.  
Object>>#description uses this default builder. Subclasses might  
override #description to use a different one, if necessary.

> Notes: Includes a performance enhancement to double speed of  
> description building. (selectors of Object are ignored)

I have some projects that add descriptions to Object, so this  
optimization would break certain things.

Lukas

-- 
Lukas Renggli
http://www.lukas-renggli.ch

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 3774 bytes
Desc: not available
Url : http://www.iam.unibe.ch/pipermail/smallwiki/attachments/20070818/df678730/smime.bin

From renggli at iam.unibe.ch  Sat Aug 18 14:58:31 2007
From: renggli at iam.unibe.ch (Lukas Renggli)
Date: Sat, 18 Aug 2007 14:58:31 +0200
Subject: Error in PUUsersWidget
In-Reply-To: <BAY130-W13B49692071A6BCFE84A3FD2D90@phx.gbl>
References: <BAY130-W13B49692071A6BCFE84A3FD2D90@phx.gbl>
Message-ID: <B383CA76-3925-4188-A13D-8A9159C33469@iam.unibe.ch>

> is it just a matter of adding <= to Boolean so it can respond to  
> the binary message <= , for orderedcollection not sure
> any ideas ?

Yes, that would help in this case, but I don't know if this is  
reasonable to do so. I will just commit a small change to Magritte  
that does not allow sorting for columns other than magnitues and  
strings by default. This should prevent these errors.

Lukas

-- 
Lukas Renggli
http://www.lukas-renggli.ch

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 3774 bytes
Desc: not available
Url : http://www.iam.unibe.ch/pipermail/smallwiki/attachments/20070818/2432417d/smime.bin

From keith_hodges at yahoo.co.uk  Sun Aug 19 02:24:40 2007
From: keith_hodges at yahoo.co.uk (Keith Hodges)
Date: Sun, 19 Aug 2007 01:24:40 +0100
Subject: [Enh] MACustomDescriptionBuilder
In-Reply-To: <9F41F462-979A-4C89-A0F8-DE129F06F910@iam.unibe.ch>
References: <46C6533F.9070502@yahoo.co.uk>
	<9F41F462-979A-4C89-A0F8-DE129F06F910@iam.unibe.ch>
Message-ID: <46C78DC8.3080900@yahoo.co.uk>


>
> I have some projects that add descriptions to Object, so this 
> optimization would break certain things.
>
> Lukas
Following your mail I checked a couple of things, and found a 
significant bug in my optimization.....

Putting a breakpoint in #build: when it calls allSelectors shows that it 
works through the selectors of the following chain of classes!

PRComponent class
PRStructure class
PRDecorated class
PRObject class
Object class

and also it continues to

ProtoObject class
Class
ClassDescription
Behavior
Object
ProtoObject

If we are only looking at the class side then thats a lot of selectors 
which we are not interested in

myPRComponent class allSelectors size 1075.
(myPRComponent class allSelectorsBelow: Object) size. 577
(myPRComponent class allSelectorsBelow: ProtoObject class) size  50.

before

[ 1000 timesRepeat: [MADescriptionBuilder default build: (ProductData) ] 
] timeToRun. 21622

after
[ 1000 timesRepeat: [MADescriptionBuilder default build: (ProductData) ] 
] timeToRun. 1717

best regards

Keith

From matthias.berth at googlemail.com  Tue Aug 21 22:22:57 2007
From: matthias.berth at googlemail.com (Matthias Berth)
Date: Tue, 21 Aug 2007 22:22:57 +0200
Subject: Magritte: adding a condition on the whole object in its description
Message-ID: <9aa6ae270708211322g1398b0d6i7c8acaffd13c32e1@mail.gmail.com>

Hello,


I want to add a condition to a Magritte-described object that checks the
whole object for consistency (and not just a single attribute). Is that
possible?

Example: Think about creating a workgroup from some of the users of your
site. The workgroup has a manager. The condition is that the manager must be
one of the members of the workgroup.

I've tried to do this. I have two descriptions on the workgroup class side:
one for the manager (saying that it can be any of the site's users) and one
for the users (saying that this is a subset of all users). I can add a
method #description to modify the standard description that is assembled
from the description* methods on the class side. So I tried:

   Workgroup>>description
    ^super description
        addCondition: [:workgroup | workgroup users includes: workgroup
manager]
        labelled: 'Manager must be a member of the workgroup';
        yourself

Alas, trying to edit an existing workgroup, I learned that the argument of
the condition block is a MACheckedMemento, so I can't ask it for users.

What would be the best way to have this check everytime the work group is
edited? Or should I just write the #manager: method such that the condition
will be satisfied automatically?

Cheers

Matthias


PS: I want to say a big thank you, especially to Lukas, for the great
answers on this list. I have learned something every time, and I was
impressed when my questions on how to do something were answered with a
patch.
PPS: No, this email is not a request for a patch :-)
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.iam.unibe.ch/pipermail/smallwiki/attachments/20070821/f38aae4d/attachment.html

From ramon.leon at allresnet.com  Tue Aug 21 22:34:19 2007
From: ramon.leon at allresnet.com (Ramon Leon)
Date: Tue, 21 Aug 2007 13:34:19 -0700
Subject: Magritte: adding a condition on the whole object in its
	description
In-Reply-To: <9aa6ae270708211322g1398b0d6i7c8acaffd13c32e1@mail.gmail.com>
References: <9aa6ae270708211322g1398b0d6i7c8acaffd13c32e1@mail.gmail.com>
Message-ID: <002f01c7e432$a6b2c560$f000a8c0@hq.allresnet.com>

> Subject: Magritte: adding a condition on the whole object in 
> its description
> 
> Hello,
> 
> I want to add a condition to a Magritte-described object that 
> checks the whole object for consistency (and not just a 
> single attribute). Is that possible? 
> 
> Example: Think about creating a workgroup from some of the 
> users of your site. The workgroup has a manager. The 
> condition is that the manager must be one of the members of 
> the workgroup. 
> 
> I've tried to do this. I have two descriptions on the 
> workgroup class side: one for the manager (saying that it can 
> be any of the site's users) and one for the users (saying 
> that this is a subset of all users). I can add a method 
> #description to modify the standard description that is 
> assembled from the description* methods on the class side. So 
> I tried: 
> 
>    Workgroup>>description
>     ^super description
>         addCondition: [:workgroup | workgroup users includes: 
> workgroup manager] 
>         labelled: 'Manager must be a member of the workgroup'; 
>         yourself
> 
> Alas, trying to edit an existing workgroup, I learned that 
> the argument of the condition block is a MACheckedMemento, so 
> I can't ask it for users. 
> 
> What would be the best way to have this check everytime the 
> work group is edited? Or should I just write the #manager: 
> method such that the condition will be satisfied automatically? 
> 
> Cheers
> Matthias

See
http://onsmalltalk.com/programming/smalltalk/multiple-field-validation-rules
-in-magritte/ to see how to get at the values you're looking for, very
similar situation.

Ramon Leon
http://onsmalltalk.com


From matthias.berth at googlemail.com  Tue Aug 21 23:16:55 2007
From: matthias.berth at googlemail.com (Matthias Berth)
Date: Tue, 21 Aug 2007 23:16:55 +0200
Subject: Magritte: adding a condition on the whole object in its
	description
In-Reply-To: <002f01c7e432$a6b2c560$f000a8c0@hq.allresnet.com>
References: <9aa6ae270708211322g1398b0d6i7c8acaffd13c32e1@mail.gmail.com>
	<002f01c7e432$a6b2c560$f000a8c0@hq.allresnet.com>
Message-ID: <9aa6ae270708211416y6531396dh368bbe4c2de3f235@mail.gmail.com>

Thanks, that looks like exactly what I need :-)

Matthias

On 8/21/07, Ramon Leon <ramon.leon at allresnet.com> wrote:
>
> > Subject: Magritte: adding a condition on the whole object in
> > its description
> >
> > Hello,
> >
> > I want to add a condition to a Magritte-described object that
> > checks the whole object for consistency (and not just a
> > single attribute). Is that possible?
> >
> > Example: Think about creating a workgroup from some of the
> > users of your site. The workgroup has a manager. The
> > condition is that the manager must be one of the members of
> > the workgroup.
> >
> > I've tried to do this. I have two descriptions on the
> > workgroup class side: one for the manager (saying that it can
> > be any of the site's users) and one for the users (saying
> > that this is a subset of all users). I can add a method
> > #description to modify the standard description that is
> > assembled from the description* methods on the class side. So
> > I tried:
> >
> >    Workgroup>>description
> >     ^super description
> >         addCondition: [:workgroup | workgroup users includes:
> > workgroup manager]
> >         labelled: 'Manager must be a member of the workgroup';
> >         yourself
> >
> > Alas, trying to edit an existing workgroup, I learned that
> > the argument of the condition block is a MACheckedMemento, so
> > I can't ask it for users.
> >
> > What would be the best way to have this check everytime the
> > work group is edited? Or should I just write the #manager:
> > method such that the condition will be satisfied automatically?
> >
> > Cheers
> > Matthias
>
> See
>
> http://onsmalltalk.com/programming/smalltalk/multiple-field-validation-rules
> -in-magritte/ to see how to get at the values you're looking for, very
> similar situation.
>
> Ramon Leon
> http://onsmalltalk.com
>
>
> _______________________________________________
> SmallWiki, Magritte, Pier and Related Tools ...
> https://www.iam.unibe.ch/mailman/listinfo/smallwiki
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.iam.unibe.ch/pipermail/smallwiki/attachments/20070821/fe4aac36/attachment.html

From franzl at fjkonrad.de  Thu Aug 23 14:35:59 2007
From: franzl at fjkonrad.de (Franz Josef)
Date: Thu, 23 Aug 2007 14:35:59 +0200
Subject: Pier v. Plone
In-Reply-To: <C3A30441-EBAF-48C6-AAB5-D7E3EABD7AA8@iam.unibe.ch>
References: <46C21A71.9020204@cyberhaus.us>	<46C23065.9080702@moedae.com>	<46C2D0C5.8020702@fjkonrad.de>	<E0845684-C2D6-4837-9B83-71D804F15F7E@iam.unibe.ch>	<46C430E5.8060701@fjkonrad.de>
	<C3A30441-EBAF-48C6-AAB5-D7E3EABD7AA8@iam.unibe.ch>
Message-ID: <46CD7F2F.7010301@fjkonrad.de>


>> - Full-text Searching - all the content is searchable even Word or
>> OpenOffice documents and PDF files (with add ons under Linux also for
>> Excel and PowerPoint and other file types).
>>     
>
> There is already a full-text search. It does not included uploaded  
> files though. If somebody would implement readers for those strange  
> binary formats, that would certainly be an easy thing to add to the  
> search results.
>
> Cheers,
> Lukas
>
>   
Lukas,
I've created a page on Swiki <http://wiki.squeak.org/squeak/572> that
shows how you can use extrnal converters to get the plain text out from
various file formats like pdf, postscript, excel, word and powerpoint.
There are also some small examples to evaluate. Perhaps this will help
you or other Pier developer to implement a full-text search for some
external file formats.

Regards,
Franz Josef

From j.squeak at cyberhaus.us  Fri Aug 24 00:20:12 2007
From: j.squeak at cyberhaus.us (Jimmie Houchin)
Date: Thu, 23 Aug 2007 17:20:12 -0500
Subject: Magritte book
Message-ID: <46CE081C.2040006@cyberhaus.us>

I am reading Lukas book Magritte ? Meta-Described Web Application 
Development.

I have tried to type in the code and run it.
But it doesn't run. :(

DateDescription selector: #birthday label: 'Birthday')
	between: (Date year: 1900) and: Date today;
	yourself.

This is in the latest squeak-dev image.

It complains that DateDescription doesn't exist and offers 
MADateDescription instead. Okay.

But then #year: isn't a method in Date either. :(

It would be nice if Pier/Magritte/Seaside had current documentation.

Would it be possible, reasonable to put up an OO.org doc or some such of 
the original book so that it could be kept up to date by the community?

I know this is asking a lot. But I also hate to ask Lukas to do all of 
the work by himself.

I don't know exactly what would be the best process for community 
stewardship over material like this. But it would be nice if we had an 
excellent place to start. And I think this book is. It would be nice if 
it was current with the current favored images.

If this isn't doable for whatever reason, I certainly respect Lukas' 
wishes concerning his material and authorship. Next best thing until a 
different book or such material came out would be an excellent errata 
page for this book.

Lets make choosing Squeak/Seaside/Pier as easy as possible.

Any way, some thoughts for consideration.

Thanks.

Jimmie



From matthias.berth at googlemail.com  Fri Aug 24 10:41:58 2007
From: matthias.berth at googlemail.com (Matthias Berth)
Date: Fri, 24 Aug 2007 10:41:58 +0200
Subject: Magritte book
In-Reply-To: <46CE081C.2040006@cyberhaus.us>
References: <46CE081C.2040006@cyberhaus.us>
Message-ID: <9aa6ae270708240141w10700309ndd683a288942751c@mail.gmail.com>

I think that an "errata" or "comments" page is an excellent idea. How about
a wiki page on http://smallwiki.unibe.ch/smallwiki/pier/ and a hint on
http://smallwiki.unibe.ch/smallwiki/pier/pierfaq/ ? You can use
anonymous/xopu to login.

Cheers

Matthias


On 8/24/07, Jimmie Houchin <j.squeak at cyberhaus.us> wrote:
>
> I am reading Lukas book Magritte ? Meta-Described Web Application
> Development.
>
> I have tried to type in the code and run it.
> But it doesn't run. :(
>
> DateDescription selector: #birthday label: 'Birthday')
>         between: (Date year: 1900) and: Date today;
>         yourself.
>
> This is in the latest squeak-dev image.
>
> It complains that DateDescription doesn't exist and offers
> MADateDescription instead. Okay.
>
> But then #year: isn't a method in Date either. :(
>
> It would be nice if Pier/Magritte/Seaside had current documentation.
>
> Would it be possible, reasonable to put up an OO.org doc or some such of
> the original book so that it could be kept up to date by the community?
>
> I know this is asking a lot. But I also hate to ask Lukas to do all of
> the work by himself.
>
> I don't know exactly what would be the best process for community
> stewardship over material like this. But it would be nice if we had an
> excellent place to start. And I think this book is. It would be nice if
> it was current with the current favored images.
>
> If this isn't doable for whatever reason, I certainly respect Lukas'
> wishes concerning his material and authorship. Next best thing until a
> different book or such material came out would be an excellent errata
> page for this book.
>
> Lets make choosing Squeak/Seaside/Pier as easy as possible.
>
> Any way, some thoughts for consideration.
>
> Thanks.
>
> Jimmie
>
>
>
> _______________________________________________
> SmallWiki, Magritte, Pier and Related Tools ...
> https://www.iam.unibe.ch/mailman/listinfo/smallwiki
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.iam.unibe.ch/pipermail/smallwiki/attachments/20070824/20454489/attachment-0001.html

From cappadonza at hotmail.com  Fri Aug 24 10:57:52 2007
From: cappadonza at hotmail.com (i p)
Date: Fri, 24 Aug 2007 08:57:52 +0000
Subject: Magritte book
Message-ID: <BAY130-W226AA11E8D6AE57AE1330BD2D70@phx.gbl>


"Magritte ? Meta-Described Web ApplicationDevelopment" is not a book. but its Lukas's Masters degree thesis work. The thesis documents the the design and overall architecture of magritte, the benefits of magritte and what problems it solves, this is not going to be updated (i dont think ) as it is not inteded to be a book for magritte, but this document will give you a good understanding of the overall architecture and design decisions made in designing magritte . 
The wiki would be a good place to have up to date information and current documentation, up to date class names etc


Date: Fri, 24 Aug 2007 10:41:58 +0200From: matthias.berth at googlemail.comTo: smallwiki at iam.unibe.chSubject: Re: Magritte bookI think that an "errata" or "comments" page is an excellent idea. How about a wiki page on http://smallwiki.unibe.ch/smallwiki/pier/ and a hint on http://smallwiki.unibe.ch/smallwiki/pier/pierfaq/ ? You can use anonymous/xopu to login.Cheers Matthias
On 8/24/07, Jimmie Houchin <j.squeak at cyberhaus.us> wrote: 
I am reading Lukas book Magritte ? Meta-Described Web ApplicationDevelopment.I have tried to type in the code and run it.But it doesn't run. :(DateDescription selector: #birthday label: 'Birthday')         between: (Date year: 1900) and: Date today;        yourself.This is in the latest squeak-dev image.It complains that DateDescription doesn't exist and offersMADateDescription instead. Okay. But then #year: isn't a method in Date either. :(It would be nice if Pier/Magritte/Seaside had current documentation.Would it be possible, reasonable to put up an OO.org doc or some such ofthe original book so that it could be kept up to date by the community?I know this is asking a lot. But I also hate to ask Lukas to do all ofthe work by himself.I don't know exactly what would be the best process for community stewardship over material like this. But it would be nice if we had anexcellent place to start. And I think this book is. It would be nice ifit was current with the current favored images.If this isn't doable for whatever reason, I certainly respect Lukas' wishes concerning his material and authorship. Next best thing until adifferent book or such material came out would be an excellent erratapage for this book.Lets make choosing Squeak/Seaside/Pier as easy as possible. Any way, some thoughts for consideration.Thanks.Jimmie_______________________________________________SmallWiki, Magritte, Pier and Related Tools ...https://www.iam.unibe.ch/mailman/listinfo/smallwiki
_________________________________________________________________
Discover the new Windows Vista
http://search.msn.com/results.aspx?q=windows+vista&mkt=en-US&form=QBRE
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.iam.unibe.ch/pipermail/smallwiki/attachments/20070824/1bba079e/attachment.html

From j.squeak at cyberhaus.us  Fri Aug 24 14:47:12 2007
From: j.squeak at cyberhaus.us (Jimmie Houchin)
Date: Fri, 24 Aug 2007 07:47:12 -0500
Subject: Magritte book
In-Reply-To: <BAY130-W226AA11E8D6AE57AE1330BD2D70@phx.gbl>
References: <BAY130-W226AA11E8D6AE57AE1330BD2D70@phx.gbl>
Message-ID: <46CED350.3050500@cyberhaus.us>

i p wrote:
> 
> "Magritte ? Meta-Described Web Application
> Development" is not a book. but its Lukas's Masters degree thesis work. 
> The thesis documents the the design and overall architecture of 
> magritte, the benefits of magritte and what problems it solves, this is 
> not going to be updated (i dont think ) as it is not inteded to be a 
> book for magritte, but this document will give you a good understanding 
> of the overall architecture and design decisions made in designing 
> magritte .
> The wiki would be a good place to have up to date information and 
> current documentation, up to date class names etc
[snip original message]

Yes, you are technically correct. It is not a book technically. I 
thought it was his thesis, but missed inside where it said so.

Nevertheless. It can provide an excellent foundation for such a book. It 
is being provided as documentation for Magritte. It would still be nice 
if this documentation was current. If Lukas wishes for this be to be 
left alone and remain a thesis and a historic artifact it can be so and 
with great respect to him for the work he put into his thesis and the 
value it currently provides. But over time it will diminish as 
documentation. In any living system if the documentation isn't also 
changing, then the documentation will at some point become a historic 
artifact of the program at a point in time.

So the question still remains on whether this thesis can form the 
foundation of a documentation book or booklet which can and will remain 
current documentation of Magritte, Pier, etc? Or do we from the 
knowledge gained from all of the materials start from scratch to create 
such a work?

His thesis will always be his thesis. That is a historical and 
documented fact. But if it is made available as the beginning of some 
versioned Magritte/Pier documentation then that document can be kept 
current with the state of Squeak, Seaside, Magritte, Pier.
By versioned I mean it can have a version number which states the 
versions of Squeak, Seaside, Magritte, Pier that are used in its pages.

Stale documentation always frustrates and hinders beginners and people 
exploring new technology. We need to enable the adoption of these tools. 
Not provide another excuse for passing them by and instead using 
"industry standard" and accepted tools which in many ways are less 
productive than the Squeak toolset.

Any way just some thoughts.

Jimmie



From cappadonza at hotmail.com  Fri Aug 24 14:58:22 2007
From: cappadonza at hotmail.com (i p)
Date: Fri, 24 Aug 2007 12:58:22 +0000
Subject: Magritte book
Message-ID: <BAY130-W1819955336875490173B2D2D70@phx.gbl>


yeh thats a good point, it would be great if we could use the thesis as a foundation to create an up to date versioned Magritte/Pier documentation > Date: Fri, 24 Aug 2007 07:47:12 -0500> From: j.squeak at cyberhaus.us> To: smallwiki at iam.unibe.ch> Subject: Re: Magritte book> > i p wrote:> > > > "Magritte ? Meta-Described Web Application> > Development" is not a book. but its Lukas's Masters degree thesis work. > > The thesis documents the the design and overall architecture of > > magritte, the benefits of magritte and what problems it solves, this is > > not going to be updated (i dont think ) as it is not inteded to be a > > book for magritte, but this document will give you a good understanding > > of the overall architecture and design decisions made in designing > > magritte .> > The wiki would be a good place to have up to date information and > > current documentation, up to date class names etc> [snip original message]> > Yes, you are technically correct. It is not a book technically. I > thought it was his thesis, but missed inside where it said so.> > Nevertheless. It can provide an excellent foundation for such a book. It > is being provided as documentation for Magritte. It would still be nice > if this documentation was current. If Lukas wishes for this be to be > left alone and remain a thesis and a historic artifact it can be so and > with great respect to him for the work he put into his thesis and the > value it currently provides. But over time it will diminish as > documentation. In any living system if the documentation isn't also > changing, then the documentation will at some point become a historic > artifact of the program at a point in time.> > So the question still remains on whether this thesis can form the > foundation of a documentation book or booklet which can and will remain > current documentation of Magritte, Pier, etc? Or do we from the > knowledge gained from all of the materials start from scratch to create > such a work?> > His thesis will always be his thesis. That is a historical and > documented fact. But if it is made available as the beginning of some > versioned Magritte/Pier documentation then that document can be kept > current with the state of Squeak, Seaside, Magritte, Pier.> By versioned I mean it can have a version number which states the > versions of Squeak, Seaside, Magritte, Pier that are used in its pages.> > Stale documentation always frustrates and hinders beginners and people > exploring new technology. We need to enable the adoption of these tools. > Not provide another excuse for passing them by and instead using > "industry standard" and accepted tools which in many ways are less > productive than the Squeak toolset.> > Any way just some thoughts.> > Jimmie> > > > _______________________________________________> SmallWiki, Magritte, Pier and Related Tools ...> https://www.iam.unibe.ch/mailman/listinfo/smallwiki
_________________________________________________________________
Discover the new Windows Vista
http://search.msn.com/results.aspx?q=windows+vista&mkt=en-US&form=QBRE
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.iam.unibe.ch/pipermail/smallwiki/attachments/20070824/69d7e896/attachment.html

From renggli at iam.unibe.ch  Fri Aug 24 17:23:04 2007
From: renggli at iam.unibe.ch (Lukas Renggli)
Date: Fri, 24 Aug 2007 17:23:04 +0200
Subject: Pier v. Plone
In-Reply-To: <46CD7F2F.7010301@fjkonrad.de>
References: <46C21A71.9020204@cyberhaus.us>	<46C23065.9080702@moedae.com>	<46C2D0C5.8020702@fjkonrad.de>	<E0845684-C2D6-4837-9B83-71D804F15F7E@iam.unibe.ch>	<46C430E5.8060701@fjkonrad.de>
	<C3A30441-EBAF-48C6-AAB5-D7E3EABD7AA8@iam.unibe.ch>
	<46CD7F2F.7010301@fjkonrad.de>
Message-ID: <DF171F52-101F-49EE-B625-4324BFDE81FC@iam.unibe.ch>

> I've created a page on Swiki <http://wiki.squeak.org/squeak/572> that
> shows how you can use extrnal converters to get the plain text out  
> from
> various file formats like pdf, postscript, excel, word and powerpoint.
> There are also some small examples to evaluate. Perhaps this will help
> you or other Pier developer to implement a full-text search for some
> external file formats.

Thanks for writing down this summary.

I am afraid that I don't have the time to work on this. Maybe a  
Google Summer of Code project? The stemming library is also  
interesting, but I am afraid that this only works for english  
sources? Maybe there are multilingual stemming libraries where we can  
just pipe the results from the converters?

Lukas

-- 
Lukas Renggli
http://www.lukas-renggli.ch


From dave at thedesignexperience.org  Fri Aug 24 17:26:55 2007
From: dave at thedesignexperience.org (Dave Bauer)
Date: Fri, 24 Aug 2007 11:26:55 -0400
Subject: Pier v. Plone
In-Reply-To: <DF171F52-101F-49EE-B625-4324BFDE81FC@iam.unibe.ch>
References: <46C21A71.9020204@cyberhaus.us> <46C23065.9080702@moedae.com>
	<46C2D0C5.8020702@fjkonrad.de>
	<E0845684-C2D6-4837-9B83-71D804F15F7E@iam.unibe.ch>
	<46C430E5.8060701@fjkonrad.de>
	<C3A30441-EBAF-48C6-AAB5-D7E3EABD7AA8@iam.unibe.ch>
	<46CD7F2F.7010301@fjkonrad.de>
	<DF171F52-101F-49EE-B625-4324BFDE81FC@iam.unibe.ch>
Message-ID: <6c4e07720708240826rd645d1fybcba793b6bc1d098@mail.gmail.com>

On 8/24/07, Lukas Renggli <renggli at iam.unibe.ch> wrote:

 The stemming library is also
> interesting, but I am afraid that this only works for english
> sources? Maybe there are multilingual stemming libraries where we can
> just pipe the results from the converters?

Snowball can do stemming for multiple languages
http://snowball.tartarus.org/

The PostgreSQL full text indexer tsearch2 uses it.

Dave

>
> Lukas
>
> --
> Lukas Renggli
> http://www.lukas-renggli.ch
>
>
> _______________________________________________
> SmallWiki, Magritte, Pier and Related Tools ...
> https://www.iam.unibe.ch/mailman/listinfo/smallwiki
>


-- 
Dave Bauer
dave at solutiongrove.com
http://www.solutiongrove.com

From renggli at iam.unibe.ch  Fri Aug 24 19:57:02 2007
From: renggli at iam.unibe.ch (Lukas Renggli)
Date: Fri, 24 Aug 2007 19:57:02 +0200
Subject: Magritte book
In-Reply-To: <46CE081C.2040006@cyberhaus.us>
References: <46CE081C.2040006@cyberhaus.us>
Message-ID: <08E42E90-CAD7-45EA-8E5C-26F181E0E922@iam.unibe.ch>

> DateDescription selector: #birthday label: 'Birthday')
> 	between: (Date year: 1900) and: Date today;
> 	yourself.

Mhh, yes there are several problems with this code:

	MADateDescription new
		selectorAccessor: #birthday;
		label: 'Birthday';
		min: (Date year: 1900 day: 1);
		max: Date today;
		yourself

> Would it be possible, reasonable to put up an OO.org doc or some  
> such of
> the original book so that it could be kept up to date by the  
> community?

What is this? The link doesn't work.

> I know this is asking a lot. But I also hate to ask Lukas to do all of
> the work by himself.

I also thought about the need of updating the documentation, however  
I have to check with the University in what form this is possible. I  
will do that when I am back from ESUG. Ping me in case I forget.

The sources are all in LaTeX. The documentation in the back is  
automatically built from the class and method comments, but I don't  
know if this is useful at all?

> Yes, you are technically correct. It is not a book technically. I
> thought it was his thesis, but missed inside where it said so.

It tells so on the front page, but it is in German because of a  
strange requirement of the University.

There is also a paper that is more recent (but less practical):

	http://www.iam.unibe.ch/~scg/cgi-bin/scgbib.cgi/abstract=yes?Reng07a

I will present it at MODELS 2007 in Nashville. In case anybody feels  
like chatting about Magritte, Pier, Seaside, Smalltalk or programming  
in general don't hesitate to send a mail.

Cheers,
Lukas

-- 
Lukas Renggli
http://www.lukas-renggli.ch


From j.squeak at cyberhaus.us  Fri Aug 24 20:57:56 2007
From: j.squeak at cyberhaus.us (Jimmie Houchin)
Date: Fri, 24 Aug 2007 13:57:56 -0500
Subject: Magritte book
In-Reply-To: <08E42E90-CAD7-45EA-8E5C-26F181E0E922@iam.unibe.ch>
References: <46CE081C.2040006@cyberhaus.us>
	<08E42E90-CAD7-45EA-8E5C-26F181E0E922@iam.unibe.ch>
Message-ID: <46CF2A34.9060002@cyberhaus.us>

Lukas Renggli wrote:
>> DateDescription selector: #birthday label: 'Birthday')
>> 	between: (Date year: 1900) and: Date today;
>> 	yourself.
> 
> Mhh, yes there are several problems with this code:
> 
> 	MADateDescription new
> 		selectorAccessor: #birthday;
> 		label: 'Birthday';
> 		min: (Date year: 1900 day: 1);
> 		max: Date today;
> 		yourself
> 
>> Would it be possible, reasonable to put up an OO.org doc  
>> or some such of the original book so that it could be   
>> kept up to date by the community?
> 
> What is this? The link doesn't work.

My apologies. It wasn't meant to be a link but rather to an 
OpenOffice.org word processing document. Which then would be 
http://www.openoffice.org but you probably already know that when I put 
it into proper context.

>> I know this is asking a lot. But I also hate to ask Lukas to do all of
>> the work by himself.
> 
> I also thought about the need of updating the documentation, however  
> I have to check with the University in what form this is possible. I  
> will do that when I am back from ESUG. Ping me in case I forget.

Super. That would be great. You put a lot of work into that. And I think 
it forms an excellent basis of understanding and philosophy behind Pier 
and its design.

I would love see the Seaside/Pier community grow and prosper. Good docs 
can help tremendously. Poor docs can definitely hurt.

> The sources are all in LaTeX. The documentation in the back is  
> automatically built from the class and method comments, but I don't  
> know if this is useful at all?

Someone here may know how to use LaTeX sources. If not then we can 
attempt any number of things to extract the textual sources.

In fact I just did a "select all" in Acrobat and copied the contents to 
a OO.org doc and it isn't perfect but it is very usable. It would 
require some reformatting and the images imported into the document. But 
it is definitely workable.

What we would need are the images and licensing permissions from the 
copyright holders. So if you are wanting to proceed on something like 
this. You will need to release it under an appropriate license or your 
choosing. And if the University has copyrights also, they would also.

You would then probably want to rename this document to something 
appropriate and then clean up some of the "thesis" type text inside and 
giving proper attribution to your original thesis as the foundation and 
origination of this document. Possibly, even providing a url to your 
original thesis.

>> Yes, you are technically correct. It is not a book technically. I
>> thought it was his thesis, but missed inside where it said so.
> 
> It tells so on the front page, but it is in German because of a  
> strange requirement of the University.

Yes. That scared me at first. I just printed out the document and then 
the Title page is in German. Uh oh. :)

But then I turned the page. I could read that. ;)

We USAmericans are a wimpy monolingual bunch. :)

But in your Acknowledgments you also state this:
Thanks for writing with me the two papers, [Duca05] and [Reng06b], which 
were an important point of reference for this master thesis.

Which is what I missed.

> There is also a paper that is more recent (but less practical):
> 
> http://www.iam.unibe.ch/~scg/cgi-bin/scgbib.cgi/abstract=yes?Reng07a
> 
> I will present it at MODELS 2007 in Nashville. In case anybody feels  
> like chatting about Magritte, Pier, Seaside, Smalltalk or programming  
> in general don't hesitate to send a mail.

Thanks for the link. I'll definitely read it also.

Anyway, enjoy ESUG.

Jimmie

From matthias.berth at googlemail.com  Fri Aug 24 22:57:41 2007
From: matthias.berth at googlemail.com (Matthias Berth)
Date: Fri, 24 Aug 2007 22:57:41 +0200
Subject: Magritte book
In-Reply-To: <08E42E90-CAD7-45EA-8E5C-26F181E0E922@iam.unibe.ch>
References: <46CE081C.2040006@cyberhaus.us>
	<08E42E90-CAD7-45EA-8E5C-26F181E0E922@iam.unibe.ch>
Message-ID: <9aa6ae270708241357r5c3c052dq875be9793cbb6e8f@mail.gmail.com>

Hi Lukas,

interesting paper, I did not notice that I can add attributes to a
pier page through the web. I'll try that and read up on the user
extensibility of Magritte. There is one more thing:

"As expected SmallWiki performed better with a cumulated search time
of 2456 ms. In Pier
the search took 8190 ms, so the meta-driven search is about 30% slower than
the hard coded one."

I guess you mean that the search in SmallWiki takes only 30% of the
time it takes in Pier?


Cheers

Matthias

PS: If you want to release the LaTeX source for your thesis, I could
try the automated conversion to html. From there one would go to wiki
markup, right? :-)

> There is also a paper that is more recent (but less practical):
>
>         http://www.iam.unibe.ch/~scg/cgi-bin/scgbib.cgi/abstract=yes?Reng07a
>
> I will present it at MODELS 2007 in Nashville. In case anybody feels
> like chatting about Magritte, Pier, Seaside, Smalltalk or programming
> in general don't hesitate to send a mail.

From rbb at techgame.net  Fri Aug 24 22:59:31 2007
From: rbb at techgame.net (Brian Brown)
Date: Fri, 24 Aug 2007 14:59:31 -0600
Subject: Magritte book
In-Reply-To: <46CF2A34.9060002@cyberhaus.us>
References: <46CE081C.2040006@cyberhaus.us>
	<08E42E90-CAD7-45EA-8E5C-26F181E0E922@iam.unibe.ch>
	<46CF2A34.9060002@cyberhaus.us>
Message-ID: <1CF77731-A4A1-470F-9455-AC6A9680E07D@techgame.net>


On Aug 24, 2007, at 12:57 PM, Jimmie Houchin wrote:
>
> Someone here may know how to use LaTeX sources. If not then we can
> attempt any number of things to extract the textual sources.
>

Well, we can certainly use LaTeX and then we could post the PDF, and  
if some wanted a more "word-processy" interface without a bunch of  
manual work, we could import it into LyX (www.lyx.org), which works  
quite well.

My .02 cents.

- Brian