From me at rpatterson.net  Tue Jan  2 04:47:01 2007
From: me at rpatterson.net (Ross Patterson)
Date: Mon, 01 Jan 2007 19:47:01 -0800
Subject: [z3-five] Using vocabularies for Choice fields where the
	vocabulary is acquired
References: <emu9h4$m8e$1@sea.gmane.org> <4592CE19.7000202@hannosch.info>
	<emunok$v7$1@sea.gmane.org>
Message-ID: <87bqlit7ii.fsf@superfluity.lefae.org>

Martin Aspeli <optilude at gmx.net> writes:

> Hanno Schlichting wrote:
>
>> I'm not sure that my approach is by any means good or advisable but you
>> could take a look at the PortalTypesVocabulary in
>> plone.app.controlpanel.vocabularies and its usage on the search control
>> panel defined by the ISearchSchema in search.py in the same package.
>> 
>> The code can be found at
>> http://dev.plone.org/plone/browser/plone.app.controlpanel/trunk/plone/app/controlpanel
>> 
>>> Can anyone point me to some code or documentation that shows how to wait 
>>> until run-time to define the vocabulary, and gives me the current 
>>> context from which to do so?
>> 
>> Maybe you can even just use the vocabulary I've already written in which
>> case we might need to move it to a more general package.
>
> Ah, excellent Hanno. :)
>
> I can use this just by name, I guess. Having a package 
> (plone.app.vocabularies?) with common vocabularies (portal type, 
> workflow states, roles, groups, users...) may make sense.
>
> I'm adding roles and groups to plone.app.contentrules.vocabularies, if 
> you're up for the task of consolidating (I don't have time right now).
>
> Martin

Ive been getting up to speed on Z3 schemas and I was wondering why
plone.app.vocabularies uses vocabularies instead of sources?

Ross


From plone at hannosch.info  Tue Jan  2 07:43:41 2007
From: plone at hannosch.info (Hanno Schlichting)
Date: Tue, 02 Jan 2007 07:43:41 +0100
Subject: [z3-five] Using vocabularies for Choice fields where the
 vocabulary is acquired
In-Reply-To: <87bqlit7ii.fsf@superfluity.lefae.org>
References: <emu9h4$m8e$1@sea.gmane.org>
	<4592CE19.7000202@hannosch.info>	<emunok$v7$1@sea.gmane.org>
	<87bqlit7ii.fsf@superfluity.lefae.org>
Message-ID: <4599FF1D.6010800@hannosch.info>

Hi Ross,

Ross Patterson wrote:
>> Hanno Schlichting wrote:
>>
>>> I'm not sure that my approach is by any means good or advisable but you
>>> could take a look at the PortalTypesVocabulary in
>>> plone.app.controlpanel.vocabularies and its usage on the search control
>>> panel defined by the ISearchSchema in search.py in the same package.
> 
> Ive been getting up to speed on Z3 schemas and I was wondering why
> plone.app.vocabularies uses vocabularies instead of sources?

As my first sentence should have told you, I'm by no means an expert or
even an advanced user of formlib. So vocabularies were just what I found
working for me quite easily.

I know there have been some talk on the Zope3 mailing list a while back
that vocabularies are going to be deprecated in favor of sources
someday, but it also sounded like sources are not yet as well documented
and easy to use.

If you could enlighten us how to use sources instead for one of the
vocabularies currently in plone.app.vocabularies that would be awesome.

Thx,
Hanno


From me at rpatterson.net  Tue Jan  2 10:15:41 2007
From: me at rpatterson.net (Ross Patterson)
Date: Tue, 02 Jan 2007 01:15:41 -0800
Subject: [z3-five] Using vocabularies for Choice fields where the
	vocabulary is acquired
References: <emu9h4$m8e$1@sea.gmane.org> <4592CE19.7000202@hannosch.info>
	<emunok$v7$1@sea.gmane.org> <87bqlit7ii.fsf@superfluity.lefae.org>
	<4599FF1D.6010800@hannosch.info>
Message-ID: <87zm91ssaq.fsf@superfluity.lefae.org>

Hanno Schlichting <plone at hannosch.info>
writes:

> Hi Ross,
>
> Ross Patterson wrote:
>>> Hanno Schlichting wrote:
>>>
>>>> I'm not sure that my approach is by any means good or advisable but you
>>>> could take a look at the PortalTypesVocabulary in
>>>> plone.app.controlpanel.vocabularies and its usage on the search control
>>>> panel defined by the ISearchSchema in search.py in the same package.
>> 
>> Ive been getting up to speed on Z3 schemas and I was wondering why
>> plone.app.vocabularies uses vocabularies instead of sources?
>
> As my first sentence should have told you, I'm by no means an expert or
> even an advanced user of formlib. So vocabularies were just what I found
> working for me quite easily.
>
> I know there have been some talk on the Zope3 mailing list a while back
> that vocabularies are going to be deprecated in favor of sources
> someday, but it also sounded like sources are not yet as well documented
> and easy to use.
>
> If you could enlighten us how to use sources instead for one of the
> vocabularies currently in plone.app.vocabularies that would be awesome.
>
> Thx,
> Hanno

Ok, as I educate myself I'll try and submit some patches back.
Thanks!

Ross


From y.2007- at wcm-solutions.de  Sun Jan  7 14:48:43 2007
From: y.2007- at wcm-solutions.de (yuppie)
Date: Sun, 07 Jan 2007 14:48:43 +0100
Subject: [z3-five] Zope 2.10.2 coming up
In-Reply-To: <C36356B26D8B8C1FE9A626CE@suxmac2.local>
References: <C36356B26D8B8C1FE9A626CE@suxmac2.local>
Message-ID: <45A0FA3B.5070901@wcm-solutions.de>

Hi!


Andreas Jung wrote:
> I plan to release Zope 2.10.2 in a week or so (don't nail
> me to a particular date).

The Zope 2.10 branch still uses Five 1.5.1. There are a few changes on 
the Five 1.5 branch (including a security fix) which should become part 
of Zope 2.10.2.

I volunteer to resolve this as follows:

Five releases are now linked tightly to specific Zope releases, so it 
doesn't make much sense to do still separate Five releases. See this thread:
http://codespeak.net/pipermail/z3-five/2006q4/001889.html

If there are no objections, I'll stitch the Five 1.5 branch HEAD 
directly into the Zope 2.10 branch (without using svn externals) and 
Five trunk HEAD into the Zope trunk. This would mean that from now on 
Five 1.5 and 1.6 are maintained as part of Zope 2. In fact there will be 
no Five 1.5.2 or 1.6 because there will no longer be any separate releases.

Thoughts?


Cheers,

	Yuppie


From lists at zopyx.com  Sun Jan  7 15:28:59 2007
From: lists at zopyx.com (Andreas Jung)
Date: Sun, 07 Jan 2007 15:28:59 +0100
Subject: [z3-five] Zope 2.10.2 coming up
In-Reply-To: <45A0FA3B.5070901@wcm-solutions.de>
References: <C36356B26D8B8C1FE9A626CE@suxmac2.local>
	<45A0FA3B.5070901@wcm-solutions.de>
Message-ID: <5F17C43487CFA815EF07CEEC@suxmac2.local>



--On 7. Januar 2007 14:48:43 +0100 yuppie <y.2007- at wcm-solutions.de> wrote:

> If there are no objections, I'll stitch the Five 1.5 branch HEAD directly
> into the Zope 2.10 branch (without using svn externals) and Five trunk
> HEAD into the Zope trunk. This would mean that from now on Five 1.5 and
> 1.6 are maintained as part of Zope 2. In fact there will be no Five 1.5.2
> or 1.6 because there will no longer be any separate releases.

I don't have a problem with this approach. In fact the change would make 
clear which version belongs to what.

Andreas
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 186 bytes
Desc: not available
Url : http://codespeak.net/pipermail/z3-five/attachments/20070107/7aba69f4/attachment.pgp 

From georgeleejr at gmail.com  Sun Jan  7 20:57:56 2007
From: georgeleejr at gmail.com (George Lee)
Date: Sun, 7 Jan 2007 19:57:56 +0000 (UTC)
Subject: [z3-five] Customizing ZopeTwoPageTemplates through the web
Message-ID: <loom.20070107T205454-766@post.gmane.org>

Hi,

What is the progress on work being done to make Zope3's ZopeTwoPageTemplates
customizable through the web? I remember seeing a post or changeset about work
being done on this but can't find it now.

The beauty of calling templates as methods on a view, i.e.
  <tal:block replace="view/subthread" />
and then defining a subthread.pt, is wonderful but I'm wondering when it'll be
able to customize these templates TTW.

Peace,
George


From philipp at weitershausen.de  Mon Jan  8 09:58:08 2007
From: philipp at weitershausen.de (Philipp von Weitershausen)
Date: Mon, 08 Jan 2007 09:58:08 +0100
Subject: [z3-five] Zope 2.10.2 coming up
In-Reply-To: <45A0FA3B.5070901@wcm-solutions.de>
References: <C36356B26D8B8C1FE9A626CE@suxmac2.local>
	<45A0FA3B.5070901@wcm-solutions.de>
Message-ID: <45A207A0.8080903@weitershausen.de>

yuppie wrote:
> Hi!
> 
> 
> Andreas Jung wrote:
>> I plan to release Zope 2.10.2 in a week or so (don't nail
>> me to a particular date).
> 
> The Zope 2.10 branch still uses Five 1.5.1. There are a few changes on 
> the Five 1.5 branch (including a security fix) which should become part 
> of Zope 2.10.2.
> 
> I volunteer to resolve this as follows:
> 
> Five releases are now linked tightly to specific Zope releases, so it 
> doesn't make much sense to do still separate Five releases. See this 
> thread:
> http://codespeak.net/pipermail/z3-five/2006q4/001889.html
> 
> If there are no objections, I'll stitch the Five 1.5 branch HEAD 
> directly into the Zope 2.10 branch (without using svn externals) and 
> Five trunk HEAD into the Zope trunk. This would mean that from now on 
> Five 1.5 and 1.6 are maintained as part of Zope 2. In fact there will be 
> no Five 1.5.2 or 1.6 because there will no longer be any separate releases.
> 
> Thoughts?

+ sys.maxint

-- 
http://worldcookery.com -- Professional Zope documentation and training
2nd edition of Web Component Development with Zope 3 is now shipping!


From philipp at weitershausen.de  Tue Jan  9 00:12:18 2007
From: philipp at weitershausen.de (Philipp von Weitershausen)
Date: Tue, 09 Jan 2007 00:12:18 +0100
Subject: [z3-five] MailHosts as utilities
In-Reply-To: <86D9AF3A9BF3B999CD88A467@suxmac2.local>
References: <4577E35D.6030805@simplistix.co.uk>	<319e029f0612070320y5e6e84c0ye466d68acfb0b5d0@mail.gmail.com>
	<86D9AF3A9BF3B999CD88A467@suxmac2.local>
Message-ID: <45A2CFD2.4050309@weitershausen.de>

Andreas Jung wrote:
>> On 12/7/06, Chris Withers <chris at simplistix.co.uk> wrote:
>>> Anyone done any work on writing an interface for MailHost-like things
>>> and then getting MailHost, MaildropHost, SecureMailHost and friends to
>>> implement the interface and be registered as local utilities?
>> I think Zope3 actually may have some stuff like that. CPSMailAccess
>> uses it, I think.
> 
> zope.sendmail?

Yes, zope.sendmail (previously known as zope.app.mail) works fine on 
Zope 2. I would suggest using this one.


-- 
http://worldcookery.com -- Professional Zope documentation and training
2nd edition of Web Component Development with Zope 3 is now shipping!

From georgeleejr at gmail.com  Tue Jan  9 00:57:07 2007
From: georgeleejr at gmail.com (George Lee)
Date: Mon, 8 Jan 2007 23:57:07 +0000 (UTC)
Subject: [z3-five] Customizing ZopeTwoPageTemplates through the web
References: <loom.20070107T205454-766@post.gmane.org>
Message-ID: <loom.20070109T005638-289@post.gmane.org>

George Lee <georgeleejr <at> gmail.com> writes:

> 
> Hi,
> 
> What is the progress on work being done to make Zope3's ZopeTwoPageTemplates
> customizable through the web?


Ah: http://thread.gmane.org/gmane.comp.web.zope.plone.teams.framework/783

And also, I think work is being done on this in plone.app.layout

Peace,
George


From me at rpatterson.net  Tue Jan  9 08:54:17 2007
From: me at rpatterson.net (Ross Patterson)
Date: Mon, 08 Jan 2007 23:54:17 -0800
Subject: [z3-five] Persistent Interfaces
Message-ID: <87vejgzlcm.fsf@superfluity.lefae.org>

I'm building a Zope2.10/Plone3 project using the CA wherever I can.
I'm relying heavily on marker interfaces and I'm at a point where I
need to allow a site administrator to create new marker interfaces to
be used to mark content.

I've looked at zodbcode, zope.app.module, and then finally found
srichter's SchemaUtility work in zope.app.schema that was removed by
jim in r69203.  I'm only minimally able to make sense of this stuff,
but none of it seems like what I'm looking for. I'd prefer to
programatically create new interfaces than fake modules or other code
object.  Something like::

    >>> from new import classobj
    >>> from zope.app.interface import PersistentInterface
    >>> app.IFoo = classobj('IFoo', (PersistentInterface,), dict(__doc__='IFoo'))
    >>> from zope.interface import directlyProvides
    >>> directlyProvides(app.index_html, app.IFoo)

This actually all works except when the instance is stopped and
restarted.  When restarted, viewing index_html raises an error becuase
the app.IFoo.depdndents PersistentDict is missing it's
app.IFoo.depdndents.data attribute.  I see that there's a
zodbcode.patch.Wrapper used in zope.app.interface that seems like it's
intended to resolve this issue but I have no idea how to use it.

Actually, I don't have much idea what I'm doing at all, but I know
what I want.  :)  I want to create marker interface programatically in
response to a site administrators TTW actions and have them persist.
I want to mark persistent objects as providing those interfaces (and,
of course, have those declarations persist).  In response to other
site administrator TTW actions, I want to register adapters for
objects that provide those marker interfaces and have those
registrations persist (the adapters themselves need not persist, but
that would be nice).

I want this because the CA registry does exactly the kind of efficient
lookup I'd like to do in this case.  The only difference is that I
don't want the marker interfaces and adapter registrations to be
restricted to global code and I want them created through TTW actions.

srichter, did SchemaUtility do what I'm talking about here?  I see how
it managed the persistent state of an interface established through
global code, but I don't see how it could have resulted in the
arbitrary creation of new interfaces TTW.  Can you help me understand
a little more what's going on there?

Can anyone else advise on implementation directions or better design
patterns?

Ross


From chris at simplistix.co.uk  Tue Jan  9 09:21:52 2007
From: chris at simplistix.co.uk (Chris Withers)
Date: Tue, 09 Jan 2007 08:21:52 +0000
Subject: [z3-five] MailHosts as utilities
In-Reply-To: <45A2CFD2.4050309@weitershausen.de>
References: <4577E35D.6030805@simplistix.co.uk>	<319e029f0612070320y5e6e84c0ye466d68acfb0b5d0@mail.gmail.com>	<86D9AF3A9BF3B999CD88A467@suxmac2.local>
	<45A2CFD2.4050309@weitershausen.de>
Message-ID: <45A350A0.2020901@simplistix.co.uk>

Philipp von Weitershausen wrote:
> Andreas Jung wrote:
>>> On 12/7/06, Chris Withers <chris at simplistix.co.uk> wrote:
>>>> Anyone done any work on writing an interface for MailHost-like things
>>>> and then getting MailHost, MaildropHost, SecureMailHost and friends to
>>>> implement the interface and be registered as local utilities?
>>> I think Zope3 actually may have some stuff like that. CPSMailAccess
>>> uses it, I think.
>> zope.sendmail?
> 
> Yes, zope.sendmail (previously known as zope.app.mail) works fine on 
> Zope 2. I would suggest using this one.

This is kind of missing my point. There's currently MailHost, 
MaildropHost and one or two others that provide useful mailing 
functionality.

However, selecting from available options is hard because you have to 
pass all the possible meta_types to superValues.

It's to avoid this problem that I'd like to find a way to move mailing 
in Zope 2 to something interface-based, not because I want to use 
another implementation...

Chris

-- 
Simplistix - Content Management, Zope & Python Consulting
            - http://www.simplistix.co.uk


From optilude at gmx.net  Tue Jan  9 14:07:36 2007
From: optilude at gmx.net (Martin Aspeli)
Date: Tue, 09 Jan 2007 13:07:36 +0000
Subject: [z3-five] Customizing ZopeTwoPageTemplates through the web
In-Reply-To: <loom.20070109T005638-289@post.gmane.org>
References: <loom.20070107T205454-766@post.gmane.org>
	<loom.20070109T005638-289@post.gmane.org>
Message-ID: <eo0437$r97$4@sea.gmane.org>

George Lee wrote:
> George Lee <georgeleejr <at> gmail.com> writes:
> 
>> Hi,
>>
>> What is the progress on work being done to make Zope3's ZopeTwoPageTemplates
>> customizable through the web?
> 
> 
> Ah: http://thread.gmane.org/gmane.comp.web.zope.plone.teams.framework/783
> 
> And also, I think work is being done on this in plone.app.layout

No, plone.app.layout has nothing to do with this, though 
five.customerize is what you want. Philipp also just posted to the 
framework team list asking for help completing it. If you have bandwidth 
to help polish it up that would be major way of contributing to Plone 3!

Martin


From philipp at weitershausen.de  Tue Jan  9 14:59:54 2007
From: philipp at weitershausen.de (Philipp von Weitershausen)
Date: Tue, 9 Jan 2007 14:59:54 +0100
Subject: [z3-five] MailHosts as utilities
In-Reply-To: <45A350A0.2020901@simplistix.co.uk>
References: <4577E35D.6030805@simplistix.co.uk>	<319e029f0612070320y5e6e84c0ye466d68acfb0b5d0@mail.gmail.com>	<86D9AF3A9BF3B999CD88A467@suxmac2.local>
	<45A2CFD2.4050309@weitershausen.de>
	<45A350A0.2020901@simplistix.co.uk>
Message-ID: <CC1FF7F1-343C-4383-B65B-864D1F9B1B08@weitershausen.de>

On 9 Jan 2007, at 09:21 , Chris Withers wrote:
> Philipp von Weitershausen wrote:
>> Andreas Jung wrote:
>>>> On 12/7/06, Chris Withers <chris at simplistix.co.uk> wrote:
>>>>> Anyone done any work on writing an interface for MailHost-like  
>>>>> things
>>>>> and then getting MailHost, MaildropHost, SecureMailHost and  
>>>>> friends to
>>>>> implement the interface and be registered as local utilities?
>>>> I think Zope3 actually may have some stuff like that. CPSMailAccess
>>>> uses it, I think.
>>> zope.sendmail?
>> Yes, zope.sendmail (previously known as zope.app.mail) works fine  
>> on Zope 2. I would suggest using this one.
>
> This is kind of missing my point. There's currently MailHost,  
> MaildropHost and one or two others that provide useful mailing  
> functionality.
>
> However, selecting from available options is hard because you have  
> to pass all the possible meta_types to superValues.
>
> It's to avoid this problem that I'd like to find a way to move  
> mailing in Zope 2 to something interface-based, not because I want  
> to use another implementation...

I see. Well, it'd be useful, for starters, to implement a common  
IMailHost interface that is shared among those different  
implementations. They could then be registered as local persistent  
utilities for IMailHost.


From rocky at serverzen.com  Tue Jan  9 15:40:17 2007
From: rocky at serverzen.com (Rocky Burt)
Date: Tue, 09 Jan 2007 11:10:17 -0330
Subject: [z3-five] MailHosts as utilities
In-Reply-To: <CC1FF7F1-343C-4383-B65B-864D1F9B1B08@weitershausen.de>
References: <4577E35D.6030805@simplistix.co.uk>
	<319e029f0612070320y5e6e84c0ye466d68acfb0b5d0@mail.gmail.com>
	<86D9AF3A9BF3B999CD88A467@suxmac2.local>
	<45A2CFD2.4050309@weitershausen.de> <45A350A0.2020901@simplistix.co.uk>
	<CC1FF7F1-343C-4383-B65B-864D1F9B1B08@weitershausen.de>
Message-ID: <1168353614.21802.331.camel@localhost>

On Tue, 2007-09-01 at 14:59 +0100, Philipp von Weitershausen wrote:
> I see. Well, it'd be useful, for starters, to implement a common  
> IMailHost interface that is shared among those different  
> implementations. They could then be registered as local persistent  
> utilities for IMailHost.

Well, whatever interface we end up using the need I see is as follows:

  1) use existing zope3 mail interface (or create new one like
IMailHost)
  2) make sure global utility exists for this interface
  3) make persistent version of utility that can be registered locally
and stores it's settings internally
  4) introduce MailHost BBB class for CMF that proxies for the new local
utility and spews deprecation warnings

We've been toying with this plan for Plone but no one has managed to get
time to do it.

- Rocky

-- 
Rocky Burt
ServerZen Software -- http://www.serverzen.com
News About The Server (blog) -- http://www.serverzen.net
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part
Url : http://codespeak.net/pipermail/z3-five/attachments/20070109/60931ef2/attachment.pgp 

From georgeleejr at gmail.com  Tue Jan  9 16:55:20 2007
From: georgeleejr at gmail.com (George Lee)
Date: Tue, 9 Jan 2007 15:55:20 +0000 (UTC)
Subject: [z3-five] Customizing ZopeTwoPageTemplates through the web
References: <loom.20070107T205454-766@post.gmane.org>
	<loom.20070109T005638-289@post.gmane.org>
	<eo0437$r97$4@sea.gmane.org>
Message-ID: <loom.20070109T165433-915@post.gmane.org>

Martin Aspeli <optilude <at> gmx.net> writes:

> 
> George Lee wrote:
> > George Lee <georgeleejr <at> gmail.com> writes:
> > 
> >> Hi,
> >>
> >> What is the progress on work being done to make Zope3's ZopeTwoPageTemplates
> >> customizable through the web?
> 
> No, plone.app.layout has nothing to do with this


Whoops, I meant plone.app.form:

http://dev.plone.org/plone/browser/plone.app.form/trunk/plone/app/form/_named.py

But five.customerize is really the one to look at.


Peace,
George


From chris at simplistix.co.uk  Wed Jan 10 12:55:06 2007
From: chris at simplistix.co.uk (Chris Withers)
Date: Wed, 10 Jan 2007 11:55:06 +0000
Subject: [z3-five] [Fwd: [Zope-dev] [Warning] Zope 3 component architecture
 (CA) not reliably usable for registrations from Python]
Message-ID: <45A4D41A.8080008@simplistix.co.uk>

Does this belong on this list or zope3-dev?

Chris

-- 
Simplistix - Content Management, Zope & Python Consulting
            - http://www.simplistix.co.uk
-------------- next part --------------
An embedded message was scrubbed...
From: "Dieter Maurer" <dieter at handshake.de>
Subject: [Zope-dev] [Warning] Zope 3 component architecture (CA) not
	reliably usable for registrations from Python
Date: Wed, 10 Jan 2007 10:01:23 +0100
Size: 3958
Url: http://codespeak.net/pipermail/z3-five/attachments/20070110/3d660617/attachment-0001.eml 

From chris at simplistix.co.uk  Wed Jan 10 12:53:13 2007
From: chris at simplistix.co.uk (Chris Withers)
Date: Wed, 10 Jan 2007 11:53:13 +0000
Subject: [z3-five] MailHosts as utilities
In-Reply-To: <1168353614.21802.331.camel@localhost>
References: <4577E35D.6030805@simplistix.co.uk>	<319e029f0612070320y5e6e84c0ye466d68acfb0b5d0@mail.gmail.com>	<86D9AF3A9BF3B999CD88A467@suxmac2.local>	<45A2CFD2.4050309@weitershausen.de>
	<45A350A0.2020901@simplistix.co.uk>	<CC1FF7F1-343C-4383-B65B-864D1F9B1B08@weitershausen.de>
	<1168353614.21802.331.camel@localhost>
Message-ID: <45A4D3A9.9010408@simplistix.co.uk>

Rocky Burt wrote:
> Well, whatever interface we end up using the need I see is as follows:
> 
>   1) use existing zope3 mail interface (or create new one like
> IMailHost)

I wonder what the existing zope3 mail interfaces looks like? I'd be 
pleasantly surprised if MailHost already has the required methods :-)

>   2) make sure global utility exists for this interface

Really? I'd prefer no global utility were available by default - 
possible spam sources should not exist by default...

>   3) make persistent version of utility that can be registered locally
> and stores it's settings internally

I'd like to see existing MailHost-ish things fulfill this roll.

>   4) introduce MailHost BBB class for CMF that proxies for the new local
> utility and spews deprecation warnings

I don't see a need for this.

cheers,

Chris

-- 
Simplistix - Content Management, Zope & Python Consulting
            - http://www.simplistix.co.uk


From philipp at weitershausen.de  Wed Jan 10 18:18:20 2007
From: philipp at weitershausen.de (Philipp von Weitershausen)
Date: Wed, 10 Jan 2007 18:18:20 +0100
Subject: [z3-five] [Fwd: [Zope-dev] [Warning] Zope 3 component
 architecture (CA) not reliably usable for registrations from Python]
In-Reply-To: <45A4D41A.8080008@simplistix.co.uk>
References: <45A4D41A.8080008@simplistix.co.uk>
Message-ID: <45A51FDC.4090604@weitershausen.de>

Chris Withers wrote:
> Does this belong on this list or zope3-dev?

If at all, it should've been on zope3-dev.


> Subject:
> [Zope-dev] [Warning] Zope 3 component architecture (CA) not reliably 
> usable for registrations from Python
> From:
> "Dieter Maurer" <dieter at handshake.de>
> Date:
> Wed, 10 Jan 2007 10:01:23 +0100
> To:
> zope at zope.org, zope-dev at zope.org
> 
> To:
> zope at zope.org, zope-dev at zope.org
> CC:
> 
> 
> I tried to use Zope3 events to get informed when requests start and end.
> 
> One of our modules (the interface module to "jpype") requires such
> a notification for reliable work. Therefore, it tried to register
> the corresponding subscriptions on import of this module.
> Unfortunately, this works very unreliably -- for the following reasons:
> 
> 
>   The CA performs by itself only extremely minimal initialization.
>   Especially, it does not register the "Adapters" service, necessary
>   for the registration of event subscriptions.
> 
>   The full CA initialization is only performed quite late, in the call
>   to "Products.Five.initialize".
>   Event notification registration attempts before this time will fail
>   (with a "ComponentLookupError: 'Adapters'"), 
>   those performed after this time will succeed.
> 
> The "Five" initialization is stupid enough that it will fail
> when the service registration for 'Adapters' has already been performed.
> 
> 
> At the current state, event notifications can reliably only be defined 
> via ZCML. This is unfortunate for our use case where the registration
> should be bound to the import of a given module.
> 
> 
> 


-- 
http://worldcookery.com -- Professional Zope documentation and training
2nd edition of Web Component Development with Zope 3 is now shipping!


From y.2007- at wcm-solutions.de  Wed Jan 10 18:28:42 2007
From: y.2007- at wcm-solutions.de (yuppie)
Date: Wed, 10 Jan 2007 18:28:42 +0100
Subject: [z3-five] Zope 2.10.2 coming up
In-Reply-To: <45A207A0.8080903@weitershausen.de>
References: <C36356B26D8B8C1FE9A626CE@suxmac2.local>	<45A0FA3B.5070901@wcm-solutions.de>
	<45A207A0.8080903@weitershausen.de>
Message-ID: <45A5224A.5030805@wcm-solutions.de>

Hi!


Philipp von Weitershausen wrote:
> yuppie wrote:
>>
>> The Zope 2.10 branch still uses Five 1.5.1. There are a few changes on 
>> the Five 1.5 branch (including a security fix) which should become 
>> part of Zope 2.10.2.
>>
>> I volunteer to resolve this as follows:
>>
>> Five releases are now linked tightly to specific Zope releases, so it 
>> doesn't make much sense to do still separate Five releases. See this 
>> thread:
>> http://codespeak.net/pipermail/z3-five/2006q4/001889.html
>>
>> If there are no objections, I'll stitch the Five 1.5 branch HEAD 
>> directly into the Zope 2.10 branch (without using svn externals) and 
>> Five trunk HEAD into the Zope trunk. This would mean that from now on 
>> Five 1.5 and 1.6 are maintained as part of Zope 2. In fact there will 
>> be no Five 1.5.2 or 1.6 because there will no longer be any separate 
>> releases.
>>
>> Thoughts?
> 
> + sys.maxint
> 

For the records:

After some more off-list discussion with Philipp I decided to move only 
the trunk. This is done now.

The 1.5 branch still exists and Philipp tagged Five 1.5.2 for Zope 2.10.2.


Cheers,

	Yuppie


From lists at zopyx.com  Wed Jan 10 18:31:29 2007
From: lists at zopyx.com (Andreas Jung)
Date: Wed, 10 Jan 2007 18:31:29 +0100
Subject: [z3-five] [Zope-dev] Re: Zope 2.10.2 coming up
In-Reply-To: <45A5224A.5030805@wcm-solutions.de>
References: <C36356B26D8B8C1FE9A626CE@suxmac2.local>
	<45A0FA3B.5070901@wcm-solutions.de>	<45A207A0.8080903@weitershausen.de>
	<45A5224A.5030805@wcm-solutions.de>
Message-ID: <07F09054C101046229406752@217-22-205-96.rev-ptr.fmsweb.de>



--On 10. Januar 2007 18:28:42 +0100 yuppie <y.2007- at wcm-solutions.de> wrote:

> Hi!
>
>
> Philipp von Weitershausen wrote:
>> yuppie wrote:
>>>
>>> The Zope 2.10 branch still uses Five 1.5.1. There are a few changes on
>>> the Five 1.5 branch (including a security fix) which should become
>>> part of Zope 2.10.2.
>>>
>>> I volunteer to resolve this as follows:
>>>
>>> Five releases are now linked tightly to specific Zope releases, so it
>>> doesn't make much sense to do still separate Five releases. See this
>>> thread:
>>> http://codespeak.net/pipermail/z3-five/2006q4/001889.html
>>>
>>> If there are no objections, I'll stitch the Five 1.5 branch HEAD
>>> directly into the Zope 2.10 branch (without using svn externals) and
>>> Five trunk HEAD into the Zope trunk. This would mean that from now on
>>> Five 1.5 and 1.6 are maintained as part of Zope 2. In fact there will
>>> be no Five 1.5.2 or 1.6 because there will no longer be any separate
>>> releases.
>>>
>>> Thoughts?
>>
>> + sys.maxint
>>
>
> For the records:
>
> After some more off-list discussion with Philipp I decided to move only
> the trunk. This is done now.
>
> The 1.5 branch still exists and Philipp tagged Five 1.5.2 for Zope 2.10.2.
>
>
>

Thank you both!

Andreas
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 186 bytes
Desc: not available
Url : http://codespeak.net/pipermail/z3-five/attachments/20070110/f11eaecc/attachment.pgp 

From philipp at weitershausen.de  Wed Jan 10 19:33:32 2007
From: philipp at weitershausen.de (Philipp von Weitershausen)
Date: Wed, 10 Jan 2007 19:33:32 +0100
Subject: [z3-five] MailHosts as utilities
In-Reply-To: <45A4D3A9.9010408@simplistix.co.uk>
References: <4577E35D.6030805@simplistix.co.uk>	<319e029f0612070320y5e6e84c0ye466d68acfb0b5d0@mail.gmail.com>	<86D9AF3A9BF3B999CD88A467@suxmac2.local>	<45A2CFD2.4050309@weitershausen.de>	<45A350A0.2020901@simplistix.co.uk>	<CC1FF7F1-343C-4383-B65B-864D1F9B1B08@weitershausen.de>	<1168353614.21802.331.camel@localhost>
	<45A4D3A9.9010408@simplistix.co.uk>
Message-ID: <45A5317C.2020302@weitershausen.de>

Chris Withers wrote:
> Rocky Burt wrote:
>> Well, whatever interface we end up using the need I see is as follows:
>>
>>   1) use existing zope3 mail interface (or create new one like
>> IMailHost)
> 
> I wonder what the existing zope3 mail interfaces looks like? I'd be 
> pleasantly surprised if MailHost already has the required methods :-)

Gee, Chris. If only you could see the source code of Zope 3 or there was 
a tool that would let you inspect every interface in Zope 3 or there was 
a book that covered this... Shame, shame...

>>   2) make sure global utility exists for this interface
> 
> Really? I'd prefer no global utility were available by default - 
> possible spam sources should not exist by default...

Right, that doesn't seem like a necessity. It's certainly not the case 
in Zope 3. Sending emails is highly deployment specific.

>>   3) make persistent version of utility that can be registered locally
>> and stores it's settings internally
> 
> I'd like to see existing MailHost-ish things fulfill this roll.

Sure, though it'd be nice to have the same interface for both global 
utilities that can be defined in Zope 3 and local ones. Then code 
wanting to send emails wouldn't have to bother which one it wnats.


-- 
http://worldcookery.com -- Professional Zope documentation and training
2nd edition of Web Component Development with Zope 3 is now shipping!

From philipp at weitershausen.de  Wed Jan 10 19:39:02 2007
From: philipp at weitershausen.de (Philipp von Weitershausen)
Date: Wed, 10 Jan 2007 19:39:02 +0100
Subject: [z3-five] HEADS UP: Five trunk no longer exists
Message-ID: <eo3bqq$rhu$1@sea.gmane.org>

Just wanted to let you know that Yvo has integrated Five trunk into the 
Zope 2 trunk as planned. Thank you Yvo.

Five 1.5.x was not moved and probably won't be.

Bugfixes now need to go into Five 1.3, 1.4, 1.5 and the Zope 2 trunk.


-- 
http://worldcookery.com -- Professional Zope documentation and training
2nd edition of Web Component Development with Zope 3 is now shipping!


From chris at simplistix.co.uk  Wed Jan 10 22:06:55 2007
From: chris at simplistix.co.uk (Chris Withers)
Date: Wed, 10 Jan 2007 21:06:55 +0000
Subject: [z3-five] HEADS UP: Five trunk no longer exists
In-Reply-To: <eo3bqq$rhu$1@sea.gmane.org>
References: <eo3bqq$rhu$1@sea.gmane.org>
Message-ID: <45A5556F.4040009@simplistix.co.uk>

Philipp von Weitershausen wrote:
> Just wanted to let you know that Yvo has integrated Five trunk into the 
> Zope 2 trunk as planned. Thank you Yvo.

I'd recommend actually deleting the trunk then, if this hasn't already 
been done...

> Five 1.5.x was not moved and probably won't be.
> 
> Bugfixes now need to go into Five 1.3, 1.4, 1.5 and the Zope 2 trunk.

Wow, that's a lot of maintenance branches :-(

Chris

-- 
Simplistix - Content Management, Zope & Python Consulting
            - http://www.simplistix.co.uk


From chris at simplistix.co.uk  Wed Jan 10 22:05:17 2007
From: chris at simplistix.co.uk (Chris Withers)
Date: Wed, 10 Jan 2007 21:05:17 +0000
Subject: [z3-five] MailHosts as utilities
In-Reply-To: <45A5317C.2020302@weitershausen.de>
References: <4577E35D.6030805@simplistix.co.uk>	<319e029f0612070320y5e6e84c0ye466d68acfb0b5d0@mail.gmail.com>	<86D9AF3A9BF3B999CD88A467@suxmac2.local>	<45A2CFD2.4050309@weitershausen.de>	<45A350A0.2020901@simplistix.co.uk>	<CC1FF7F1-343C-4383-B65B-864D1F9B1B08@weitershausen.de>	<1168353614.21802.331.camel@localhost>
	<45A4D3A9.9010408@simplistix.co.uk>
	<45A5317C.2020302@weitershausen.de>
Message-ID: <45A5550D.8080108@simplistix.co.uk>

Philipp von Weitershausen wrote:
>>>   3) make persistent version of utility that can be registered locally
>>> and stores it's settings internally
>>
>> I'd like to see existing MailHost-ish things fulfill this roll.
> 
> Sure, though it'd be nice to have the same interface for both global 
> utilities that can be defined in Zope 3 and local ones. Then code 
> wanting to send emails wouldn't have to bother which one it wnats.

Isn't this what adapters are for?

It'd be unpleasantly surprised if an adapter from Zope 3's mailer 
interface to IMailHost isn't quite easy to right.

Sadly, I'm guessing if you try to get a utility that implements 
IMailHost, it won't pick up global zope3 mailers, even though there is 
an adapter for them...

cheers,

Chris

-- 
Simplistix - Content Management, Zope & Python Consulting
            - http://www.simplistix.co.uk


From philipp at weitershausen.de  Thu Jan 11 00:10:04 2007
From: philipp at weitershausen.de (Philipp von Weitershausen)
Date: Thu, 11 Jan 2007 00:10:04 +0100
Subject: [z3-five] HEADS UP: Five trunk no longer exists
In-Reply-To: <45A5556F.4040009@simplistix.co.uk>
References: <eo3bqq$rhu$1@sea.gmane.org> <45A5556F.4040009@simplistix.co.uk>
Message-ID: <B993EF19-8D22-4629-BDA5-2D7F0056C437@weitershausen.de>

On 10 Jan 2007, at 22:06 , Chris Withers wrote:
> Philipp von Weitershausen wrote:
>> Just wanted to let you know that Yvo has integrated Five trunk  
>> into the Zope 2 trunk as planned. Thank you Yvo.
>
> I'd recommend actually deleting the trunk then, if this hasn't  
> already been done...

It's been done. It's just a README.txt pointing to the new location now.

>> Five 1.5.x was not moved and probably won't be.
>> Bugfixes now need to go into Five 1.3, 1.4, 1.5 and the Zope 2 trunk.
>
> Wow, that's a lot of maintenance branches :-(

Tell me about it.




From faassen at startifact.com  Thu Jan 11 19:38:10 2007
From: faassen at startifact.com (Martijn Faassen)
Date: Thu, 11 Jan 2007 19:38:10 +0100
Subject: [z3-five] registering Browser Skins
Message-ID: <eo606d$kdk$1@sea.gmane.org>

Hi there,

I think I just noticed a missing utility registration in Five (in Zope 
2.10.1). I am unable to access the 'Browser Skins' vocabulary unless I 
add this to my own ZCML:

   <utility
      name="Browser Skins"
  component="zope.app.publisher.browser.vocabulary.BrowserSkinsVocabulary"
     />

I think that this can safely be part of Five. What do you think?

Regards,

Martijn


From philipp at weitershausen.de  Fri Jan 12 00:52:58 2007
From: philipp at weitershausen.de (Philipp von Weitershausen)
Date: Fri, 12 Jan 2007 00:52:58 +0100
Subject: [z3-five] registering Browser Skins
In-Reply-To: <eo606d$kdk$1@sea.gmane.org>
References: <eo606d$kdk$1@sea.gmane.org>
Message-ID: <45A6CDDA.5070300@weitershausen.de>

Martijn Faassen wrote:
> Hi there,
> 
> I think I just noticed a missing utility registration in Five (in Zope 
> 2.10.1). I am unable to access the 'Browser Skins' vocabulary unless I 
> add this to my own ZCML:
> 
>    <utility
>       name="Browser Skins"
>   component="zope.app.publisher.browser.vocabulary.BrowserSkinsVocabulary"
>      />
> 
> I think that this can safely be part of Five. What do you think?

+1

-- 
http://worldcookery.com -- Professional Zope documentation and training
2nd edition of Web Component Development with Zope 3 is now shipping!


From d.schulz81 at gmx.net  Tue Jan 16 10:40:29 2007
From: d.schulz81 at gmx.net (Dennis Schulz)
Date: Tue, 16 Jan 2007 10:40:29 +0100
Subject: [z3-five] local utility with zope 2.9.5
Message-ID: <45AC9D8D.2020808@gmx.net>

Hi,

I would like (no I not really like, I have to) to register a local 
utility with zope 2.9.5 and Five 1.4.2.
(I know that it is more trouble than with the new zope version but I 
really have to finish something on the current Plone 2.5.)

I found the following example:

https://svn.openplans.org/svn/topp.rose/trunk/utility.py


I tried it out (see below code) and I get always the message 
"IPossibleSite" not provided.
Do I have to register the utility on my Application or on the Plone 
Site? Both of them do not provide IPossibleSite.
When I explicitly provide IPossibleSite 
(directlyProvides(app,IPossibleSite)) the call  sm = 
app.getSiteManager() raises an attribute error afterwards.....

Below there code of the registration is attached.
I call it in the Extensions/install.py of my archetype based product. 
Anything wrong with this?

Could anyone tell me if this is the right approach or point me to some 
other example?

Dennis





from ore.annotation import 
UtilityAnnotation,annotationAdapter,AnnotationUtility
from ore.annotation.interfaces import IAnnotationUtility,IUtilityAnnotatable
try:
    from OFS.Application import Application
    from OFS.SimpleItem import SimpleItem

    class OFSAnnotationUtility(SimpleItem, AnnotationUtility):
        """
        Zope2 friendly Util
        """
except ImportError:
    pass

def install(self):
   
    out = StringIO()
    installTypes(self, out, listTypes(PROJECTNAME), PROJECTNAME)
    install_subskin(self, out, GLOBALS)
    """
    add local utility in zope2
    """
    app = self
 
    print "before"
    print app
       
    from zope.app.component.interfaces import IPossibleSite
    from zope.interface import directlyProvides
    from Products.Five.site.localsite import enableLocalSiteHook
    findroot = True

    # from alecm's 'listen'
    if findroot:
        # find the zope application root
        app = aq_inner(app)
        while app is not None and not isinstance(app, Application):
          
            app = aq_parent(app)

    print "after"
    print app

    # If we have the zope Application and the utility is not yet
    # registered, then register it.
    if (app is not None and
        zapi.queryUtility(IAnnotationUtility, default=None) is None):
        enableLocalSiteHook(app)
        setSite(app)
        sm = app.getSiteManager()
        sm.registerUtility(IAnnotationUtility, AnnotationUtility(NAME))

    out.write("Successfully Yeah Yeah installed %s." % PROJECTNAME)
    return out.getvalue()









From rocky at serverzen.com  Tue Jan 16 12:47:27 2007
From: rocky at serverzen.com (Rocky Burt)
Date: Tue, 16 Jan 2007 08:17:27 -0330
Subject: [z3-five] local utility with zope 2.9.5
In-Reply-To: <45AC9D8D.2020808@gmx.net>
References: <45AC9D8D.2020808@gmx.net>
Message-ID: <1168948044.7724.43.camel@localhost>

On Tue, 2007-16-01 at 10:40 +0100, Dennis Schulz wrote:
> I tried it out (see below code) and I get always the message 
> "IPossibleSite" not provided.
> Do I have to register the utility on my Application or on the Plone 
> Site? Both of them do not provide IPossibleSite.
> When I explicitly provide IPossibleSite 
> (directlyProvides(app,IPossibleSite)) the call  sm = 
> app.getSiteManager() raises an attribute error afterwards.....
> 
> Below there code of the registration is attached.
> I call it in the Extensions/install.py of my archetype based product. 
> Anything wrong with this?

Your install code looks good and yes on Zope 2.10 that would just work.
The problem is that none of the container classes provided by Zope 2.9
are actually proper IPossibleSite's.  But Five provides a way to make
them IPossibleSite's properly.

I believe the two steps you're missing are better suited to ZCML.

1) make the class implement IPossibleSite
    <class class="OFS.Application.Application">
        <implements
            interface="zope.app.site.interfaces.IPossibleSite"
            />
    </class>

2) monkey patch Application class to properly meet the IPossibleSite
contract
    <five:localsite
        class="OFS.Application.Application"
        />

Basically in Zope 2.9 you need to do those steps before you can try to
treat any class instance as an IPossibleSite (in my apps I tend to make
the plone site a local site by doing something similar).

- Rocky

-- 
Rocky Burt
ServerZen Software -- http://www.serverzen.com
News About The Server (blog) -- http://www.serverzen.net
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part
Url : http://codespeak.net/pipermail/z3-five/attachments/20070116/800c8d40/attachment.pgp 

From d.schulz81 at gmx.net  Tue Jan 16 13:15:28 2007
From: d.schulz81 at gmx.net (Dennis Schulz)
Date: Tue, 16 Jan 2007 13:15:28 +0100
Subject: [z3-five]  local utility with zope 2.9.5
Message-ID: <45ACC1E0.1060805@gmx.net>

I registed Plone Site as Five Site and tried to register the local utility.

   <five:localsite class="Products.CMFPlone.Portal.PloneSite" />

now the install script goes further, but ends up with

TypeError: ('Could not adapt', <PloneSite at /dd>, <InterfaceClass 
Products.Five.site.interfaces.IFiveUtilityRegistry>)


I also made my own content type that subclasses FiveSite and implementes 
IPossibleSite and for my needs it would be perfectly sufficient to 
register the utility only inside the content type.........

.......but how can I reach it from the install method of my zope 2 product?
.......or how to do it in general?


2007-01-16 12:55:36 ERROR Zope.SiteErrorLog 
http://localhost:8080/dd/portal_quickinstaller/installProducts
Traceback (innermost last):
   Module Products.CMFQuickInstallerTool.QuickInstallerTool, line 326, 
in installProduct
    - __traceback_info__: ('PurchaseRequisition',)
   Module transaction._manager, line 96, in commit
   Module transaction._transaction, line 366, in commit
   Module transaction._transaction, line 295, in savepoint
   Module transaction._transaction, line 292, in savepoint
   Module transaction._transaction, line 675, in __init__
   Module ZODB.Connection, line 1012, in savepoint
   Module ZODB.Connection, line 526, in _commit
   Module ZODB.Connection, line 553, in _store_objects
   Module ZODB.serialize, line 407, in serialize
   Module ZODB.serialize, line 415, in _dump
PicklingError: Can't pickle <class 'OFSAnnotationUtility'>: attribute 
lookup __builtin__.OFSAnnotationUtility failed
C:\Python\Zope\Instance\Products\CMFQuickInstallerTool\QuickInstallerTool.py:348: 
DeprecationWarning: This will be removed in ZODB 3.7:
subtransactions are deprecated; use sp.rollback() instead of 
transaction.abort(1), where `sp` is the corresponding savepoint captured 
earlier
   transaction.abort(1)   #this is very naughty
2007-01-16 12:55:36 ERROR Zope.SiteErrorLog 
http://localhost:8080/dd/portal_quickinstaller/installProducts
Traceback (innermost last):
   Module Products.CMFQuickInstallerTool.QuickInstallerTool, line 413, 
in installProduct
    - __traceback_info__: ('PurchaseRequisition',)
   Module Products.CMFQuickInstallerTool.InstalledProduct, line 99, in 
update
   Module ZODB.Connection, line 857, in register
   Module ZODB.Connection, line 867, in _register
   Module transaction._transaction, line 256, in join
   Module transaction._transaction, line 250, in _prior_operation_failed
TransactionFailedError: An operation previously failed, with traceback:

   File 
"C:\Python\Zope\Zope\lib\python\ZServer\PubCore\ZServerPublisher.py", 
line 23, in __init__
     response=response)
   File "C:\Python\Zope\Zope\lib\python\ZPublisher\Publish.py", line 
395, in publish_module
     environ, debug, request, response)
   File "C:\Python\Zope\Zope\lib\python\ZPublisher\Publish.py", line 
196, in publish_module_standard
     response = publish(request, module_name, after_list, debug=debug)
   File 
"C:\Python\Zope\Instance\Products\PlacelessTranslationService\PatchStringIO.py", 
line 34, in new_publish
     x = Publish.old_publish(request, module_name, after_list, debug)
   File "C:\Python\Zope\Zope\lib\python\ZPublisher\Publish.py", line 
115, in publish
     request, bind=1)
   File "C:\Python\Zope\Zope\lib\python\ZPublisher\mapply.py", line 88, 
in mapply
     if debug is not None: return debug(object,args,context)
   File "C:\Python\Zope\Zope\lib\python\ZPublisher\Publish.py", line 41, 
in call_object
     result=apply(object,args) # Type s<cr> to step into published object.
   File 
"C:\Python\Zope\Instance\Products\CMFQuickInstallerTool\QuickInstallerTool.py", 
line 454, in installProducts
     reinstall=reinstall)
   File 
"C:\Python\Zope\Instance\Products\CMFQuickInstallerTool\QuickInstallerTool.py", 
line 326, in installProduct
     transaction.commit(1)
   File "C:\Python\Zope\Zope\lib\python\transaction\_manager.py", line 
96, in commit
     return self.get().commit(sub, deprecation_wng=False)
   File "C:\Python\Zope\Zope\lib\python\transaction\_transaction.py", 
line 366, in commit
     self._subtransaction_savepoint = self.savepoint(optimistic=True)
   File "C:\Python\Zope\Zope\lib\python\transaction\_transaction.py", 
line 295, in savepoint
     self._saveCommitishError() # reraises!
   File "C:\Python\Zope\Zope\lib\python\transaction\_transaction.py", 
line 292, in savepoint
     savepoint = Savepoint(self, optimistic, *self._resources)
   File "C:\Python\Zope\Zope\lib\python\transaction\_transaction.py", 
line 675, in __init__
     savepoint = savepoint()
   File "C:\Python\Zope\Zope\lib\python\ZODB\Connection.py", line 1012, 
in savepoint
     self._commit(None)
   File "C:\Python\Zope\Zope\lib\python\ZODB\Connection.py", line 526, 
in _commit
     self._store_objects(ObjectWriter(obj), transaction)
   File "C:\Python\Zope\Zope\lib\python\ZODB\Connection.py", line 553, 
in _store_objects
     p = writer.serialize(obj)  # This calls __getstate__ of obj
   File "C:\Python\Zope\Zope\lib\python\ZODB\serialize.py", line 407, in 
serialize
     return self._dump(meta, obj.__getstate__())
   File "C:\Python\Zope\Zope\lib\python\ZODB\serialize.py", line 415, in 
_dump
     self._p.dump(classmeta)
PicklingError: Can't pickle <class 'OFSAnnotationUtility'>: attribute 
lookup __builtin__.OFSAnnotationUtility failed
C:\Python\Zope\Instance\Products\CMFQuickInstallerTool\QuickInstallerTool.py:426: 
DeprecationWarning: This will be removed in ZODB 3.7:
subtransactions are deprecated; use sp.rollback() instead of 
transaction.abort(1), where `sp` is the corresponding savepoint captured 
earlier
   transaction.abort(1)   #this is very naughty
2007-01-16 12:55:36 ERROR Zope.SiteErrorLog 
http://localhost:8080/dd/portal_quickinstaller/installProducts
Traceback (innermost last):
   Module ZPublisher.Publish, line 121, in publish
   Module Zope2.App.startup, line 240, in commit
   Module transaction._manager, line 96, in commit
   Module transaction._transaction, line 370, in commit
   Module transaction._transaction, line 250, in _prior_operation_failed
TransactionFailedError: An operation previously failed, with traceback:

   File 
"C:\Python\Zope\Zope\lib\python\ZServer\PubCore\ZServerPublisher.py", 
line 23, in __init__
     response=response)
   File "C:\Python\Zope\Zope\lib\python\ZPublisher\Publish.py", line 
395, in publish_module
     environ, debug, request, response)
   File "C:\Python\Zope\Zope\lib\python\ZPublisher\Publish.py", line 
196, in publish_module_standard
     response = publish(request, module_name, after_list, debug=debug)
   File 
"C:\Python\Zope\Instance\Products\PlacelessTranslationService\PatchStringIO.py", 
line 34, in new_publish
     x = Publish.old_publish(request, module_name, after_list, debug)
   File "C:\Python\Zope\Zope\lib\python\ZPublisher\Publish.py", line 
115, in publish
     request, bind=1)
   File "C:\Python\Zope\Zope\lib\python\ZPublisher\mapply.py", line 88, 
in mapply
     if debug is not None: return debug(object,args,context)
   File "C:\Python\Zope\Zope\lib\python\ZPublisher\Publish.py", line 41, 
in call_object
     result=apply(object,args) # Type s<cr> to step into published object.
   File 
"C:\Python\Zope\Instance\Products\CMFQuickInstallerTool\QuickInstallerTool.py", 
line 454, in installProducts
     reinstall=reinstall)
   File 
"C:\Python\Zope\Instance\Products\CMFQuickInstallerTool\QuickInstallerTool.py", 
line 326, in installProduct
     transaction.commit(1)
   File "C:\Python\Zope\Zope\lib\python\transaction\_manager.py", line 
96, in commit
     return self.get().commit(sub, deprecation_wng=False)
   File "C:\Python\Zope\Zope\lib\python\transaction\_transaction.py", 
line 366, in commit
     self._subtransaction_savepoint = self.savepoint(optimistic=True)
   File "C:\Python\Zope\Zope\lib\python\transaction\_transaction.py", 
line 295, in savepoint
     self._saveCommitishError() # reraises!
   File "C:\Python\Zope\Zope\lib\python\transaction\_transaction.py", 
line 292, in savepoint
     savepoint = Savepoint(self, optimistic, *self._resources)
   File "C:\Python\Zope\Zope\lib\python\transaction\_transaction.py", 
line 675, in __init__
     savepoint = savepoint()
   File "C:\Python\Zope\Zope\lib\python\ZODB\Connection.py", line 1012, 
in savepoint
     self._commit(None)
   File "C:\Python\Zope\Zope\lib\python\ZODB\Connection.py", line 526, 
in _commit
     self._store_objects(ObjectWriter(obj), transaction)
   File "C:\Python\Zope\Zope\lib\python\ZODB\Connection.py", line 553, 
in _store_objects
     p = writer.serialize(obj)  # This calls __getstate__ of obj
   File "C:\Python\Zope\Zope\lib\python\ZODB\serialize.py", line 407, in 
serialize
     return self._dump(meta, obj.__getstate__())
   File "C:\Python\Zope\Zope\lib\python\ZODB\serialize.py", line 415, in 
_dump
     self._p.dump(classmeta)
PicklingError: Can't pickle <class 'OFSAnnotationUtility'>: attribute 
lookup __builtin__.OFSAnnotationUtility failed
2007-01-16 12:55:36 ERROR root Exception while rendering an error message
Traceback (most recent call last):
   File "C:\Python\Zope\Zope\lib\python\OFS\SimpleItem.py", line 223, in 
raise_standardErrorMessage
     v = s(**kwargs)
   File "C:\Python\Zope\Instance\Products\CMFCore\FSPythonScript.py", 
line 108, in __call__
     return Script.__call__(self, *args, **kw)
   File "C:\Python\Zope\Zope\lib\python\Shared\DC\Scripts\Bindings.py", 
line 311, in __call__
     return self._bindAndExec(args, kw, None)
   File "C:\Python\Zope\Zope\lib\python\Shared\DC\Scripts\Bindings.py", 
line 348, in _bindAndExec
     return self._exec(bound_data, args, kw)
   File "C:\Python\Zope\Instance\Products\CMFCore\FSPythonScript.py", 
line 164, in _exec
     result = f(*args, **kw)
   File "Script (Python)", line 16, in standard_error_message
   File "C:\Python\Zope\Zope\lib\python\Shared\DC\Scripts\Bindings.py", 
line 311, in __call__
     return self._bindAndExec(args, kw, None)
   File "C:\Python\Zope\Zope\lib\python\Shared\DC\Scripts\Bindings.py", 
line 348, in _bindAndExec
     return self._exec(bound_data, args, kw)
   File "C:\Python\Zope\Instance\Products\CMFCore\FSPageTemplate.py", 
line 195, in _exec
     result = self.pt_render(extra_context=bound_names)
   File "C:\Python\Zope\Instance\Products\CMFCore\FSPageTemplate.py", 
line 134, in pt_render
     result = FSPageTemplate.inheritedAttribute('pt_render')(
   File 
"C:\Python\Zope\Zope\lib\python\Products\PageTemplates\PageTemplate.py", 
line 104, in pt_render
     tal=not source, strictinsert=0)()
   File "C:\Python\Zope\Zope\lib\python\TAL\TALInterpreter.py", line 
238, in __call__
     self.interpret(self.program)
   File "C:\Python\Zope\Zope\lib\python\TAL\TALInterpreter.py", line 
281, in interpret
     handlers[opcode](self, args)
   File "C:\Python\Zope\Zope\lib\python\TAL\TALInterpreter.py", line 
749, in do_useMacro
     self.interpret(macro)
   File "C:\Python\Zope\Zope\lib\python\TAL\TALInterpreter.py", line 
281, in interpret
     handlers[opcode](self, args)
   File "C:\Python\Zope\Zope\lib\python\TAL\TALInterpreter.py", line 
457, in do_optTag_tal
     self.do_optTag(stuff)
   File "C:\Python\Zope\Zope\lib\python\TAL\TALInterpreter.py", line 
442, in do_optTag
     return self.no_tag(start, program)
   File "C:\Python\Zope\Zope\lib\python\TAL\TALInterpreter.py", line 
437, in no_tag
     self.interpret(program)
   File "C:\Python\Zope\Zope\lib\python\TAL\TALInterpreter.py", line 
281, in interpret
     handlers[opcode](self, args)
   File "C:\Python\Zope\Zope\lib\python\TAL\TALInterpreter.py", line 
749, in do_useMacro
     self.interpret(macro)
   File "C:\Python\Zope\Zope\lib\python\TAL\TALInterpreter.py", line 
281, in interpret
     handlers[opcode](self, args)
   File "C:\Python\Zope\Zope\lib\python\TAL\TALInterpreter.py", line 
507, in do_setLocal_tal
     self.engine.setLocal(name, self.engine.evaluateValue(expr))
   File 
"C:\Python\Zope\Zope\lib\python\Products\PageTemplates\TALES.py", line 
221, in evaluate
     return expression(self)
   File 
"C:\Python\Zope\Zope\lib\python\Products\PageTemplates\Expressions.py", 
line 185, in __call__
     return self._eval(econtext)
   File 
"C:\Python\Zope\Zope\lib\python\Products\PageTemplates\Expressions.py", 
line 180, in _eval
     return render(ob, econtext.vars)
   File 
"C:\Python\Zope\Zope\lib\python\Products\PageTemplates\Expressions.py", 
line 85, in render
     ob = ob()
   File "C:\Python\Zope\Instance\Products\CMFPlone\browser\plone.py", 
line 66, in globalize
     self._initializeData(options=options)
   File "C:\Python\Zope\Instance\Products\CMFPlone\browser\plone.py", 
line 118, in _initializeData
     self._data['actions'] = actions = (options.get('actions', None) or
   File "C:\Python\Zope\Instance\Products\CMFPlone\ActionsTool.py", line 
46, in listFilteredActionsFor
     [self._getActions(provider_name, actions, object) \
   File "C:\Python\Zope\Instance\Products\CMFPlone\ActionsTool.py", line 
34, in _getActions
     actions.extend( provider.listActionInfos(object=object) )
   File 
"C:\Python\Zope\Instance\Products\CMFCore\ActionProviderBase.py", line 
86, in listActionInfos
     actions = self.listActions(object=object)
   File "C:\PLONE\Data\Products\CMFonFive\fiveactionstool.py", line 139, 
in listActions
     for menu_id in _listMenuIds():
   File "C:\PLONE\Data\Products\CMFonFive\fiveactionstool.py", line 32, 
in _listMenuIds
     return [id for id, utility in zapi.getUtilitiesFor(IBrowserMenu)]
   File "C:\Python\Zope\Zope\lib\python\zope\component\__init__.py", 
line 263, in getUtilitiesFor
     return getSiteManager(context).getUtilitiesFor(interface)
   File "C:\PLONE\Data\Products\Five\site\localsite.py", line 142, in 
getUtilitiesFor
   File "C:\PLONE\Data\Products\Five\site\localsite.py", line 121, in 
utilities
   File "C:\Python\Zope\Zope\lib\python\zope\interface\interface.py", 
line 682, in __call__
     raise TypeError("Could not adapt", obj, self)
TypeError: ('Could not adapt', <PloneSite at /dd>, <InterfaceClass 
Products.Five.site.interfaces.IFiveUtilityRegistry>)


From d.schulz81 at gmx.net  Tue Jan 16 14:05:40 2007
From: d.schulz81 at gmx.net (Dennis Schulz)
Date: Tue, 16 Jan 2007 14:05:40 +0100
Subject: [z3-five] local utility with zope 2.9.5
In-Reply-To: <1168948044.7724.43.camel@localhost>
References: <45AC9D8D.2020808@gmx.net> <1168948044.7724.43.camel@localhost>
Message-ID: <45ACCDA4.9040800@gmx.net>

Thanks for the help,
 
now I money patched OFS.Application as a site, but unfortunately the 
same error as when I try the register on the Plone Site. (See traceback 
at the end.)

Can't pickle <class 'OFSAnnotationUtility'>, maybe is it bad to mix it 
with OFS.SimpleItem to get it zope2 compatible? Or IOBTree can't be pickled?

Here is the utility I want to register:

class AnnotationUtility( Persistent ):
   
    def __init__(self):
        self.annotations = IOBTree()

    def get( self, context ):
        utility = component.getUtility( IIntIds, context=self )
        iid = utility.getId( context )
        if not self.annotations.has_key( iid ):
            self.create( iid )
        return iid, self.annotations.get( iid )

    def create( self, intid ):
        state = OOBTree()
        self.annotations[ intid ] = state

2007-01-16 14:00:46 ERROR Zope.SiteErrorLog 
http://localhost:8080/dd/portal_quickinstaller/installProducts
Traceback (innermost last):
  Module Products.CMFQuickInstallerTool.QuickInstallerTool, line 326, in 
installProduct
   - __traceback_info__: ('PurchaseRequisition',)
  Module transaction._manager, line 96, in commit
  Module transaction._transaction, line 366, in commit
  Module transaction._transaction, line 295, in savepoint
  Module transaction._transaction, line 292, in savepoint
  Module transaction._transaction, line 675, in __init__
  Module ZODB.Connection, line 1012, in savepoint
  Module ZODB.Connection, line 526, in _commit
  Module ZODB.Connection, line 553, in _store_objects
  Module ZODB.serialize, line 407, in serialize
  Module ZODB.serialize, line 415, in _dump
PicklingError: Can't pickle <class 'OFSAnnotationUtility'>: attribute 
lookup __builtin__.OFSAnnotationUtility failed
C:\Python\Zope\Instance\Products\CMFQuickInstallerTool\QuickInstallerTool.py:348: 
DeprecationWarning: This will be removed in ZODB 3.7:
subtransactions are deprecated; use sp.rollback() instead of 
transaction.abort(1), where `sp` is the corresponding savepoint captured 
earlier
  transaction.abort(1)   #this is very naughty
2007-01-16 14:00:46 ERROR Zope.SiteErrorLog 
http://localhost:8080/dd/portal_quickinstaller/installProducts
Traceback (innermost last):
  Module Products.CMFQuickInstallerTool.QuickInstallerTool, line 413, in 
installProduct
   - __traceback_info__: ('PurchaseRequisition',)
  Module Products.CMFQuickInstallerTool.InstalledProduct, line 99, in update
  Module ZODB.Connection, line 857, in register
  Module ZODB.Connection, line 867, in _register
  Module transaction._transaction, line 256, in join
  Module transaction._transaction, line 250, in _prior_operation_failed
TransactionFailedError: An operation previously failed, with traceback:

  File 
"C:\Python\Zope\Zope\lib\python\ZServer\PubCore\ZServerPublisher.py", 
line 23, in __init__
    response=response)
  File "C:\Python\Zope\Zope\lib\python\ZPublisher\Publish.py", line 395, 
in publish_module
    environ, debug, request, response)
  File "C:\Python\Zope\Zope\lib\python\ZPublisher\Publish.py", line 196, 
in publish_module_standard
    response = publish(request, module_name, after_list, debug=debug)
  File 
"C:\Python\Zope\Instance\Products\PlacelessTranslationService\PatchStringIO.py", 
line 34, in new_publish
    x = Publish.old_publish(request, module_name, after_list, debug)
  File "C:\Python\Zope\Zope\lib\python\ZPublisher\Publish.py", line 115, 
in publish
    request, bind=1)
  File "C:\Python\Zope\Zope\lib\python\ZPublisher\mapply.py", line 88, 
in mapply
    if debug is not None: return debug(object,args,context)
  File "C:\Python\Zope\Zope\lib\python\ZPublisher\Publish.py", line 41, 
in call_object
    result=apply(object,args) # Type s<cr> to step into published object.
  File 
"C:\Python\Zope\Instance\Products\CMFQuickInstallerTool\QuickInstallerTool.py", 
line 454, in installProducts
    reinstall=reinstall)
  File 
"C:\Python\Zope\Instance\Products\CMFQuickInstallerTool\QuickInstallerTool.py", 
line 326, in installProduct
    transaction.commit(1)
  File "C:\Python\Zope\Zope\lib\python\transaction\_manager.py", line 
96, in commit
    return self.get().commit(sub, deprecation_wng=False)
  File "C:\Python\Zope\Zope\lib\python\transaction\_transaction.py", 
line 366, in commit
    self._subtransaction_savepoint = self.savepoint(optimistic=True)
  File "C:\Python\Zope\Zope\lib\python\transaction\_transaction.py", 
line 295, in savepoint
    self._saveCommitishError() # reraises!
  File "C:\Python\Zope\Zope\lib\python\transaction\_transaction.py", 
line 292, in savepoint
    savepoint = Savepoint(self, optimistic, *self._resources)
  File "C:\Python\Zope\Zope\lib\python\transaction\_transaction.py", 
line 675, in __init__
    savepoint = savepoint()
  File "C:\Python\Zope\Zope\lib\python\ZODB\Connection.py", line 1012, 
in savepoint
    self._commit(None)
  File "C:\Python\Zope\Zope\lib\python\ZODB\Connection.py", line 526, in 
_commit
    self._store_objects(ObjectWriter(obj), transaction)
  File "C:\Python\Zope\Zope\lib\python\ZODB\Connection.py", line 553, in 
_store_objects
    p = writer.serialize(obj)  # This calls __getstate__ of obj
  File "C:\Python\Zope\Zope\lib\python\ZODB\serialize.py", line 407, in 
serialize
    return self._dump(meta, obj.__getstate__())
  File "C:\Python\Zope\Zope\lib\python\ZODB\serialize.py", line 415, in 
_dump
    self._p.dump(classmeta)
PicklingError: Can't pickle <class 'OFSAnnotationUtility'>: attribute 
lookup __builtin__.OFSAnnotationUtility failed
C:\Python\Zope\Instance\Products\CMFQuickInstallerTool\QuickInstallerTool.py:426: 
DeprecationWarning: This will be removed in ZODB 3.7:
subtransactions are deprecated; use sp.rollback() instead of 
transaction.abort(1), where `sp` is the corresponding savepoint captured 
earlier
  transaction.abort(1)   #this is very naughty
2007-01-16 14:00:46 ERROR Zope.SiteErrorLog 
http://localhost:8080/dd/portal_quickinstaller/installProducts
Traceback (innermost last):
  Module ZPublisher.Publish, line 121, in publish
  Module Zope2.App.startup, line 240, in commit
  Module transaction._manager, line 96, in commit
  Module transaction._transaction, line 370, in commit
  Module transaction._transaction, line 250, in _prior_operation_failed
TransactionFailedError: An operation previously failed, with traceback:

  File 
"C:\Python\Zope\Zope\lib\python\ZServer\PubCore\ZServerPublisher.py", 
line 23, in __init__
    response=response)
  File "C:\Python\Zope\Zope\lib\python\ZPublisher\Publish.py", line 395, 
in publish_module
    environ, debug, request, response)
  File "C:\Python\Zope\Zope\lib\python\ZPublisher\Publish.py", line 196, 
in publish_module_standard
    response = publish(request, module_name, after_list, debug=debug)
  File 
"C:\Python\Zope\Instance\Products\PlacelessTranslationService\PatchStringIO.py", 
line 34, in new_publish
    x = Publish.old_publish(request, module_name, after_list, debug)
  File "C:\Python\Zope\Zope\lib\python\ZPublisher\Publish.py", line 115, 
in publish
    request, bind=1)
  File "C:\Python\Zope\Zope\lib\python\ZPublisher\mapply.py", line 88, 
in mapply
    if debug is not None: return debug(object,args,context)
  File "C:\Python\Zope\Zope\lib\python\ZPublisher\Publish.py", line 41, 
in call_object
    result=apply(object,args) # Type s<cr> to step into published object.
  File 
"C:\Python\Zope\Instance\Products\CMFQuickInstallerTool\QuickInstallerTool.py", 
line 454, in installProducts
    reinstall=reinstall)
  File 
"C:\Python\Zope\Instance\Products\CMFQuickInstallerTool\QuickInstallerTool.py", 
line 326, in installProduct
    transaction.commit(1)
  File "C:\Python\Zope\Zope\lib\python\transaction\_manager.py", line 
96, in commit
    return self.get().commit(sub, deprecation_wng=False)
  File "C:\Python\Zope\Zope\lib\python\transaction\_transaction.py", 
line 366, in commit
    self._subtransaction_savepoint = self.savepoint(optimistic=True)
  File "C:\Python\Zope\Zope\lib\python\transaction\_transaction.py", 
line 295, in savepoint
    self._saveCommitishError() # reraises!
  File "C:\Python\Zope\Zope\lib\python\transaction\_transaction.py", 
line 292, in savepoint
    savepoint = Savepoint(self, optimistic, *self._resources)
  File "C:\Python\Zope\Zope\lib\python\transaction\_transaction.py", 
line 675, in __init__
    savepoint = savepoint()
  File "C:\Python\Zope\Zope\lib\python\ZODB\Connection.py", line 1012, 
in savepoint
    self._commit(None)
  File "C:\Python\Zope\Zope\lib\python\ZODB\Connection.py", line 526, in 
_commit
    self._store_objects(ObjectWriter(obj), transaction)
  File "C:\Python\Zope\Zope\lib\python\ZODB\Connection.py", line 553, in 
_store_objects
    p = writer.serialize(obj)  # This calls __getstate__ of obj
  File "C:\Python\Zope\Zope\lib\python\ZODB\serialize.py", line 407, in 
serialize
    return self._dump(meta, obj.__getstate__())
  File "C:\Python\Zope\Zope\lib\python\ZODB\serialize.py", line 415, in 
_dump
    self._p.dump(classmeta)
PicklingError: Can't pickle <class 'OFSAnnotationUtility'>: attribute 
lookup __builtin__.OFSAnnotationUtility failed
2007-01-16 14:00:46 ERROR root Exception while rendering an error message
Traceback (most recent call last):
  File "C:\Python\Zope\Zope\lib\python\OFS\SimpleItem.py", line 223, in 
raise_standardErrorMessage
    v = s(**kwargs)
  File "C:\Python\Zope\Instance\Products\CMFCore\FSPythonScript.py", 
line 108, in __call__
    return Script.__call__(self, *args, **kw)
  File "C:\Python\Zope\Zope\lib\python\Shared\DC\Scripts\Bindings.py", 
line 311, in __call__
    return self._bindAndExec(args, kw, None)
  File "C:\Python\Zope\Zope\lib\python\Shared\DC\Scripts\Bindings.py", 
line 348, in _bindAndExec
    return self._exec(bound_data, args, kw)
  File "C:\Python\Zope\Instance\Products\CMFCore\FSPythonScript.py", 
line 164, in _exec
    result = f(*args, **kw)
  File "Script (Python)", line 16, in standard_error_message
  File "C:\Python\Zope\Zope\lib\python\Shared\DC\Scripts\Bindings.py", 
line 311, in __call__
    return self._bindAndExec(args, kw, None)
  File "C:\Python\Zope\Zope\lib\python\Shared\DC\Scripts\Bindings.py", 
line 348, in _bindAndExec
    return self._exec(bound_data, args, kw)
  File "C:\Python\Zope\Instance\Products\CMFCore\FSPageTemplate.py", 
line 195, in _exec
    result = self.pt_render(extra_context=bound_names)
  File "C:\Python\Zope\Instance\Products\CMFCore\FSPageTemplate.py", 
line 134, in pt_render
    result = FSPageTemplate.inheritedAttribute('pt_render')(
  File 
"C:\Python\Zope\Zope\lib\python\Products\PageTemplates\PageTemplate.py", 
line 104, in pt_render
    tal=not source, strictinsert=0)()
  File "C:\Python\Zope\Zope\lib\python\TAL\TALInterpreter.py", line 238, 
in __call__
    self.interpret(self.program)
  File "C:\Python\Zope\Zope\lib\python\TAL\TALInterpreter.py", line 281, 
in interpret
    handlers[opcode](self, args)
  File "C:\Python\Zope\Zope\lib\python\TAL\TALInterpreter.py", line 749, 
in do_useMacro
    self.interpret(macro)
  File "C:\Python\Zope\Zope\lib\python\TAL\TALInterpreter.py", line 281, 
in interpret
    handlers[opcode](self, args)
  File "C:\Python\Zope\Zope\lib\python\TAL\TALInterpreter.py", line 457, 
in do_optTag_tal
    self.do_optTag(stuff)
  File "C:\Python\Zope\Zope\lib\python\TAL\TALInterpreter.py", line 442, 
in do_optTag
    return self.no_tag(start, program)
  File "C:\Python\Zope\Zope\lib\python\TAL\TALInterpreter.py", line 437, 
in no_tag
    self.interpret(program)
  File "C:\Python\Zope\Zope\lib\python\TAL\TALInterpreter.py", line 281, 
in interpret
    handlers[opcode](self, args)
  File "C:\Python\Zope\Zope\lib\python\TAL\TALInterpreter.py", line 749, 
in do_useMacro
    self.interpret(macro)
  File "C:\Python\Zope\Zope\lib\python\TAL\TALInterpreter.py", line 281, 
in interpret
    handlers[opcode](self, args)
  File "C:\Python\Zope\Zope\lib\python\TAL\TALInterpreter.py", line 507, 
in do_setLocal_tal
    self.engine.setLocal(name, self.engine.evaluateValue(expr))
  File "C:\Python\Zope\Zope\lib\python\Products\PageTemplates\TALES.py", 
line 221, in evaluate
    return expression(self)
  File 
"C:\Python\Zope\Zope\lib\python\Products\PageTemplates\Expressions.py", 
line 185, in __call__
    return self._eval(econtext)
  File 
"C:\Python\Zope\Zope\lib\python\Products\PageTemplates\Expressions.py", 
line 180, in _eval
    return render(ob, econtext.vars)
  File 
"C:\Python\Zope\Zope\lib\python\Products\PageTemplates\Expressions.py", 
line 85, in render
    ob = ob()
  File "C:\Python\Zope\Instance\Products\CMFPlone\browser\plone.py", 
line 66, in globalize
    self._initializeData(options=options)
  File "C:\Python\Zope\Instance\Products\CMFPlone\browser\plone.py", 
line 118, in _initializeData
    self._data['actions'] = actions = (options.get('actions', None) or
  File "C:\Python\Zope\Instance\Products\CMFPlone\ActionsTool.py", line 
46, in listFilteredActionsFor
    [self._getActions(provider_name, actions, object) \
  File "C:\Python\Zope\Instance\Products\CMFPlone\ActionsTool.py", line 
34, in _getActions
    actions.extend( provider.listActionInfos(object=object) )
  File "C:\Python\Zope\Instance\Products\CMFCore\ActionProviderBase.py", 
line 86, in listActionInfos
    actions = self.listActions(object=object)
  File "C:\PLONE\Data\Products\CMFonFive\fiveactionstool.py", line 139, 
in listActions
    for menu_id in _listMenuIds():
  File "C:\PLONE\Data\Products\CMFonFive\fiveactionstool.py", line 32, 
in _listMenuIds
    return [id for id, utility in zapi.getUtilitiesFor(IBrowserMenu)]
  File "C:\Python\Zope\Zope\lib\python\zope\component\__init__.py", line 
263, in getUtilitiesFor
    return getSiteManager(context).getUtilitiesFor(interface)
  File "C:\PLONE\Data\Products\Five\site\localsite.py", line 142, in 
getUtilitiesFor
  File "C:\PLONE\Data\Products\Five\site\localsite.py", line 121, in 
utilities
  File "C:\Python\Zope\Zope\lib\python\zope\interface\interface.py", 
line 682, in __call__
    raise TypeError("Could not adapt", obj, self)
TypeError: ('Could not adapt', <Application at >, <InterfaceClass 
Products.Five.site.interfaces.IFiveUtilityRegistry>)



From d.schulz81 at gmx.net  Thu Jan 18 12:21:34 2007
From: d.schulz81 at gmx.net (Dennis Schulz)
Date: Thu, 18 Jan 2007 12:21:34 +0100
Subject: [z3-five] Adaption of workflow tool
Message-ID: <45AF583E.6090403@gmx.net>

Hello Plone List,

I have a database driven application that creates temporary mapped 
objects not intended to be persistent content types (mainly Zope 3 with 
Alchemist Product)
That worked quite well until now, but completly without workflow.
Now I would like to give workflow to these objects. (well I have to, 
that application depends heavily on it)
As all my users and groups are mapped from LDAP inside Plone, so it 
would be nice to use the Plone workflow.

To achieve this, I will have use some kind of persistent proxy object to 
store the workflow state and the reference.

I looked at Zope 3 ways of doing this, means to have a persistent 
annotation with workflow state and IKeyReference adapter that refers to 
a resolvable url and the unique database key.
But the bad and messy local utility and annotation support of Zope 2.9 
brought me out of this track................I think zope 2.9 is not 
mature enough to support this. dont even know if 2.10 is.

Now I found this document and have a look at the way the CPS proxy 
object is done, but it also relies on portal types:
http://www.nuxeo.com/docs/europython2004/cps_proxies_workflows_events.pdf

So my question is, has anyone done an adaption of plone workflow that 
does not use a portal type, but maybe a OFS.SimpleItem (that just stores 
the reference to the data base and workflow state) or can point me to 
additional material?

I dont know if this solution is good style and easily possible in any 
way, any additional help or recommendations would be appreciated.

If such an adaption is not easily possible, the fallback solution is to 
write my own little workflow engine, but I would weep a litte doing this.

Dennis


From sidnei at enfoldsystems.com  Thu Jan 18 13:57:33 2007
From: sidnei at enfoldsystems.com (Sidnei da Silva)
Date: Thu, 18 Jan 2007 10:57:33 -0200
Subject: [z3-five] Adaption of workflow tool
In-Reply-To: <45AF583E.6090403@gmx.net>
References: <45AF583E.6090403@gmx.net>
Message-ID: <a7a2b76b0701180457n4a31b306mb26c40da8d3cfa12@mail.gmail.com>

Hi Dennis,

Kapil T. send a patch to the CMF mailing list that makes the
WorkflowTool use adapters to lookup the workflow and also to store the
workflow information I believe. It should be possible, with that
patch, to write an adapter that can even store the workflow
information in your RDB.

-- 
Sidnei da Silva
Enfold Systems                http://enfoldsystems.com
Fax +1 832 201 8856     Office +1 713 942 2377 Ext 214

From d.schulz81 at gmx.net  Thu Jan 18 13:59:53 2007
From: d.schulz81 at gmx.net (Dennis Schulz)
Date: Thu, 18 Jan 2007 13:59:53 +0100
Subject: [z3-five] Adaption of workflow tool
In-Reply-To: <B72907C7-CB39-41EB-ADC8-A9E366D4A5C6@zscout.de>
References: <45AF583E.6090403@gmx.net>
	<B72907C7-CB39-41EB-ADC8-A9E366D4A5C6@zscout.de>
Message-ID: <45AF6F49.8090008@gmx.net>

There is a product called alchemist that does integration into zope 2.
it also contains an example.

http://plone.org/products/alchemist

I am not completely sure if it has been tested with plain zope 3 but I 
think it's possible too.

Dennis




Janko Hauser escribi?:
>
> Am 18.01.2007 um 12:21 schrieb Dennis Schulz:
>
>> I have a database driven application that creates temporary mapped
>> objects not intended to be persistent content types (mainly Zope 3 with
>> Alchemist Product)
>
> May I ask, which products you are using to connect vie sqlalchemy to 
> the database? Specifically how the connection pooling does work there? 
> I looked into zalchemy, but do not see, how connections are persisted.
>
> Regarding your workflow question, if you use zope-2.10 it is possible 
> to use hurry.workflow from Martin Faasen.
> This is a simple workflowtool, which could perhaps be adapted to your 
> needs.
>
> With regards,
>
> __Janko Hauser
>
> --Janko Hauser  email:  jhauser at zscout.de
>               mobile: +49 1721 641552
>
>



From d.schulz81 at gmx.net  Thu Jan 18 14:07:14 2007
From: d.schulz81 at gmx.net (Dennis Schulz)
Date: Thu, 18 Jan 2007 14:07:14 +0100
Subject: [z3-five] Adaption of workflow tool
In-Reply-To: <a7a2b76b0701180457n4a31b306mb26c40da8d3cfa12@mail.gmail.com>
References: <45AF583E.6090403@gmx.net>
	<a7a2b76b0701180457n4a31b306mb26c40da8d3cfa12@mail.gmail.com>
Message-ID: <45AF7102.3040905@gmx.net>

Aha, thank you ;-)

seems like Kapil is not available in these days....
may I ask how can I get this patch. I have never been on the CMS mailing 
list.

Regards, Dennis


Sidnei da Silva escribi?:
> Hi Dennis,
>
> Kapil T. send a patch to the CMF mailing list that makes the
> WorkflowTool use adapters to lookup the workflow and also to store the
> workflow information I believe. It should be possible, with that
> patch, to write an adapter that can even store the workflow
> information in your RDB.
>



From sidnei at enfoldsystems.com  Thu Jan 18 14:55:33 2007
From: sidnei at enfoldsystems.com (Sidnei da Silva)
Date: Thu, 18 Jan 2007 11:55:33 -0200
Subject: [z3-five] Adaption of workflow tool
In-Reply-To: <45AF7102.3040905@gmx.net>
References: <45AF583E.6090403@gmx.net>
	<a7a2b76b0701180457n4a31b306mb26c40da8d3cfa12@mail.gmail.com>
	<45AF7102.3040905@gmx.net>
Message-ID: <a7a2b76b0701180555g77788ab8la5023aba7b5bb296@mail.gmail.com>

On 1/18/07, Dennis Schulz <d.schulz81 at gmx.net> wrote:
> Aha, thank you ;-)
>
> seems like Kapil is not available in these days....
> may I ask how can I get this patch. I have never been on the CMS mailing
> list.

http://www.mail-archive.com/zope-cmf at lists.zope.org/msg02995.html

-- 
Sidnei da Silva
Enfold Systems                http://enfoldsystems.com
Fax +1 832 201 8856     Office +1 713 942 2377 Ext 214

From chris at simplistix.co.uk  Sun Jan 28 12:16:31 2007
From: chris at simplistix.co.uk (Chris Withers)
Date: Sun, 28 Jan 2007 11:16:31 +0000
Subject: [z3-five] (re)using zpt macros from view page templates
Message-ID: <45BC860F.9040302@simplistix.co.uk>

Hi All,

I have a page template that's being used for Five view. I now want to 
re-use part of that template as a zpt macro in another Five view.

How should I go about doing this?

cheers,

Chris

-- 
Simplistix - Content Management, Zope & Python Consulting
            - http://www.simplistix.co.uk

From lists at zopyx.com  Sun Jan 28 12:43:34 2007
From: lists at zopyx.com (Andreas Jung)
Date: Sun, 28 Jan 2007 12:43:34 +0100
Subject: [z3-five] (re)using zpt macros from view page templates
In-Reply-To: <45BC860F.9040302@simplistix.co.uk>
References: <45BC860F.9040302@simplistix.co.uk>
Message-ID: <5ACA5A1F87992FAEBF789311@suxmac2.local>



--On 28. Januar 2007 11:16:31 +0000 Chris Withers <chris at simplistix.co.uk> 
wrote:

> Hi All,
>
> I have a page template that's being used for Five view. I now want to
> re-use part of that template as a zpt macro in another Five view.
>
> How should I go about doing this?
>
>

By registering the template with the macros as views?

Andreas
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 186 bytes
Desc: not available
Url : http://codespeak.net/pipermail/z3-five/attachments/20070128/ef0dec7d/attachment.pgp 

From chris at simplistix.co.uk  Sun Jan 28 14:30:37 2007
From: chris at simplistix.co.uk (Chris Withers)
Date: Sun, 28 Jan 2007 13:30:37 +0000
Subject: [z3-five] (re)using zpt macros from view page templates
In-Reply-To: <5ACA5A1F87992FAEBF789311@suxmac2.local>
References: <45BC860F.9040302@simplistix.co.uk>
	<5ACA5A1F87992FAEBF789311@suxmac2.local>
Message-ID: <45BCA57D.1000207@simplistix.co.uk>

Andreas Jung wrote:
> 
>> I have a page template that's being used for Five view. I now want to
>> re-use part of that template as a zpt macro in another Five view.
>>
>> How should I go about doing this?
> 
> By registering the template with the macros as views?

Is that the canonical way? it seems pretty heavyweight in terms of 
instantiating a whole view, with request and all, just to get a macro...

How about:

my_macro.pt
-----------
<html metal:define-macro="template">
...etc...
<metal:x define-slot="foo"/>
...etc...
</html>

macros.py
---------
from Products.Five.browser.pagetemplatefile import \
     ZopeTwoPageTemplateFile
my_macro = ZopeTwoPageTemplateFile('macros.pt')

myview.py
---------

from macros import my_macro

class MyView(BrowserView):

     def getMacro(self):
         return my_macro.macros['template']

myview.pt
---------
<html metal:use-macro="view/getMacro">
<div fill-slot="foo">
blah
</div>
</html>

Would the above work?
Is there anything wrong with it?

cheers,

Chris

-- 
Simplistix - Content Management, Zope & Python Consulting
            - http://www.simplistix.co.uk

From lists at zopyx.com  Sun Jan 28 14:41:32 2007
From: lists at zopyx.com (Andreas Jung)
Date: Sun, 28 Jan 2007 14:41:32 +0100
Subject: [z3-five] (re)using zpt macros from view page templates
In-Reply-To: <45BCA57D.1000207@simplistix.co.uk>
References: <45BC860F.9040302@simplistix.co.uk>
	<5ACA5A1F87992FAEBF789311@suxmac2.local>
	<45BCA57D.1000207@simplistix.co.uk>
Message-ID: <55D50FEBB2534351B99F2D3E@suxmac2.local>



--On 28. Januar 2007 13:30:37 +0000 Chris Withers <chris at simplistix.co.uk> 
wrote:

> Andreas Jung wrote:
>>
>>> I have a page template that's being used for Five view. I now want to
>>> re-use part of that template as a zpt macro in another Five view.
>>>
>>> How should I go about doing this?
>>
>> By registering the template with the macros as views?
>
> Is that the canonical way? it seems pretty heavyweight in terms of
> instantiating a whole view, with request and all, just to get a macro...
>

Isn't it sufficient to register the template through ZCML?

-aj
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 186 bytes
Desc: not available
Url : http://codespeak.net/pipermail/z3-five/attachments/20070128/ba3733d7/attachment.pgp 

From chris at simplistix.co.uk  Sun Jan 28 15:34:38 2007
From: chris at simplistix.co.uk (Chris Withers)
Date: Sun, 28 Jan 2007 14:34:38 +0000
Subject: [z3-five] (re)using zpt macros from view page templates
In-Reply-To: <55D50FEBB2534351B99F2D3E@suxmac2.local>
References: <45BC860F.9040302@simplistix.co.uk>
	<5ACA5A1F87992FAEBF789311@suxmac2.local>
	<45BCA57D.1000207@simplistix.co.uk>
	<55D50FEBB2534351B99F2D3E@suxmac2.local>
Message-ID: <45BCB47E.9030409@simplistix.co.uk>

Andreas Jung wrote:
> 
>> Is that the canonical way? it seems pretty heavyweight in terms of
>> instantiating a whole view, with request and all, just to get a macro...
>>
> Isn't it sufficient to register the template through ZCML?

Yeah, that's not what I was talking about though...

If I did:

myview.pt
---------
<html metal:use-macro="context/@@my_macro">
<div fill-slot="foo">
blah
</div>
</html>

...then we're looking at least at an adapter looking, probably some 
security checks in the traversal, and who knows what else.

Is there anything wrong with just having the page template as a 
module-level global and importing from there?

cheers,

Chris

-- 
Simplistix - Content Management, Zope & Python Consulting
            - http://www.simplistix.co.uk

From regebro at gmail.com  Mon Jan 29 10:20:34 2007
From: regebro at gmail.com (Lennart Regebro)
Date: Mon, 29 Jan 2007 10:20:34 +0100
Subject: [z3-five] (re)using zpt macros from view page templates
In-Reply-To: <45BCB47E.9030409@simplistix.co.uk>
References: <45BC860F.9040302@simplistix.co.uk>
	<5ACA5A1F87992FAEBF789311@suxmac2.local>
	<45BCA57D.1000207@simplistix.co.uk>
	<55D50FEBB2534351B99F2D3E@suxmac2.local>
	<45BCB47E.9030409@simplistix.co.uk>
Message-ID: <319e029f0701290120v72d6a91fu8a3d3a17a4dc43fa@mail.gmail.com>

On 1/28/07, Chris Withers <chris at simplistix.co.uk> wrote:
> Is there anything wrong with just having the page template as a
> module-level global and importing from there?

Wrong? Is this a moral question? ;)

Seriously though: It's hard to override, otherwise I can't see
anything wrong with it, assuming it works.
-- 
Lennart Regebro: Python, Zope, CPS, Plone consulting.
+33 661 58 14 64

From maciej.wisniowski at coig.katowice.pl  Mon Jan 29 10:52:00 2007
From: maciej.wisniowski at coig.katowice.pl (Maciej Wisniowski)
Date: Mon, 29 Jan 2007 10:52:00 +0100
Subject: [z3-five] (re)using zpt macros from view page templates
In-Reply-To: <45BCB47E.9030409@simplistix.co.uk>
References: <45BC860F.9040302@simplistix.co.uk>	<5ACA5A1F87992FAEBF789311@suxmac2.local>	<45BCA57D.1000207@simplistix.co.uk>	<55D50FEBB2534351B99F2D3E@suxmac2.local>
	<45BCB47E.9030409@simplistix.co.uk>
Message-ID: <45BDC3C0.8020202@coig.katowice.pl>

>
>> Isn't it sufficient to register the template through ZCML?
>>     
> ...then we're looking at least at an adapter looking, probably some 
> security checks in the traversal, and who knows what else.
>   
Even default zope skins are views that consist
of other views (zope.app.basicskin.standardmacros).

> Is there anything wrong with just having the page template as a 
> module-level global and importing from there?
And what about base view like:

class MyBaseView(BrowserView):
    my_macro = ZopeTwoPageTemplateFile('macros.pt')

class MyView(MyBaseView):

     def getMacro(self):
         return self.my_macro.macros['template']


-- 
Maciej Wisniowski

From chris at simplistix.co.uk  Mon Jan 29 12:16:11 2007
From: chris at simplistix.co.uk (Chris Withers)
Date: Mon, 29 Jan 2007 11:16:11 +0000
Subject: [z3-five] (re)using zpt macros from view page templates
In-Reply-To: <319e029f0701290120v72d6a91fu8a3d3a17a4dc43fa@mail.gmail.com>
References: <45BC860F.9040302@simplistix.co.uk>	
	<5ACA5A1F87992FAEBF789311@suxmac2.local>	
	<45BCA57D.1000207@simplistix.co.uk>	
	<55D50FEBB2534351B99F2D3E@suxmac2.local>	
	<45BCB47E.9030409@simplistix.co.uk>
	<319e029f0701290120v72d6a91fu8a3d3a17a4dc43fa@mail.gmail.com>
Message-ID: <45BDD77B.1020705@simplistix.co.uk>

Lennart Regebro wrote:
> Seriously though: It's hard to override, 

What do you mean by that?

cheers,

Chris

-- 
Simplistix - Content Management, Zope & Python Consulting
            - http://www.simplistix.co.uk

From chris at simplistix.co.uk  Mon Jan 29 12:17:17 2007
From: chris at simplistix.co.uk (Chris Withers)
Date: Mon, 29 Jan 2007 11:17:17 +0000
Subject: [z3-five] (re)using zpt macros from view page templates
In-Reply-To: <45BDC3C0.8020202@coig.katowice.pl>
References: <45BC860F.9040302@simplistix.co.uk>	<5ACA5A1F87992FAEBF789311@suxmac2.local>	<45BCA57D.1000207@simplistix.co.uk>	<55D50FEBB2534351B99F2D3E@suxmac2.local>
	<45BCB47E.9030409@simplistix.co.uk>
	<45BDC3C0.8020202@coig.katowice.pl>
Message-ID: <45BDD7BD.2050703@simplistix.co.uk>

Maciej Wisniowski wrote:
>>> Isn't it sufficient to register the template through ZCML?
>>>     
>> ...then we're looking at least at an adapter looking, probably some 
>> security checks in the traversal, and who knows what else.
>>   
> Even default zope skins are views that consist
> of other views (zope.app.basicskin.standardmacros).

You say that like that means it's a good thing... ;-)

>> Is there anything wrong with just having the page template as a 
>> module-level global and importing from there?
> And what about base view like:
> 
> class MyBaseView(BrowserView):
>     my_macro = ZopeTwoPageTemplateFile('macros.pt')
> 
> class MyView(MyBaseView):
> 
>      def getMacro(self):
>          return self.my_macro.macros['template']

Yup, that's another possibility :-)

cheers,

Chris

-- 
Simplistix - Content Management, Zope & Python Consulting
            - http://www.simplistix.co.uk

From maciej.wisniowski at coig.katowice.pl  Mon Jan 29 13:00:45 2007
From: maciej.wisniowski at coig.katowice.pl (Maciej Wisniowski)
Date: Mon, 29 Jan 2007 13:00:45 +0100
Subject: [z3-five] (re)using zpt macros from view page templates
In-Reply-To: <45BDD7BD.2050703@simplistix.co.uk>
References: <45BC860F.9040302@simplistix.co.uk>	<5ACA5A1F87992FAEBF789311@suxmac2.local>	<45BCA57D.1000207@simplistix.co.uk>	<55D50FEBB2534351B99F2D3E@suxmac2.local>
	<45BCB47E.9030409@simplistix.co.uk>
	<45BDC3C0.8020202@coig.katowice.pl>
	<45BDD7BD.2050703@simplistix.co.uk>
Message-ID: <45BDE1ED.2090806@coig.katowice.pl>


>
> You say that like that means it's a good thing... ;-)
In fact it was not my intention to say whether it is good or bad
but in general I think that because even zope core uses
this kind of macros then we may say that, at last, it is not
very bad way of doing such things.  But of course this
depends on particular use case too.

-- 
Maciej Wisniowski

From chris at simplistix.co.uk  Mon Jan 29 22:35:36 2007
From: chris at simplistix.co.uk (Chris Withers)
Date: Mon, 29 Jan 2007 21:35:36 +0000
Subject: [z3-five] Unexpectedly unprotected code
Message-ID: <45BE68A8.5030403@simplistix.co.uk>

Hi All,

I have a view registered as follows:

   <browser:page
       for="ISomething"
       name="myview"
       class="MyView"
       template="myview.pt"
       permission="cmf.ModifyPortalContent"
       />

The view class looks like this:

from Products.Five.browser import BrowserView

class MyView(BrowserView):

     def __init__(self,context,request):
         self.context = context
         self.request = request
         # point A


I've noticed two things about code running at point A by going to the
following url:

http://localhost:7001/mysite/anobject/@@myview

1. Code running at this point has no security context, so anything that 
uses getSecurityManager will get an anonymous user. This is annoying. 
Why is it like this?

2. More worrying, the code running at point A is "trusted". Am I being
unreasonable to expect that code only to run if the current user has
cmf.ModifyPortalContent?

The first implementation of this view did it's form processing in the
__init__ method. Because of point 2, that meant that any anonymous user
could edit objects they shouldn't have been able to. The code needed a 
user object though, so thanks to point 1, it always did the wrong thing, 
so even legitimate users ended up screwed ;-)

Can anyone explain a bit more about this and how you're supposed to get
around this?

Cheers,

Chris


From sidnei at enfoldsystems.com  Mon Jan 29 22:57:02 2007
From: sidnei at enfoldsystems.com (Sidnei da Silva)
Date: Mon, 29 Jan 2007 15:57:02 -0600
Subject: [z3-five] Unexpectedly unprotected code
In-Reply-To: <45BE68A8.5030403@simplistix.co.uk>
References: <45BE68A8.5030403@simplistix.co.uk>
Message-ID: <a7a2b76b0701291357k5a5b0499w87d00f98332316a5@mail.gmail.com>

On 1/29/07, Chris Withers <chris at simplistix.co.uk> wrote:
> 1. Code running at this point has no security context, so anything that
> uses getSecurityManager will get an anonymous user. This is annoying.
> Why is it like this?

It's because the traversal machinery needs to get the final object to
be published to find the innermost acl_users object, to validate the
possibly local user.

So what happens is that the validate() method of the security
machinery is only called after the traversal finds the last object in
the chain.

Your __init__ method is called right before that, during traversal,
while *creating* the view object, which most likely will be the last
object in the chain.

> 2. More worrying, the code running at point A is "trusted". Am I being
> unreasonable to expect that code only to run if the current user has
> cmf.ModifyPortalContent?

Yes you are? :)

> Can anyone explain a bit more about this and how you're supposed to get
> around this?

By implementing your stuff into a proper method of the view class that
gets called *after* the traversal has taken place, and not in the
'constructor'.

-- 
Sidnei da Silva
Enfold Systems                http://enfoldsystems.com
Fax +1 832 201 8856     Office +1 713 942 2377 Ext 214

From optilude at gmx.net  Mon Jan 29 22:57:49 2007
From: optilude at gmx.net (Martin Aspeli)
Date: Mon, 29 Jan 2007 21:57:49 +0000
Subject: [z3-five] Unexpectedly unprotected code
In-Reply-To: <45BE68A8.5030403@simplistix.co.uk>
References: <45BE68A8.5030403@simplistix.co.uk>
Message-ID: <eplqmb$bs3$1@sea.gmane.org>

Chris Withers wrote:

> class MyView(BrowserView):
> 
>      def __init__(self,context,request):
>          self.context = context
>          self.request = request
>          # point A
> 
> 
> I've noticed two things about code running at point A by going to the
> following url:
> 
> http://localhost:7001/mysite/anobject/@@myview
> 
> 1. Code running at this point has no security context, so anything that 
> uses getSecurityManager will get an anonymous user. This is annoying. 
> Why is it like this?

Because you're in __init__(). Try the same thing in a content object, 
you won't have one there either.

> 2. More worrying, the code running at point A is "trusted". Am I being
> unreasonable to expect that code only to run if the current user has
> cmf.ModifyPortalContent?

All code inside views is trusted. It's not TTW code, so I think that's 
reasonable. This is Zope 2, remember, so security proxies start at the 
TTW/filesystem boundary.

> The first implementation of this view did it's form processing in the
> __init__ method. Because of point 2, that meant that any anonymous user
> could edit objects they shouldn't have been able to.

That doesn't sound sensible... The anonymous user shouldn't be allowed 
to instantiate the view at all, because it's protected by 
permission="cmf.ModifyPortalContent". At least that's my understanding. 
It may be that __init__() gets called first, though...

In any case, form processing in __init__() is not very pretty, I think 
it makes much more sense to do that in __call__().

> Can anyone explain a bit more about this and how you're supposed to get
> around this?

Put code in __call__(). Don't use 'template' in ZCML, but instead do

from Products.Five.browser import BrowserView
from Products.Five.browser.pagetemplatefile import ViewPageTemplateFile

class Foo(BrowserView):

   template = ViewPageTemplateFile('myview.pt')

   def __call__(self):
       # process from self.context and self.request
       return self.template()

Martin


From philipp at weitershausen.de  Mon Jan 29 23:40:08 2007
From: philipp at weitershausen.de (Philipp von Weitershausen)
Date: Mon, 29 Jan 2007 23:40:08 +0100
Subject: [z3-five] Unexpectedly unprotected code
In-Reply-To: <45BE68A8.5030403@simplistix.co.uk>
References: <45BE68A8.5030403@simplistix.co.uk>
Message-ID: <45BE77C8.1030909@weitershausen.de>

Chris Withers wrote:
> 1. Code running at this point has no security context, so anything that 
> uses getSecurityManager will get an anonymous user. This is annoying. 
> Why is it like this?

Sidnei already said it: it's traversal, there's no security context 
determined yet.

> 2. More worrying, the code running at point A is "trusted". Am I being
> unreasonable to expect that code only to run if the current user has
> cmf.ModifyPortalContent?

No, you're simply not getting the difference between __init__ and a 
__call__ on a view :).

> The first implementation of this view did it's form processing in the
> __init__ method.

That's typically not considered good style. As Sidnei said already, the 
__init__ of a view is invoked when the view is looked up. That happens 
during traversal. There's no certainty that the view will be the last 
thing in the traversal chain or that the view will actually end up being 
called. For example, the view might not be accessible to the current 
user due to insufficient privileges.

Therefore doing stuff in __init__ that changes the state of the app is 
asking for trouble.

> Because of point 2, that meant that any anonymous user
> could edit objects they shouldn't have been able to. The code needed a 
> user object though, so thanks to point 1, it always did the wrong thing, 
> so even legitimate users ended up screwed ;-)
> 
> Can anyone explain a bit more about this and how you're supposed to get
> around this?

Actual view code belongs in __call__.

-- 
http://worldcookery.com -- Professional Zope documentation and training
Next Zope 3 training at Camp5: http://trizpug.org/boot-camp/camp5

From philipp at weitershausen.de  Mon Jan 29 23:45:31 2007
From: philipp at weitershausen.de (Philipp von Weitershausen)
Date: Mon, 29 Jan 2007 23:45:31 +0100
Subject: [z3-five] (re)using zpt macros from view page templates
In-Reply-To: <45BCA57D.1000207@simplistix.co.uk>
References: <45BC860F.9040302@simplistix.co.uk>	<5ACA5A1F87992FAEBF789311@suxmac2.local>
	<45BCA57D.1000207@simplistix.co.uk>
Message-ID: <45BE790B.5080302@weitershausen.de>

Chris Withers wrote:
> Andreas Jung wrote:
>>> I have a page template that's being used for Five view. I now want to
>>> re-use part of that template as a zpt macro in another Five view.
>>>
>>> How should I go about doing this?
>> By registering the template with the macros as views?
> 
> Is that the canonical way? it seems pretty heavyweight in terms of 
> instantiating a whole view, with request and all, just to get a macro...
> 
> How about:
> 
> my_macro.pt
> -----------
> <html metal:define-macro="template">
> ...etc...
> <metal:x define-slot="foo"/>
> ...etc...
> </html>

<browser:page name="mymacro" template="my_macro.pt" ... />

> myview.pt
> ---------
> <html metal:use-macro="view/getMacro">

use-macro="context/@@mymacro/template"

> <div fill-slot="foo">
> blah
> </div>
> </html>
> 
> Would the above work?

Why dontcha try it out?!?

Anyway, it looks like lots of unnecessary typing. Just register the 
tempalte with the macro as a page and access the macro as outlined above.

-- 
http://worldcookery.com -- Professional Zope documentation and training
Next Zope 3 training at Camp5: http://trizpug.org/boot-camp/camp5

From chris at simplistix.co.uk  Mon Jan 29 23:55:06 2007
From: chris at simplistix.co.uk (Chris Withers)
Date: Mon, 29 Jan 2007 22:55:06 +0000
Subject: [z3-five] (re)using zpt macros from view page templates
In-Reply-To: <45BE790B.5080302@weitershausen.de>
References: <45BC860F.9040302@simplistix.co.uk>	<5ACA5A1F87992FAEBF789311@suxmac2.local>
	<45BCA57D.1000207@simplistix.co.uk>
	<45BE790B.5080302@weitershausen.de>
Message-ID: <45BE7B4A.4060601@simplistix.co.uk>

Philipp von Weitershausen wrote:
>> my_macro.pt
>> -----------
>> <html metal:define-macro="template">
>> ...etc...
>> <metal:x define-slot="foo"/>
>> ...etc...
>> </html>
> 
> <browser:page name="mymacro" template="my_macro.pt" ... />

Okay, but what interface do I bind it to? *, so it's available on all 
objects? It just seems a bit wasteful, a view is essentially a 
multiadapter for a context and a request. A macro needs neither context 
nor request, and feels a lot close to a utility to me.

I want to get it from one define place, but only be accessible either 
from trusted code, or by people who are allowed to view it.

> Anyway, it looks like lots of unnecessary typing. Just register the 
> tempalte with the macro as a page and access the macro as outlined above.

Well, my way wasn't much typing, but I hit a problem.

<browser:page name="mymacro" template="my_macro.pt" ... /> seems to have 
a TAL version of 1.5.

Products.Five.browser.pagetemplatefile.ViewPageTemplateFile seems to 
instantiate templates with a TAL version of 1.6.

So, if I try and use my macro from a normal Five view, I get a 
METALError: macro 'mymacro' has incompatible version '1.6'.

Why is that? How can I fix it and get them both to use the same TAL version?

cheers,

Chris

-- 
Simplistix - Content Management, Zope & Python Consulting
            - http://www.simplistix.co.uk

From chris at simplistix.co.uk  Mon Jan 29 23:57:03 2007
From: chris at simplistix.co.uk (Chris Withers)
Date: Mon, 29 Jan 2007 22:57:03 +0000
Subject: [z3-five] Unexpectedly unprotected code
In-Reply-To: <eplqmb$bs3$1@sea.gmane.org>
References: <45BE68A8.5030403@simplistix.co.uk> <eplqmb$bs3$1@sea.gmane.org>
Message-ID: <45BE7BBF.7060805@simplistix.co.uk>

Martin Aspeli wrote:
>> The first implementation of this view did it's form processing in the
>> __init__ method. 
> 
> That doesn't sound sensible... 

Indeed, that's why I moved it to its own method. That method still seems 
to need a doctstring to placate Zope 2's publisher, but am I right in 
thinking that method is protected by the view's permissions?

> The anonymous user shouldn't be allowed 
> to instantiate the view at all, because it's protected by 
> permission="cmf.ModifyPortalContent". 

Yes, I think this is the bit that surprised me.

Why can an anonymous user cause a view they have no rights to see to be 
instantiated?

cheers,

Chris

-- 
Simplistix - Content Management, Zope & Python Consulting
            - http://www.simplistix.co.uk

From philipp at weitershausen.de  Mon Jan 29 23:59:55 2007
From: philipp at weitershausen.de (Philipp von Weitershausen)
Date: Mon, 29 Jan 2007 23:59:55 +0100
Subject: [z3-five] (re)using zpt macros from view page templates
In-Reply-To: <45BE7B4A.4060601@simplistix.co.uk>
References: <45BC860F.9040302@simplistix.co.uk>	<5ACA5A1F87992FAEBF789311@suxmac2.local>
	<45BCA57D.1000207@simplistix.co.uk>
	<45BE790B.5080302@weitershausen.de>
	<45BE7B4A.4060601@simplistix.co.uk>
Message-ID: <84F4CD03-7C18-4B6D-AF17-029FFDAB1777@weitershausen.de>

On 29 Jan 2007, at 23:55 , Chris Withers wrote:
> Philipp von Weitershausen wrote:
>>> my_macro.pt
>>> -----------
>>> <html metal:define-macro="template">
>>> ...etc...
>>> <metal:x define-slot="foo"/>
>>> ...etc...
>>> </html>
>> <browser:page name="mymacro" template="my_macro.pt" ... />
>
> Okay, but what interface do I bind it to? *, so it's available on  
> all objects? It just seems a bit wasteful, a view is essentially a  
> multiadapter for a context and a request. A macro needs neither  
> context nor request, and feels a lot close to a utility to me.

It's not always about actually *using* context or request. These two  
can also serve as keys for the lookup. The skin is expressed as a  
marker interface on the request, so for different skins you could  
have different versions of the 'mymacro' view, hence different macros  
for your templates.

> I want to get it from one define place, but only be accessible  
> either from trusted code, or by people who are allowed to view it.

You don't have to do it the way I outlined it...

>> Anyway, it looks like lots of unnecessary typing. Just register  
>> the tempalte with the macro as a page and access the macro as  
>> outlined above.
>
> Well, my way wasn't much typing, but I hit a problem.
>
> <browser:page name="mymacro" template="my_macro.pt" ... /> seems to  
> have a TAL version of 1.5.
>
> Products.Five.browser.pagetemplatefile.ViewPageTemplateFile seems  
> to instantiate templates with a TAL version of 1.6.

That's hard to believe. Couldn't it be you're confusing Zope3's  
*zope.app.pagetemplate*.ViewPageTemplateFile with Five's one?

> So, if I try and use my macro from a normal Five view, I get a  
> METALError: macro 'mymacro' has incompatible version '1.6'.
>
> Why is that? How can I fix it and get them both to use the same TAL  
> version?

Use everything from Five. It works.




From sidnei at enfoldsystems.com  Tue Jan 30 00:02:20 2007
From: sidnei at enfoldsystems.com (Sidnei da Silva)
Date: Mon, 29 Jan 2007 17:02:20 -0600
Subject: [z3-five] Unexpectedly unprotected code
In-Reply-To: <45BE7BBF.7060805@simplistix.co.uk>
References: <45BE68A8.5030403@simplistix.co.uk> <eplqmb$bs3$1@sea.gmane.org>
	<45BE7BBF.7060805@simplistix.co.uk>
Message-ID: <a7a2b76b0701291502o2127b3e9we5ab8524b82e6ca6@mail.gmail.com>

On 1/29/07, Chris Withers <chris at simplistix.co.uk> wrote:
> Why can an anonymous user cause a view they have no rights to see to be
> instantiated?

I guess that because you need acquisition context to check rights, and
you need an instance to have acquisition context.

-- 
Sidnei da Silva
Enfold Systems                http://enfoldsystems.com
Fax +1 832 201 8856     Office +1 713 942 2377 Ext 214

From philipp at weitershausen.de  Tue Jan 30 00:06:16 2007
From: philipp at weitershausen.de (Philipp von Weitershausen)
Date: Tue, 30 Jan 2007 00:06:16 +0100
Subject: [z3-five] Unexpectedly unprotected code
In-Reply-To: <a7a2b76b0701291502o2127b3e9we5ab8524b82e6ca6@mail.gmail.com>
References: <45BE68A8.5030403@simplistix.co.uk>
	<eplqmb$bs3$1@sea.gmane.org>	<45BE7BBF.7060805@simplistix.co.uk>
	<a7a2b76b0701291502o2127b3e9we5ab8524b82e6ca6@mail.gmail.com>
Message-ID: <45BE7DE8.7000706@weitershausen.de>

Sidnei da Silva wrote:
> On 1/29/07, Chris Withers <chris at simplistix.co.uk> wrote:
>> Why can an anonymous user cause a view they have no rights to see to be
>> instantiated?
> 
> I guess that because you need acquisition context to check rights, and
> you need an instance to have acquisition context.

Absolutely correct.

-- 
http://worldcookery.com -- Professional Zope documentation and training
Next Zope 3 training at Camp5: http://trizpug.org/boot-camp/camp5

From chris at simplistix.co.uk  Tue Jan 30 08:50:00 2007
From: chris at simplistix.co.uk (Chris Withers)
Date: Tue, 30 Jan 2007 07:50:00 +0000
Subject: [z3-five] Unexpectedly unprotected code
In-Reply-To: <45BE7DE8.7000706@weitershausen.de>
References: <45BE68A8.5030403@simplistix.co.uk>
	<eplqmb$bs3$1@sea.gmane.org>	<45BE7BBF.7060805@simplistix.co.uk>
	<a7a2b76b0701291502o2127b3e9we5ab8524b82e6ca6@mail.gmail.com>
	<45BE7DE8.7000706@weitershausen.de>
Message-ID: <45BEF8A8.8040806@simplistix.co.uk>

Philipp von Weitershausen wrote:
> Sidnei da Silva wrote:
>> On 1/29/07, Chris Withers <chris at simplistix.co.uk> wrote:
>>> Why can an anonymous user cause a view they have no rights to see to be
>>> instantiated?
>>
>> I guess that because you need acquisition context to check rights, and
>> you need an instance to have acquisition context.
> 
> Absolutely correct.

Am I right in thinking Zope 3's security machinery doesn't have this 
problem?

Are there any plans to move Zope 2 to Zope 3's security machinery?

cheers,

Chris

-- 
Simplistix - Content Management, Zope & Python Consulting
            - http://www.simplistix.co.uk

From chris at simplistix.co.uk  Tue Jan 30 09:44:57 2007
From: chris at simplistix.co.uk (Chris Withers)
Date: Tue, 30 Jan 2007 08:44:57 +0000
Subject: [z3-five] (re)using zpt macros from view page templates
In-Reply-To: <84F4CD03-7C18-4B6D-AF17-029FFDAB1777@weitershausen.de>
References: <45BC860F.9040302@simplistix.co.uk>	<5ACA5A1F87992FAEBF789311@suxmac2.local>
	<45BCA57D.1000207@simplistix.co.uk>
	<45BE790B.5080302@weitershausen.de>
	<45BE7B4A.4060601@simplistix.co.uk>
	<84F4CD03-7C18-4B6D-AF17-029FFDAB1777@weitershausen.de>
Message-ID: <45BF0589.1040406@simplistix.co.uk>

Philipp von Weitershausen wrote:
> It's not always about actually *using* context or request. These two can 
> also serve as keys for the lookup. The skin is expressed as a marker 
> interface on the request, so for different skins you could have 
> different versions of the 'mymacro' view, hence different macros for 
> your templates.

Can you give me an example of how this would work or where I can look to 
see real code? It sounds like a reasonable use case, but I can't quite 
get my head around it...

>> Well, my way wasn't much typing, but I hit a problem.
>>
>> <browser:page name="mymacro" template="my_macro.pt" ... /> seems to 
>> have a TAL version of 1.5.
>>
>> Products.Five.browser.pagetemplatefile.ViewPageTemplateFile seems to 
>> instantiate templates with a TAL version of 1.6.
> 
> That's hard to believe. Couldn't it be you're confusing Zope3's 
> *zope.app.pagetemplate*.ViewPageTemplateFile with Five's one?

Please find attached a minimal example. I'm using Zope 2.9.4 and all the 
files were in lib/python of the instance. The zcml file was included 
with the following in site.zcml:

   <include files="../lib/python/view.zcml" />

Any help greatfully received!

cheers,

Chris

PS: Putting this together, I noticed another thing... am I right in 
thinking that:

<five:traversable class="OFS.Application.Application" />

...doesn't work? As you'll see in the minimal example, I had to bind to 
OFS.Folder.Folder instead and then create a dummy folder so I could test 
the view in a browser.

-- 
Simplistix - Content Management, Zope & Python Consulting
            - http://www.simplistix.co.uk
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: view.zcml
Url: http://codespeak.net/pipermail/z3-five/attachments/20070130/bcc03061/attachment.diff 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://codespeak.net/pipermail/z3-five/attachments/20070130/bcc03061/attachment.htm 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://codespeak.net/pipermail/z3-five/attachments/20070130/bcc03061/attachment-0001.htm 
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: test.py
Url: http://codespeak.net/pipermail/z3-five/attachments/20070130/bcc03061/attachment-0001.diff 

From chris at simplistix.co.uk  Tue Jan 30 13:05:57 2007
From: chris at simplistix.co.uk (Chris Withers)
Date: Tue, 30 Jan 2007 12:05:57 +0000
Subject: [z3-five] (re)using zpt macros from view page templates
In-Reply-To: <45BF0589.1040406@simplistix.co.uk>
References: <45BC860F.9040302@simplistix.co.uk>	<5ACA5A1F87992FAEBF789311@suxmac2.local>	<45BCA57D.1000207@simplistix.co.uk>	<45BE790B.5080302@weitershausen.de>	<45BE7B4A.4060601@simplistix.co.uk>	<84F4CD03-7C18-4B6D-AF17-029FFDAB1777@weitershausen.de>
	<45BF0589.1040406@simplistix.co.uk>
Message-ID: <45BF34A5.8090005@simplistix.co.uk>

Chris Withers wrote:

> Any help greatfully received!

> from Products.Five.browser.pagetemplatefile import ViewPageTemplateFile

This was my mistake, that should have been:

from Products.Five.browser.pagetemplatefile import \
     ZopeTwoPageTemplateFile

However, now using the right class...

my_macro = ZopeTwoPageTemplateFile('my_macro.pt')

my_macro.macros just gives me a ComputedAttribute object (rather than 
the macros I'd expect).

If I do, my_macro.pt_macros(), I get the macros.

Why isn't ComputedAttribute working in this example?

cheers,

Chris

PS: Still interested in this too:
> PS: Putting this together, I noticed another thing... am I right in 
> thinking that:
> 
> <five:traversable class="OFS.Application.Application" />
> 
> ...doesn't work? As you'll see in the minimal example, I had to bind to 
> OFS.Folder.Folder instead and then create a dummy folder so I could test 
> the view in a browser.
-- 
Simplistix - Content Management, Zope & Python Consulting
            - http://www.simplistix.co.uk

From philipp at weitershausen.de  Tue Jan 30 14:35:31 2007
From: philipp at weitershausen.de (Philipp von Weitershausen)
Date: Tue, 30 Jan 2007 14:35:31 +0100
Subject: [z3-five] (re)using zpt macros from view page templates
In-Reply-To: <45BF34A5.8090005@simplistix.co.uk>
References: <45BC860F.9040302@simplistix.co.uk>	<5ACA5A1F87992FAEBF789311@suxmac2.local>	<45BCA57D.1000207@simplistix.co.uk>	<45BE790B.5080302@weitershausen.de>	<45BE7B4A.4060601@simplistix.co.uk>	<84F4CD03-7C18-4B6D-AF17-029FFDAB1777@weitershausen.de>	<45BF0589.1040406@simplistix.co.uk>
	<45BF34A5.8090005@simplistix.co.uk>
Message-ID: <45BF49A3.70101@weitershausen.de>

Chris Withers wrote:
> Chris Withers wrote:
> 
>> Any help greatfully received!
> 
>> from Products.Five.browser.pagetemplatefile import ViewPageTemplateFile
> 
> This was my mistake, that should have been:
> 
> from Products.Five.browser.pagetemplatefile import \
>      ZopeTwoPageTemplateFile

They're the same. It shouldn't make a difference.

> However, now using the right class...
> 
> my_macro = ZopeTwoPageTemplateFile('my_macro.pt')
> 
> my_macro.macros just gives me a ComputedAttribute object (rather than 
> the macros I'd expect).
> 
> If I do, my_macro.pt_macros(), I get the macros.
> 
> Why isn't ComputedAttribute working in this example?

Coz ComputedAttribute is a stupid Zope2ism that only works in an 
acquisition context.

I still think you're overcomplicating things. You don't really need any 
python code. Just register your macro template as a browser:page and 
then access the macro in the other template via 
context/@@viewname_of_the_macrotemplate/macroname. I've spelt this out 
already, it's your choice to do it or not. I won't give any predictions 
on what else might work or not since this is the standard way of 
accessing macros...

-- 
http://worldcookery.com -- Professional Zope documentation and training
Next Zope 3 training at Camp5: http://trizpug.org/boot-camp/camp5

From philipp at weitershausen.de  Tue Jan 30 15:12:28 2007
From: philipp at weitershausen.de (Philipp von Weitershausen)
Date: Tue, 30 Jan 2007 15:12:28 +0100
Subject: [z3-five] Unexpectedly unprotected code
In-Reply-To: <45BEF8A8.8040806@simplistix.co.uk>
References: <45BE68A8.5030403@simplistix.co.uk>
	<eplqmb$bs3$1@sea.gmane.org>	<45BE7BBF.7060805@simplistix.co.uk>
	<a7a2b76b0701291502o2127b3e9we5ab8524b82e6ca6@mail.gmail.com>
	<45BE7DE8.7000706@weitershausen.de>
	<45BEF8A8.8040806@simplistix.co.uk>
Message-ID: <F94802BB-6FD3-4A8C-8979-4A37A250BD1D@weitershausen.de>

On 30 Jan 2007, at 08:50 , Chris Withers wrote:
> Philipp von Weitershausen wrote:
>> Sidnei da Silva wrote:
>>> On 1/29/07, Chris Withers <chris at simplistix.co.uk> wrote:
>>>> Why can an anonymous user cause a view they have no rights to  
>>>> see to be
>>>> instantiated?
>>>
>>> I guess that because you need acquisition context to check  
>>> rights, and
>>> you need an instance to have acquisition context.
>> Absolutely correct.
>
> Am I right in thinking Zope 3's security machinery doesn't have  
> this problem?

Zope 3's security machinery certainly can't work around the fact that  
you first have to instantiate an adapter (=view) before you can make  
security assertions on it. In Zope3, therefore, app code doesn't  
belong in a view's __init__ either.

As for having acquisition contexts, the default security policy in  
Zope 3 wants views to have __parent__ so that it can acquire security  
information. This decision isn't part of the security machinery of  
Zope3, it's a policy choice.

> Are there any plans to move Zope 2 to Zope 3's security machinery?

Jim has remote plans and Martijn has some wild ideas. I don't see it  
happening any time soon, though. It's also not the issue here.

From regebro at gmail.com  Tue Jan 30 15:24:13 2007
From: regebro at gmail.com (Lennart Regebro)
Date: Tue, 30 Jan 2007 15:24:13 +0100
Subject: [z3-five] (re)using zpt macros from view page templates
In-Reply-To: <45BDD77B.1020705@simplistix.co.uk>
References: <45BC860F.9040302@simplistix.co.uk>
	<5ACA5A1F87992FAEBF789311@suxmac2.local>
	<45BCA57D.1000207@simplistix.co.uk>
	<55D50FEBB2534351B99F2D3E@suxmac2.local>
	<45BCB47E.9030409@simplistix.co.uk>
	<319e029f0701290120v72d6a91fu8a3d3a17a4dc43fa@mail.gmail.com>
	<45BDD77B.1020705@simplistix.co.uk>
Message-ID: <319e029f0701300624p78de8c99hff28d6d5115a12cf@mail.gmail.com>

On 1/29/07, Chris Withers <chris at simplistix.co.uk> wrote:
> Lennart Regebro wrote:
> > Seriously though: It's hard to override,
>
> What do you mean by that?

What I mean is that if you want to override that macro in another
product, for say a customer site, you will have to monkey-patch it. If
you instead register the macro so it becomes a view, it's easy to
override.

-- 
Lennart Regebro: Python, Zope, CPS, Plone consulting.
+33 661 58 14 64

From rocky at serverzen.com  Tue Jan 30 15:54:14 2007
From: rocky at serverzen.com (Rocky Burt)
Date: Tue, 30 Jan 2007 11:24:14 -0330
Subject: [z3-five] (re)using zpt macros from view page templates
In-Reply-To: <319e029f0701300624p78de8c99hff28d6d5115a12cf@mail.gmail.com>
References: <45BC860F.9040302@simplistix.co.uk>
	<5ACA5A1F87992FAEBF789311@suxmac2.local>
	<45BCA57D.1000207@simplistix.co.uk>
	<55D50FEBB2534351B99F2D3E@suxmac2.local>
	<45BCB47E.9030409@simplistix.co.uk>
	<319e029f0701290120v72d6a91fu8a3d3a17a4dc43fa@mail.gmail.com>
	<45BDD77B.1020705@simplistix.co.uk>
	<319e029f0701300624p78de8c99hff28d6d5115a12cf@mail.gmail.com>
Message-ID: <1170168849.7670.41.camel@localhost>

On Tue, 2007-30-01 at 15:24 +0100, Lennart Regebro wrote:
> On 1/29/07, Chris Withers <chris at simplistix.co.uk> wrote:
> > Lennart Regebro wrote:
> > > Seriously though: It's hard to override,
> >
> > What do you mean by that?
> 
> What I mean is that if you want to override that macro in another
> product, for say a customer site, you will have to monkey-patch it. If
> you instead register the macro so it becomes a view, it's easy to
> override.

Right, by specifying an override either in overrides.zcml (zope >= 2.8)
or providing a local adapter (zope >= 2.10).

- Rocky

-- 
Rocky Burt
ServerZen Software -- http://www.serverzen.com
News About The Server (blog) -- http://www.serverzen.net
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part
Url : http://codespeak.net/pipermail/z3-five/attachments/20070130/1c4762ce/attachment-0001.pgp 

From philipp at weitershausen.de  Tue Jan 30 16:19:58 2007
From: philipp at weitershausen.de (Philipp von Weitershausen)
Date: Tue, 30 Jan 2007 16:19:58 +0100
Subject: [z3-five] (re)using zpt macros from view page templates
In-Reply-To: <1170168849.7670.41.camel@localhost>
References: <45BC860F.9040302@simplistix.co.uk>	<5ACA5A1F87992FAEBF789311@suxmac2.local>	<45BCA57D.1000207@simplistix.co.uk>	<55D50FEBB2534351B99F2D3E@suxmac2.local>	<45BCB47E.9030409@simplistix.co.uk>	<319e029f0701290120v72d6a91fu8a3d3a17a4dc43fa@mail.gmail.com>	<45BDD77B.1020705@simplistix.co.uk>	<319e029f0701300624p78de8c99hff28d6d5115a12cf@mail.gmail.com>
	<1170168849.7670.41.camel@localhost>
Message-ID: <45BF621E.2090101@weitershausen.de>

Rocky Burt wrote:
> On Tue, 2007-30-01 at 15:24 +0100, Lennart Regebro wrote:
>> On 1/29/07, Chris Withers <chris at simplistix.co.uk> wrote:
>>> Lennart Regebro wrote:
>>>> Seriously though: It's hard to override,
>>> What do you mean by that?
>> What I mean is that if you want to override that macro in another
>> product, for say a customer site, you will have to monkey-patch it. If
>> you instead register the macro so it becomes a view, it's easy to
>> override.
> 
> Right, by specifying an override either in overrides.zcml (zope >= 2.8)
> or providing a local adapter (zope >= 2.10).

Actually, the way we typically override views, especially the ones that 
are about the appearance of the app, is through layers and skins.


-- 
http://worldcookery.com -- Professional Zope documentation and training
Next Zope 3 training at Camp5: http://trizpug.org/boot-camp/camp5

From faassen at startifact.com  Tue Jan 30 18:32:39 2007
From: faassen at startifact.com (Martijn Faassen)
Date: Tue, 30 Jan 2007 18:32:39 +0100
Subject: [z3-five] Unexpectedly unprotected code
In-Reply-To: <F94802BB-6FD3-4A8C-8979-4A37A250BD1D@weitershausen.de>
References: <45BE68A8.5030403@simplistix.co.uk>	<eplqmb$bs3$1@sea.gmane.org>	<45BE7BBF.7060805@simplistix.co.uk>	<a7a2b76b0701291502o2127b3e9we5ab8524b82e6ca6@mail.gmail.com>	<45BE7DE8.7000706@weitershausen.de>	<45BEF8A8.8040806@simplistix.co.uk>
	<F94802BB-6FD3-4A8C-8979-4A37A250BD1D@weitershausen.de>
Message-ID: <epnvfk$vbi$1@sea.gmane.org>

Philipp von Weitershausen wrote:
> On 30 Jan 2007, at 08:50 , Chris Withers wrote: 
>> Are there any plans to move Zope 2 to Zope 3's security machinery?
> 
> Jim has remote plans and Martijn has some wild ideas. I don't see it  
> happening any time soon, though. It's also not the issue here.

And you have some other ideas to make the Zope 2 security model more 
Zope 3 compatible so it can look at __parent__.

My ideas have evolved to the point that I like trusted code more and 
more, and I'm not sure it's worth the effort to expend a lot of time to 
make untrusted code work.

Regards,

Martijn


From chris at simplistix.co.uk  Tue Jan 30 19:55:26 2007
From: chris at simplistix.co.uk (Chris Withers)
Date: Tue, 30 Jan 2007 18:55:26 +0000
Subject: [z3-five] (re)using zpt macros from view page templates
In-Reply-To: <45BF49A3.70101@weitershausen.de>
References: <45BC860F.9040302@simplistix.co.uk>	<5ACA5A1F87992FAEBF789311@suxmac2.local>	<45BCA57D.1000207@simplistix.co.uk>	<45BE790B.5080302@weitershausen.de>	<45BE7B4A.4060601@simplistix.co.uk>	<84F4CD03-7C18-4B6D-AF17-029FFDAB1777@weitershausen.de>	<45BF0589.1040406@simplistix.co.uk>
	<45BF34A5.8090005@simplistix.co.uk>
	<45BF49A3.70101@weitershausen.de>
Message-ID: <45BF949E.8060204@simplistix.co.uk>

Philipp von Weitershausen wrote:
>>> from Products.Five.browser.pagetemplatefile import ViewPageTemplateFile
>>
>> This was my mistake, that should have been:
>>
>> from Products.Five.browser.pagetemplatefile import \
>>      ZopeTwoPageTemplateFile
> 
> They're the same. It shouldn't make a difference.

No, they're not...

Products.Five.browser.pagetemplatefile import ViewPageTemplateFile

...just imports ViewPageTemplateFile from there, it's origin, if you 
read pagetemplatefile.py is:

from zope.app.pagetemplate.viewpagetemplatefile import \
     ViewPageTemplateFile

...and it appears to be a total red herring, since it's not actually 
used anywhere inside pagetemplatefile.py :-(

Maybe it could get removed?

>> However, now using the right class...
>>
>> my_macro = ZopeTwoPageTemplateFile('my_macro.pt')
>>
>> my_macro.macros just gives me a ComputedAttribute object (rather than 
>> the macros I'd expect).
>>
>> If I do, my_macro.pt_macros(), I get the macros.
>>
>> Why isn't ComputedAttribute working in this example?
> 
> Coz ComputedAttribute is a stupid Zope2ism that only works in an 
> acquisition context.

Yup, I got that after much nashing...

> I still think you're overcomplicating things. 

Indeed, here's what I've ended up with, which I think is just fine...

mymacro.pt
----------
<html metal:define-macro="template">
...
</html>

macros.py
---------
from Products.Five.browser.pagetemplatefile import \
      ZopeTwoPageTemplateFile
mymacro = ZopeTwoPageTemplateFile('mymacro.pt')

myview.py
---------
from macros import mymacro

class MyView(BrowserView):
    mymacro = mymacro

myview.pt:
<html metal:use-macro="view/mymacros/macros/template">
...
</html>

I find this much more lightweight, not to mention secure, than 
registering a full blown browser page just for some macros...

Chris

-- 
Simplistix - Content Management, Zope & Python Consulting
            - http://www.simplistix.co.uk

From philipp at weitershausen.de  Tue Jan 30 20:02:50 2007
From: philipp at weitershausen.de (Philipp von Weitershausen)
Date: Tue, 30 Jan 2007 20:02:50 +0100
Subject: [z3-five] (re)using zpt macros from view page templates
In-Reply-To: <45BF949E.8060204@simplistix.co.uk>
References: <45BC860F.9040302@simplistix.co.uk>	<5ACA5A1F87992FAEBF789311@suxmac2.local>	<45BCA57D.1000207@simplistix.co.uk>	<45BE790B.5080302@weitershausen.de>	<45BE7B4A.4060601@simplistix.co.uk>	<84F4CD03-7C18-4B6D-AF17-029FFDAB1777@weitershausen.de>	<45BF0589.1040406@simplistix.co.uk>
	<45BF34A5.8090005@simplistix.co.uk>
	<45BF49A3.70101@weitershausen.de>
	<45BF949E.8060204@simplistix.co.uk>
Message-ID: <C8D8FE69-F3AF-427F-BD19-80F33B74F34B@weitershausen.de>

On 30 Jan 2007, at 19:55 , Chris Withers wrote:
> Philipp von Weitershausen wrote:
>>>> from Products.Five.browser.pagetemplatefile import  
>>>> ViewPageTemplateFile
>>>
>>> This was my mistake, that should have been:
>>>
>>> from Products.Five.browser.pagetemplatefile import \
>>>      ZopeTwoPageTemplateFile
>> They're the same. It shouldn't make a difference.
>
> No, they're not...
>
> Products.Five.browser.pagetemplatefile import ViewPageTemplateFile
>
> ...just imports ViewPageTemplateFile from there, it's origin, if  
> you read pagetemplatefile.py is:
>
> from zope.app.pagetemplate.viewpagetemplatefile import \
>     ViewPageTemplateFile
>
> ...and it appears to be a total red herring, since it's not  
> actually used anywhere inside pagetemplatefile.py :-(

Aaah. In more recent Five versions, I aliased ZopeTwoPageTemplateFile  
(what a stupid name) to ViewPageTemplateFile.

> Maybe it could get removed?

Aaah, yes, the unused import should be removed.





From d.schulz81 at gmx.net  Wed Jan 31 11:19:15 2007
From: d.schulz81 at gmx.net (Dennis Schulz)
Date: Wed, 31 Jan 2007 11:19:15 +0100
Subject: [z3-five] using getToolByName in utility
Message-ID: <45C06D23.8040103@gmx.net>

Hello...

I made an utility (vocabulary factory) that returns all users. It 
queries the membership tool. (see code at the end)
That works fine for displaying it as a dropdown in a formlib form.
However, when change a value and submit the form, the utility gets 
called in a weird context by the formlib validation machinery.
When I print out aq_chain then I only have one object without base request.
Within this weird context the getToolByName loopup is not possible.

Is there any workaround except trying to skip validation (how?)?

Dennis

class UserVocabulary(object):
    implements(IVocabularyFactory)
    def __call__(self, context):
            context = getattr(context, 'context', context)
            context = aq_inner(context)
            print aq_chain(context)
            mt = getToolByName(context, 'portal_membership')
            memberids = mt.listMemberIds()
            result=[( mt.getMemberById(i).getProperty('fullname'),i) for 
i in memberids]
            return SimpleVocabulary.fromItems(result)
UserVocabularyFactory = UserVocabulary()







From chris at simplistix.co.uk  Wed Jan 31 11:15:34 2007
From: chris at simplistix.co.uk (Chris Withers)
Date: Wed, 31 Jan 2007 10:15:34 +0000
Subject: [z3-five] Unexpectedly unprotected code
In-Reply-To: <epnvfk$vbi$1@sea.gmane.org>
References: <45BE68A8.5030403@simplistix.co.uk>	<eplqmb$bs3$1@sea.gmane.org>	<45BE7BBF.7060805@simplistix.co.uk>	<a7a2b76b0701291502o2127b3e9we5ab8524b82e6ca6@mail.gmail.com>	<45BE7DE8.7000706@weitershausen.de>	<45BEF8A8.8040806@simplistix.co.uk>	<F94802BB-6FD3-4A8C-8979-4A37A250BD1D@weitershausen.de>
	<epnvfk$vbi$1@sea.gmane.org>
Message-ID: <45C06C46.8040103@simplistix.co.uk>

Martijn Faassen wrote:
> My ideas have evolved to the point that I like trusted code more and 
> more, and I'm not sure it's worth the effort to expend a lot of time to 
> make untrusted code work.

Oh I dunno, I think this an exceptionally important use case which the 
Zope community seems to be ignoring more and more and which used to be 
the main thing that brought people to Zope: the ability for a trusted 
but not necessarily fully competent user to write code while protecting 
them from accessing data they shouldn't and trying to help them not 
shoot themselves in the foot...

I think that's still well worth doing...

Chris

-- 
Simplistix - Content Management, Zope & Python Consulting
            - http://www.simplistix.co.uk


From y.2007- at wcm-solutions.de  Wed Jan 31 13:19:50 2007
From: y.2007- at wcm-solutions.de (yuppie)
Date: Wed, 31 Jan 2007 13:19:50 +0100
Subject: [z3-five] using getToolByName in utility
In-Reply-To: <45C06D23.8040103@gmx.net>
References: <45C06D23.8040103@gmx.net>
Message-ID: <45C08966.9060905@wcm-solutions.de>

Hi Dennis!


Dennis Schulz wrote:
> I made an utility (vocabulary factory) that returns all users. It 
> queries the membership tool. (see code at the end)
> That works fine for displaying it as a dropdown in a formlib form.
> However, when change a value and submit the form, the utility gets 
> called in a weird context by the formlib validation machinery.
> When I print out aq_chain then I only have one object without base request.
> Within this weird context the getToolByName loopup is not possible.

IIRC the validation machinery uses the context bound to the field. 
Something seems to be wrong with the bind() call.

> class UserVocabulary(object):
>     implements(IVocabularyFactory)
>     def __call__(self, context):

What is context and aq_chain(context) here?

>             context = getattr(context, 'context', context)

And what here?

>             context = aq_inner(context)
>             print aq_chain(context)
>             mt = getToolByName(context, 'portal_membership')
>             memberids = mt.listMemberIds()
>             result=[( mt.getMemberById(i).getProperty('fullname'),i) for 
> i in memberids]
>             return SimpleVocabulary.fromItems(result)
> UserVocabularyFactory = UserVocabulary()


HTH,

	Yuppie


From d.schulz81 at gmx.net  Wed Jan 31 14:56:29 2007
From: d.schulz81 at gmx.net (Dennis Schulz)
Date: Wed, 31 Jan 2007 14:56:29 +0100
Subject: [z3-five] using getToolByName in utility
In-Reply-To: <45C08966.9060905@wcm-solutions.de>
References: <45C06D23.8040103@gmx.net> <45C08966.9060905@wcm-solutions.de>
Message-ID: <45C0A00D.3010608@gmx.net>

**Hello Yuppie,**

**No matter where I print context (all places that you suggested) , 
aq_base(context) it is always the same, only my object without 
acquisition chain:**

<PurchaseRequisition at 3>
[<PurchaseRequisition at 3>]


I am using the applychanges of alchemist and the alchmist bind method.
Could this be the reason?

Here is the applychanges

def applyChanges(context, form_fields, data, adapters=None):
    """
    the default applyChanges from formlib, doesn't play nice with subobjects
    unless a new object is created on edit, we want to edit the value in 
place
    so to detect value changes on subobject changes we manually introspect
    the state of the mapped object to detect changes.
    """
    if adapters is None:
        adapters = {}

    changed = False

    for form_field in form_fields:
        field = form_field.field
        # Adapt context, if necessary
        interface = field.interface
        adapter = adapters.get(interface)
        if adapter is None:
            if interface is None:
                adapter = context
            else:
                adapter = interface(context)
            adapters[interface] = adapter

        name = form_field.__name__
        newvalue = data.get(name, form_field) # using form_field as marker
      
        if IObject.providedBy( field ) and ITableSchema.providedBy( 
newvalue ):
            # we expunge objects from session, to prevent spurious db 
changes on caught
            # form exceptions. if we get to the application of form 
data, then we passed
            # through validation, and values should be reattached to the 
session, for
            # persistence.
            if not attribute_manager.is_modified( newvalue ):
                continue
           
            newvalue = aq_base( newvalue )
            session_id = getattr( newvalue, '_sa_session_id', None)
           
            if session_id is None:
                session = get_session()
                if getattr( newvalue, '_instance_key', None):
                    session._attach( newvalue )
                    # this doesn't work as well we would expect, it 
resets state on updates
                    # session.save_or_update( newvalue )
                    assert newvalue in session.dirty
                    changed = True
                    continue
                # else its new
                session.save( newvalue )                   
                changed = True

        if (newvalue is not form_field) and (field.get(adapter) != 
newvalue):
            changed = True
            field.set(adapter, newvalue)

    return changed


****

** **

** **

**Here the exception I get:**

**ConstraintNotSatisfied**

Sorry, a site error occurred.

Traceback (innermost last):

    * Module ZPublisher.Publish, line 196, in publish_module_standard
    * Module Products.PlacelessTranslationService.PatchStringIO, line
      34, in new_publish
    * Module ZPublisher.Publish, line 146, in publish
    * Module Zope2.App.startup, line 222, in zpublisher_exception_hook
    * Module ZPublisher.Publish, line 115, in publish
    * Module ZPublisher.mapply, line 88, in mapply
    * Module ZPublisher.Publish, line 41, in call_object
    * Module zope.formlib.form, line 757, in __call__
    * Module Products.PurchaseRequisition.browser.CELLSformlib, line
      571, in update
    * Module Products.Five.formlib.formbase, line 54, in update
    * Module zope.formlib.form, line 738, in update
    * Module zope.formlib.form, line 582, in success
    * Module Products.PurchaseRequisition.browser.CELLSformlib, line
      458, in handle_save_action
    * Module Products.PurchaseRequisition.browser.CELLSformlib, line
      1021, in applyChanges
    * Module zope.schema._bootstrapfields, line 183, in set
    * Module ore.alchemist.bind, line 53, in __set__
    * Module zope.schema._bootstrapfields, line 138, in validate
    * Module zope.schema._field, line 271, in _validate

ConstraintNotSatisfied: klora (Also, the following error occurred while 
attempting to render the standard error message, please see the event 
log for full details: 'str' object has no attribute 'args')
yuppie escribi?:
> Hi Dennis!
>
>
> Dennis Schulz wrote:
>   
>> I made an utility (vocabulary factory) that returns all users. It 
>> queries the membership tool. (see code at the end)
>> That works fine for displaying it as a dropdown in a formlib form.
>> However, when change a value and submit the form, the utility gets 
>> called in a weird context by the formlib validation machinery.
>> When I print out aq_chain then I only have one object without base request.
>> Within this weird context the getToolByName loopup is not possible.
>>     
>
> IIRC the validation machinery uses the context bound to the field. 
> Something seems to be wrong with the bind() call.
>
>   
>> class UserVocabulary(object):
>>     implements(IVocabularyFactory)
>>     def __call__(self, context):
>>     
>
> What is context and aq_chain(context) here?
>
>   
>>             context = getattr(context, 'context', context)
>>     
>
> And what here?
>
>   
>>             context = aq_inner(context)
>>             print aq_chain(context)
>>             mt = getToolByName(context, 'portal_membership')
>>             memberids = mt.listMemberIds()
>>             result=[( mt.getMemberById(i).getProperty('fullname'),i) for 
>> i in memberids]
>>             return SimpleVocabulary.fromItems(result)
>> UserVocabularyFactory = UserVocabulary()
>>     
>
>
> HTH,
>
> 	Yuppie
>
> _______________________________________________
> z3-five mailing list
> z3-five at codespeak.net
> http://codespeak.net/mailman/listinfo/z3-five
>
>
>   



From y.2007- at wcm-solutions.de  Wed Jan 31 15:33:48 2007
From: y.2007- at wcm-solutions.de (yuppie)
Date: Wed, 31 Jan 2007 15:33:48 +0100
Subject: [z3-five] using getToolByName in utility
In-Reply-To: <45C0A00D.3010608@gmx.net>
References: <45C06D23.8040103@gmx.net> <45C08966.9060905@wcm-solutions.de>
	<45C0A00D.3010608@gmx.net>
Message-ID: <45C0A8CC.8010508@wcm-solutions.de>

Dennis Schulz wrote:
> **No matter where I print context (all places that you suggested) , 
> aq_base(context) it is always the same, only my object without 
> acquisition chain:**
> 
> <PurchaseRequisition at 3>
> [<PurchaseRequisition at 3>]
> 
> 
> I am using the applychanges of alchemist and the alchmist bind method.
> Could this be the reason?

>     * Module ore.alchemist.bind, line 53, in __set__

__set__() is called with your object as argument. You have to figure out 
why that object is not acquisition wrapped. AFAICS your code only works 
if your object is wrapped in that place.

HTH, Yuppie


From d.schulz81 at gmx.net  Wed Jan 31 17:12:03 2007
From: d.schulz81 at gmx.net (Dennis Schulz)
Date: Wed, 31 Jan 2007 17:12:03 +0100
Subject: [z3-five] using getToolByName in utility
In-Reply-To: <45C0A8CC.8010508@wcm-solutions.de>
References: <45C06D23.8040103@gmx.net> <45C08966.9060905@wcm-solutions.de>
	<45C0A00D.3010608@gmx.net> <45C0A8CC.8010508@wcm-solutions.de>
Message-ID: <45C0BFD3.6060801@gmx.net>

In the applychanges the object is still wrapped into acquisition correctly.
and gets passed to field.set(adapter,newvalue) with the right context.

In the set of the field the aq chain is ok as well
it calls then setattr(object, self.__name__, value)

now the set of the alchemist ValidatedProperty is invoked.
in the set method there the aq_chain is destroyed.

I dont know why the set is called without acquisition context.
So you see any way to change something here?

Dennis


yuppie escribi?:
> Dennis Schulz wrote:
>> **No matter where I print context (all places that you suggested) , 
>> aq_base(context) it is always the same, only my object without 
>> acquisition chain:**
>>
>> <PurchaseRequisition at 3>
>> [<PurchaseRequisition at 3>]
>>
>>
>> I am using the applychanges of alchemist and the alchmist bind method.
>> Could this be the reason?
>
>>     * Module ore.alchemist.bind, line 53, in __set__
>
> __set__() is called with your object as argument. You have to figure 
> out why that object is not acquisition wrapped. AFAICS your code only 
> works if your object is wrapped in that place.
>
> HTH, Yuppie
>
>



From faassen at startifact.com  Wed Jan 31 17:12:54 2007
From: faassen at startifact.com (Martijn Faassen)
Date: Wed, 31 Jan 2007 17:12:54 +0100
Subject: [z3-five] Unexpectedly unprotected code
In-Reply-To: <45C06C46.8040103@simplistix.co.uk>
References: <45BE68A8.5030403@simplistix.co.uk>	<eplqmb$bs3$1@sea.gmane.org>	<45BE7BBF.7060805@simplistix.co.uk>	<a7a2b76b0701291502o2127b3e9we5ab8524b82e6ca6@mail.gmail.com>	<45BE7DE8.7000706@weitershausen.de>	<45BEF8A8.8040806@simplistix.co.uk>	<F94802BB-6FD3-4A8C-8979-4A37A250BD1D@weitershausen.de>	<epnvfk$vbi$1@sea.gmane.org>
	<45C06C46.8040103@simplistix.co.uk>
Message-ID: <epqf66$dck$1@sea.gmane.org>

Chris Withers wrote:
> Martijn Faassen wrote:
>> My ideas have evolved to the point that I like trusted code more and 
>> more, and I'm not sure it's worth the effort to expend a lot of time to 
>> make untrusted code work.
> 
> Oh I dunno, I think this an exceptionally important use case which the 
> Zope community seems to be ignoring more and more and which used to be 
> the main thing that brought people to Zope: the ability for a trusted 
> but not necessarily fully competent user to write code while protecting 
> them from accessing data they shouldn't and trying to help them not 
> shoot themselves in the foot...
> 
> I think that's still well worth doing...

I agree the use case exists. I'm not sure how important it is, though 
traditionally it's been quite important to Zope 2.

I think there are a lot of *other* things we should be doing first to 
make an inexperienced developer happier with Zope 3. Some of those 
things we've been trying to do with Grok.

One of the things that bugs me even as an *experienced* developer is 
that Zope 3's pervasive security has a heavy cost during development. It 
happened to me quite frequently I had to debug why Zope 3 didn't let me 
do something I should do, and I had to dig through ZCML files and add 
security declarations quite often, and mess about with __parent__ quite 
often, and use removeAllSecurityProxies() and such quite often. I 
consider this very off-putting to any developer, experienced or not.

Regards,

Martijn


From chrism at plope.com  Wed Jan 31 17:37:14 2007
From: chrism at plope.com (Chris McDonough)
Date: Wed, 31 Jan 2007 11:37:14 -0500
Subject: [z3-five] Unexpectedly unprotected code
In-Reply-To: <epqf66$dck$1@sea.gmane.org>
References: <45BE68A8.5030403@simplistix.co.uk>	<eplqmb$bs3$1@sea.gmane.org>	<45BE7BBF.7060805@simplistix.co.uk>	<a7a2b76b0701291502o2127b3e9we5ab8524b82e6ca6@mail.gmail.com>	<45BE7DE8.7000706@weitershausen.de>	<45BEF8A8.8040806@simplistix.co.uk>	<F94802BB-6FD3-4A8C-8979-4A37A250BD1D@weitershausen.de>	<epnvfk$vbi$1@sea.gmane.org>
	<45C06C46.8040103@simplistix.co.uk> <epqf66$dck$1@sea.gmane.org>
Message-ID: <911BAE4F-6AF2-4D15-A0CE-5C3EF8CFDCBD@plope.com>

As always, a big +1 to anything that helps experienced developers not  
need to care about unstrusted code.

- C

On Jan 31, 2007, at 11:12 AM, Martijn Faassen wrote:

> Chris Withers wrote:
>> Martijn Faassen wrote:
>>> My ideas have evolved to the point that I like trusted code more and
>>> more, and I'm not sure it's worth the effort to expend a lot of  
>>> time to
>>> make untrusted code work.
>>
>> Oh I dunno, I think this an exceptionally important use case which  
>> the
>> Zope community seems to be ignoring more and more and which used  
>> to be
>> the main thing that brought people to Zope: the ability for a trusted
>> but not necessarily fully competent user to write code while  
>> protecting
>> them from accessing data they shouldn't and trying to help them not
>> shoot themselves in the foot...
>>
>> I think that's still well worth doing...
>
> I agree the use case exists. I'm not sure how important it is, though
> traditionally it's been quite important to Zope 2.
>
> I think there are a lot of *other* things we should be doing first to
> make an inexperienced developer happier with Zope 3. Some of those
> things we've been trying to do with Grok.
>
> One of the things that bugs me even as an *experienced* developer is
> that Zope 3's pervasive security has a heavy cost during  
> development. It
> happened to me quite frequently I had to debug why Zope 3 didn't  
> let me
> do something I should do, and I had to dig through ZCML files and add
> security declarations quite often, and mess about with __parent__  
> quite
> often, and use removeAllSecurityProxies() and such quite often. I
> consider this very off-putting to any developer, experienced or not.
>
> Regards,
>
> Martijn
>
> _______________________________________________
> z3-five mailing list
> z3-five at codespeak.net
> http://codespeak.net/mailman/listinfo/z3-five
>


From y.2007- at wcm-solutions.de  Wed Jan 31 18:14:40 2007
From: y.2007- at wcm-solutions.de (yuppie)
Date: Wed, 31 Jan 2007 18:14:40 +0100
Subject: [z3-five] using getToolByName in utility
In-Reply-To: <45C0BFD3.6060801@gmx.net>
References: <45C06D23.8040103@gmx.net>
	<45C08966.9060905@wcm-solutions.de>	<45C0A00D.3010608@gmx.net>
	<45C0A8CC.8010508@wcm-solutions.de> <45C0BFD3.6060801@gmx.net>
Message-ID: <45C0CE80.4060202@wcm-solutions.de>

Hi Dennis!


Dennis Schulz wrote:
> In the applychanges the object is still wrapped into acquisition correctly.
> and gets passed to field.set(adapter,newvalue) with the right context.
> 
> In the set of the field the aq chain is ok as well
> it calls then setattr(object, self.__name__, value)
> 
> now the set of the alchemist ValidatedProperty is invoked.
> in the set method there the aq_chain is destroyed.
> 
> I dont know why the set is called without acquisition context.

Acquisition isn't available in __setattr__ :(

CMFCalendar.browser.event (on SVN trunk) does something similar, but in 
that case an adapter is used. The wrapped object is the context of the 
adapter, not the adapter itself. So the aq_chain is not destroyed in 
__setattr__.

> So you see any way to change something here?

No idea. (Besides using an adapter.) The 'tools as local utilities' work 
currently done on CMF trunk will make this much easier. You no longer 
need Acquisition to get the tool (=utility).


Cheers,

	Yuppie


From pw_lists at slinkp.com  Wed Jan 31 19:16:25 2007
From: pw_lists at slinkp.com (Paul Winkler)
Date: Wed, 31 Jan 2007 13:16:25 -0500
Subject: [z3-five] Unexpectedly unprotected code
In-Reply-To: <45C06C46.8040103@simplistix.co.uk>
References: <45BE68A8.5030403@simplistix.co.uk> <eplqmb$bs3$1@sea.gmane.org>
	<45BE7BBF.7060805@simplistix.co.uk>
	<a7a2b76b0701291502o2127b3e9we5ab8524b82e6ca6@mail.gmail.com>
	<45BE7DE8.7000706@weitershausen.de>
	<45BEF8A8.8040806@simplistix.co.uk>
	<F94802BB-6FD3-4A8C-8979-4A37A250BD1D@weitershausen.de>
	<epnvfk$vbi$1@sea.gmane.org> <45C06C46.8040103@simplistix.co.uk>
Message-ID: <20070131181625.GC7183@slinkp.com>

On Wed, Jan 31, 2007 at 10:15:34AM +0000, Chris Withers wrote:
> Martijn Faassen wrote:
> > My ideas have evolved to the point that I like trusted code more and 
> > more, and I'm not sure it's worth the effort to expend a lot of time to 
> > make untrusted code work.
> 
> Oh I dunno, I think this an exceptionally important use case which the 
> Zope community seems to be ignoring more and more and which used to be 
> the main thing that brought people to Zope: the ability for a trusted 
> but not necessarily fully competent user to write code while protecting 
> them from accessing data they shouldn't and trying to help them not 
> shoot themselves in the foot...

Hmm, was that really a big draw to Zope?  Considering the relative
popularity of various alternatives (php, rails,...) none of which have
a concept of "untrusted" code, I've sometimes wondered if the whole
idea of untrusted code is a solution in search of a problem.
But I'd be curious to hear evidence otherwise.


-- 

Paul Winkler
http://www.slinkp.com

From chris at simplistix.co.uk  Wed Jan 31 21:22:16 2007
From: chris at simplistix.co.uk (Chris Withers)
Date: Wed, 31 Jan 2007 20:22:16 +0000
Subject: [z3-five] Unexpectedly unprotected code
In-Reply-To: <20070131181625.GC7183@slinkp.com>
References: <45BE68A8.5030403@simplistix.co.uk>
	<eplqmb$bs3$1@sea.gmane.org>	<45BE7BBF.7060805@simplistix.co.uk>	<a7a2b76b0701291502o2127b3e9we5ab8524b82e6ca6@mail.gmail.com>	<45BE7DE8.7000706@weitershausen.de>	<45BEF8A8.8040806@simplistix.co.uk>	<F94802BB-6FD3-4A8C-8979-4A37A250BD1D@weitershausen.de>	<epnvfk$vbi$1@sea.gmane.org>
	<45C06C46.8040103@simplistix.co.uk>
	<20070131181625.GC7183@slinkp.com>
Message-ID: <45C0FA78.9020208@simplistix.co.uk>

Paul Winkler wrote:
> Hmm, was that really a big draw to Zope? 

When Zope was growing rapidly (1999-2003 for me, others may disagree) 
everyone who came to Zope came to it by installing it and writing TTW 
code. That was predicated on what I'm talking about ;-)

> Considering the relative
> popularity of various alternatives (php, rails,...) 

I dunno about rails, but considering how heinously insecure php is, I 
would quite happilly sacrifice popularity for security.

> none of which have
> a concept of "untrusted" code, I've sometimes wondered if the whole
> idea of untrusted code is a solution in search of a problem.
> But I'd be curious to hear evidence otherwise.

In Zope 2, almost all components had to worry about "untrusted" code.
I'd imagine in Zope 3 that should change to being:

- only publication objects, which must surely control all access based 
on security declarations of the objects involved and deny all access to 
objects without any assertions

- only code that is designed to be scripted by Zope 2's semi-trusted 
users. I'd imagine this would be a much smaller amount of code, and I'd 
imagine it would be designed to be highly encapsulated versus the 
pervasiveness of Zope 2's "untrusted code"

However, security is also about resricting access to data objects. Zope 
2's integrated security policy was a huge huge win for this, I hope that 
doesn't ever get lost in Zope 3.

cheers,

Chris

-- 
Simplistix - Content Management, Zope & Python Consulting
            - http://www.simplistix.co.uk


From chris at simplistix.co.uk  Wed Jan 31 21:10:02 2007
From: chris at simplistix.co.uk (Chris Withers)
Date: Wed, 31 Jan 2007 20:10:02 +0000
Subject: [z3-five] Unexpectedly unprotected code
In-Reply-To: <epqf66$dck$1@sea.gmane.org>
References: <45BE68A8.5030403@simplistix.co.uk>	<eplqmb$bs3$1@sea.gmane.org>	<45BE7BBF.7060805@simplistix.co.uk>	<a7a2b76b0701291502o2127b3e9we5ab8524b82e6ca6@mail.gmail.com>	<45BE7DE8.7000706@weitershausen.de>	<45BEF8A8.8040806@simplistix.co.uk>	<F94802BB-6FD3-4A8C-8979-4A37A250BD1D@weitershausen.de>	<epnvfk$vbi$1@sea.gmane.org>	<45C06C46.8040103@simplistix.co.uk>
	<epqf66$dck$1@sea.gmane.org>
Message-ID: <45C0F79A.6060908@simplistix.co.uk>

Martijn Faassen wrote:
> I agree the use case exists. I'm not sure how important it is, though 
> traditionally it's been quite important to Zope 2.

I think the userbase who used to thrive on this have, for the whole, 
moved over to Plone.

> I think there are a lot of *other* things we should be doing first to 
> make an inexperienced developer happier with Zope 3. Some of those 
> things we've been trying to do with Grok.

Agreed. I think supporting the use case I mention is an application of 
zope 3, not zope 3 itself...

> One of the things that bugs me even as an *experienced* developer is 
> that Zope 3's pervasive security has a heavy cost during development.

I don't agree.

Zope's fundamental premise is publishing a tree of objects by various 
protocols. If you do that, you need to make sure you're not exposing 
things to abuse by untrusted users, and the only way to do that is to 
deny-by-default everything to publishing protocols.

Perhaps the way this is currently done is Zope 3 isn't as developer 
friendly as it should be, but you really really want a big 
deny-by-default hammer somewhere between "trusted code" and any 
publishing protocol.

> It 
> happened to me quite frequently I had to debug why Zope 3 didn't let me 
> do something I should do, and I had to dig through ZCML files 

I thought we'd established that zcml sucks? ;-)

> and add 
> security declarations quite often,

I think you need to positively declare the security assertions relating 
to objects (be they simple python types, functions, methods or anything 
else) somewhere...

> and mess about with __parent__ quite 
> often, 

Yes, the containment constraint of Zope 2 and 3 security policies is 
annoying. I hope to implement a security policy for zope 3 at some point 
that is purely context-based rather than containment based.

> and use removeAllSecurityProxies() and such quite often. 

It's a shame you have to do that :-(

cheers,

Chris

-- 
Simplistix - Content Management, Zope & Python Consulting
            - http://www.simplistix.co.uk


From d.schulz81 at gmx.net  Thu Feb  1 11:04:33 2007
From: d.schulz81 at gmx.net (Dennis Schulz)
Date: Thu, 01 Feb 2007 11:04:33 +0100
Subject: [z3-five] using getToolByName in utility
In-Reply-To: <45C0CE80.4060202@wcm-solutions.de>
References: <45C06D23.8040103@gmx.net> <45C08966.9060905@wcm-solutions.de>
	<45C0A00D.3010608@gmx.net> <45C0A8CC.8010508@wcm-solutions.de>
	<45C0BFD3.6060801@gmx.net> <45C0CE80.4060202@wcm-solutions.de>
Message-ID: <45C1BB31.3070908@gmx.net>

Hello Yuppie,

I looked at EventTypeVocabulary in CMFCalendar.browser.event. 
In the Utitily itself I cant see any difference to mine.
Can you tell where it is used?

I my case does it make I create an adapter in the applychanges and then use the context of the adapter instead of the adapter?

set(adapter,newvalue)

Thanks,
Dennis

yuppie escribi?:
> Hi Dennis!
>
>
> Dennis Schulz wrote:
>> In the applychanges the object is still wrapped into acquisition 
>> correctly.
>> and gets passed to field.set(adapter,newvalue) with the right context.
>>
>> In the set of the field the aq chain is ok as well
>> it calls then setattr(object, self.__name__, value)
>>
>> now the set of the alchemist ValidatedProperty is invoked.
>> in the set method there the aq_chain is destroyed.
>>
>> I dont know why the set is called without acquisition context.
>
> Acquisition isn't available in __setattr__ :(
>
> CMFCalendar.browser.event (on SVN trunk) does something similar, but 
> in that case an adapter is used. The wrapped object is the context of 
> the adapter, not the adapter itself. So the aq_chain is not destroyed 
> in __setattr__.
>
>> So you see any way to change something here?
>
> No idea. (Besides using an adapter.) The 'tools as local utilities' 
> work currently done on CMF trunk will make this much easier. You no 
> longer need Acquisition to get the tool (=utility).
>
>
> Cheers,
>
>     Yuppie
>
>



From y.2007- at wcm-solutions.de  Thu Feb  1 11:43:59 2007
From: y.2007- at wcm-solutions.de (yuppie)
Date: Thu, 01 Feb 2007 11:43:59 +0100
Subject: [z3-five] using getToolByName in utility
In-Reply-To: <45C1BB31.3070908@gmx.net>
References: <45C06D23.8040103@gmx.net>
	<45C08966.9060905@wcm-solutions.de>	<45C0A00D.3010608@gmx.net>
	<45C0A8CC.8010508@wcm-solutions.de>	<45C0BFD3.6060801@gmx.net>
	<45C0CE80.4060202@wcm-solutions.de> <45C1BB31.3070908@gmx.net>
Message-ID: <45C1C46F.2030009@wcm-solutions.de>

Hi Dennis!


Dennis Schulz wrote:
> I looked at EventTypeVocabulary in CMFCalendar.browser.event. 
> In the Utitily itself I cant see any difference to mine.
> Can you tell where it is used?

The name of the utility is 'cmf.calendar.AvailableEventTypes'. It is 
used in the 'categories' field of IEventSchema. The EventSchemaAdapter 
implements this interface, not the persistent Event object itself.

> I my case does it make I create an adapter in the applychanges and then use the context of the adapter instead of the adapter?
> 
> set(adapter,newvalue)

In CMFCalendar we have to adapt the old Event class anyway. Don't know 
if it makes sense to implement your stuff in a similar way. Can't help 
you with that.


Cheers,

	Yuppie


From regebro at gmail.com  Thu Feb  1 11:51:11 2007
From: regebro at gmail.com (Lennart Regebro)
Date: Thu, 1 Feb 2007 11:51:11 +0100
Subject: [z3-five] Unexpectedly unprotected code
In-Reply-To: <45C0FA78.9020208@simplistix.co.uk>
References: <45BE68A8.5030403@simplistix.co.uk>
	<45BE7BBF.7060805@simplistix.co.uk>
	<a7a2b76b0701291502o2127b3e9we5ab8524b82e6ca6@mail.gmail.com>
	<45BE7DE8.7000706@weitershausen.de>
	<45BEF8A8.8040806@simplistix.co.uk>
	<F94802BB-6FD3-4A8C-8979-4A37A250BD1D@weitershausen.de>
	<epnvfk$vbi$1@sea.gmane.org> <45C06C46.8040103@simplistix.co.uk>
	<20070131181625.GC7183@slinkp.com> <45C0FA78.9020208@simplistix.co.uk>
Message-ID: <319e029f0702010251i243638cese22b4425ddce868b@mail.gmail.com>

On 1/31/07, Chris Withers <chris at simplistix.co.uk> wrote:
> When Zope was growing rapidly (1999-2003 for me, others may disagree)
> everyone who came to Zope came to it by installing it and writing TTW
> code. That was predicated on what I'm talking about ;-)

Aboslutely. But I'm not sure that the draw was (or is) that
unexperiencd users could do things in a untrusted context. The draw
was, I think, always the ease and speed of development, and the way
that you could do just one little thing and immediately see it
working,  without the need to write loads of code first. The untrusted
security context was the price you had to pay for the speedy
development. You needed to dance around the security restrictions and
do easy things in a complicated way.

-- 
Lennart Regebro: Python, Zope, CPS, Plone consulting.
+33 661 58 14 64

From d.schulz81 at gmx.net  Thu Feb  1 15:15:34 2007
From: d.schulz81 at gmx.net (Dennis Schulz)
Date: Thu, 01 Feb 2007 15:15:34 +0100
Subject: [z3-five] using getToolByName in utility
In-Reply-To: <45C1C46F.2030009@wcm-solutions.de>
References: <45C06D23.8040103@gmx.net> <45C08966.9060905@wcm-solutions.de>
	<45C0A00D.3010608@gmx.net> <45C0A8CC.8010508@wcm-solutions.de>
	<45C0BFD3.6060801@gmx.net> <45C0CE80.4060202@wcm-solutions.de>
	<45C1BB31.3070908@gmx.net> <45C1C46F.2030009@wcm-solutions.de>
Message-ID: <45C1F606.1060800@gmx.net>

Hi Yuppie,

To make use of the adapter like you did the Products.CMFDefault.formlib 
is necessary?
The Products.CMFDefault.formlib package is not included into Plone 2.5.2 
I am working with. I guess it will be part of Plone 3.0.
Can I use it in Plone 2.5.2?

Dennis


yuppie escribi?:
> Hi Dennis!
>
>
> Dennis Schulz wrote:
>> I looked at EventTypeVocabulary in CMFCalendar.browser.event. In the 
>> Utitily itself I cant see any difference to mine.
>> Can you tell where it is used?
>
> The name of the utility is 'cmf.calendar.AvailableEventTypes'. It is 
> used in the 'categories' field of IEventSchema. The EventSchemaAdapter 
> implements this interface, not the persistent Event object itself.
>
>> I my case does it make I create an adapter in the applychanges and 
>> then use the context of the adapter instead of the adapter?
>>
>> set(adapter,newvalue)
>
> In CMFCalendar we have to adapt the old Event class anyway. Don't know 
> if it makes sense to implement your stuff in a similar way. Can't help 
> you with that.
>
>
> Cheers,
>
>     Yuppie
>
>



From y.2007- at wcm-solutions.de  Thu Feb  1 16:16:09 2007
From: y.2007- at wcm-solutions.de (yuppie)
Date: Thu, 01 Feb 2007 16:16:09 +0100
Subject: [z3-five] using getToolByName in utility
In-Reply-To: <45C1F606.1060800@gmx.net>
References: <45C06D23.8040103@gmx.net>
	<45C08966.9060905@wcm-solutions.de>	<45C0A00D.3010608@gmx.net>
	<45C0A8CC.8010508@wcm-solutions.de>	<45C0BFD3.6060801@gmx.net>
	<45C0CE80.4060202@wcm-solutions.de>	<45C1BB31.3070908@gmx.net>
	<45C1C46F.2030009@wcm-solutions.de> <45C1F606.1060800@gmx.net>
Message-ID: <45C20439.8080102@wcm-solutions.de>

Hi Dennis!


Dennis Schulz wrote:
> To make use of the adapter like you did the Products.CMFDefault.formlib 
> is necessary?
> The Products.CMFDefault.formlib package is not included into Plone 2.5.2 
> I am working with. I guess it will be part of Plone 3.0.
> Can I use it in Plone 2.5.2?

I never used alchemist, but I doubt using Products.CMFDefault.formlib in 
combination with alchemist makes much sense. And without modifications 
you can't use that code in Plone 2.5.2 anyway.

Other possible approaches:

- The right way to solve your problem would be using a local site 
manager. Even if your tools are not registered as utilities, you can get 
hold of the site root and therefor of the tool without using 
acquisition. Plone 3.0 might give you some hints how to set this up.

- If you don't need the validation at field level, the easiest solution 
might be subclassing the field and disabling the validation by 
overriding validate().


HTH,

	Yuppie


From pw_lists at slinkp.com  Thu Feb  1 16:41:19 2007
From: pw_lists at slinkp.com (Paul Winkler)
Date: Thu, 1 Feb 2007 10:41:19 -0500
Subject: [z3-five] Unexpectedly unprotected code
In-Reply-To: <45C0FA78.9020208@simplistix.co.uk>
References: <eplqmb$bs3$1@sea.gmane.org> <45BE7BBF.7060805@simplistix.co.uk>
	<a7a2b76b0701291502o2127b3e9we5ab8524b82e6ca6@mail.gmail.com>
	<45BE7DE8.7000706@weitershausen.de>
	<45BEF8A8.8040806@simplistix.co.uk>
	<F94802BB-6FD3-4A8C-8979-4A37A250BD1D@weitershausen.de>
	<epnvfk$vbi$1@sea.gmane.org> <45C06C46.8040103@simplistix.co.uk>
	<20070131181625.GC7183@slinkp.com>
	<45C0FA78.9020208@simplistix.co.uk>
Message-ID: <20070201154119.GA7431@slinkp.com>


On Wed, Jan 31, 2007 at 08:22:16PM +0000, Chris Withers wrote:
> Paul Winkler wrote:
> >Hmm, was that really a big draw to Zope? 
> 
> When Zope was growing rapidly (1999-2003 for me, others may disagree) 
> everyone who came to Zope came to it by installing it and writing TTW 
> code. That was predicated on what I'm talking about ;-)

Yeah, but nowhere is it written in stone that the security model of
TTW code needs to be different than the security model of filesystem
code.  I'm heretically suggesting that we consider whether this
distinction does more harm than good. For a trivial example, the old
"import re is not allowed" FAQ.  IIRC the motivation for disallowing
re is that regular expressions can be indeterminately slow and a naive
scripter might create code that's an easy DOS target - or a malicious
scripter might do so deliberately.  In that case at least, I'm now
pretty firmly on the side of "give 'em the rope and if they hang
themselves, it's their own problem".

Where this hypothesis obviously falls down is the "multiple customers
(each with their own engineering staff) sharing one Zope"
scenario. Some Zope hosting providers run their low-budget plans this
way. You can't leave customer A a giant back door into customer B's
data. In this case, the default zope 2 security model makes a lot of
sense.

But since as far as I know Zope is the only framework that supports
that hosting model, again I wonder if the complexity is worth it.  If
the only option for hosting multiple customers is to give each its own
instance, things get much simpler. OTOH, being in the Zope hosting
business seems to be difficult enough already without exploding the
number of long-running processes required. I doubt
http://www.objectis.org/about-en/objectis/index_html would have the
resources for the model I'm suggesting.

-- 

Paul Winkler
http://www.slinkp.com

From chrism at plope.com  Thu Feb  1 16:47:10 2007
From: chrism at plope.com (Chris McDonough)
Date: Thu, 1 Feb 2007 10:47:10 -0500
Subject: [z3-five] Unexpectedly unprotected code
In-Reply-To: <20070201154119.GA7431@slinkp.com>
References: <eplqmb$bs3$1@sea.gmane.org> <45BE7BBF.7060805@simplistix.co.uk>
	<a7a2b76b0701291502o2127b3e9we5ab8524b82e6ca6@mail.gmail.com>
	<45BE7DE8.7000706@weitershausen.de>
	<45BEF8A8.8040806@simplistix.co.uk>
	<F94802BB-6FD3-4A8C-8979-4A37A250BD1D@weitershausen.de>
	<epnvfk$vbi$1@sea.gmane.org> <45C06C46.8040103@simplistix.co.uk>
	<20070131181625.GC7183@slinkp.com>
	<45C0FA78.9020208@simplistix.co.uk>
	<20070201154119.GA7431@slinkp.com>
Message-ID: <B7401D86-3444-4244-B995-29B82109EF37@plope.com>

On Feb 1, 2007, at 10:41 AM, Paul Winkler wrote:

>
> On Wed, Jan 31, 2007 at 08:22:16PM +0000, Chris Withers wrote:
>> Paul Winkler wrote:
>>> Hmm, was that really a big draw to Zope?
>>
>> When Zope was growing rapidly (1999-2003 for me, others may disagree)
>> everyone who came to Zope came to it by installing it and writing TTW
>> code. That was predicated on what I'm talking about ;-)
>
> Yeah, but nowhere is it written in stone that the security model of
> TTW code needs to be different than the security model of filesystem
> code.  I'm heretically suggesting that we consider whether this
> distinction does more harm than good. For a trivial example, the old
> "import re is not allowed" FAQ.  IIRC the motivation for disallowing
> re is that regular expressions can be indeterminately slow and a naive
> scripter might create code that's an easy DOS target - or a malicious
> scripter might do so deliberately.  In that case at least, I'm now
> pretty firmly on the side of "give 'em the rope and if they hang
> themselves, it's their own problem".

FTR, I'm in the camp of "lets just have one programming model"...

http://www.plope.com/Members/chrism/ttb_programming

FWIW, I haven't yet used "grok" but from what I've seen of it, it  
seems consistent with the ideas in that blog entry.

- C


From regebro at gmail.com  Fri Feb  2 10:15:10 2007
From: regebro at gmail.com (Lennart Regebro)
Date: Fri, 2 Feb 2007 10:15:10 +0100
Subject: [z3-five] Unexpectedly unprotected code
In-Reply-To: <20070201154119.GA7431@slinkp.com>
References: <eplqmb$bs3$1@sea.gmane.org>
	<a7a2b76b0701291502o2127b3e9we5ab8524b82e6ca6@mail.gmail.com>
	<45BE7DE8.7000706@weitershausen.de>
	<45BEF8A8.8040806@simplistix.co.uk>
	<F94802BB-6FD3-4A8C-8979-4A37A250BD1D@weitershausen.de>
	<epnvfk$vbi$1@sea.gmane.org> <45C06C46.8040103@simplistix.co.uk>
	<20070131181625.GC7183@slinkp.com> <45C0FA78.9020208@simplistix.co.uk>
	<20070201154119.GA7431@slinkp.com>
Message-ID: <319e029f0702020115h2bb47990i7ec6ed38d7a4b926@mail.gmail.com>

On 2/1/07, Paul Winkler <pw_lists at slinkp.com> wrote:
> Yeah, but nowhere is it written in stone that the security model of
> TTW code needs to be different than the security model of filesystem
> code.  I'm heretically suggesting that we consider whether this
> distinction does more harm than good. For a trivial example, the old
> "import re is not allowed" FAQ.  IIRC the motivation for disallowing
> re is that regular expressions can be indeterminately slow and a naive
> scripter might create code that's an easy DOS target - or a malicious
> scripter might do so deliberately.  In that case at least, I'm now
> pretty firmly on the side of "give 'em the rope and if they hang
> themselves, it's their own problem".

For re, yes. But for other things, no. It's a case of multiple not
serious security holes possibly combining nto a serous one. If you can
do anything in a TTW, then all you need is to find one hole, that
allows to change a TTW-script, and all security of the system is
completely gone.

So that that TTW code needs a different security model than filesystem
code IS written in stone.

However, is is not written in stone that the quick hacky development
you can do in Zope2 must be TTW. ;)

-- 
Lennart Regebro: Python, Zope, CPS, Plone consulting.
+33 661 58 14 64

From daniel at infrae.com  Mon Feb  5 15:35:00 2007
From: daniel at infrae.com (Daniel Nouri)
Date: Mon, 05 Feb 2007 15:35:00 +0100
Subject: [z3-five] formlib acquisition error
Message-ID: <eq7faj$q3v$1@sea.gmane.org>

Hi!

I'm using Five 1.4.1 with Zope 2.9.6 and using the (base-) classes

   Products.Five.formlib.formbase.PageForm
   Products.Five.browser.pagetemplatefile.ViewPageTemplateFile

I'm trying to call self.template() in an action handler, but that fails
because self (the form) is not acquistion-wrapped (at all) in the body
of the action handler, thus

      @form.action("foo")
      def action_foo(self, action, data):
          return self.template()

is bound to fail.


-- 
Daniel Nouri
Infrae ? http://infrae.com


From maciej.wisniowski at coig.katowice.pl  Tue Feb  6 07:31:39 2007
From: maciej.wisniowski at coig.katowice.pl (Maciej Wisniowski)
Date: Tue, 06 Feb 2007 07:31:39 +0100
Subject: [z3-five] formlib acquisition error
In-Reply-To: <eq7faj$q3v$1@sea.gmane.org>
References: <eq7faj$q3v$1@sea.gmane.org>
Message-ID: <45C820CB.3020707@coig.katowice.pl>


> I'm using Five 1.4.1 with Zope 2.9.6 and using the (base-) classes
>
>    Products.Five.formlib.formbase.PageForm
>    Products.Five.browser.pagetemplatefile.ViewPageTemplateFile
>
> I'm trying to call self.template() in an action handler, but that fails
> because self (the form) is not acquistion-wrapped (at all) in the body
> of the action handler, thus
>
>       @form.action("foo")
>       def action_foo(self, action, data):
>           return self.template()
>
> is bound to fail.
>   
You have import of ViewPageTemplateFile. Do you
use this to create self.template object?
If so then maybe try with

from Products.Five.browser.pagetemplatefile import \
     ZopeTwoPageTemplateFile

Take a look at Chris Withers' thread about "(re)using zpt 
macros from view page templates" from 2007-01-28.
Maybe this will help somehow.

-- 
Maciej Wisniowski


From tim at sitefusion.co.uk  Tue Feb  6 11:27:12 2007
From: tim at sitefusion.co.uk (Tim Hicks)
Date: Tue, 06 Feb 2007 10:27:12 +0000
Subject: [z3-five] formlib acquisition error
In-Reply-To: <eq7faj$q3v$1@sea.gmane.org>
References: <eq7faj$q3v$1@sea.gmane.org>
Message-ID: <45C85800.3010108@sitefusion.co.uk>

Daniel Nouri wrote:
> Hi!
> 
> I'm using Five 1.4.1 with Zope 2.9.6 and using the (base-) classes
> 
>    Products.Five.formlib.formbase.PageForm
>    Products.Five.browser.pagetemplatefile.ViewPageTemplateFile
> 
> I'm trying to call self.template() in an action handler, but that fails
> because self (the form) is not acquistion-wrapped (at all) in the body
> of the action handler, thus
> 
>       @form.action("foo")
>       def action_foo(self, action, data):
>           return self.template()
> 
> is bound to fail.

I had a problem like this a few weeks ago.  When I discussed it with
alecm on #plone, he seemed to know what was causing the problem
(something to do with the form.action decorator storing objects in such
a way that acquisition got stripped).  Unfortunately, the appropriate
fix was, apparently, not obvious or easy.

The workaround that I employed was to simply declare my template as a
full-blown view and then look it up with getMultiAdapter instead of as
an attribute of the form.

Maybe alecm can remember more...


Tim

From chris at simplistix.co.uk  Wed Feb  7 17:06:28 2007
From: chris at simplistix.co.uk (Chris Withers)
Date: Wed, 07 Feb 2007 16:06:28 +0000
Subject: [z3-five] testings event firing
Message-ID: <45C9F904.4090904@simplistix.co.uk>

Hi All,

This starts of with me wanting to test that I've inserted the correct 
calls to event.notify in my code.

So, the natural way would be to have a dummy subscriber, and then 
exercise the code where event.notify should be present, right?

Okay, but if I just do:

class mySubscriber:

   def __init__(self):
     self.events = []

   def __call__(self,*args,**kw):
     self.events.append((args,kw))

s = mySubscriber()
provideHandler(s,IWhatever)

....then that handler will get called for every event in the test run, 
right?

How do I get rid of it once I'm done with a particular test?

I tried playing with subclassing zope.component.testing.PlacelessSetup 
but that trashed all my normal zcml setup.

I guess a secondary question then becomes how do you unit test that the 
assignments in your .zcml files are correct?

Is this the realm of functional tests? If so, how do you run functional 
and unit tests in the same run? Are there any docs or examples of all this?

cheers,

Chris

-- 
Simplistix - Content Management, Zope & Python Consulting
            - http://www.simplistix.co.uk

From philipp at weitershausen.de  Wed Feb  7 17:39:35 2007
From: philipp at weitershausen.de (Philipp von Weitershausen)
Date: Wed, 07 Feb 2007 17:39:35 +0100
Subject: [z3-five] testings event firing
In-Reply-To: <45C9F904.4090904@simplistix.co.uk>
References: <45C9F904.4090904@simplistix.co.uk>
Message-ID: <45CA00C7.7040700@weitershausen.de>

Chris Withers wrote:
> This starts of with me wanting to test that I've inserted the correct 
> calls to event.notify in my code.
> 
> So, the natural way would be to have a dummy subscriber, and then 
> exercise the code where event.notify should be present, right?

Sure, or simply use an already existing setup from 
zope.component.eventtesting. There are a few tests (also in Five, I 
think) that demonstrate how to use it.

> Okay, but if I just do:
> 
> class mySubscriber:
> 
>    def __init__(self):
>      self.events = []
> 
>    def __call__(self,*args,**kw):
>      self.events.append((args,kw))
> 
> s = mySubscriber()
> provideHandler(s,IWhatever)
> 
> ....then that handler will get called for every event in the test run, 
> right?

Yup.

> How do I get rid of it once I'm done with a particular test?
> 
> I tried playing with subclassing zope.component.testing.PlacelessSetup 
> but that trashed all my normal zcml setup.

Well, yeah, of course it does. PlacelessSetup is a bad name and it's 
actually nowadays just a small shim for zope.testing.cleanup.CleanUp. 
"CleanUp" says it all: it cleans up after each test. Completely. No 
compromises. That's important for unit tests.

So, the question is, are you writing a fucntional test or a unit test? 
In a unit test, you might not actually want to (or have to!) load all of 
your ZCML, but only register those components necessary for the test. 
And then it's not so much of a problem that everything gets torn down 
after each test (in fact, you'd rather like that in a unit test).

In a functional test, you load the ZCML once for a whole bunch of tests 
and then execute all your tests within that setup. Individual tests 
shouldn't really modify the registries then. The common way to do this 
is to use zope.testing's test layer support. You put those tests that 
require common setup routines that may take long to load (e.g. booting 
up all of the ZCML you have) into one layer and let the layer do that 
setup (and tear down). zope.testing has some doctests on layers.

> I guess a secondary question then becomes how do you unit test that the 
> assignments in your .zcml files are correct?

That's not a unit test, that's an integration / functional test.

> Is this the realm of functional tests? If so, how do you run functional 
> and unit tests in the same run? Are there any docs or examples of all this?

Five has unit tests and funtional tests, though they are sometimes hard 
to distinguish because

a) even some unit tests do their setup by calling ZCML, which is mostly 
overkill and slows them down

b) functional tests do their own setup and don't rely on a common setup 
layer like Zope 3 functional tests do.

For Zope 2.11 I'd like to refactor Five's functional tests to use a 
common setup layer (one that simply bootstraps Five's ZCML) and cleanup 
unit tests not to use ZCML to make them even more light-weight.


-- 
http://worldcookery.com -- Professional Zope documentation and training
Next Zope 3 training at Camp5: http://trizpug.org/boot-camp/camp5

From d.schulz81 at gmx.net  Thu Feb  8 00:19:00 2007
From: d.schulz81 at gmx.net (Dennis Schulz)
Date: Thu, 08 Feb 2007 00:19:00 +0100
Subject: [z3-five] Plone Tool without acquisition
Message-ID: <45CA5E64.4040801@gmx.net>

> - The right way to solve your problem would be using a local site 
> manager. Even if your tools are not registered as utilities, you can 
> get hold of the site root and therefor of the tool without using 
> acquisition. Plone 3.0 might give you some hints how to set this up.

I am not familiar with the Plone 3 code.
Can anyone point me to an example where  I can see how to set up a local 
site manager?

I registered

 <five:localsite
         class="Products.CMFPlone.Portal.PloneSite" />
 
What is the easiest way to get hold of the site root so I can use the 
Plone tool without acquisition?

Regards, Dennis




From philipp at weitershausen.de  Thu Feb  8 02:33:16 2007
From: philipp at weitershausen.de (Philipp von Weitershausen)
Date: Thu, 08 Feb 2007 02:33:16 +0100
Subject: [z3-five] Plone Tool without acquisition
In-Reply-To: <45CA5E64.4040801@gmx.net>
References: <45CA5E64.4040801@gmx.net>
Message-ID: <45CA7DDC.1080107@weitershausen.de>

Dennis Schulz wrote:
>> - The right way to solve your problem would be using a local site 
>> manager. Even if your tools are not registered as utilities, you can 
>> get hold of the site root and therefor of the tool without using 
>> acquisition. Plone 3.0 might give you some hints how to set this up.
> 
> I am not familiar with the Plone 3 code.
> Can anyone point me to an example where  I can see how to set up a local 
> site manager?
> 
> I registered
> 
>  <five:localsite
>          class="Products.CMFPlone.Portal.PloneSite" />
>  
> What is the easiest way to get hold of the site root so I can use the 
> Plone tool without acquisition?

Why would you want to get hold of the site root? Assuming that the CMF 
tools are registered as local utilities, you should simply be able to say:

   from zope.component import getUtility
   from Products.CMFPlone.interfaces import IPloneTool # verify this

   plone_tool = getUtility(IPloneTool)

Zope will automatically figure out that where it should get that tool 
from (because it traversed over the site).

-- 
http://worldcookery.com -- Professional Zope documentation and training
Next Zope 3 training at Camp5: http://trizpug.org/boot-camp/camp5

From chris at simplistix.co.uk  Thu Feb  8 09:35:35 2007
From: chris at simplistix.co.uk (Chris Withers)
Date: Thu, 08 Feb 2007 08:35:35 +0000
Subject: [z3-five] Plone Tool without acquisition
In-Reply-To: <45CA7DDC.1080107@weitershausen.de>
References: <45CA5E64.4040801@gmx.net> <45CA7DDC.1080107@weitershausen.de>
Message-ID: <45CAE0D7.1000800@simplistix.co.uk>

Philipp von Weitershausen wrote:

> Why would you want to get hold of the site root? Assuming that the CMF 
> tools are registered as local utilities, you should simply be able to say:
> 
>    from zope.component import getUtility
>    from Products.CMFPlone.interfaces import IPloneTool # verify this
> 
>    plone_tool = getUtility(IPloneTool)
> 
> Zope will automatically figure out that where it should get that tool 
> from (because it traversed over the site).

Really? That feels a bit DWIM :-(

I thought you had to pass in a context to get local utilities?

cheers,

Chris

-- 
Simplistix - Content Management, Zope & Python Consulting
            - http://www.simplistix.co.uk


From chris at simplistix.co.uk  Thu Feb  8 09:33:37 2007
From: chris at simplistix.co.uk (Chris Withers)
Date: Thu, 08 Feb 2007 08:33:37 +0000
Subject: [z3-five] testings event firing
In-Reply-To: <45CA00C7.7040700@weitershausen.de>
References: <45C9F904.4090904@simplistix.co.uk>
	<45CA00C7.7040700@weitershausen.de>
Message-ID: <45CAE061.7080008@simplistix.co.uk>

Philipp von Weitershausen wrote:
> Sure, or simply use an already existing setup from 
> zope.component.eventtesting.

...which is only available in 2.10 ;-)

> In a functional test, you load the ZCML once for a whole bunch of tests 
> and then execute all your tests within that setup. 

What's the canonical way to load all the zcml that Zope normally does?

> shouldn't really modify the registries then. The common way to do this 
> is to use zope.testing's test layer support. You put those tests that 
> require common setup routines that may take long to load (e.g. booting 
> up all of the ZCML you have) into one layer and let the layer do that 
> setup (and tear down). zope.testing has some doctests on layers.

Yup, I've read all those I could find, but I'm still left wondering how 
to fire up all the ZCML ;-)

cheers,

Chris

-- 
Simplistix - Content Management, Zope & Python Consulting
            - http://www.simplistix.co.uk


From daniel at infrae.com  Thu Feb  8 10:50:22 2007
From: daniel at infrae.com (Daniel Nouri)
Date: Thu, 08 Feb 2007 10:50:22 +0100
Subject: [z3-five] testings event firing
In-Reply-To: <45CAE061.7080008@simplistix.co.uk>
References: <45C9F904.4090904@simplistix.co.uk>	<45CA00C7.7040700@weitershausen.de>
	<45CAE061.7080008@simplistix.co.uk>
Message-ID: <45CAF25E.3070309@infrae.com>

Chris Withers wrote:
> Philipp von Weitershausen wrote:
>> Sure, or simply use an already existing setup from 
>> zope.component.eventtesting.
> 
> ...which is only available in 2.10 ;-)
> 
>> In a functional test, you load the ZCML once for a whole bunch of tests 
>> and then execute all your tests within that setup. 
> 
> What's the canonical way to load all the zcml that Zope normally does?
> 
>> shouldn't really modify the registries then. The common way to do this 
>> is to use zope.testing's test layer support. You put those tests that 
>> require common setup routines that may take long to load (e.g. booting 
>> up all of the ZCML you have) into one layer and let the layer do that 
>> setup (and tear down). zope.testing has some doctests on layers.
> 
> Yup, I've read all those I could find, but I'm still left wondering how 
> to fire up all the ZCML ;-)

I know you wouldn't look at Plone for an example :p so here's an example
from Silva:
https://infrae.com/svn/Silva/branch/zope-2.10/tests/layer.py

Look at SilvaZCMLLayer, which does the setup for all of Silva's
integration tests.  It has ``zcml.load_site()`` and it also registers a
subscriber for anything (for which you can use your own list just as well).


-- 
Daniel Nouri
Infrae ? http://infrae.com


From philipp at weitershausen.de  Thu Feb  8 11:02:11 2007
From: philipp at weitershausen.de (Philipp von Weitershausen)
Date: Thu, 8 Feb 2007 11:02:11 +0100
Subject: [z3-five] Plone Tool without acquisition
In-Reply-To: <45CAE0D7.1000800@simplistix.co.uk>
References: <45CA5E64.4040801@gmx.net> <45CA7DDC.1080107@weitershausen.de>
	<45CAE0D7.1000800@simplistix.co.uk>
Message-ID: <FC82A0AD-D35B-4F7C-902A-9A628832A05F@weitershausen.de>

On 8 Feb 2007, at 09:35 , Chris Withers wrote:
> Philipp von Weitershausen wrote:
>
>> Why would you want to get hold of the site root? Assuming that the  
>> CMF tools are registered as local utilities, you should simply be  
>> able to say:
>>    from zope.component import getUtility
>>    from Products.CMFPlone.interfaces import IPloneTool # verify this
>>    plone_tool = getUtility(IPloneTool)
>> Zope will automatically figure out that where it should get that  
>> tool from (because it traversed over the site).
>
> Really? That feels a bit DWIM :-(

Yes.

> I thought you had to pass in a context to get local utilities?

No. That's the point, the context is found during traversal and other  
code doesn't have to be aware of the context. In fact, why would code  
*care* where the utility is from? All it wants is a utility, be it  
local or global.




From y.2007- at wcm-solutions.de  Thu Feb  8 11:06:18 2007
From: y.2007- at wcm-solutions.de (yuppie)
Date: Thu, 08 Feb 2007 11:06:18 +0100
Subject: [z3-five] Plone Tool without acquisition
In-Reply-To: <45CA7DDC.1080107@weitershausen.de>
References: <45CA5E64.4040801@gmx.net> <45CA7DDC.1080107@weitershausen.de>
Message-ID: <45CAF61A.20000@wcm-solutions.de>

Hi Philipp!


Philipp von Weitershausen wrote:
> Dennis Schulz wrote:
>>> - The right way to solve your problem would be using a local site 
>>> manager. Even if your tools are not registered as utilities, you can 
>>> get hold of the site root and therefor of the tool without using 
>>> acquisition. Plone 3.0 might give you some hints how to set this up.
>> I am not familiar with the Plone 3 code.
>> Can anyone point me to an example where  I can see how to set up a local 
>> site manager?
>>
>> I registered
>>
>>  <five:localsite
>>          class="Products.CMFPlone.Portal.PloneSite" />
>>  
>> What is the easiest way to get hold of the site root so I can use the 
>> Plone tool without acquisition?
> 
> Why would you want to get hold of the site root? Assuming that the CMF 
> tools are registered as local utilities, you should simply be able to say:
> 
>    from zope.component import getUtility
>    from Products.CMFPlone.interfaces import IPloneTool # verify this
> 
>    plone_tool = getUtility(IPloneTool)
> 
> Zope will automatically figure out that where it should get that tool 
> from (because it traversed over the site).

You missed the start of this thread. We were not talking about a clean 
solution that requires properly registered and acquisition wrapped tools.

We were talking about a hack for using tools in vocabularies without 
acquisition context. Right now, in Zope 2.9, without the not yet 
existing five.localsitemanager.

In Zope 2.10 I can use zope.app.component.hooks.getSite() to get the 
acquisition wrapped site. Don't know what needs to be set up in Zope 2.9 
to have this available.


Cheers,

	Yuppie


From philipp at weitershausen.de  Thu Feb  8 12:00:54 2007
From: philipp at weitershausen.de (Philipp von Weitershausen)
Date: Thu, 8 Feb 2007 12:00:54 +0100
Subject: [z3-five] Plone Tool without acquisition
In-Reply-To: <45CAF61A.20000@wcm-solutions.de>
References: <45CA5E64.4040801@gmx.net> <45CA7DDC.1080107@weitershausen.de>
	<45CAF61A.20000@wcm-solutions.de>
Message-ID: <1EF3E0C7-3070-4C79-88D3-1EB118AF2690@weitershausen.de>

On 8 Feb 2007, at 11:06 , yuppie wrote:
> Philipp von Weitershausen wrote:
>> Dennis Schulz wrote:
>>>> - The right way to solve your problem would be using a local  
>>>> site manager. Even if your tools are not registered as  
>>>> utilities, you can get hold of the site root and therefor of the  
>>>> tool without using acquisition. Plone 3.0 might give you some  
>>>> hints how to set this up.
>>> I am not familiar with the Plone 3 code.
>>> Can anyone point me to an example where  I can see how to set up  
>>> a local site manager?
>>>
>>> I registered
>>>
>>>  <five:localsite
>>>          class="Products.CMFPlone.Portal.PloneSite" />
>>>  What is the easiest way to get hold of the site root so I can  
>>> use the Plone tool without acquisition?
>> Why would you want to get hold of the site root? Assuming that the  
>> CMF tools are registered as local utilities, you should simply be  
>> able to say:
>>    from zope.component import getUtility
>>    from Products.CMFPlone.interfaces import IPloneTool # verify this
>>    plone_tool = getUtility(IPloneTool)
>> Zope will automatically figure out that where it should get that  
>> tool from (because it traversed over the site).
>
> You missed the start of this thread.

Possibly. As far as I'm concerned, the thread started with Dennis's  
email, though, didn't it?

> We were not talking about a clean solution that requires properly  
> registered and acquisition wrapped tools.
>
> We were talking about a hack for using tools in vocabularies  
> without acquisition context. Right now, in Zope 2.9, without the  
> not yet existing five.localsitemanager.

Gotcha.

> In Zope 2.10 I can use zope.app.component.hooks.getSite() to get  
> the acquisition wrapped site. Don't know what needs to be set up in  
> Zope 2.9 to have this available.

Well, the canonical way in both Zope 2.10 and 2.9 is to make your  
plone site an ISite. However, in  Zope 2.10 that's a lot cleaner and  
nicer (though I still want to make it even nicer for Zope 2.11).

In Five 1.3/1.4 on Zope 2.9, this can be done using the  
<five:localsite class="...PloneSite" /> directive. This will add the  
IPossibleSite interface to the PloneSite class. That means plone  
sites *can* be turned into ISites, you just need to perform the  
actual operation on each individual object. To do that, you need to  
call Products.Five.site.localsite.enableLocalSiteHook(the_plone_site)  
on the plone site object. This will install a traversal hook so the  
site is remembered during traversal.

As said, don't use this code past Zope 2.9.


From y.2007- at wcm-solutions.de  Thu Feb  8 12:10:05 2007
From: y.2007- at wcm-solutions.de (yuppie)
Date: Thu, 08 Feb 2007 12:10:05 +0100
Subject: [z3-five] Plone Tool without acquisition
In-Reply-To: <1EF3E0C7-3070-4C79-88D3-1EB118AF2690@weitershausen.de>
References: <45CA5E64.4040801@gmx.net>
	<45CA7DDC.1080107@weitershausen.de>	<45CAF61A.20000@wcm-solutions.de>
	<1EF3E0C7-3070-4C79-88D3-1EB118AF2690@weitershausen.de>
Message-ID: <45CB050D.80001@wcm-solutions.de>

Philipp von Weitershausen wrote:
> On 8 Feb 2007, at 11:06 , yuppie wrote:
>> You missed the start of this thread.
> 
> Possibly. As far as I'm concerned, the thread started with Dennis's  
> email, though, didn't it?

Oops! I didn't notice Dennis started a new thread. But he did follow up 
to an older thread where he explained what he wants to do.

Cheers, Yuppie


From d.schulz81 at gmx.net  Thu Feb  8 20:24:08 2007
From: d.schulz81 at gmx.net (Dennis Schulz)
Date: Thu, 08 Feb 2007 20:24:08 +0100
Subject: [z3-five] Plone Tool without acquisition
In-Reply-To: <45CB050D.80001@wcm-solutions.de>
References: <45CA5E64.4040801@gmx.net>	<45CA7DDC.1080107@weitershausen.de>	<45CAF61A.20000@wcm-solutions.de>	<1EF3E0C7-3070-4C79-88D3-1EB118AF2690@weitershausen.de>
	<45CB050D.80001@wcm-solutions.de>
Message-ID: <45CB78D8.8050200@gmx.net>

Hi,

sorry for the confusion by the new thread. I think for my current 
project in Zope 2.9.6 it does not make sense to use local utilities.
I have to find workarounds for the moment and wait for plone 3 to work 
like that.

I just would have been nice......

Dennis


yuppie escribi?:
> Philipp von Weitershausen wrote:
>   
>> On 8 Feb 2007, at 11:06 , yuppie wrote:
>>     
>>> You missed the start of this thread.
>>>       
>> Possibly. As far as I'm concerned, the thread started with Dennis's  
>> email, though, didn't it?
>>     
>
> Oops! I didn't notice Dennis started a new thread. But he did follow up 
> to an older thread where he explained what he wants to do.
>
> Cheers, Yuppie
>
> _______________________________________________
> z3-five mailing list
> z3-five at codespeak.net
> http://codespeak.net/mailman/listinfo/z3-five
>
>
>   


From d.schulz81 at gmx.net  Thu Feb  8 21:47:04 2007
From: d.schulz81 at gmx.net (Dennis Schulz)
Date: Thu, 08 Feb 2007 21:47:04 +0100
Subject: [z3-five] Security in Content Classes
Message-ID: <45CB8C48.7090701@gmx.net>

Here a short question about security,

In Zope 2 style product classes I have to protect the methods against 
being called.
With Zope 3 views the whole view is protected by a permission in zcml 
and this is not necessary?
But the content object still should be protected (is derived from 
SimpleItem)


Cheers Dennis

From lists at zopyx.com  Fri Feb  9 13:00:09 2007
From: lists at zopyx.com (Andreas Jung)
Date: Fri, 09 Feb 2007 13:00:09 +0100
Subject: [z3-five] Marker interfaces in Five 1.0.2
Message-ID: <45D61439174DCEDEBA4C3543@vg180585.grp.haufemg.com>

Hi,

don't stone me to death but is there a way to attach marker interfaces
of objects in Five 1.0.2?

Andreas
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 186 bytes
Desc: not available
Url : http://codespeak.net/pipermail/z3-five/attachments/20070209/544a9797/attachment.pgp 

From chris at simplistix.co.uk  Fri Feb  9 14:26:43 2007
From: chris at simplistix.co.uk (Chris Withers)
Date: Fri, 09 Feb 2007 13:26:43 +0000
Subject: [z3-five] Plone Tool without acquisition
In-Reply-To: <FC82A0AD-D35B-4F7C-902A-9A628832A05F@weitershausen.de>
References: <45CA5E64.4040801@gmx.net> <45CA7DDC.1080107@weitershausen.de>
	<45CAE0D7.1000800@simplistix.co.uk>
	<FC82A0AD-D35B-4F7C-902A-9A628832A05F@weitershausen.de>
Message-ID: <45CC7693.6040502@simplistix.co.uk>

Philipp von Weitershausen wrote:
>>> Why would you want to get hold of the site root? Assuming that the 
>>> CMF tools are registered as local utilities, you should simply be 
>>> able to say:
>>>    from zope.component import getUtility
>>>    from Products.CMFPlone.interfaces import IPloneTool # verify this
>>>    plone_tool = getUtility(IPloneTool)
>>> Zope will automatically figure out that where it should get that tool 
>>> from (because it traversed over the site).
>>
>> Really? That feels a bit DWIM :-(
> 
> Yes.

Do you see that as a good or a bad thing?

>> I thought you had to pass in a context to get local utilities?
> 
> No. That's the point, the context is found during traversal and other 
> code doesn't have to be aware of the context. In fact, why would code 
> *care* where the utility is from? All it wants is a utility, be it local 
> or global.

Well, I'd imagine if you went to the trouble of registering local 
utilities, you did so for a reason?

If that's the case, then how do I make sure the right utilities are used 
when I'm traversing around several objects in different locations in the 
same piece of code?

I guess I'm interested in what couldn't as "traversed over" and how I 
can, if I need to, override where "Zope" thinks I've been traversing...

Chris

-- 
Simplistix - Content Management, Zope & Python Consulting
            - http://www.simplistix.co.uk

From philipp at weitershausen.de  Fri Feb  9 14:36:39 2007
From: philipp at weitershausen.de (Philipp von Weitershausen)
Date: Fri, 9 Feb 2007 14:36:39 +0100
Subject: [z3-five] Plone Tool without acquisition
In-Reply-To: <45CC7693.6040502@simplistix.co.uk>
References: <45CA5E64.4040801@gmx.net> <45CA7DDC.1080107@weitershausen.de>
	<45CAE0D7.1000800@simplistix.co.uk>
	<FC82A0AD-D35B-4F7C-902A-9A628832A05F@weitershausen.de>
	<45CC7693.6040502@simplistix.co.uk>
Message-ID: <E001152F-A8D3-4DB9-8444-66A2682A04C3@weitershausen.de>

On 9 Feb 2007, at 14:26 , Chris Withers wrote:
> Philipp von Weitershausen wrote:
>>>> Why would you want to get hold of the site root? Assuming that  
>>>> the CMF tools are registered as local utilities, you should  
>>>> simply be able to say:
>>>>    from zope.component import getUtility
>>>>    from Products.CMFPlone.interfaces import IPloneTool # verify  
>>>> this
>>>>    plone_tool = getUtility(IPloneTool)
>>>> Zope will automatically figure out that where it should get that  
>>>> tool from (because it traversed over the site).
>>>
>>> Really? That feels a bit DWIM :-(
>> Yes.
>
> Do you see that as a good or a bad thing?

The yes was for the "really". If you don't believe me, read the  
chapter on Sites in my book. It's all explained in detail there.

IMO not having to specify a context is a good thing. Sometimes code  
doesnt' have access to a context, but it still wants to lookup a  
utility. Should it not be possible to let that utility come from a  
local component registry? I as the one who makes up application  
policy should be able to say where that utility comes from.

>>> I thought you had to pass in a context to get local utilities?
>> No. That's the point, the context is found during traversal and  
>> other code doesn't have to be aware of the context. In fact, why  
>> would code *care* where the utility is from? All it wants is a  
>> utility, be it local or global.
>
> Well, I'd imagine if you went to the trouble of registering local  
> utilities, you did so for a reason?

You're missing the point. The code that's looking up the utility  
might be some completely different piece of software than the piece  
of code that provides the utility. Where that utility is registered  
is an implementation detail of the utility in the end. The code  
looking up the utility just wants to get an ISomething, no matter  
where it's from.

> If that's the case, then how do I make sure the right utilities are  
> used when I'm traversing around several objects in different  
> locations in the same piece of code?

There's only one kind of traversal -- URL traversal -- and it only  
happens once per request. The other thing sometimes referred to as  
"traversal" is resolving object graphs. That doesnt' trigger the site  
remembering code.

> I guess I'm interested in what couldn't as "traversed over" and how  
> I can, if I need to, override where "Zope" thinks I've been  
> traversing...

I don't think there's anything to worry about.

From d.w.morriss at gmail.com  Fri Feb  9 16:44:18 2007
From: d.w.morriss at gmail.com (whit)
Date: Fri, 09 Feb 2007 09:44:18 -0600
Subject: [z3-five] Marker interfaces in Five 1.0.2
In-Reply-To: <45D61439174DCEDEBA4C3543@vg180585.grp.haufemg.com>
References: <45D61439174DCEDEBA4C3543@vg180585.grp.haufemg.com>
Message-ID: <45CC96D2.8090005@gmail.com>

Andreas Jung wrote:
> Hi,
> 
> don't stone me to death but is there a way to attach marker interfaces
> of objects in Five 1.0.2?
> 
> Andreas
> 

does this not work?

from zope.interface import directlyProvidedBy
from zope.interface import directlyProvides

def mark(ob, interface):
     directlyProvides(ob, directlyProvidedBy(ob), interface)

-w

-- 

------ d. whit morriss ------
- senior engineer, opencore -
- http://www.openplans.org  -
- m: 415-710-8975           -

"If you don't know where you are,
  you don't know anything at all" 


Dr. Edgar Spencer, Ph.D., 1995


From philipp at weitershausen.de  Fri Feb  9 21:42:54 2007
From: philipp at weitershausen.de (Philipp von Weitershausen)
Date: Fri, 09 Feb 2007 21:42:54 +0100
Subject: [z3-five] Marker interfaces in Five 1.0.2
In-Reply-To: <45CC96D2.8090005@gmail.com>
References: <45D61439174DCEDEBA4C3543@vg180585.grp.haufemg.com>
	<45CC96D2.8090005@gmail.com>
Message-ID: <45CCDCCE.4050708@weitershausen.de>

whit wrote:
> Andreas Jung wrote:
>> Hi,
>>
>> don't stone me to death but is there a way to attach marker interfaces
>> of objects in Five 1.0.2?
>>
>> Andreas
>>
> 
> does this not work?
> 
> from zope.interface import directlyProvidedBy
> from zope.interface import directlyProvides
> 
> def mark(ob, interface):
>      directlyProvides(ob, directlyProvidedBy(ob), interface)

Actually, you probably might want to stick the marker interfaces to the 
beginning:

   def mark(ob, *interfaces):
       directlyProvides(ob, *interfaces, directlyProvidedBy(ob))

Since Zope 3.2 this function is part of the zope.interface package and 
called "alsoProvides".


-- 
http://worldcookery.com -- Professional Zope documentation and training
Next Zope 3 training at Camp5: http://trizpug.org/boot-camp/camp5

From d.w.morriss at gmail.com  Fri Feb  9 23:31:56 2007
From: d.w.morriss at gmail.com (whit)
Date: Fri, 09 Feb 2007 16:31:56 -0600
Subject: [z3-five] Marker interfaces in Five 1.0.2
In-Reply-To: <45CCDCCE.4050708@weitershausen.de>
References: <45D61439174DCEDEBA4C3543@vg180585.grp.haufemg.com>	<45CC96D2.8090005@gmail.com>
	<45CCDCCE.4050708@weitershausen.de>
Message-ID: <45CCF65C.4020404@gmail.com>

Philipp von Weitershausen wrote:
> whit wrote:
>> Andreas Jung wrote:
>>> Hi,
>>>
>>> don't stone me to death but is there a way to attach marker interfaces
>>> of objects in Five 1.0.2?
>>>
>>> Andreas
>>>
>> does this not work?
>>
>> from zope.interface import directlyProvidedBy
>> from zope.interface import directlyProvides
>>
>> def mark(ob, interface):
>>      directlyProvides(ob, directlyProvidedBy(ob), interface)
> 
> Actually, you probably might want to stick the marker interfaces to the 
> beginning:
> 
>    def mark(ob, *interfaces):
>        directlyProvides(ob, *interfaces, directlyProvidedBy(ob))
> 
> Since Zope 3.2 this function is part of the zope.interface package and 
> called "alsoProvides".
> 
> 
yup... otherwise the other interfaces will resolve first.

-w

-- 

------ d. whit morriss ------
- senior engineer, opencore -
- http://www.openplans.org  -
- m: 415-710-8975           -

"If you don't know where you are,
  you don't know anything at all" 


Dr. Edgar Spencer, Ph.D., 1995


From chris at simplistix.co.uk  Mon Feb 12 18:21:38 2007
From: chris at simplistix.co.uk (Chris Withers)
Date: Mon, 12 Feb 2007 17:21:38 +0000
Subject: [z3-five] context-based utility lookup
In-Reply-To: <E001152F-A8D3-4DB9-8444-66A2682A04C3@weitershausen.de>
References: <45CA5E64.4040801@gmx.net> <45CA7DDC.1080107@weitershausen.de>
	<45CAE0D7.1000800@simplistix.co.uk>
	<FC82A0AD-D35B-4F7C-902A-9A628832A05F@weitershausen.de>
	<45CC7693.6040502@simplistix.co.uk>
	<E001152F-A8D3-4DB9-8444-66A2682A04C3@weitershausen.de>
Message-ID: <45D0A222.8030104@simplistix.co.uk>

Philipp von Weitershausen wrote:
>> Do you see that as a good or a bad thing?
> 
> The yes was for the "really". If you don't believe me, read the chapter 
> on Sites in my book. It's all explained in detail there.

Ah, okay, haven't got to that section in the 2nd Ed yet ;-)

> IMO not having to specify a context is a good thing. Sometimes code 
> doesnt' have access to a context, but it still wants to lookup a 
> utility. Should it not be possible to let that utility come from a local 
> component registry? I as the one who makes up application policy should 
> be able to say where that utility comes from.

Indeed, but can you override the context that gets used to figure out 
what utility gets picked up?

>> Well, I'd imagine if you went to the trouble of registering local 
>> utilities, you did so for a reason?
> 
> You're missing the point. The code that's looking up the utility might 
> be some completely different piece of software than the piece of code 
> that provides the utility. 

No, I understand that.

>> If that's the case, then how do I make sure the right utilities are 
>> used when I'm traversing around several objects in different locations 
>> in the same piece of code?
> 
> There's only one kind of traversal -- URL traversal -- and it only 
> happens once per request.

Ah, okay. I suspect I'm going to learn lots more about this when I come 
to refactoring the publications stuff...

> The other thing sometimes referred to as 
> "traversal" is resolving object graphs. That doesnt' trigger the site 
> remembering code.

Good ;-)

cheers,

Chris

-- 
Simplistix - Content Management, Zope & Python Consulting
            - http://www.simplistix.co.uk

From philipp at weitershausen.de  Tue Feb 13 00:59:03 2007
From: philipp at weitershausen.de (Philipp von Weitershausen)
Date: Tue, 13 Feb 2007 00:59:03 +0100
Subject: [z3-five] context-based utility lookup
In-Reply-To: <45D0A222.8030104@simplistix.co.uk>
References: <45CA5E64.4040801@gmx.net> <45CA7DDC.1080107@weitershausen.de>
	<45CAE0D7.1000800@simplistix.co.uk>
	<FC82A0AD-D35B-4F7C-902A-9A628832A05F@weitershausen.de>
	<45CC7693.6040502@simplistix.co.uk>
	<E001152F-A8D3-4DB9-8444-66A2682A04C3@weitershausen.de>
	<45D0A222.8030104@simplistix.co.uk>
Message-ID: <7889AC38-27CD-4AFE-918F-A5644CD77073@weitershausen.de>

On 12 Feb 2007, at 18:21 , Chris Withers wrote:
>> IMO not having to specify a context is a good thing. Sometimes  
>> code doesnt' have access to a context, but it still wants to  
>> lookup a utility. Should it not be possible to let that utility  
>> come from a local component registry? I as the one who makes up  
>> application policy should be able to say where that utility comes  
>> from.
>
> Indeed, but can you override the context that gets used to figure  
> out what utility gets picked up?

Yes, you can always pass in a context explicitly in which case the  
closest site to that context will be used for the lookup.

>>> If that's the case, then how do I make sure the right utilities  
>>> are used when I'm traversing around several objects in different  
>>> locations in the same piece of code?
>> There's only one kind of traversal -- URL traversal -- and it only  
>> happens once per request.
>
> Ah, okay. I suspect I'm going to learn lots more about this when I  
> come to refactoring the publications stuff...

Actually, the publication is quite agnostic of sites, fortunately.  
All the publication does is send out an IBeforeTraverseEvent for each  
object it is about to traverse. zope.app.component has a subscriber  
for that event on ISites that "remembers" the site in a thread local.

In Zope 2, it works similarly, except that there's no publication  
that sends this event. Currently, we install a BeforeTraversalHook  
(something that the ZPublisher allows us to do) on an object-per- 
object basis; this hook sends the event (and then the same subscriber  
as in Zope 3 picks it up). This I'd actually like to change for Zope  
2.11 so that we get before traverse events for all objects and don't  
have to install hooks on persistent objects anymore.




From chris at simplistix.co.uk  Tue Feb 13 09:14:49 2007
From: chris at simplistix.co.uk (Chris Withers)
Date: Tue, 13 Feb 2007 08:14:49 +0000
Subject: [z3-five] context-based utility lookup
In-Reply-To: <7889AC38-27CD-4AFE-918F-A5644CD77073@weitershausen.de>
References: <45CA5E64.4040801@gmx.net> <45CA7DDC.1080107@weitershausen.de>
	<45CAE0D7.1000800@simplistix.co.uk>
	<FC82A0AD-D35B-4F7C-902A-9A628832A05F@weitershausen.de>
	<45CC7693.6040502@simplistix.co.uk>
	<E001152F-A8D3-4DB9-8444-66A2682A04C3@weitershausen.de>
	<45D0A222.8030104@simplistix.co.uk>
	<7889AC38-27CD-4AFE-918F-A5644CD77073@weitershausen.de>
Message-ID: <45D17379.5020701@simplistix.co.uk>

Philipp von Weitershausen wrote:
>> Indeed, but can you override the context that gets used to figure out 
>> what utility gets picked up?
> 
> Yes, you can always pass in a context explicitly in which case the 
> closest site to that context will be used for the lookup.

Good stuff :-)

Out of curiosity, is it acquisition or __parent__ that's used to find 
the "closest" site?

> Actually, the publication is quite agnostic of sites, fortunately. All 
> the publication does is send out an IBeforeTraverseEvent for each object 
> it is about to traverse. zope.app.component has a subscriber for that 
> event on ISites that "remembers" the site in a thread local.

Hmm, that brings me to another minor bug bear: I hadto step through the 
event process a week or so ago in pdb and was pretty horrified as to how 
many function calls were involved from the time event.notify() got 
called to when my event subscriber code was actually called. Not only 
does it make pdb'ing a pain, but I also wonder how much of a performance 
hurt that's going to be in a high volume application with a lot of 
events being fired?

cheers,

Chris

-- 
Simplistix - Content Management, Zope & Python Consulting
            - http://www.simplistix.co.uk

From philipp at weitershausen.de  Tue Feb 13 09:28:21 2007
From: philipp at weitershausen.de (Philipp von Weitershausen)
Date: Tue, 13 Feb 2007 09:28:21 +0100
Subject: [z3-five] context-based utility lookup
In-Reply-To: <45D17379.5020701@simplistix.co.uk>
References: <45CA5E64.4040801@gmx.net> <45CA7DDC.1080107@weitershausen.de>
	<45CAE0D7.1000800@simplistix.co.uk>
	<FC82A0AD-D35B-4F7C-902A-9A628832A05F@weitershausen.de>
	<45CC7693.6040502@simplistix.co.uk>
	<E001152F-A8D3-4DB9-8444-66A2682A04C3@weitershausen.de>
	<45D0A222.8030104@simplistix.co.uk>
	<7889AC38-27CD-4AFE-918F-A5644CD77073@weitershausen.de>
	<45D17379.5020701@simplistix.co.uk>
Message-ID: <CB73EA23-CF0C-4A1F-8B10-90692033DB96@weitershausen.de>

On 13 Feb 2007, at 09:14 , Chris Withers wrote:
> Philipp von Weitershausen wrote:
>>> Indeed, but can you override the context that gets used to figure  
>>> out what utility gets picked up?
>> Yes, you can always pass in a context explicitly in which case the  
>> closest site to that context will be used for the lookup.
>
> Good stuff :-)
>
> Out of curiosity, is it acquisition or __parent__ that's used to  
> find the "closest" site?

Lacking acquisition, Zope 3 uses __parent__. In Zope 2 we use  
aq_parent if there's no __parent__.

>> Actually, the publication is quite agnostic of sites, fortunately.  
>> All the publication does is send out an IBeforeTraverseEvent for  
>> each object it is about to traverse. zope.app.component has a  
>> subscriber for that event on ISites that "remembers" the site in a  
>> thread local.
>
> Hmm, that brings me to another minor bug bear: I hadto step through  
> the event process a week or so ago in pdb and was pretty horrified  
> as to how many function calls were involved from the time  
> event.notify() got called to when my event subscriber code was  
> actually called.

For a standard event it's 2 functions before your subscriber gets  
called:

   1. zope.event.notify
   2. zope.component.event.dispatch

For an object event there's a third:

   3. zope.component.event.objectEventNotify

I guess it depends on one's point of view if that's "many" or not.

> Not only does it make pdb'ing a pain, but I also wonder how much of  
> a performance hurt that's going to be in a high volume application  
> with a lot of events being fired?

Well, if you wonder, I suggest some profiling. I'd be surprised if  
this were actually a big performance problem, but there's only one  
way to find that out for sure.

As for pdb'ing, I cannot exactly recommend stepping through  
zope.component lookup calls. I've found it more effective to query  
the component registries for their registrations (getSiteManager 
().registeredAdapters(), registeredUtilities(), etc.). This typically  
reveals the problem much quicker, without getting lost in the  
(sometimes optimized and therefore inscrutable) CA code. The only  
thing I'm missing there is a set of filters that one can apply when  
querying the registries, though those shouldn't be too hard to write.

From maciej.wisniowski at coig.katowice.pl  Thu Feb 15 23:57:43 2007
From: maciej.wisniowski at coig.katowice.pl (Maciej Wisniowski)
Date: Thu, 15 Feb 2007 23:57:43 +0100
Subject: [z3-five] Formlib custom template - AttributeError: debug.
Message-ID: <45D4E567.7030005@coig.katowice.pl>

Hi

I want to create custom template for Formlib based form.
My code is at the bottom of this post. Zope 2.9.6, Five 1.4.2
In general I tried to define template with:

template = namedtemplate.NamedTemplate('mytemp')
template = ViewPageTemplateFile('mytemp.pt')

In both cases, when I try to visit @@mypage.html I get:

Traceback (innermost last):
  Module ZPublisher.Publish, line 115, in publish
  Module ZPublisher.mapply, line 88, in mapply
  Module ZPublisher.Publish, line 41, in call_object
  Module zope.formlib.form, line 758, in __call__
  Module zope.formlib.form, line 752, in render
  Module zope.app.pagetemplate.viewpagetemplatefile, line 83, in __call__
  Module zope.app.pagetemplate.viewpagetemplatefile, line 47, in __call__
  Module ZPublisher.HTTPRequest, line 1219, in __getattr__
AttributeError: debug

mytemp.pt is basically a copy of pageform.pt from formlib. Without
custom template assignment everything works.

What is wrong? Code is run at Plone but it's plain five with
formlib here. Possibly lack of acquisition wrapper?


File: browser/views.py:
-----------------------
from zope.app.i18n import ZopeMessageFactory as _
from zope.formlib import form
from Products.Five.formlib import formbase
from Products.myproduct import interfaces
from zope.app.pagetemplate import ViewPageTemplateFile
from zope.formlib import namedtemplate
from zope.formlib import interfaces as form_interfaces

class StartForm(formbase.EditForm):
    form_fields = form.FormFields(interfaces.IInter).select('blablabla')

    template = namedtemplate.NamedTemplate('mytemp')  # option 1
    #template = ViewPageTemplateFile('mytemp.pt')     # option 2

    def __init__(self, *args, **kwargs):
        formbase.EditForm.__init__(self, *args, **kwargs)

mytemp_page_template = namedtemplate.NamedTemplateImplementation(
    ViewPageTemplateFile('mytemp.pt'), form_interfaces.IPageForm)


File: browser/configure.zcml:
----------------------------

<configure
    xmlns="http://namespaces.zope.org/zope"
    xmlns:browser="http://namespaces.zope.org/browser"
    xmlns:five="http://namespaces.zope.org/five">

  <adapter factory=".views.mytemp_page_template" name="mytemp" />

  <browser:page
      for="..interfaces.IInter"
      name="mypage.html"
      class=".views.StartForm"
      permission="zope2.View"
      />
</configure>


-- 
Maciej Wisniowski

From tim at sitefusion.co.uk  Fri Feb 16 00:35:21 2007
From: tim at sitefusion.co.uk (Tim Hicks)
Date: Thu, 15 Feb 2007 23:35:21 +0000
Subject: [z3-five] Formlib custom template - AttributeError: debug.
In-Reply-To: <45D4E567.7030005@coig.katowice.pl>
References: <45D4E567.7030005@coig.katowice.pl>
Message-ID: <45D4EE39.6090307@sitefusion.co.uk>

Maciej Wisniowski wrote:
> Hi
> 
> I want to create custom template for Formlib based form.
> My code is at the bottom of this post. Zope 2.9.6, Five 1.4.2
> In general I tried to define template with:
> 
> template = namedtemplate.NamedTemplate('mytemp')
> template = ViewPageTemplateFile('mytemp.pt')

I managed to get things working by using the following:

from Products.Five.browser import pagetemplatefile
from plone.app.form._named import named_template_adapter
_template = pagetemplatefile.ViewPageTemplateFile('pageform.pt')
default_named_template_adapter = named_template_adapter(_template)


Not sure if it's the "right" way of doing it, but it works for me.


hth,


Tim

From optilude at gmx.net  Fri Feb 16 02:28:24 2007
From: optilude at gmx.net (Martin Aspeli)
Date: Fri, 16 Feb 2007 01:28:24 +0000
Subject: [z3-five] Is browser:addMenuItem and addingInfo() supposed to work?
Message-ID: <er31ec$1nd$1@sea.gmane.org>

Hi,

I've seen reports of people making this work, but I'm struggling. I've 
got an add view (formlib AddForm) and an add menu item like so:

     <browser:addMenuItem
         title="Foo"
         factory="my.package.Foo"
         view="my.package.Foo"
         permission="cmf.AddPortalContent"
         />

I then want to list all addable views (in fact, I'm trying to make Plone 
respect add-menus and ultimately add forms, at least in the case where 
an add menu item can be linked to an old-fashioned CMF).

However, I get ComponentLookupErrors, because adding.py in Five does this:

     def addingInfo(self):
         """Return menu data.

         This is sorted by title.
         """
         container = self.context
         result = []
         for menu_id in (self.menu_id, 'zope.app.container.add'):
             if not menu_id:
                 continue
             for item in getMenu(menu_id, self, self.request):
                 extra = item.get('extra')
                 if extra:
                     factory = extra.get('factory')
                     if factory:
                         factory = zope.component.getUtility(IFactory, 
factory)
                         if not checkFactory(container, None, factory):
                             continue
                         elif item['extra']['factory'] != item['action']:
                             item['has_custom_add_view']=True
                 result.append(item)

         result.sort(lambda a, b: cmp(a['title'], b['title']))
         return result

self.menu_id is from here:

class ContentAdding(Adding, SimpleItem):

     menu_id = "add_content"

Now, neither add_content nor zope.app.container.add is registered 
anywhere (Products.Five or zope.*) that I can see. The latter is 
provided in a test in zope.app.publisher. The former is nowhere to be 
found. Unless I provide both, I can't see how this won't fail every time.

I can add some registrations in Plone, or send patch to Five if we can 
get it in time for Plone 3.

Martin


From tim at se.linux.org  Fri Feb 16 12:06:35 2007
From: tim at se.linux.org (Tim =?iso-8859-1?Q?Terleg=E5rd?=)
Date: Fri, 16 Feb 2007 12:06:35 +0100
Subject: [z3-five] Formlib custom template - AttributeError: debug.
In-Reply-To: <45D4E567.7030005@coig.katowice.pl>
References: <45D4E567.7030005@coig.katowice.pl>
Message-ID: <20070216110635.GA18351@kraftur.se.linux.org>

> Traceback (innermost last):
>   Module ZPublisher.Publish, line 115, in publish
>   Module ZPublisher.mapply, line 88, in mapply
>   Module ZPublisher.Publish, line 41, in call_object
>   Module zope.formlib.form, line 758, in __call__
>   Module zope.formlib.form, line 752, in render
>   Module zope.app.pagetemplate.viewpagetemplatefile, line 83, in __call__
>   Module zope.app.pagetemplate.viewpagetemplatefile, line 47, in __call__
>   Module ZPublisher.HTTPRequest, line 1219, in __getattr__
> AttributeError: debug

I got this yesterday as well. Zope 3 requests has a debug attribute, but
Zope 2 requests have not. So there's a problem when Zope 3 code uses the
Zope 2 request and expects there to be a debug attribute.

I fixed this by adding the debug attribute to the ZPlublisher.HTTPRequest
class inside __init__:

    from zope.publisher.base import DebugFlags
    self.debug = DebugFlags()    # inside __init__

Philipp said it will be fixed in Zope 2.11 or something. But there's no
way we can wait that long to get our stuff working  :)

Tim

From maciej.wisniowski at coig.katowice.pl  Fri Feb 16 12:26:17 2007
From: maciej.wisniowski at coig.katowice.pl (Maciej Wisniowski)
Date: Fri, 16 Feb 2007 12:26:17 +0100
Subject: [z3-five] Formlib custom template - AttributeError: debug.
In-Reply-To: <20070216110635.GA18351@kraftur.se.linux.org>
References: <45D4E567.7030005@coig.katowice.pl>
	<20070216110635.GA18351@kraftur.se.linux.org>
Message-ID: <45D594D9.1080202@coig.katowice.pl>


> I got this yesterday as well. Zope 3 requests has a debug attribute, but
> Zope 2 requests have not. So there's a problem when Zope 3 code uses the
> Zope 2 request and expects there to be a debug attribute.
>
> I fixed this by adding the debug attribute to the ZPlublisher.HTTPRequest
> class inside __init__:
>
>     from zope.publisher.base import DebugFlags
>     self.debug = DebugFlags()    # inside __init__
>
> Philipp said it will be fixed in Zope 2.11 or something. But there's no
> way we can wait that long to get our stuff working  :)
>   
Good to know but I wonder how default formlib form works then?
It also uses NamedTemplates and rest of the stuff just like I used
in my form.

Where is the point that original formlib 'pageform.pt' doesn't
raise this error? I didn't find anything about 'debug' in
formlib code (but maybe I missed something)?

-- 
Maciej Wisniowski

From maciej.wisniowski at coig.katowice.pl  Fri Feb 16 12:30:46 2007
From: maciej.wisniowski at coig.katowice.pl (Maciej Wisniowski)
Date: Fri, 16 Feb 2007 12:30:46 +0100
Subject: [z3-five] Formlib custom template - AttributeError: debug.
In-Reply-To: <45D4EE39.6090307@sitefusion.co.uk>
References: <45D4E567.7030005@coig.katowice.pl>
	<45D4EE39.6090307@sitefusion.co.uk>
Message-ID: <45D595E6.6040208@coig.katowice.pl>


>
> I managed to get things working by using the following:
>
> from Products.Five.browser import pagetemplatefile
> from plone.app.form._named import named_template_adapter
> _template = pagetemplatefile.ViewPageTemplateFile('pageform.pt')
> default_named_template_adapter = named_template_adapter(_template)
>   
Huh... I've even seen named_template_adapter code
in plone core before but was not sure how to use this.
Thanks

-- 
Maciej Wisniowski

From tim at se.linux.org  Fri Feb 16 12:34:33 2007
From: tim at se.linux.org (Tim =?iso-8859-1?Q?Terleg=E5rd?=)
Date: Fri, 16 Feb 2007 12:34:33 +0100
Subject: [z3-five] traversing path expression
Message-ID: <20070216113433.GB18351@kraftur.se.linux.org>

Hi,

I'm trying to use Five and plone.app.form to get an automatic edit form
which uses zope.formlib. There seems to be a problem with traversal of the
path expressions in the template. I get back None for the 'view/xxx'
expression for any value of xxx.

The OrderedMultiSelectWidget (orderedSelectionList.pt) is used for showing
my List field. After debugging the widget, tal etc I finally came to the
part where traversing apapters are used for finding 'xxx' of the view. The
widget (which is the view) is adapted to ITraversable in
zope.app.traversing.adapers.py. It finds Five.traversable.FiveTraversable
which doesn't look for the 'xxx' attribute, it just looks for an adapter
with the name 'xxx'. It can't find one and the traversing returns None
and so does the path expression.

FiveTraversable inherits from DefaultTraversable which seems to look at
the attribute, but FiveTraversable overrides the traverse method with
code that doesn't consider the attributes, it only search for an adapter.

I'm very confused. Am I missing some five zcml directives or why can't
my path expression be traversed correctly? I'm using Five 1.4,2.

Tim

From tim at se.linux.org  Fri Feb 16 15:42:37 2007
From: tim at se.linux.org (Tim =?iso-8859-1?Q?Terleg=E5rd?=)
Date: Fri, 16 Feb 2007 15:42:37 +0100
Subject: [z3-five] traversing path expression
In-Reply-To: <20070216113433.GB18351@kraftur.se.linux.org>
References: <20070216113433.GB18351@kraftur.se.linux.org>
Message-ID: <20070216144237.GC18351@kraftur.se.linux.org>

On Fri, Feb 16, 2007 at 12:34:33PM +0100, Tim Terleg?rd wrote:
> 
> I'm trying to use Five and plone.app.form to get an automatic edit form
> which uses zope.formlib. There seems to be a problem with traversal of the
> path expressions in the template. I get back None for the 'view/xxx'
> expression for any value of xxx.
> 
> The OrderedMultiSelectWidget (orderedSelectionList.pt) is used for showing
> my List field. After debugging the widget, tal etc I finally came to the
> part where traversing apapters are used for finding 'xxx' of the view. The
> widget (which is the view) is adapted to ITraversable in
> zope.app.traversing.adapers.py. It finds Five.traversable.FiveTraversable
> which doesn't look for the 'xxx' attribute, it just looks for an adapter
> with the name 'xxx'. It can't find one and the traversing returns None
> and so does the path expression.
> 
> FiveTraversable inherits from DefaultTraversable which seems to look at
> the attribute, but FiveTraversable overrides the traverse method with
> code that doesn't consider the attributes, it only search for an adapter.

This workaround worked.

  <adapter
      for="zope.app.form.browser.interfaces.IInputWidget"
      provides="zope.app.traversing.interfaces.ITraversable"
      factory="zope.app.traversing.adapters.DefaultTraversable"
      />

It seems that the widget I use (OrderedMultiSelectWidget) renders itself
and it therefore uses zope3 page template engine, which is not fully
compatible with zope2 (atleast when it comes to path expression
traversing). So problem solved.

Tim

From maciej.wisniowski at coig.katowice.pl  Fri Feb 16 22:37:17 2007
From: maciej.wisniowski at coig.katowice.pl (Maciej Wisniowski)
Date: Fri, 16 Feb 2007 22:37:17 +0100
Subject: [z3-five] Formlib custom template - AttributeError: debug.
In-Reply-To: <45D4E567.7030005@coig.katowice.pl>
References: <45D4E567.7030005@coig.katowice.pl>
Message-ID: <45D6240D.9010600@coig.katowice.pl>


> What is wrong? Code is run at Plone but it's plain five with
> formlib here. Possibly lack of acquisition wrapper?
I found the bug. In general I looked at lib/python/zope/formlib
instead of Products/Five/formlib.

ViewPageTemplateFile should be imported like:

from Products.Five.browser.pagetemplatefile import ViewPageTemplateFile

not like:

from zope.app.pagetemplate import ViewPageTemplateFile.

Also template definition (pageform.pt) is (re)defined in Five package.

-- 
Maciej Wisniowski

From d.schulz81 at gmx.net  Sat Feb 17 17:32:13 2007
From: d.schulz81 at gmx.net (Dennis Schulz)
Date: Sat, 17 Feb 2007 17:32:13 +0100
Subject: [z3-five] view on several permissions?
Message-ID: <45D72E0D.90402@gmx.net>

Hi,

is it possible to register a view on several permissions in zcml?
I tried the following notations but didn't work:

permission="cmf.ListFolderContents; cells.pr.approve"
permission="cmf.ListFolderContents, cells.pr.approve"

What is the proper way of doing this?
Do I have to register the view two times?

Cheers, Dennis


From philipp at weitershausen.de  Sun Feb 18 12:57:11 2007
From: philipp at weitershausen.de (Philipp von Weitershausen)
Date: Sun, 18 Feb 2007 12:57:11 +0100
Subject: [z3-five] view on several permissions?
In-Reply-To: <45D72E0D.90402@gmx.net>
References: <45D72E0D.90402@gmx.net>
Message-ID: <45D83F17.90102@weitershausen.de>

Dennis Schulz wrote:
> is it possible to register a view on several permissions in zcml?
> I tried the following notations but didn't work:
> 
> permission="cmf.ListFolderContents; cells.pr.approve"
> permission="cmf.ListFolderContents, cells.pr.approve"
> 
> What is the proper way of doing this?
> Do I have to register the view two times?

Anything, including a view, can only have one unique permission. That's 
how Zope's security system works. You will have to make up your mind and 
decide which permission to protect the view with.


-- 
http://worldcookery.com -- Professional Zope documentation and training
Next Zope 3 training at Camp5: http://trizpug.org/boot-camp/camp5

From philipp at weitershausen.de  Sun Feb 18 13:04:39 2007
From: philipp at weitershausen.de (Philipp von Weitershausen)
Date: Sun, 18 Feb 2007 13:04:39 +0100
Subject: [z3-five] Is browser:addMenuItem and addingInfo() supposed to
	work?
In-Reply-To: <er31ec$1nd$1@sea.gmane.org>
References: <er31ec$1nd$1@sea.gmane.org>
Message-ID: <45D840D7.4000002@weitershausen.de>

Martin Aspeli wrote:
> I've seen reports of people making this work, but I'm struggling. I've 
> got an add view (formlib AddForm) and an add menu item like so:
> 
>      <browser:addMenuItem
>          title="Foo"
>          factory="my.package.Foo"
>          view="my.package.Foo"
>          permission="cmf.AddPortalContent"
>          />
> 
> I then want to list all addable views (in fact, I'm trying to make Plone 
> respect add-menus and ultimately add forms, at least in the case where 
> an add menu item can be linked to an old-fashioned CMF).
> 
> However, I get ComponentLookupErrors, because adding.py in Five does this:
[snip]

Next time, I suggest you paste the error traceback, because that's way 
more informative than a code snippet from Five that everybody can lookup 
by himself. Right now I had to guess what component it failed to lookup...

> Now, neither add_content nor zope.app.container.add is registered 
> anywhere (Products.Five or zope.*) that I can see.

It's registered in zope.app/menus.zcml. This file is probably not 
available in Zope 2 (anymore):

   <menu
       id="zope.app.container.add"
       title="Add"
       description="Menu for objects to be added according to
                    containment constraints"
       interface="zope.app.publisher.interfaces.browser.AddMenu"
       />

> The latter is 
> provided in a test in zope.app.publisher. The former is nowhere to be 
> found. Unless I provide both, I can't see how this won't fail every time.

add_content is old and no longer used in Zope 3.

*Sigh*. This is *exactly* why I want to get rid of all of this useless 
code duplication in Five.

> I can add some registrations in Plone, or send patch to Five if we can 
> get it in time for Plone 3.

For now I suggest you guys register zope.app.container.add yourself. Or, 
perhaps you register a variant and also regsiter your own adding view 
that refers to your variant.


-- 
http://worldcookery.com -- Professional Zope documentation and training
Next Zope 3 training at Camp5: http://trizpug.org/boot-camp/camp5

From optilude at gmx.net  Sun Feb 18 13:12:56 2007
From: optilude at gmx.net (Martin Aspeli)
Date: Sun, 18 Feb 2007 12:12:56 +0000
Subject: [z3-five] Is browser:addMenuItem and addingInfo() supposed to
	work?
In-Reply-To: <45D840D7.4000002@weitershausen.de>
References: <er31ec$1nd$1@sea.gmane.org> <45D840D7.4000002@weitershausen.de>
Message-ID: <er9fv1$7l1$1@sea.gmane.org>

Philipp von Weitershausen wrote:
> Martin Aspeli wrote:
>> I've seen reports of people making this work, but I'm struggling. I've 
>> got an add view (formlib AddForm) and an add menu item like so:
>>
>>      <browser:addMenuItem
>>          title="Foo"
>>          factory="my.package.Foo"
>>          view="my.package.Foo"
>>          permission="cmf.AddPortalContent"
>>          />
>>
>> I then want to list all addable views (in fact, I'm trying to make Plone 
>> respect add-menus and ultimately add forms, at least in the case where 
>> an add menu item can be linked to an old-fashioned CMF).
>>
>> However, I get ComponentLookupErrors, because adding.py in Five does this:
> [snip]
> 
> Next time, I suggest you paste the error traceback, because that's way 
> more informative than a code snippet from Five that everybody can lookup 
> by himself. Right now I had to guess what component it failed to lookup...

Yeah, sorry, was being lazy. The error is looking up the utility for the 
IBrowserMenu.

>> Now, neither add_content nor zope.app.container.add is registered 
>> anywhere (Products.Five or zope.*) that I can see.
> 
> It's registered in zope.app/menus.zcml. This file is probably not 
> available in Zope 2 (anymore):
> 
>    <menu
>        id="zope.app.container.add"
>        title="Add"
>        description="Menu for objects to be added according to
>                     containment constraints"
>        interface="zope.app.publisher.interfaces.browser.AddMenu"
>        />
> 
>> The latter is 
>> provided in a test in zope.app.publisher. The former is nowhere to be 
>> found. Unless I provide both, I can't see how this won't fail every time.
> 
> add_content is old and no longer used in Zope 3.
> 
> *Sigh*. This is *exactly* why I want to get rid of all of this useless 
> code duplication in Five.

:)

>> I can add some registrations in Plone, or send patch to Five if we can 
>> get it in time for Plone 3.
> 
> For now I suggest you guys register zope.app.container.add yourself. Or, 
> perhaps you register a variant and also regsiter your own adding view 
> that refers to your variant.

Actually, I think it can't work anyway, because it first tries to look 
for "add_content" which isn't registered anywhere either.

We made it work by looking up items registered for the 
zope.app.publisher.interfaces.browser.AddMenu directly, i.e. not by 
name. The <addMenuItem> directive falls back on this interface.

Martin


From optilude at gmx.net  Sat Feb 24 04:19:20 2007
From: optilude at gmx.net (Martin Aspeli)
Date: Sat, 24 Feb 2007 03:19:20 +0000
Subject: [z3-five] FieldProperty, acquisition and Zope 2
Message-ID: <eroas3$gnc$1@sea.gmane.org>

Hi all,

I was using FieldProperty in something like this:

class ITask(Interface):
      assigned_to = schema.Choice(title=_(u"Assigned to"),
                                  vocabulary="assignees")

The assignees vocabulary is defined as such:

def assigneeVocabularyFactory(context):
      """Get possible assignees from the immediate parent project
      """
      team = ()
      parent = context
      while parent is not None and not team:
          project = IProject(parent, None)
          if project is not None:
              team = project.team
          else:
              parent = aq_parent(parent)
      return SimpleVocabulary.fromValues(team)
alsoProvides(assigneeVocabularyFactory, IVocabularyFactory)

I then have a class which uses a FieldProperty:

class Task(SimpleItem):
      implements(ITask)
      assigned_to = FieldProperty(IDeliverable['assigned_to'])

I've got an add form which uses form.applyChanges(task,
self.form_fields, data)

Now, this fails horribly, miserably and deep-deep inside zope.schema or
thereabouts. To be honest, the traceback is pretty useless, but I've
worked out that the error occurs because FieldProperty is a property,
with __get__() and __set__() methods, and as soon as I go there, the
task loses its acquisition context.

Now, http://zope.org/Collectors/Zope/2163 tells me that we can't have
acquisition inside property accessors.

I'm wondering, though, if we have other options? Could Five provide
something similar to FieldProperty but which is acquisition-aware?

Martin


From philipp at weitershausen.de  Sat Feb 24 12:04:49 2007
From: philipp at weitershausen.de (Philipp von Weitershausen)
Date: Sat, 24 Feb 2007 12:04:49 +0100
Subject: [z3-five] FieldProperty, acquisition and Zope 2
In-Reply-To: <eroas3$gnc$1@sea.gmane.org>
References: <eroas3$gnc$1@sea.gmane.org>
Message-ID: <45E01BD1.1030404@weitershausen.de>

Martin Aspeli wrote:
> Now, this fails horribly, miserably and deep-deep inside zope.schema or
> thereabouts. To be honest, the traceback is pretty useless, but I've
> worked out that the error occurs because FieldProperty is a property,
> with __get__() and __set__() methods, and as soon as I go there, the
> task loses its acquisition context.
> 
> Now, http://zope.org/Collectors/Zope/2163 tells me that we can't have
> acquisition inside property accessors.
> 
> I'm wondering, though, if we have other options? Could Five provide
> something similar to FieldProperty but which is acquisition-aware?

Unlikely. I find FieldProperty is YAGNI most of the time. You describe 
that you're using formlib forms to apply the data to the object. Formlib 
does all the validation already. Why do you need FieldProperty?


-- 
http://worldcookery.com -- Professional Zope documentation and training
Next Zope 3 training at Camp5: http://trizpug.org/boot-camp/camp5

From optilude at gmx.net  Sat Feb 24 13:12:14 2007
From: optilude at gmx.net (Martin Aspeli)
Date: Sat, 24 Feb 2007 12:12:14 +0000
Subject: [z3-five] FieldProperty, acquisition and Zope 2
In-Reply-To: <45E01BD1.1030404@weitershausen.de>
References: <eroas3$gnc$1@sea.gmane.org> <45E01BD1.1030404@weitershausen.de>
Message-ID: <erpa39$kv6$1@sea.gmane.org>

Philipp von Weitershausen wrote:
> Martin Aspeli wrote:
>> Now, this fails horribly, miserably and deep-deep inside zope.schema or
>> thereabouts. To be honest, the traceback is pretty useless, but I've
>> worked out that the error occurs because FieldProperty is a property,
>> with __get__() and __set__() methods, and as soon as I go there, the
>> task loses its acquisition context.
>>
>> Now, http://zope.org/Collectors/Zope/2163 tells me that we can't have
>> acquisition inside property accessors.
>>
>> I'm wondering, though, if we have other options? Could Five provide
>> something similar to FieldProperty but which is acquisition-aware?
> 
> Unlikely. I find FieldProperty is YAGNI most of the time. You describe 
> that you're using formlib forms to apply the data to the object. Formlib 
> does all the validation already. Why do you need FieldProperty?

Hey, I learned about them from *your* book, you tell me. :-)

I agree it's a bit YAGNI. :)

Martin


From rocky at serverzen.com  Mon Mar  5 16:13:24 2007
From: rocky at serverzen.com (Rocky)
Date: Mon, 05 Mar 2007 07:13:24 -0800
Subject: [z3-five] first public five.localsitemanager release
Message-ID: <1173107604.363714.229090@30g2000cwc.googlegroups.com>

Just wanted to let you all know that I've just published the first
public five.localsitemanager release to the cheeseshop.  You can find
it's info at:
  http://cheeseshop.python.org/pypi/five.localsitemanager/

This should probably mean that the Five/site stuff in Products.Five in
zope2 svn trunk should be deprecated.

The one additional feature I'd personally like to see before 1.0 final
is a full management screen located at mysite/++etc++site that would
allow one to register/unregister components.  Right now that traversal
does exist but it basically just points to the ObjectManager zmi
screens for the component registry.

If anyone else has any feature requests, be sure to post them.

Regards,
Rocky


From philipp at weitershausen.de  Mon Mar  5 17:38:36 2007
From: philipp at weitershausen.de (Philipp von Weitershausen)
Date: Mon, 05 Mar 2007 11:38:36 -0500
Subject: [z3-five] first public five.localsitemanager release
In-Reply-To: <1173107604.363714.229090@30g2000cwc.googlegroups.com>
References: <1173107604.363714.229090@30g2000cwc.googlegroups.com>
Message-ID: <45EC478C.5050805@weitershausen.de>

Rocky wrote:
> Just wanted to let you all know that I've just published the first
> public five.localsitemanager release to the cheeseshop.  You can find
> it's info at:
>   http://cheeseshop.python.org/pypi/five.localsitemanager/

Yay, great stuff!

> This should probably mean that the Five/site stuff in Products.Five in
> zope2 svn trunk should be deprecated.

At least it should be discouraged and five.localsitemanager should be 
encouraged at the same time. In a future version of Zope2/Five, we might 
give Five some long-awaited cleanup and ship five.localsitemanager with 
Zope 2 as a replacement. BBB will have to be maintained.

> The one additional feature I'd personally like to see before 1.0 final
> is a full management screen located at mysite/++etc++site that would
> allow one to register/unregister components.  Right now that traversal
> does exist but it basically just points to the ObjectManager zmi
> screens for the component registry.

Sounds good.


-- 
http://worldcookery.com -- Professional Zope documentation and training
Next Zope 3 training at Camp5: http://trizpug.org/boot-camp/camp5

From lists at zopyx.com  Mon Mar  5 17:47:40 2007
From: lists at zopyx.com (Andreas Jung)
Date: Mon, 05 Mar 2007 17:47:40 +0100
Subject: [z3-five] first public five.localsitemanager release
In-Reply-To: <1173107604.363714.229090@30g2000cwc.googlegroups.com>
References: <1173107604.363714.229090@30g2000cwc.googlegroups.com>
Message-ID: <51767FDE9E1C668F1E69397B@[192.168.0.24]>



--On 5. M?rz 2007 07:13:24 -0800 Rocky <rocky at serverzen.com> wrote:

> Just wanted to let you all know that I've just published the first
> public five.localsitemanager release to the cheeseshop.  You can find
> it's info at:
>   http://cheeseshop.python.org/pypi/five.localsitemanager/


Nice. But I wonder why it has to be a dedicated project and isn't part
of further Five releases?

Andreas
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 186 bytes
Desc: not available
Url : http://codespeak.net/pipermail/z3-five/attachments/20070305/d46c751b/attachment.pgp 

From philipp at weitershausen.de  Mon Mar  5 18:00:51 2007
From: philipp at weitershausen.de (Philipp von Weitershausen)
Date: Mon, 05 Mar 2007 12:00:51 -0500
Subject: [z3-five] first public five.localsitemanager release
In-Reply-To: <51767FDE9E1C668F1E69397B@[192.168.0.24]>
References: <1173107604.363714.229090@30g2000cwc.googlegroups.com>
	<51767FDE9E1C668F1E69397B@[192.168.0.24]>
Message-ID: <45EC4CC3.4080104@weitershausen.de>

Andreas Jung wrote:
>> Just wanted to let you all know that I've just published the first
>> public five.localsitemanager release to the cheeseshop.  You can find
>> it's info at:
>>   http://cheeseshop.python.org/pypi/five.localsitemanager/
> 
> 
> Nice. But I wonder why it has to be a dedicated project and isn't part
> of further Five releases?

Because this is a new feature that we now want to use in Zope 2.10. 
Adding smaller packages, independently distributed packages that 
everybody can install if they want is much easier than maintaining a 
whole new Five release, telling people to install that, etc.

I expect that much of Five will actually go away once we make some tiny 
but important changes in Zope 2 (make Acquisition aware of __parent__).


-- 
http://worldcookery.com -- Professional Zope documentation and training
Next Zope 3 training at Camp5: http://trizpug.org/boot-camp/camp5

From lists at zopyx.com  Wed Mar 21 06:59:27 2007
From: lists at zopyx.com (Andreas Jung)
Date: Wed, 21 Mar 2007 06:59:27 +0100
Subject: [z3-five] Programmatic initialization of utilites during the Zope
 startup phase
Message-ID: <256A8DFABE9AD1ADDB5E2840@suxmac2.local>

Hi,

I have the need to register some utilities dynamically during the startup 
phase of Zope. In detail: I have a product that uses z3c.sqlalchemy (a Zope 
3 database adapter for Zope2/3). During the startup I need to read some
database configurations from a configuration file and setup a named utility
for each of the database configurations. I can't and don't want to do that
in ZCML. However the Zope product initialization happens before the Zope 3
registry stuff is available. I have a workaround with a setup utility
registered through ZCML that performs the registration of the utilities
after Zope is up and before answering to the first product specific request.
This works but I consider this to be an ugly workaround. Any ideas
for a better solution?

Andreas
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 186 bytes
Desc: not available
Url : http://codespeak.net/pipermail/z3-five/attachments/20070321/27661cf9/attachment.pgp 

From sidnei at enfoldsystems.com  Wed Mar 21 13:53:59 2007
From: sidnei at enfoldsystems.com (Sidnei da Silva)
Date: Wed, 21 Mar 2007 09:53:59 -0300
Subject: [z3-five] Programmatic initialization of utilites during the
	Zope startup phase
In-Reply-To: <256A8DFABE9AD1ADDB5E2840@suxmac2.local>
References: <256A8DFABE9AD1ADDB5E2840@suxmac2.local>
Message-ID: <a7a2b76b0703210553r154ef598v4a4971a355040c9e@mail.gmail.com>

I had a similar issue here, and I was able to just add some code on
the inside the initialize() method of my Zope product:

    from zope.interface import providedBy
    from zope.component import getGlobalSiteManager
    for iface in providedBy(utility):
        getGlobalSiteManager().provideUtility(iface, utility, u'')

Are you saying that in your case the global site manager is not
available yet by the time the product's initialize() is called?
-- 
Sidnei da Silva
Enfold Systems                http://enfoldsystems.com
Fax +1 832 201 8856     Office +1 713 942 2377 Ext 214

From lists at zopyx.com  Wed Mar 21 14:39:12 2007
From: lists at zopyx.com (Andreas Jung)
Date: Wed, 21 Mar 2007 14:39:12 +0100
Subject: [z3-five] Programmatic initialization of utilites during the
 Zope startup phase
In-Reply-To: <a7a2b76b0703210553r154ef598v4a4971a355040c9e@mail.gmail.com>
References: <256A8DFABE9AD1ADDB5E2840@suxmac2.local>
	<a7a2b76b0703210553r154ef598v4a4971a355040c9e@mail.gmail.com>
Message-ID: <AD738115CD2E732F6D6EE3C5@suxmac2.local>



--On 21. M?rz 2007 09:53:59 -0300 Sidnei da Silva 
<sidnei at enfoldsystems.com> wrote:

> I had a similar issue here, and I was able to just add some code on
> the inside the initialize() method of my Zope product:
>
>     from zope.interface import providedBy
>     from zope.component import getGlobalSiteManager
>     for iface in providedBy(utility):
>         getGlobalSiteManager().provideUtility(iface, utility, u'')
>
> Are you saying that in your case the global site manager is not
> available yet by the time the product's initialize() is called?
>

We're still on Zope 2.8.something...I think I tried all that with
the related APIs that are available in Zope 3.0.

Andreas
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 186 bytes
Desc: not available
Url : http://codespeak.net/pipermail/z3-five/attachments/20070321/08d67ada/attachment.pgp 

From ra at burningman.com  Fri Mar 23 02:16:28 2007
From: ra at burningman.com (Rob Miller)
Date: Thu, 22 Mar 2007 18:16:28 -0700
Subject: [z3-five] weird ZEO cache error during five:registerPackage
	processing
Message-ID: <etv9p7$7va$1@sea.gmane.org>

hi all,

i've got a very strange problem happening.  whenever i start my Zope server 
against a brand new ZODB being served up by ZEO, Zope dies while in the ZCML 
processing for a five:registerPackage call.  here's the traceback:

Traceback (most recent call last):
   File "/home/rob/openplans/lib/zope/lib/python/Zope2/Startup/run.py", line 
56, in ?
     run()
   File "/home/rob/openplans/lib/zope/lib/python/Zope2/Startup/run.py", line 
21, in run
     starter.prepare()
   File 
"/home/rob/openplans/lib/Zope-2.9.6/lib/python/Zope2/Startup/__init__.py", 
line 98, in prepare
     self.startZope()
   File 
"/home/rob/openplans/lib/Zope-2.9.6/lib/python/Zope2/Startup/__init__.py", 
line 257, in startZope
     Zope2.startup()
   File "/home/rob/openplans/lib/Zope-2.9.6/lib/python/Zope2/__init__.py", 
line 47, in startup
     _startup()
   File "/home/rob/openplans/lib/Zope-2.9.6/lib/python/Zope2/App/startup.py", 
line 103, in startup
     OFS.Application.initialize(application)
   File "/home/rob/openplans/lib/Zope-2.9.6/lib/python/OFS/Application.py", 
line 304, in initialize
     initializer.initialize()
   File "/home/rob/openplans/lib/Zope-2.9.6/lib/python/OFS/Application.py", 
line 333, in initialize
     self.install_products()
   File "/home/rob/openplans/lib/Zope-2.9.6/lib/python/OFS/Application.py", 
line 605, in install_products
     return install_products(app)
   File "/home/rob/openplans/lib/Zope-2.9.6/lib/python/OFS/Application.py", 
line 636, in install_products
     folder_permissions, raise_exc=debug_mode)
   File "/home/rob/openplans/lib/Zope-2.9.6/lib/python/OFS/Application.py", 
line 783, in install_product
     initmethod(context)
   File "/home/rob/openplans/zope/Products/Five/__init__.py", line 31, in 
initialize
     zcml.load_site()
   File "/home/rob/openplans/zope/Products/Five/zcml.py", line 41, in load_site
     _context = xmlconfig.file(file)
   File 
"/home/rob/openplans/lib/zope/lib/python/zope/configuration/xmlconfig.py", 
line 559, in file
     context.execute_actions()
   File 
"/home/rob/openplans/lib/Zope-2.9.6/lib/python/zope/configuration/config.py", 
line 612, in execute_actions
     callable(*args, **kw)
   File "/home/rob/openplans/zope/Products/Five/fiveconfigure.py", line 284, 
in _registerPackage
     transaction.commit()
   File 
"/home/rob/openplans/lib/Zope-2.9.6/lib/python/transaction/_manager.py", line 
96, in commit
     return self.get().commit(sub, deprecation_wng=False)
   File 
"/home/rob/openplans/lib/Zope-2.9.6/lib/python/transaction/_transaction.py", 
line 380, in commit
     self._saveCommitishError() # This raises!
   File 
"/home/rob/openplans/lib/Zope-2.9.6/lib/python/transaction/_transaction.py", 
line 378, in commit
     self._commitResources()
   File 
"/home/rob/openplans/lib/Zope-2.9.6/lib/python/transaction/_transaction.py", 
line 441, in _commitResources
     rm.tpc_finish(self)
   File "/home/rob/openplans/lib/Zope-2.9.6/lib/python/ZODB/Connection.py", 
line 679, in tpc_finish
     self._storage.tpc_finish(transaction, callback)
   File "/home/rob/openplans/lib/Zope-2.9.6/lib/python/ZEO/ClientStorage.py", 
line 974, in tpc_finish
     self._update_cache(tid)
   File "/home/rob/openplans/lib/Zope-2.9.6/lib/python/ZEO/ClientStorage.py", 
line 999, in _update_cache
     self._cache.invalidate(oid, version, tid)
   File "/home/rob/openplans/lib/Zope-2.9.6/lib/python/ZEO/cache.py", line 
367, in invalidate
     assert tid is not None and cur_tid < tid
zope.configuration.config.ConfigurationExecutionError: exceptions.AssertionError:
   in:
   File 
"/home/rob/openplans/src/opencore_trunk/opencore/siteui/configure.zcml", line 
9.2-12.7
     <five:registerPackage
        package="opencore.siteui"
        initialize="opencore.siteui.zope2.initialize"
        />


this is using Zope 2.9.6 and the Five 1.4 branch.  the tag that is causing the 
problem can be found at http://tinyurl.com/yqoaw7.

i've absolutely verified that the registerPackage directive is causing the 
problem; if i remove that tag Zope starts with no problem.  similarly, if i 
don't use ZEO it starts w/o trouble.  and subsequent starts, once there is a 
Data.fs file to connect to, also seem to work.  it's only this one combination 
that causes the problem.  it's a bit of a headache, though, b/c it means that 
our buildbot processes, which construct an environment from scratch every 
time, fail before they even begin.

i know very little about the ZEO caching infrastructure, and don't even know 
where to being in debugging this.  anyone have any clues what might be going on?

-r


From chrism at plope.com  Fri Mar 23 03:09:45 2007
From: chrism at plope.com (Chris McDonough)
Date: Thu, 22 Mar 2007 22:09:45 -0400
Subject: [z3-five] weird ZEO cache error during five:registerPackage
	processing
In-Reply-To: <etv9p7$7va$1@sea.gmane.org>
References: <etv9p7$7va$1@sea.gmane.org>
Message-ID: <C0B348E2-162D-4570-9C7A-E669AD5E6F80@plope.com>

This seems to be isolated to packages which are not in the Products  
namespace.  Another person sent a similar bugreport to the Zope (or  
ZODB?) list.  The problem seems to be that the registerPackage  
handler (in fiveconfigure.py) attempts to commit a transaction in the  
midst of a finally clause, which smells a little weird.  If you  
comment the transaction.commit out in that directive, what happens?

On Mar 22, 2007, at 9:16 PM, Rob Miller wrote:

> registerPackage


From chrism at plope.com  Sat Mar 24 00:34:42 2007
From: chrism at plope.com (Chris McDonough)
Date: Fri, 23 Mar 2007 19:34:42 -0400
Subject: [z3-five] weird ZEO cache error during five:registerPackage
	processing
In-Reply-To: <4604620B.4050301@burningman.com>
References: <etv9p7$7va$1@sea.gmane.org>
	<C0B348E2-162D-4570-9C7A-E669AD5E6F80@plope.com>
	<4604620B.4050301@burningman.com>
Message-ID: <265661DF-8DD5-4016-A68D-467F5E9DDF80@plope.com>

On Mar 23, 2007, at 7:26 PM, Rob Miller wrote:

> Chris McDonough wrote:
>> This seems to be isolated to packages which are not in the  
>> Products namespace.  Another person sent a similar bugreport to  
>> the Zope (or ZODB?) list.  The problem seems to be that the  
>> registerPackage handler (in fiveconfigure.py) attempts to commit a  
>> transaction in the midst of a finally clause, which smells a  
>> little weird.  If you comment the transaction.commit out in that  
>> directive, what happens?
>
> if i do that, then i get the same traceback as before, but every  
> time instead of just during the initial start.

I think it's broke, mon.  I think the utility of this directive might  
a little suspect to begin with (if you have a package that needs  
startup initialization, you could just put it in the Products  
namespace), but the only other thing I can think to try is to take  
all the exception handling out of this the  
fiveconfigure._registerPackage function and track down why it's  
barfing to start with.

Sorry Rob. :-(

From ra at burningman.com  Sat Mar 24 01:04:09 2007
From: ra at burningman.com (Rob Miller)
Date: Fri, 23 Mar 2007 17:04:09 -0700
Subject: [z3-five] weird ZEO cache error during five:registerPackage
	processing
In-Reply-To: <265661DF-8DD5-4016-A68D-467F5E9DDF80@plope.com>
References: <etv9p7$7va$1@sea.gmane.org>	<C0B348E2-162D-4570-9C7A-E669AD5E6F80@plope.com>	<4604620B.4050301@burningman.com>
	<265661DF-8DD5-4016-A68D-467F5E9DDF80@plope.com>
Message-ID: <46046AF9.7070704@burningman.com>

Chris McDonough wrote:
> On Mar 23, 2007, at 7:26 PM, Rob Miller wrote:
> 
>> Chris McDonough wrote:
>>> This seems to be isolated to packages which are not in the  
>>> Products namespace.  Another person sent a similar bugreport to  
>>> the Zope (or ZODB?) list.  The problem seems to be that the  
>>> registerPackage handler (in fiveconfigure.py) attempts to commit a  
>>> transaction in the midst of a finally clause, which smells a  
>>> little weird.  If you comment the transaction.commit out in that  
>>> directive, what happens?
>> if i do that, then i get the same traceback as before, but every  
>> time instead of just during the initial start.
> 
> I think it's broke, mon.  I think the utility of this directive might  
> a little suspect to begin with (if you have a package that needs  
> startup initialization, you could just put it in the Products  
> namespace), but the only other thing I can think to try is to take  
> all the exception handling out of this the  
> fiveconfigure._registerPackage function and track down why it's  
> barfing to start with.
> 
> Sorry Rob. :-(

'saright.  as you said, there's always the Products namespace.  :-).

-r


From d.w.morriss at gmail.com  Mon Mar 26 17:20:09 2007
From: d.w.morriss at gmail.com (whit)
Date: Mon, 26 Mar 2007 10:20:09 -0500
Subject: [z3-five] weird ZEO cache error during five:registerPackage
	processing
In-Reply-To: <46046AF9.7070704@burningman.com>
References: <etv9p7$7va$1@sea.gmane.org>	<C0B348E2-162D-4570-9C7A-E669AD5E6F80@plope.com>	<4604620B.4050301@burningman.com>	<265661DF-8DD5-4016-A68D-467F5E9DDF80@plope.com>
	<46046AF9.7070704@burningman.com>
Message-ID: <eu8ob9$j25$1@sea.gmane.org>

Rob Miller wrote:
> Chris McDonough wrote:
>> On Mar 23, 2007, at 7:26 PM, Rob Miller wrote:
>>
>>> Chris McDonough wrote:
>>>> This seems to be isolated to packages which are not in the  
>>>> Products namespace.  Another person sent a similar bugreport to  
>>>> the Zope (or ZODB?) list.  The problem seems to be that the  
>>>> registerPackage handler (in fiveconfigure.py) attempts to commit a  
>>>> transaction in the midst of a finally clause, which smells a  
>>>> little weird.  If you comment the transaction.commit out in that  
>>>> directive, what happens?
>>> if i do that, then i get the same traceback as before, but every  
>>> time instead of just during the initial start.
>> I think it's broke, mon.  I think the utility of this directive might  
>> a little suspect to begin with (if you have a package that needs  
>> startup initialization, you could just put it in the Products  
>> namespace), but the only other thing I can think to try is to take  
>> all the exception handling out of this the  
>> fiveconfigure._registerPackage function and track down why it's  
>> barfing to start with.
>>
>> Sorry Rob. :-(
> 
> 'saright.  as you said, there's always the Products namespace.  :-).
> 
> -r
> 

rocky?

-w

-- 

------ d. whit morriss ------
- senior engineer, opencore -
- http://www.openplans.org  -
- m: 415-710-8975           -

"If you don't know where you are,
  you don't know anything at all" 


Dr. Edgar Spencer, Ph.D., 1995