- Zope-Dev - Zope lists

Re: [Zope-dev] aq_.* names reserved?
by Dieter Maurer 15 Sep '00

15 Sep '00

Steve Alexander writes: > I'm hacking around with some external methods called aq_containment and > aq_context. > > I just found out that I can't call them from DTML. I can call them from > the URL line of a browser just fine. > > If I rename them to a_containment and a_context, they work from DTML. > > I guess there's something in Acquisistion.c that reserves all aq_.* > names. The code is in "AccessControl.ZopeSecurityPolicy.validate". It allows access to "aq_explicit" and "aq_parent" only. I am a bit astonished that URL traversal is possible. Probably, this was not intended. Dieter

1 0

Re: [Zope-dev] ZOracleDA not working with 2.2.1?
by Matthew T. Kromer 15 Sep '00

15 Sep '00

fog(a)mixadlive.com wrote: > > Hi *, > > after upgrading to latest Zope we are having problems with the > ZOracleDA product. The first SQLMethod **after** a call to an external > method using DCOracle to update the db, fails with an error saying that > the connection is closed. (connection can't be re-opened, the only way > out is to restart zope, sigh.) > > anybody experienced the same problem? is ZOracleDA non-compatible > with Zope 2.2.1? > > ciao and tahnk you very much (as always), > federico > > -- > Federico Di Gregorio > MIXAD LIVE System Programmer fog(a)mixadlive.com > Debian GNU/Linux Developer & Italian Press Contact fog(a)debian.org > A short story: I want you. I love you. I'll miss you. -- Me Please try using DCOracle 1.3b1 to see if that fixes things; there are some bugs with connection handling in versions of DCOracle where if the connection dies it will not recover.

1 0

Making copies of ExtensionClass instances
by Greg Ward 15 Sep '00

15 Sep '00

Hi all -- I don't imagine I'm the first one to have this problem, but I don't see anything obvious in either the source or docs for ExtensionClass that helps: how do you make a copy of an ExtensionClass instance? The 'copy' module can't handle it, because it dispatches on type, and 1) it doesn't know about the type ExtensionClass, and 2) the type of an extension class instances is its extension class, so I can't just add an entry to the 'copy' module's dispatch table. So I tried to write my own 'copy()' method... but that foundered when I couldn't find a way to create a new instance of a particular extension class without using its constructor. The 'new' module can't do it, as it's 'instance()' function doesn't know about extension classes either. Argh! So what's the trick? Surely pickle and cPickle have a way to create new extension class instances when unpickling -- copying is pretty close to pickle/unpickle, so how I do it? Thanks -- Greg

1 0

ZOracleDA not working with 2.2.1?
by fog＠mixadlive.com 15 Sep '00

15 Sep '00

Hi *, after upgrading to latest Zope we are having problems with the ZOracleDA product. The first SQLMethod **after** a call to an external method using DCOracle to update the db, fails with an error saying that the connection is closed. (connection can't be re-opened, the only way out is to restart zope, sigh.) anybody experienced the same problem? is ZOracleDA non-compatible with Zope 2.2.1? ciao and tahnk you very much (as always), federico -- Federico Di Gregorio MIXAD LIVE System Programmer fog(a)mixadlive.com Debian GNU/Linux Developer & Italian Press Contact fog(a)debian.org A short story: I want you. I love you. I'll miss you. -- Me

1 0

AUTHENTICATED_USER.has_role !!
by Maulin Doshi 15 Sep '00

15 Sep '00

Hello all, Please read the following code: ********************************************************************** <dtml-if "AUTHENTICATED_USER.has_role('Vorstand')"> <dtml-with vorstand> hello <dtml-in neuanmeldungen_sql start=query_start> <dtml-if sequence-start> <p><a href="vorstand/freigeben_html">Neue Mitgliedsanträge sind eingegangen!</a></p> </dtml-if> </dtml-in> <dtml-in newNews_sql start=query_start> <dtml-if sequence-start> <p><a href="vorstand/newNewsAuthoForm_html" >New News are waiting for your authorisation!</a></p> </dtml-if> </dtml-in> <p><a href="vorstand/bankdaten_report">Bankdaten abfragen.</a></p> <p><a href="vorstand/mass_email_form" >Send Mass Email to all Members</a></p> <hr size="1" /> </dtml-with> </dtml-if> ********************************************************************** The above code use to worked very OK with Zope version: Zope 2.1.6 (binary release, python 1.5.2, linux2-x86). But now when I have upgraded to Zope version: Zope 2.2.1 (binary release, python 1.5.2, linux2-x86), I dont find "<dtml-if "AUTHENTICATED_USER.has_role('Vorstand')">" working the way it use to!! Please help me with the reason for it and the way i can solve my purpose may be in the other way. Looking forward to hear from u as soon as possible, Maulin Doshi Germany _________________________________________________________________________ Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com. Share information about yourself, create your own public profile at http://profiles.msn.com.

1 0

Re: [Zope-dev] ODBC Error
by Hung Jung Lu 15 Sep '00

15 Sep '00

>From: Rik Hoekstra <rik.hoekstra(a)inghist.nl> > >There _was_ a patch available (can't find it now, search the maillist >archives) Read my postings around April. I include two of my messages. The patch was submitted in March. I believe it even went into the Collector. But Digicool people simply refused to update the ZODBC database adapter, citing they are too busy and that for patches they need to make things really sure. Fine, in between March and September (6 months, that is), they have not been able to include the patch in the release. Meanwhile, many people have downloaded and used ZODBC. I don't understand. Is it so hard to include the patch? 6 months is more than enough time to determine that this patch does not cause viruses or major catastrophes. If there have been disasters using this patch, people would have yelled loud and clear. I have submitted bug a report to Java's Bug Parade before, and they got back to me about 6 months later, long after my project was killed. :) Now this bug is taking about the same time. :) I am impressed. What's the fun of making _every_ newbie going through the same ritual of posting to the newsgroups and get the problem fixed only 3 days later? Is it fun to waste so many people's time??? (OK, sarcasm aside, isn't time to do something about it?) :) It _is_ a carcass piece of software until someone resuscitate it. Hung Jung ===================================================== Message: Re: [Zope] Z ODBC Error Value: unpack sequence of wrong size Author: "Hung Jung Lu" Date: 04/06 04:01 AM --- In zope(a)egroups.com, "Hung Jung Lu" <hungjunglu@h...> wrote: > --- In zope(a)egroups.com, Graham Chiu <anon_emouse@h...> wrote: > >Ioan Coman <ioan(a)usa.net> writes > > >I'm using Zope 2.1.5 (binary release, python 1.5.2, win32-x86) > >... > > > Zope Error > > > Error Type: ValueError > > > Error Value: unpack sequence of wrong size OK, I search a bit more on e-groups, and found out a patch to fix this problem. Please visit: http://www.zope.org/Members/Duncan/zodbcda_patch The ZODBC adapter actually works fine. It's just that the Test tab does not work. Hung Jung ===================================================== Message: Re: [Zope] Why so many problems with database adapters? Author: "Hung Jung Lu" Date: 04/07 12:51 AM >From: Michel Pelletier <michel(a)digicool.com> >Tino Wildenhain wrote: > > If not, so we can all hope and wish, but not _demand_ to get the > > fixes :-) I certainly wouldn't _demand_ fixes for something that is _free_. I am not that silly a person. :) But I see over and over again newbies (including myself) going through the same pain that others have already gone before. Some of the nicer people have digged into the problems and FIXED them with a patch. Sadly, I had to repeat the pain simply because I was NOT aware of the patch! So, the question is simply: do we want all future users to go down the exact same path, or do we want to save them a little bit of time and stress, and make Zope a better brand name? >in CVS? Doubt it. Approaching the rat hole at every increasing speed. I know. But a simple Wiki page with a simple HTTP link to a person's personal homepage will allow you to try his/her patched version at your own risk, without Digicool's integrity being involved at all. It's like a tree with a central trunk, if you know what I mean in Revision Control Systems like VSS. The branches eventually die out, but before the trunk gets a newer version, the users can try the patched versions from the branches. The patched versions don't even need to sit on zope.org website, they could sit in each individual home site (there are so many free web hosting sites out there.) >Right now people can send a patch to the collector, but it takes us time >time test it and understand it, especially if it's a deep issue. That's totally fine and understandable. But we need to figure out a way for people to be aware of the bugs and patches right at the moment when they download a product. This way, Digicool can even afford to release a new version only once every 6 months, say. :) Also, over time, if an outside person starts to develop a credible branch, and Digicool is too busy to take care of the trunk, we'll know it's moment to either (1) outsource the product, or (2) push the branch into zope.org and make it the trunk. >I vote Wiki. Sounds good. Anything that allows user to add HTTP links is good enough. If they can scribble comments, so much the better. :) regards, Hung Jung _________________________________________________________________________ Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com. Share information about yourself, create your own public profile at http://profiles.msn.com.

1 0

Acquisition by context first
by Steve Alexander 15 Sep '00

15 Sep '00

Here's an external method that ought to work. Well, it works for me: # acquire the object of name "attr" by context alone def aquire_by_context(self, attr): ob = self while hasattr(ob, 'aq_base'): aq_b=getattr(ob, 'aq_base', ob) if hasattr(aq_b, attr): attr_obj = getattr(aq_b, attr) break; ob = ob.aq_parent else: # not found raise KeyError, attr return self.restrictedTraverse(attr_obj.getPhysicalPath()) This should be secure, because I'm using restrictedTraverse. Use in DTML like this: <dtml-var standard_html_header> <br> <h2>Context</h2> <dtml-var "acquire_by_context(this(), 'O')"> <br> <dtml-var standard_html_footer> Assuming that you called the external method "acquire_by_context". Here, I'm assuming that you have objects like this: some_folder A C B O contains the string "some_folder->B->O" O contains the string "some_folder->O" aq dtml method containing the dtml above Typing into your browser the URL .../some_folder/A/aq will get you some_folder->O Typing into your browser the URL .../some_folder/B/A/aq will get you some_folder->B->O To verify this against normal acquisition, use this DTML: <dtml-var standard_html_header> <br> <h2>Context</h2> <dtml-var "a_context(this(), 'O')"> <br> <h2>Normal</h2> <dtml-var O> <dtml-var standard_html_footer> -- Steve Alexander Software Engineer Cat-Box limited http://www.cat-box.net

1 2

RequestContainer ?
by Steve Alexander 15 Sep '00

15 Sep '00

There's a RequestContainer instance at the top of the acquisition list, as accessed through aq_parent or through aq_inner.aq_parent. Previously, with recursive acquisition hacks, you could tell that you were at the end of the trail because aq_parent returned None. Unfortunately, RequestContainer doesn't seem to be acquisition-wrapped, and doesn't support the aq_* attributes. Here's an external method suggested by Evan Simpson in August, which doesn't work because of having the RequestContainer at the end of the trail. ---- def aq_containment(ob): context = [] while ob is not None: context.append(ob.aq_base) ob = ob.aq_inner.aq_parent ob = context.pop() while context: ob = context.pop().__of__(ob) return ob ---- What is the recommended, forwards-compatible way of knowing when to stop recursively looking through acquisition wrappers? -- Steve Alexander Software Engineer Cat-Box limited http://www.cat-box.net

1 0

aq_.* names reserved?
by Steve Alexander 15 Sep '00

15 Sep '00

I'm hacking around with some external methods called aq_containment and aq_context. I just found out that I can't call them from DTML. I can call them from the URL line of a browser just fine. If I rename them to a_containment and a_context, they work from DTML. I guess there's something in Acquisistion.c that reserves all aq_.* names. Bug or feature? -- Steve Alexander Software Engineer Cat-Box limited http://www.cat-box.net

1 0

Acquisition Confusion :S
by Chris Withers 15 Sep '00

15 Sep '00

Hi Jim, (or any zope-dev'ers who can explain this ;-) I've just been reading: http://www.zope.org/Members/jim/Info/IPC8/AcquisitionAlgebra ...and I've got confused on the 'Getting the containment wrapper' section. In anser to 'What is A.B.C.D?' I can't see how you can get: ((D o (C o A)) o (B o A)) Here's my working: A.B.C.D = (B o A) : .C.D = ((C o A) o (B o A)) : .D = ((D o C) o ((C o A) o (B o A))) You can't apply the simplification: ((Z o Y) o (Y o Z)) to the leftmost section 'cos the brackets are in the wrong place... so how to you make the jump to: = ((D o (C o A)) o (B o A)) (but should it?!) Also, is the definition "aq_inner is obtained by finding the innermost wrapped object" the same as "aq_inner is the 'self'/'object' part of the current aquisiton wrapper"? Confusedly and looking for help, Chris

5 22